Time-resolved Spectroscopy in Low Pressure Plasmas Johannes Lindén Lund Observatory Lund University
2005-EXA12
Degree project of 20 credit points (for a degree of Master) June 2005 Lund Observatory Box 43 SE-221 00 Lund Sweden
Contents
1
2
3
4
5
Introdu tion
4
The Physi s of Fluores ent Tubes
7
1.1 The Importan e of Light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Ele tri Dis harge Lamps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 The Purpose of this Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4 4 5
2.1 General Properties of Plasmas . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2 Basi Properties of Low-Pressure Dis harges . . . . . . . . . . . . . . . . . . . 9 2.3 Operating a Fluores ent Tube . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Time-resolved Spe tros opy
3.1 3.2 3.3 3.4
Something about time-s ales . The Basi Idea . . . . . . . . . Equipment . . . . . . . . . . . The Data Pro essing . . . . . . 3.4.1 Andor programming . . 3.4.2 MATLAB programming 3.5 Equipment Setup . . . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
16
16 17 17 20 20 21 21
Appli ations
25
Con lusions and Future Appli ations
41
4.1 Where to look . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 4.2 Experiments and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 4.3 Dis ussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 5.1 Con lusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 5.2 Future Appli ations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
A Manual
43
B Andor
46
C MATLAB
54
Bibliography
60
1
2 Abstra t
In this work a method for studying emission lines in the visible region from uores ent lighting tubes is developed. Fluores ent tubes are ele tri al dis harge lamps, meaning the radiating body onsists of a plasma, whi h is a partly ionized gas. The pressure in ordninary room lighting tubes are some thousandths of an atmosphere, whi h makes these plasmas low pressure plasmas. The method developed in this work is based on time-resolved spe tros opy, i.e. one studies the intensity of the emission as a fun tion of time. Sin e uores ent tubes operate on alternating urrent, the emission behavior repeats itself after ea h period. The aim of this method is to study the intensity variation of the emission from dierent elements in the plasma in a uores ent tube during one period. The emission behavior also varies from one position to another, and this is also studied in the work presented here, using opti al bres pointing too dierent positions along the tube. The purpose of the method is to gain information about the energy- onversion path leading from ele tri al energy to desirable photons. A better understanding of the energy and material ow lose to the ele trodes at the ends of the uores ent tube is essential to the development of more e ient and longer lasting uores ent tubes. The work is partly a onsequen e of a ooperation proje t between the department of Atomi Astrophysi s at Lunds Observatory and Aura Light AB in Karlskrona, Sweden, whi h is a ommer ial manufa turer of uores ent tubes. Apart from an introdu tion of plasma fundamentals and basi physi s of uores ent tubes, this report des ribes an experimental setup and its omponents, designed to investigate uores ent tubes using time-resolved spe tros opy. The method will be demonstrated, using an Aura Light 36 W T8 uores ent tube. A manual for operating the setup is presented at the end of the report (Appendix A). Key words:
Fluores ent lighting tubes, time-resolved spe tros opy, low pressure plasma
3 Sammanfattning
I detta arbete utve klas en metod för att studera emissionslinjer i det synliga området från lysrör. Lysrör är s.k. urladdningslampor, vilket innebär att den strålande kroppen utgörs av ett plasma, d.v.s. delvis joniserad gas. Try ket i vanliga lysrör för rumsbelysning är några tusendelar av en atmosfär, vilket gör dessa plasmor till lågtry ksplasmor. Metoden som utve klas i detta arbete är baserad på tidsupplöst spektroskopi, d.v.s. man studerar intensiteten av det emitterade ljuset som funktion av tiden. Eftersom lysrör drivs på växelström, upprepar sig emissionsbeteendet efter en period. Målet med metoden är att studera intensitetsvariationen hos emissionen från olika ämnen i plasmat i ett lysrör under en period. Emissionsbeteendet varierar o kså mellan olika positioner, o h även detta studeras i arbetet, med hjälp av optiska brer som riktas mot olika positioner på röret. Syftet med metoden är att tillgodose sig information om energiomvandlingen från elektrisk energi till önskvärda fotoner. En bättre förståelse av energi- o h material ödet i elektrodregionerna i lysrör är grundläggande för utve klandet av eektivare o h mera långlivade lysrör. Arbetet är initierat av ett samarbetsprojekt mellan avdelningen för Atomär Astrofysik vid Astronomiska institutionen vid Lunds universitet o h företaget Aura Light AB i Karlskrona, som är en kommersiell tillverkare av lysrör. Förutom en introduktion till plasmafysik o h elementär lysrörsfysik, beskriver denna rapport en experimentuppställning o h dess delar, avsedd för att undersöka lysrör tidsupplöst. Metoden kommer att demonstreras på ett Aura Light 36 W T8 lysrör. Längst bak i rapporten nns även en manual för uppställningen (Appendix A). Ny kelord:
Lysrör, tidsupplöst spektroskopi, lågtry ksplasma
Chapter 1 Introdu tion 1.1
The Importan e of Light
There is no doubt that light onstitutes a great ne essity in modern human life of today, but it is so ommon, that it is taken for granted. Light is today both heap and highly available. There are lamps swit hed on twenty four hours a day, and many pla es are blazed with light even though there's not a person around, like parkinglot buildings and parts of ities and roads at night. Most of this light is produ ed by ele tri dis harge lamps, unlike the lassi al in andes ent light bulb. In an ordinary in andes ent light bulb, the tungsten lament is heated by an ele tri al urrent, and glows bright be ause of the heat. This produ e not only radiation of visible light, but a large fra tion of invisible infrared ele tromagneti radiation. This radiation we experien e as heat and is the reason you burn your self if you tou h a swit hed on in andes ent light bulb. In fa t only about 5% to 8% of the provided ele tri al energy is
onverted into visible light, and the rest being onverted into heat. 1.2
Ele tri Dis harge Lamps
An ele tri al dis harge lamp is a lamp that onverts ele tri al energy into light through an atomi dis harge in a gas mixture of some sort. There are many dierent types of dis harge lamps but this thesis will deal with uores ent tubes, whi h are low pressure dis harge lamps. Compared with an in andes ent light bulb, an ele tri dis harge lamp onvert about 25% to 30% of the ele tri al energy input into light energy output. Moreover they last a longer time, 10,000 hours and up (there are 8760 hours in a year) whi h is about ten times longer than a onventional in andes ent light bulb. Of ourse there is also some disadvantages. They are omparatively expensive and they require auxiliary apparatus to regulate lamp power. Furthermore they do not fun tion well in short-time servi e, su h as being swit hed on and o frequently, and they often ontain mer ury. Fluores ent tubes are also more ommon in industries and publi rooms and pla es, and not around home environment, sin e the light is sometimes onsidered to be 'too old' or too arti ial. 4
1.3.
THE PURPOSE OF THIS WORK
5
Even though uores ent lighting tubes have been used ommer ially sin e the 1940s, there is no omplete understanding of the physi al pro esses involved. For this reason, there is mu h left to learn about how to make uores ent tubes more e ient and longer lasting. Fluores ent tubes onsumes some 10% to 15% of all ele tri ity worldwide, so even marginal improvements are important (Kitsinelis, S. et al 2004). The ele tri al dis harge, whi h is responsible for the onversion of ele tri al energy into light, o
urs in the plasma, whi h apart from neutral parti les is a mixture of negative ele trons and positive ions. Plasmas are often alled The fourth State of Matter, sin e they
an be seen as an additional further step in the ladder: solid, uid and gas. In a running lamp, the plasma onstitutes basi ally the entire inner volume of the tube. The ele tri al eld in the tube provides the ele trons with kineti energy, energy that in turn is onverted into radiation as a result of some kind of ollision pro ess. Here lies the importan e of the mer ury. By oin iden e the atomi properties of the mer ury atom is so highly suitable for this purpose that it alone is responsible for the onversion of about 60% of the supplied ele tri al energy into ultraviolet (UV) ele tromagneti radiation.
1.3
The Purpose of this Work
The life time of uores ent tubes is in prin ipal determined by the wear of the ele trodes, whi h are two laments, one at ea h end of the tube, that provides the lamp with an ele tri al eld. Sin e the plasma onsists of harged parti les, it is ele tri ally ondu tive. We therefore have a ow of ele trons from one of the ele trodes (the athode) through the plasma to the other ele trode (the anode). Sin e uores ent tubes usually operates on alternating urrent, ea h ele trode serve as a athode for half of the time, and as an anode the other half. The ele trodes onsists of a tungsten wire oated with some emitting material, a mixture of barium, strontium and al ium oxides, among other things. The purpose of the emitting material is to lower the work fun tion of the tungsten wire, making it easier for the ele trons to be emitted. The reason why the ele trodes stop fun tioning is that the emitting material basi ally is onsumed. While the lamp is operating, the urrent through the ele trodes
auses the emitting material to slowly evaporate. Furthermore, while the ele trode is in the
athode phase, it attra ts the heavy ions in the plasma. When these ions impa t on the ele trode they ause something alled sputtering, when the emitting material is kno ked o the ele trode. The evaporated and sputtered emitting material nally be omes stu k on the inside of the glass tube in the vi inity of the ele trode. The problems of low e ien y and shortened life time of uores ent tubes has been in fo us sin e the advent of dimmable uores ent lamps and systems with frequent on/oswit hing. It is known that uores ent tubes operating under these onditions experien e ex essive ele trode wear. If one an understand the hara teristi of this wear, one has a better han e of developing improved ele trodes and also improved ways of operating uores ent tubes. This diploma work aims to demonstrate a spe tros opi al method of analyzing, not only
6
CHAPTER 1.
INTRODUCTION
the wear of the ele trodes, but also the atomi pro esses in dierent positions of the plasma. The idea is thereby to determine details about the dis harge and hen e make a diagnosti for the lifetime of the lamp. Using this method under dierent operation onditions sholud make it possible to draw on lusions about more e ient ways to run uores ent lamps. There are several reasons why the development of uores ent tubes is important. As mentioned above, uores ent lighting onsumes about 10% to 15% of all ele tri ity in the world, whi h should alone be enough motivation. But there are other aspe ts too. First, the fa t that uores ent tubes ontain mer ury make them an environmental issue. To fully understand the atomi pro esses and the fun tion of the mer ury ould lead to possibilities to perhaps ome up with more environmentally friendly alternatives. Furthermore, longer lasting uores ent tubes should redu e the dis harge of mer ury. Se ond, sin e uores ent tubes are so widely used, they involve a vast industry. The
ommer ial manufa turers of uores ent light tubes have a great interest in more e onomi al and e ient methods of produ ing lamps. This require an extensive understanding of how uores ent tubes a tually works. Third, there is a publi interest in improved uores ent light sour es. Even though the modern te hnology of today oers uores ent lights in a wide variety of olor distributions, some of these te hniques are expensive. There is still mu h to be onsidered in how humans experien e dierent kinds of light. Fourth, there is the s ienti interest in developing better theoreti al models of the physi al pro esses, whi h today are not known in detail. The method presented here is based on time-resolved spe tros opy, in whi h not only the intensity as a fun tion of wavelength is onsidered, but also the intensity as a fun tion of time. This means that on e we onsider a spe i wavelength, we are able to investigate the intensity variation of this line during a ertain time span. This is required sin e lamps are operating on alternating urrent, thus time is an important dimension. The prin iple of time-resolved spe tros opy in this parti ular ase will be dis ussed later in this work. Chapter 2 will rst onsider some properties of plasmas in general, followed by some properties of low pressure dis harges and the physi al pro esses and elements involved in uores ent lighting tubes. Chapter 3 will deal with the method, the experimental setup and the method of analyzing the data. In hapter 4 the method will be demonstrated and some results will be given and dis ussed. Chapter 5 will ontain some on lusions and also dis uss some future appli ations.
Chapter 2 The Physi s of Fluores ent Tubes Plasmas as in uores ent tubes are low pressure plasmas, i.e. the pressure is some thousandth of normal air pressure (3 - 4 mbar). But before we onsider plasmas in uores ent tubes we will handle properties of plasmas in general. Last in this hapter we will study the operation of a tube, and what is going on around the ele trodes. The pro ess of ignition will not be
onsidered in this work. 2.1
General Properties of Plasmas
As mentioned in the introdu tion, a plasma an be onsidered as a fourth state of matter. With in reasing ontent of energy the normal lassi ation of the rst three states of matter is that of solid, liquid and gas. Gradually adding more energy to a system of gas will rst ex ite the atoms and mole ules of whi h it onsists to ex ited states. When the added energy is su iently high to ionize the atoms and mole ules an ionized gas is formed. The ionized gas thus onsists of neutral atoms and mole ules and positive ions and negative ele trons. The more energy that is added to the system the bigger fra tion of the gas be omes ionized. When the number of ions and ele trons are su iently numerous for their presen e to dominate the behavior of the system the plasma state is said to be established, but the word plasma is a onvenient shorthand to des ribe any vapour or gas that is partly ionized. In ordinary life we are mostly used to the rst three states of matter, and the fa t is that there are very few examples of naturally o
urring earthbound plasmas. Examples of man made plasmas would be uores ent tubes and thermonu lear fusion plasmas et . However, an ordinary ame is to be onsidered as a plasma, partly onsisting of free ele trons and ions. The ionization energy is supplied by the oxidation of the fuel. The ionosphere and the solar orona are other examples of plasmas. Getting farther away from the surfa e of the earth plasmas be ome more and more ommon. The fa t is that most of the interstellar medium onsists of free harge parti les and therefore it is believed that plasma onstitutes 99% of the known matter in the universe (Boley, 1966). Both ions and ele trons are free to move in a plasma, whi h makes it ele tri ally ondu tive. Both of the harges serve as harge arriers, but be ause of their mu h smaller mass 7
8
CHAPTER 2.
THE PHYSICS OF FLUORESCENT TUBES
the ele trons move more freely than the ions. The plasma as a whole is ele tri ally neutral, the total ioni harge being equal to the number of free ele trons. Any deviation of harge would be evened out by the enormous potential it auses. As an example, onsider a sample of plasma with ion density ni and ele tron density ne ontained in a sphere of radius r. The net harge Q of this sphere is 4 3 3 πr (ni − ne ) e and the orresponding rise in potential V is given by V =
Q e (ni − ne ) r2 = . 4πε0 r 3ε0
With a typi al value of ne , 1022 m−3 , and if ni and ne would dier by 1% the potential over a sphere of radius 1 mm would be 600 000 volts. Su h a large potential annot possible be maintained in a gas with freely moving harges. It is therefore quite safe to take the average parti le densities as given by (in the ase of singly harge ions) (2.1)
ni = ne .
It is possible to add su iently energy to totally ionize a gas, so that no neutral parti les remains. But as mentioned above the plasma state is established when the harged parti les are numerous enough to dominate the behavior of the system. That is, the plasma
an be ae ted olle tively by external ele tri and magneti elds and an ondu t ele tri
urrent. The importan e of olle tive ee ts is the distinguishing hara teristi of plasmas and the primary plasma riterion. Be ause of the long range of ele trostati for es, every harged parti le intera ts with several other parti les at the same time and will produ e interesting
onsequen es via olle tive ee ts. Individual intera tions between parti les are important only over small distan es, while over larger distan es the olle tive ee ts dominate. The distan e, that distinguish small from large distan es is known as the Debye length, ρD : it measures the distan e to whi h the ele tri eld of an individual ion or ele tron extends before it is ee tively shielded by oppositely harged parti les. Thus the Debye length an be thought of as the radius of a sphere surrounding ea h harged parti le. The names Debye radius or Debye sphere are therefore also ommon. The Debye length is determined by the temperature T and the ele tron density ne and is given by ρD =
εo kT 2e2 ne
1/2
.
(2.2)
Thus if one is to talk about olle tive ee ts in a plasma the linear dimensions L of the plasma should be large ompared to ρD : (2.3) To dedu e ρD requires a smooth de rease in the eld from a single harge, and therefore a large number of parti les in the neighborhood. This means that the number of parti les within a sphere of radius ρD must be mu h greater than one. That is ρD ≪ L.
4 3 πρ ne ≫ 1. 3 D
(2.4)
2.2.
BASIC PROPERTIES OF LOW-PRESSURE DISCHARGES
9
One of the fastest and most important of the olle tive motion is the bulk os illation of the plasma ele trons with respe t to the ions. The frequen y of this os illation is alled the plasma frequen y and is a olle tive motion of free ele trons. The plasma frequen y in Hz is given by νp =
e2 ne 4π 2 εo m
1/2
(2.5)
in other words it depends only on the ele tron density. To insure a steady-state plasma in whi h olle tive motions an o
ur the time s ale of this motion must be larger then those me hanisms tending to destroy su h olle tive motions. For example, this olle tive behavior tends to disappear if the ele tron motions are randomized by ollisions, and a ne essary ondition for its existen e is that the time between
ollisions must ex eed the os illation period: the ollision frequen y νc must be smaller than νp . νc ≪ νp . (2.6) To summarize the pre eding dis ussion we an now say that a plasma is a olle tion of
harged and neutral parti les that satis es ertain riteria. These riteria are; the plasma is approximately neutral as a whole, eq. (2.1); the Debye length is small ompared to the plasma dimensions, eq. (2.3); the number of parti les must be mu h lager then 1 inside the Debye sphere, eq. (2.4); and the plasma os illations are not strongly damped, eq. (2.6). 2.2
Basi Properties of Low-Pressure Dis harges
The pre eding part has onsidered plasmas in general. Now we will onsider low-pressure plasmas as in uores ent tubes. In these low pressure plasmas the parti le density is in order of 1016 cm−3 . When the lamp is in operation about 1 in a million of all these parti les are ionized whi h might not seem to impressive. Still, be ause of the strong Coulomb intera tion, this degree of ionization is su ient for the harge parti les to dominate the pro esses of the system. We now ontinue with onsider the elements in a uores ent tube, their properties and the reason of them being there. A uores ent tube is lled with a gas alled a buer gas. Typi al parameters for the buer gas ontent of an ordinary tubes ould be 85 % Kr and 15 % Ar. These are rare gases, i.e. they have a high ionization potential. Other rare gases like neon or xenon are also often used as buer gases in uores ent tubes. In addition a small amount of liquid mer ury (4-10 mg) is put in the tube, thus the tube ontains mer ury vapour at the saturation pressure of the urrent lamp. The plasma in an operating tube radiates mostly UV light. When this radiation rea hes the wall of the tube it is onverted to visible light by oatings of uores ent phosphor powder applied to the inside surfa e of the tube wall. By a suitable hoi e of phosphor, a wide variety
10
CHAPTER 2.
THE PHYSICS OF FLUORESCENT TUBES
of olors distribution an be a hieved, su h as warm white or old white et . The properties of phosphors in uores ent lamps will not be onsidered in this work. As mentioned in the introdu tion, the mer ury is alone responsible for the onversion of a. 60% of the ele tri al energy to UV light. The e a y of the dis harge is prin ipally determined by the fra tion of the ele tri al power dissipated in the dis harge that rea hes the phosphor in the form of 254 nm radiation emitted from the 6p 3 P1 → 6s 1 S0 transition in Hg I. This transition together with a substantial fra tion of the 6p 1 P1 → 6s 1 S0 transition, emitting another UV line at 185 nm, are responsible for most of the output of any uores ent tube (see Figure 2.1). For any given phosphor, the output of visible light is very losely proportional to the intensity of 254 nm UV light in ident upon it. There are also a number of visible emission lines from the mer ury (making the plasma look greenish blue without phosphor powder) whi h in ombination with typi al phosphors in some lamps also are important for optimum e a y and olor temperature.
Figure 2.1: Energy level diagram for mer ury, showing the states and lines responsible for the majority of the total radiation. Wavelengths given in nanometers (nm) (Waymouth, 1971).
Thus what one wants to a hieve in a uores ent lamp plasma is a high UV radiation energy loss that dominates the other two major energy losses: ionization loss and elasti ollision loss. The following will deal brie y with the energy ex hange pro esses in a dis harge plasma. The maintenan e of the plasma ondu tivity requires the produ tion of new ele tron-ion pairs by ionization as fast as they are lost. Be ause gas pressure is a small fra tion of an
2.2.
BASIC PROPERTIES OF LOW-PRESSURE DISCHARGES
11
atmosphere, the prin ipal loss pro ess for ions and ele trons is ambipolar diusion to the walls, where the ions re ombine to form neutral atoms. Hen e the loss rate and the required produ tion rate are dependent on tube diameter and gas ll pressure. The ele tri al energy input is given almost entirely to the ele trons of the plasma. A
elerated by the ele tri eld, they make elasti ollisions with gas atoms, ex hanging momentum but very little energy; they make ollisions with ea h other, whi h only ex hange energy between ele trons; and they make inelasti ollisions with gas atoms, whi h either ionizes the atoms or ex ites them to energy states from whi h the atoms may radiate. As a result, the energy input to the ele trons is shared among them in the form of an ele tron energy distribution whi h is su iently lose to a Maxwellian that it an be assigned a 'temperature'. This 'ele tron temperature', Te , is then determined by the balan e between the ele tri eld and the energy losses by the ele tron gas due to ionization, elasti s attering, and ex itation. The ele trons share the energy input to maintain a distribution of velo ities where the temperature Te is determined by the energy balan e. This distribution of velo ities in the ele tron gas is given by the Maxwell-Boltzmann formula: 2 f (U ) = √ π
U kT
1/2
(2.7)
e−U/kT .
Figure 2.2 shows this distribution for two ele tron temperatures. In a steady-state plasma, the produ tion rate of ele tron-ion pairs is the same as the rate at whi h they are lost. The produ tion rate varies greatly with ele tron temperature, the fra tion of ele trons with su iently energy to ionize an atom in reases exponentially with Te (see Figure 2.2). Ex itations of an atom requires a minimum ele tron energy, equal to the dieren e of energy of the state to whi h the atom was ex ited and the energy of its original state. Again, the fra tion of ele trons having su ient energy to ex ite atoms generally varies approximately exponentially with ele tron temperature. The outstanding e ien y of Hg dis harge in produ ing resonan e radiation is in part due to the low ex itation energy of the rst ex ited on gurations, whi h in ludes the so important resonan e level at 4.89 eV. This is slightly less then half the ionization energy of 10.4 eV. In ontrast, rare gases have rst-ex itation on gurations with resonan e levels at ex itation energies mu h greater than half the ionization potential. This is illustrated by Figure 2.3 whi h shows a part of an energy level diagram of Hg I, together with areas of the lowest ex ited states of Xe, Kr and Ar. See Table 2.1 below for some relevant energies.
Gas
Neon Argon Krypton Xenon Mer ury
Ex itation Energies
16.67/16.85 11.6/11.8 10.0/10.6 8.44/9.57 4.89/6.7
Ionization Potential
21.56 15.76 14.00 12.13 10.43
Table 2.1: Relevant energies (eV).
A
ording to Figure 2.2 the fra tion of ele trons, at an ele tron temperature of 1 eV, having energy enough to ex ite mer ury atoms is mu h greater than that of having energy
12
CHAPTER 2.
THE PHYSICS OF FLUORESCENT TUBES
Figure 2.2: Maxwell-Boltzmann distribution for two ele tron temperatures. Shown for omparison are the energies of some typi al states of mer ury and argon (Waymouth, 1971).
enough to ex ite buer gas atoms. Therefore, sin e ele tron temperature in uores ent lamps are in this vi inity, there is little or no ex itation of buer gas atoms. All the ex itation loss goes into ex iting mer ury atoms. As a result of this, most of the ele trons make elasti ollision with buer gas atoms, redu ing the mean free path of the ele trons, and inelasti ollisions with mer ury atoms, leading to ex itation or ionization. Herein lies the prin ipal fun tion of the buer gas; slowing down the diusion rate of ele trons to the wall by de reasing the mean free path of the ele trons. This then has the ee t of adjusting the ele tron temperature to a desired optimum level, whi h an be des ribed as high enough that ex itation and radiation losses greatly ex eed elasti ollision losses and low enough that ex itation of the desired state of the mer ury atom ( 3 P1 , 4.86 eV) predominates over ex itation of all higher energy states.
2.3.
OPERATING A FLUORESCENT TUBE
13
Figure 2.3: Part of energy level diagram of Hg I. Areas of the lowest ex ited states for Xe, Kr and Ar are marked out with red, green and blue respe tively.
2.3
Operating a Fluores ent Tube
As mentioned in the introdu tion, the plasma in a running uores ent lamp onstitutes basi ally the entire inner volume of the tube. But the behavior of the dis harge plasma is not the same through the entire tube. The potential, and hen e the plasma, in a lamp operated on DC shows some spe ial hara teristi s from one end to the other. Figure 2.4 is a sket h identifying the main dis harge regions with the potentials a ross them. Ele trons emerging from the athode are a
elerated through the athode fall into a region of relatively weak ele tri eld, the negative glow. There is an overprodu tion of ions in the negative glow, whi h is ompensated by a dark region of low ionization, the Faraday dark spa e. Following the Faraday dark spa e is a region of onstant ele tri eld, the positive olumn. This region is the largest and the major sour e of UV radiation in uores ent lamps. A bright anode glow ends the positive olumn due to the high potential dieren e between the positive olumn and the anode, the anode fall. For several reasons most uores ent lamps operate using an AC ballast, leading to ea h of the ele trodes alternates between serving as athode and as anode. One reason is the ee t of ataphoresis whi h o
urs in lamps operated on DC. Cataphoresis is the a
umulation of mer ury near the athode, and will ause de rease of useable UV photons. Another reason is the negative hara teristi of plasmas, i.e. the more urrent the less resistivity, as opposed to
14
CHAPTER 2.
THE PHYSICS OF FLUORESCENT TUBES
Figure 2.4: Sket h identifying the main dis harge regions and the potentials a ross them (Waymouth, 1971).
ordinary resistors. Be ause of this, an resistive ballast is needed whi h auses ohmi energy loss. To balan e this hara teristi in a AC operated lamp, an impedan e ballast is needed. For a very long time, tubes were operated on an AC of about 50/60 Hz. However, it has been known for a onsiderable time that high-frequen y operation leads to improvements in lamp luminous e a y. Therefore, newer lamps are designed to be operated on AC of about 10 kHz to 40 kHz. The main ee t of high frequen y operation is the de rease in the anode fall. High frequen y also leads to in reasing radiation e ien y in the positive olumn. The understanding of the dis harge plasma in the positive olumn is reasonably omplete. However, there remains a onsiderable la k of understanding aurond the physi s of the ele trode regions, espe ially when the ele trode is serving as a athode. Atomi and mole ular data are still missing for many important rea tions. To ompli ate things even further, a reasonable model has to be time-dependent, onsidering the AC ballast. Sin e the operating life of a uores ent lamp is determined essentially by the life of its ele trodes, a better understanding of what's going on in this region is desirable. The ele trodes in a uores ent lamp are heli es of tungsten, whi h are impregnated with alkaline-earth oxides for enhan ed ele tron emission. During normal operation, they are heated by the passage of urrent through the tungsten wire and by ion bombardment from the plasma. The oxides redu e the work fun tion of the tungsten wire, allowing the athode to supply urrent to the dis harge at an operating temperature of 1200-1400 K. In most
2.3.
OPERATING A FLUORESCENT TUBE
15
ases, the athode operates in the spot mode, a glowing spot on the wire whose position is varying during the life of the lamp as the emitting material is lo ally evaporated and sputtered. The heated ele trodes emit ele trons in an pro ess alled zero- eld thermioni emission. This apability to emit ele trons is enhan ed greatly by the presen e of a
elerating elds at the athode surfa e, the athode fall. The negative harge of the athode attra ts positive ions and the ions that rea h the athode strike it with nearly the full energy of the athode fall. Due to the thermioni emission the athode is partly prote ted from the infalling ions by a sheath of ele trons. This prote ting sheath be omes bigger the hotter the athode. On the other hand, too hot a athode will ause the emitting material to evaporate whi h drasti ally de rease the athode life time. Premature failure is aused by either too hot athode, too mu h evaporation, or too old athode, too mu h sputtering. Although, estimated life times of the ele trodes from the vapour pressure and the relevant temperature indi ate a life time mu h less then the observed one. What is the energy and material ow lose to the ele trodes? A good understanding of the pro esses near the ele trodes an be used to optimize the design. By measuring the time-resolved emission relative to the ele tri al ex itation one an re onstru t the ex itation path by omparing to simple models, and hen e map the dierent ex itation pro esses. This diploma work intends to develop a time-resolved spe tros opi al method for this purpose and to demonstrate it on some simple features.
Chapter 3 Time-resolved Spe tros opy Experimental investigations based on time-resolved spe tros opy has not been done on uores ent lamps to any great extent (Kitsinelis, S. et al 2004, Lopez, J. et al, 2005). Intensities are often only measured as a fun tion of wavelength. With time-resolved spe tros opy one measures the intensity of a spe i wavelength (or a small wavelength interval) as a fun tion of time. In uores ent lamps operated on AC, the time span of interest is that of one period. After one period the intensity behavior is repeated. Operating on 50 Hz makes this period 20 ms. Dierent elements, and therefore dierent spe tral lines, show dierent behavior during one period. This behavior also varies depending on where in the plasma one is looking, like in the positive olumn or in the ele trode region. With opti al bres and a CCD dete tor, simultaneous investigations at several dierent positions will be possible. Hen e the method developed here will make it possible to investigate emission from uores ent lamps with spe tral, temporal and spatial resolution. 3.1
Something about time-s ales
During this diploma work experiments have been done using a 36 W Aura Light T8 uores ent tube whi h was mostly operated on 50 Hz AC. Thus the period is 20 ms. During this period ea h of the ele trodes serve as an anode half the time (the anode phase) and as a
athode the other half (the athode phase). A high frequen y lamp, e.g. one that operates on 30 kHz, has a period of 33 µs. It may be of interest to onsider some dierent time-s ales on whi h dierent atomi pro esses happen. All these pro esses take part in the ex itation pro ess and by study the time dimension one might be able to draw on lusions about the importan e of ea h of the dierent pro esses. - Allowed transitions: - Metastable transitions: - Spontaneous emission: - Collisions: - Diusion:
ns
µs - s
ns
µs
ms
From the list it an be seen that parti ularly the diusion pro esses is ee ted by high 16
3.2.
THE BASIC IDEA
17
frequen y operation.
3.2
The Basi Idea
The idea is to measure the intensity as a fun tion of time during one period. This is done by measuring the intensity during a number of equidistant time sli es through the period. At ea h 'time sli e'-position a measurement is done by taking a snapshot of the spe trum line with the dete tor. These measurements will not all be possible to re ord in one single period, the exposure time for ea h snapshot is to short to get a reasonable signal. Therefor the measure of ea h position in time must be done during several periods and then added together. Thus a whole measurement of the intensity variation during one period, has to be done during a large number of periods. For example, say we want to re ord the time-resolved intensity of one period with a resolution of 100, i.e. measure the intensity at 100 dierent equidistant positions. In order to get at reasonable signal, say 20 exposures have to be made at ea h point, resulting in the total number of 100 · 20 = 2000 periods. De ning a spe i position in the period as starting time t0 = 0 (e.g. the voltage zero- rossing) ea h of the positions an be asso iated with a ertain delay-time from t0 . One of the major hallenges during this diploma work was to onstru t general routines to olle t, save and analyse the data olle ted from the dete tor and to nally plot it in a time-resolved diagram.
3.3
Equipment
The ICCD - Time resolution The dete tor is an Andor ICCD gatable dete tor model DH534-18H-03 FR (see Figure 3.1). ICCD's are Intensi ed Charge Coupled Devi es, omprising a CCD-dete tor and a gated image intensi er. As well as amplifying, the image intensi er an rapidly be swit hed on and o, allowing it to be used as a very fast shutter. These on- and o signals is send to the ICCD via a onne tor on the ICCD head. It is these on-o pulses that have to be syn hronized with the voltage zero- rossing. The dete tor is also onne ted to a omputer whi h olle ts the sampling data. A ommer ial software program, Andor MCD ver. 2.62 (AMCD), makes it possible to plot and save data. With this software there is also a possibility to setup measurement sequen es with the Andor Basi Programming Language, ver 2.62 (ABPL).
The Spe trograph - Spe tral resolution The spe trograph used is a Jarrel Ash 1 m Czerny-Turner grating spe trograph. It is shown in Figure 3.2. It is equipped with a TTL pulse ontrolled me hani al shutter whi h an be
ontrolled from the AMCD. The purpose of the me hani al shutter is to prevent too mu h infalling light on the dete tor.
18
CHAPTER 3.
TIME-RESOLVED SPECTROSCOPY
Figure 3.1: The ICCD gatable dete tor.
Figure 3.2: The 1 m Czerny-Turner grating Spe trograph.
The Fibres - Spatial resolution Figure 3.3 shows the arrangement of the opti al bres. Figure 3.3(a) shows the bre opti s input array. Instead of a image of the verti al spe tral slit, this will be seen on the CCD as a verti al row of a number of short spe tral lines. The other end of the bres an be pointed at any dire tion or at any light sour e wanted, a alibration lamp or a dierent uores ent tube. In Figure 3.3(b) they are all pointed at the same lamp. Fibre 1 (marked with one line) are pointed at the ele trode and the other bres positioned along the length axis of the tube 0.5 m apart.
Mi ropro essor ontrolled Gater and Pulser The dete tion of the voltage zero- rossing and the sending of on- and o pulses to the dete tor head at the right delay time is done by a non- onventional Mi ropro essor ontrolled Gater and Pulser (MGP), onstru ted by the supervisor of this diploma work, S. Huldt, spe ially
3.3.
19
EQUIPMENT
(a) Fibres pointed at the spe trograph entran e slit.
(b) Fibres pointed at light sour e. Fibre 1 (marked with one line) are fo used on the hot spot at the ele trode and the other bres positioned along the length axis of the tube 0.5 m apart.
Figure 3.3: The bre opti s array.
for this purpose (see Figure 3.4). This MGP re eives a trigger input signal from the AC voltage supplying the lamp under investigation. This allows the MGP to dete t the voltage zero- rossing. The MGP is onne ted to the omputer via the COM1 port. It an be send several ommands su h as exposure time for ea h snapshot (i.e the on-pulse length to the dete tor head), delay time (i.e. the time to wait after the zero- rossing), and the number of exposures, or snapshots, that are to be made. When the MGP is sent the ommand 'GO' it starts to dete t a zero- rossing and sends a pulse at the desirable delay time, then it dete ts another zero- rossing and sends the next pulse, and so on until the desirable number of exposures have been done.
Figure 3.4: The Mi ropro essor ontrolled Gater and Pulser.
The mi ropro essor in the MGP needs a ertain y le time and therefore the pulses sent to the dete tor head will not arrive the next period after the latest pulse before. After a pulse is sent, about three periods follow before the next pulse is sent. This means that if
20
CHAPTER 3.
TIME-RESOLVED SPECTROSCOPY
2000 periods is needed for a measurement (as in the example in the previous se tion) the whole measurement will take 8000 periods.
Pulse generator and Ampli er The power supply of the lamp is generated by a FLC Ele troni s Voltage Ampli er A400D1, whi h ampli es a phase signal from a HP 8116A Pulse/Fun tion Generator 50 MHz (see Figure 3.5). The pulse generator is equipped with a trigger output, whi h is onne ted to the MGP zero- rossing onne tor.
Figure 3.5: The Pulse generator and Ampli er.
The pulse generator is apable of generate sinusoidal, re tangle and sawtooth signals with a wide range of frequen ies. 3.4
The Data Pro essing
The data pro essing is divided into two parts. First the data has to be olle ted from the dete tor and saved. This is fully done by the Andor software. On e the a quisition is done, the data has to be analysed, summarized and displayed. This is the se ond part and is done by MATLAB.
3.4.1 Andor programming As mentioned above the AMCD has a possibility to setup measurement sequen es programs. The program written for this experiment is alled Simple Wave and are shown in Appendix B. On e AMCD has been started, Simple Wave an be exe uted as a program inside AMCD. ABPL is apable to ommuni ate via a COM1 port, whi h in this ase is onne ted to the
3.5.
EQUIPMENT SETUP
21
MGP. When Simple Wave is exe uted and the a quisition starts, it sends ommands about the values of delay, number of pulses and pulse length to the MGP. When the ommand 'GO' is sent the MGP dete ts the next zero- rossing and then starts to send pulses to the dete tor head. When a desirable exposures have been made, Simple Wave olle ts the data and save it away as one le. Now the delay time is updated and is sent to the MGP before a new 'GO' signal, and the pro edure is repeated. When the a quisition is nished the total number of saved les equals the value of the resolution. In other words, if one intend to do an a quisition of a period with the resolution of 200, 200 les are saved away, one for ea h delay time. The CCD- hip omprise 1024 x 1024 pixels. As was mentioned above, the use of bres results in an image of a verti al row of a number of short spe tral lines, one for ea h bre. This is the result using Full Resolution Image mode. However, using a pro ess alled binning other modes may be preferred. Binning is a pro ess that allows harge from two or more pixels to be ombined on the CCD- hip prior to readout. The Multi-Tra k mode use a verti al binning pro ess, and by hoosing a number of tra ks equal to the number of bres and a suitable position of the tra ks, instead of reading out 1024 x 1024 dierent values one reads out 1024 x the number of tra ks. Hen e, if the number of bres are six, six olumns of 1024 rows ea h are read out from the dete tor, in other words, six spe tra. Ea h and every one of the 200 saved away les in the example above ontain six olumns and 1024 rows, one olumn for ea h bre. Apart from the data les from the dete tor, an information le about the a quisition are also saved. This ontains information su h as frequen y, resolution, delay time, number of exposures, date and time et .
3.4.2 MATLAB programming In MATLAB (MATLAB 7 R14) two main s ripts are written. These are shown in Appendix C. The rst one reads in all the saved data les. The se ond one al ulates the intensity, by alling a spe ial subfun tion (also shown in Appendix C), and plots a resulting time-resolved diagram. Column no. 1 from ea h of the read in les ontains a spe trum of bre no. 1 at dierent delay times. Say, if 200 les where saved, we have 200 spe tra of the intensity from bre no. 1 whi h, when plotted together, illustrates the intensity variation through the period. Figure 3.6 shows an example. This plot was generated with the MATLAB s ript Zmatrix whi h are also to be found in Appendix C. For ea h and every one of these spe tra, a intensity value is al ulated. These values of intensity variation, result in a nal time-resolved spe trum. The resulting spe trum will
ontain six graphs, one for ea h bre. This will hen e display the behavior variation at dierent positions along the tube. 3.5
Equipment Setup
Figure 3.7 shows the whole equipment setup. When exe uting the Simple Wave s ript in AMCD the a quisition ontains the following steps:
22
CHAPTER 3.
TIME-RESOLVED SPECTROSCOPY
Fibre 6
12000
10000
8000
Rel. Intensity
6000
4000
2000
0
−2000 120 100 80 60 40 20 Wavelength
0
0
20
40
60
80
100
120
140
160
180
200
Time
Figure 3.6: A 3D-plot over the intensity variation of Hg I 546 nm over one period of 20 ms. The spe tra was taken through bre no. 6. The Wavelength and Time axis are not s aled.
• Values of delay time, pulse length and number of pulses are sent from the omputer
to the MGP via COM1.
• 'GO' is sent from the omputer to the MGP via COM1.
• The 'run' ommand in Simple Wave sends rst a Shutter pulse to open the me hani al
shutter and se ond a Fire pulse to the MGP indi ating the dete tor is ready to olle t data. Both Shutter and Fire pulse are sent through the Multi I/O Box.
• When the MGP has re eived both the 'GO' ommand and the on rming Fire pulse it
dete ts a zero- rossing of the trigger signal from the pulse generator and sends, after a hosen delay time, a pulse with desirable length to the gate input on ICCD-head. Then it again dete ts a zero- rossing and sends another pulse to the ICCD-head. This is repeated until the requested number of pulses has been sent.
• The dete tor a
umulates the infalling light during all the pulses, and when all the
3.5.
23
EQUIPMENT SETUP
pulses has been sent, the omputer olle ts the data. • The delay time is updated and sent to the MGP.
• 'GO' is sent from the omputer to the MGP and the pro edure is repeated until the
delay time has s anned a whole period or any other sele ted time span.
OPTICAL FIBRES
SPECTROGRAPH
FLUORESCENT TUBE
Mech. Shutter
Gate input AMPLIFIER ICCDDETECTOR
PULSE GENERATOR
Trig. signal Gate pulses MGP
Fire
COM1: Multi I/O Box
COMPUTER
Shutter
Figure 3.7: The equipment setup.
Figure 3.8 illustrates the ow of pulses for a ertain delay time as the 'GO' ommand is sent. Ea h of the gate pulses has to be triggered by a zero- rossing. The Transfer Time is set in Simple Wave s ript and varies with me hani al shutter type.
24
CHAPTER 3.
TIME-RESOLVED SPECTROSCOPY
'run' Ex itation pulse
Trig. signal
'GO'
??
Shutter Output
Fire Output
Gate Pulses
6
Transfer Time )
Time to open me hani al shutter (
Figure 3.8: The ow of pulses as the 'GO' ommand is sent.
Chapter 4 Appli ations 4.1
Where to look
Mer ury radiates in many dierent wavelengths. Those of greatest interest, i.e. those whi h give rise to visual radiation from the uores ent powder, are those of 185 nm and 254 nm and belongs to the UV region. Those lines never es ape the tube, sin e the tube is made of glass and absorbs wavelengths below 300 nm. Hen e we will not be able to investigate those lines. To do so, a tube with a small window of e.g. magnesium uoride, whi h is transparent to wavelengths down to 120 nm, is required. Sin e air absorbs from 200 nm and below, the system also has to be air eva uated. This is left for future experiments. In this experiment we will restri t ourselves to observe a region prin ipally orresponding to the visual region, 400 to 800 nm. Regarding the mer ury, there are a triplet of lines from the 6s7s 3 S1 to the 6s6p 3 P (see Figure 2.1). The UV 254 nm de ays from the 6s6s 3 P1 state. Restri ted to the visual region, observing these triplet lines are the losest we get to the 254 nm line. The wavelengths of these triplet lines are 405, 436 and 546 nm. In the observed wavelength region these three mer ury lines are the strongest ones emitted from the plasma throughout most of the tube. In the vi inity of the ele trodes, however, emission from other elements, e.g. the buer gases, may be dominating. The 36 W Aura Light T8 uores ent tube used has a gas mixture of 85 % Kr o h 15 % Ar. Hen e many Argon and Krypton lines may be observed around the ele trodes. In Figure 4.2 through 4.4 on pp. 27 - 29 overview spe tra of the observed wavelength region are shown taken at dierent positions along the uores ent tube (see Figure 4.1). These spe tra was taken with a USB portable bre spe trometer (O ean Opti s USB2000). The width of the little box in Figure 4.4 shows the extent of the ICCD dete tor hip, i.e. the spe trograph and ICCD setup re ord a spe tral range of about 10 nm. With a turnable handle on the spe trograph this 'box' is moveable through the entire spe tral region of interest.
25
26
CHAPTER 4.
APPLICATIONS
Figure 4.1: An Aura Light T8 uores ent tube. The outermost 10 m of the glass tube are missing uores ent powder. The positions where the overview spe tra was taken are marked out. a: at ele trode (Figure 4.4). b: at no powder (Figure 4.3). : at powder (Figure 4.2).
As was seen in Figure 3.3(b) there are no bres observing the uores ent powder. They will only onsider the region from the ele trode to the positive olumn. This is the spatial resolution we will have during our experiments. Table 4.1 below lists some interesting lines whi h are to be investigated.
Spe ies Hg I Hg I Hg I Kr I Kr I Ar I Ar I
Wavelength (nm)
404.7 435.8 546.1 760.2 769.5 696.5 706.7
Transition
6s7s 3 S1 − 6s6p 3 P0 6s7s 3 S1 − 6s6p 3 P1 6s7s 3 S1 − 6s6p 3 P2 5s [1 12 ]◦2 − 5p [1 12 ]2 5s [1 12 ]◦2 − 5p [1 12 ]1 4s [1 12 ]◦2 − 4p [ 21 ]1 4s [1 12 ]◦2 − 4p [1 12 ]2
Table 4.1: Interesting lines (www.nist.gov).
4.1.
27
WHERE TO LOOK
4000
3500
Intensity (counts)
3000
2500
2000
1500
1000
500
0
200
300
400
500 Wavelength
600
700
800
Figure 4.2: Spe trum at the uores ent powder (position in Figure 4.1). The strong intensity around 630 nm is due to the uores ent powder.
28
CHAPTER 4.
APPLICATIONS
4000
3500
Intensity (counts)
3000
2500
2000
1500
1000
500
0
200
300
400
500 Wavelength
600
700
Figure 4.3: Spe trum with no uores ent powder (position b in Figure 4.1).
800
4.1.
29
WHERE TO LOOK
4000
3500
Intensity (counts)
3000
2500
2000
1500
1000
500
0
200
300
400
500 Wavelength
600
700
800
Figure 4.4: Spe trum at the ele trode (position a in Figure 4.1). The many lines around 800 nm are mainly emission from Argon and Krypton. The little box shows the extent of the ICCD dete tor hip.
30 4.2
CHAPTER 4.
APPLICATIONS
Experiments and Results
The following time-resolved spe tra in Figure 4.5 through 4.11 on pp. 31 - 37 are re orded using the method developed in this work. The bres have the same position through all the spe tra and are des ribed in Figure 3.3(b). All the spe tra displays the intensity variation through one period under 50 Hz operation with bipolar square wave ex itation. The rst half shows the anode phase and the se ond half shows the athode phase. The rst three spe tra are Hg. The following four spe tra show lines from buer gas atoms, the rst two of these are Kr and the remaining two are Ar. The Hg shows a periodi intensity variation through the anode phase. These os illations in intensity are known as anode os illations. The athode phase initiates with a high intensity peak and through the rest of the phase the intensity from bre no. 1, i.e. the bre dire tly pointed at hot spot on the ele trode, shows the greatest intensity. The intensity from bre no. 4 is the lowest, indi ating the position of the Faraday dark spa e. The buer gases on the other hand shows intensity only at the athode phase, and then almost ex lusively from bre no. 1, i.e. the hot spot.
4.2.
4
Line: Hg I 4046
x 10
EXPERIMENTS AND RESULTS
10
Fibre 1 Fibre 2 Fibre 3 Fibre 4 Fibre 5 Fibre 6
9
8 Lamp: 36W T8 Kr 85%, Ar 15% Line: Hg I 4046 7
Phase Type: Square Frequency: 50
6
Window width: 1000 # of Exp. :50 Gain: 80 Spec: 3961 Laborant: J. Linden
Rel. intensity
Figure 4.5: Hg I 404.7 nm
Resolution: 200
5
4
Date: 6/11/2005 Time: 19:11 Background: Yes
3
2
1
0
0
0.2
0.4
0.6
0.8
1 One period (microsec.)
1.2
1.4
1.6
1.8
2 4
x 10
31
32
5
4
Line: Hg I 4358
x 10
Fibre 1 Fibre 2 Fibre 3 Fibre 4 Fibre 5 Fibre 6
3.5
Lamp: 36W T8 Kr 85%, Ar 15%
3
Line: Hg I 4358 Phase Type: Square 2.5
Resolution: 200 Window width: 1000 # of Exp. :50 Gain: 80
Rel. intensity
Figure 4.6: Hg I 435.8 nm
Frequency: 50
2
Spec: 4274 Laborant: J. Linden 1.5 Date: 6/11/2005
CHAPTER 4.
Time: 18:48 Background: Yes 1
0.5
0
0.2
0.4
0.6
0.8
1 One period (microsec.)
1.2
1.4
1.6
1.8
2 4
x 10
APPLICATIONS
0
4.2.
4
Line: Hg I 5460
x 10
EXPERIMENTS AND RESULTS
14
Fibre 1 Fibre 2 Fibre 3 Fibre 4 Fibre 5 Fibre 6
12
Lamp: 36W T8 Kr 85%, Ar 15% 10 Line: Hg I 5460 Phase Type: Square
Resolution: 200 Window width: 1000 # of Exp. :50 Gain: 80 Spec: 5380
8 Rel. intensity
Figure 4.7: Hg I 546.1 nm
Frequency: 50
6
Laborant: J. Linden Date: 6/11/2005 Time: 18:26
4
Background: Yes
2
0
0
0.2
0.4
0.6
0.8
1 One period (microsec.)
1.2
1.4
1.6
1.8
2 4
x 10
33
34
4
7
Line: Kr I 7601
x 10
Fibre 1 Fibre 2 Fibre 3 Fibre 4 Fibre 5 Fibre 6
6
Lamp: 36W T8 Kr 85%, Ar 15% 5 Line: Kr I 7601 Phase Type: Square
Resolution: 200 Window width: 1000 # of Exp. :50 Gain: 0
4 Rel. intensity
Figure 4.8: Kr I 760.2 nm
Frequency: 50
3 Spec: 7527 Laborant: J. Linden
Time: 17:29
CHAPTER 4.
Date: 6/11/2005 2
Background: Yes
1
0
0.2
0.4
0.6
0.8
1 One period (microsec.)
1.2
1.4
1.6
1.8
2 4
x 10
APPLICATIONS
0
4.2.
4
Line: Kr I 7694
x 10
EXPERIMENTS AND RESULTS
9
8
7 Lamp: 36W T8 Kr 85%, Ar 15% Fibre 1 Fibre 2 Fibre 3 Fibre 4 Fibre 5 Fibre 6
Line: Kr I 7694 6
Phase Type: Square Frequency: 50
5 Window width: 1000 # of Exp. :50 Gain: 80 Spec: 7621 Laborant: J. Linden
Rel. intensity
Figure 4.9: Kr I 769.5 nm
Resolution: 200
4
3
Date: 6/11/2005 Time: 16:42 Background: Yes
2
1
0
−1
0
0.2
0.4
0.6
0.8
1 One period (microsec.)
1.2
1.4
1.6
1.8
2 4
x 10
35
36
4
16
Line: Ar I 6965
x 10
Fibre 1 Fibre 2 Fibre 3 Fibre 4 Fibre 5 Fibre 6
14
12 Lamp: 36W T8 Kr 85%, Ar 15% Line: Ar I 6965 Phase Type: Square
10
Resolution: 200 Window width: 1000 # of Exp. :80 Gain: 200 Spec: 6890
Rel. intensity
Figure 4.10: Ar I 696.5 nm
Frequency: 50
8
6
Laborant: J. Linden Date: 6/11/2005 4
CHAPTER 4.
Time: 16:07 Background: Yes 2
−2
0
0.2
0.4
0.6
0.8
1 One period (microsec.)
1.2
1.4
1.6
1.8
2 4
x 10
APPLICATIONS
0
4.2.
4
Line: Ar I 7067
x 10
EXPERIMENTS AND RESULTS
14
Fibre 1 Fibre 2 Fibre 3 Fibre 4 Fibre 5 Fibre 6
12
Lamp: 36W T8 Kr 85%, Ar 15% Line: Ar I 7067
10
Phase Type: Square
Resolution: 200 8 Window width: 1000 # of Exp. :80 Gain: 200 Spec: 6992
Rel. intensity
Figure 4.11: Ar I 706.7 nm
Frequency: 50
6
Laborant: J. Linden Date: 6/11/2005 Time: 12:00 4 Background: Yes
2
0
0
0.2
0.4
0.6
0.8
1 One period (microsec.)
1.2
1.4
1.6
1.8
2 4
x 10
37
38
CHAPTER 4.
APPLICATIONS
Through all the spe tra, one noti es a onsistent hara teristi of spikes, espe ially in the
athode phase. This is probably due to instability in the plasma. Sin e the measurements are done during a large number of dierent periods, a ondition to get a fair signal is that the plasma behaves identi al in every period. The reason for the instability is probably that the lamp was not operated under ideal onditions. For this, a stronger ampli er or a smaller lamp is required. The ambition was to demonstrate the method using smaller lamps. Unfortunately these were not available at the time of the experiments. However, the plasma be ame more stable if the tube was driven by a trun ated sine wave. Some experiments were done under this ondition. Figure 4.12 shows the spe trum of Hg I 404 nm (the same transition as in Figure 4.5). In this ase one more learly may observe the anode os illations. Only ber no. 1, 3 and 6 are shown. A spiky hara teristi indi ates still some plasma instability in the athode phase.
4.2.
Line: Hg I 4046
x 10
EXPERIMENTS AND RESULTS
Fibre 1 Fibre 3 Fibre 6 14
Lamp: 36W T8 Kr 85%, Ar 15%
12
Line: Hg I 4046 Phase Type: Trunkated Sine Frequency: 50 10 Resolution: 200 Window width: 1000 # of Exp. :50 Gain: 80
Rel. intensity
Figure 4.12: Hg I 404.7 when lamp operated on trun ated sine. Only bre 1, 3 and 6 are plotted.
4
16
8
Spec: 3961 Laborant: J. Linden Date: 6/12/2005
6
Time: 12:40 Background: Yes 4
2
0
0
0.2
0.4
0.6
0.8
1 One period (microsec.)
1.2
1.4
1.6
1.8
2 4
x 10
39
40 4.3
CHAPTER 4.
APPLICATIONS
Dis ussion
Even though the experiments mainly were done to demonstrate the method, some of the results may be worthy of dis ussion, e.g. the anode os illations. Due to the a
elerating eld, the ele trons traveling towards the anode produ e an ex ess in ionization whi h in rease the plasma density in the sheath around the anode. The plasma density be omes high enough for the anode to olle t the ne essary ele trons without a positive anode fall, and the a
elerating eld disappears. The ex ess in plasma density stops, the ions diuse away, and the anode fall will begin to in rease on e again. When the anode fall rea h the ionization potential of mer ury, the ionization builds up very abruptly, and the anode fall again ollapses quite suddenly. This os illation in anode fall may be seen in our spe trum as anode os illations (Waymouth, p. 110, 1971). In Figure 4.12 the frequen y of the os illations an be estimated to 750 Hz. A loser look at the os illations from one bre to another, reveal a displa ement in time. The anode fall takes some time to build up. It starts at the anode and travels towards the positive olumn with a speed of 30 m/s, estimated from the displa ement between the anode os illation of bre no. 3 and no. 6 in Figure 4.12. These bres are separated with 1.5 m and the displa ement is about 400 to 500 µs. In the beginning of the athode phase the is also a high initiating peak of ionization. This does not lead to any os illation though, but stabilizes to an equilibrium. The anode fall never rea hes high above the ionization potential of mer ury, and therefore none of the buer gases are ionized. In the athode phase however, the athode fall rea hes su iently high potential, making it possible for the ele trons to ionize even the buer gases.
Chapter 5 Con lusions and Future Appli ations The main goal with this diploma work was to onstru t an experimental setup for analyzing uores ent tubes using time-resolved spe tros opy. The method has then been demonstrated on some of the strongest emission lines in the visible spe tral region from mer ury, argon and krypton. These lines were emitted from the plasma in an Aura Light 36 W T8 uores ent tube, driven mainly by a bipolar square ex itation wave with a frequen y of 50 Hz. With the use of opti al bres it has been possible to investigate several dierent positions at the same time, hen e the method also is spatially resolved. 5.1
Con lusions
The method is suitable to get an overall understanding of the behavior of dierent spe tral lines, at dierent position of the plasma and under dierent ex itation onditions. Not only does this gives an idea of how dierent elements are ee ted by dierent ex itation
onditions, but also how the plasma is omposed throughout the tube. From the demonstration experiments in this report it has been possible to al ulate the frequen y of the anode os illation, whi h is a pro ess onne ted to the diusion rate. The strongest emission lines from a plasma are not ne essarily the most interesting ones. Be ause uores ent lamps, when operated under normal or almost normal onditions, are relatively weak light sour es, to weak a signal is probably the strongest limitation for this method. A good time resolution requires short snapshots, and the shorter the snapshot the greater number of snapshots have to be made to get a reasonable signal. This involve a longer a quisition time. 5.2
Future Appli ations
The method developed in this diploma work has been applied on low frequen y operated lamps. A development is to make it suitable also for high frequen y operation, or arbitrary 41
42
CHAPTER 5.
CONCLUSIONS AND FUTURE APPLICATIONS
generated ex itation pulses. At the time of writing, this has partly been done by diploma student Joel Clementson. One line of development is to investigate a wider range of the ele tromagneti spe trum, espe ially the UV region ontaining the resonan e mer ury lines of 185 nm end 254 nm. This requires a dierent spe trograph and also tubes in luding UV transparent windows of quarts or magnesium uoride. One ambition is to develop this method, making it suitable to investigate the emitting material from the ele trodes. Hopefully this will involve a higher understanding of the wear of the ele trodes, and hen e in the future make a ontribution to the development of more e ient uores ent lighting. A knowledgements
I would like to thank my supervisors Sven Huldt and Roger Hutton, for their help, great inspiration and for sharing their great knowledge and experien e of plasma spe tros opy in general and of uores ent lighting in parti ular. I thank Sven for my always being able to ask all sorts of questions and Roger who inspired me to take part in this subje t in the rst pla e. I would also like to thank PhD student Kasper Öberg for his en ouragement, suggestions, kindness and for showing great interest. Last I would like to thank Joel Clementson, my olleague and another diploma student with whom I had the opportunity to work together with in the laboratory, for his en ouragement, feedba k and good times in the lab.
Appendix A Manual This manual des ribes the way to a hieve the type of time-resolved spe trum presented in this report. The method may be used on dierent lamps, supplied with dierent frequen ies and dierent ex itation phases like sine, square or sawtooth. One of the future appli ations is to modify this method to also look at arbitrary generated ex itation pulses, but sin e the method des ribed here only apply to no- ompli ated ex itation phases generated by a pulse generator, we all this appli ation the Simple Wave appli ation. Hen e this is the Simple Wave Manual. The manual des ribes the setup of equipment in room B148, department of atomi astrophysi s at Lunds observatory. To use the Simple Wave appli ation, it is required that the MGP is programmed with the right program (latest version named T6.hex). It is re ommended, to get a fair result, to keep the lamp running fore at least 20 min before you make an a quisition. Equipment • Make sure the equipment are onne ted as in Figure 3.7. • Swit h on the power supply to the MGP, pulse generator and ampli er. • On the pulse generator, hoose frequen y, amplitude and phase type. Make sure to
press the 'DISABLE' button.
• Turn on the lamp.
Andor
(Version 2.62.I2C Andor MCD)
• Start Andor. • Start by setting the temperature to -20◦ C by li king on the temperature bar down
in the right orner.
43
44
APPENDIX A. MANUAL
• In 'Hardware' menu, hoose 'Shutter Control'. Sele t 'Fully Auto' and 'TTL Low'.
Swit h on the me hani al shutter ontrol box (Photometri s) on top of the spe trograph.
• Cli k on the 'Setup A quisition' button. Sele t 'Multi-Tra k' and 'Multi-Tra k Setup'.
Cli k 'Custom' and 'Re all Settings'. In the 'Open' Dialog Box, sele t 'Multi-tra k-setup050207.mtr' (F:/Tidsspektroskopi/SimpleWave/AndorS ript/Multi-tra k-setup050207.mtr). Cli k 'Open' and 'OK'.
Find a Line • In the 'Setup A quisition' Dialog Box, hoose 'Real Time' and Trigger mode 'Internal'.
Cli k on the 'Setup Gate Control' button. Sele t 'Fire only'. Cli k 'OK' and in the 'Setup A quisition' Dialog Box also li k 'OK'. If the temperature is down to -20◦ C, you are now ready to look for a line. Cli k on the 'Take Signal' button (green ir le).
• To get a good view of the signal, sele t 'Image' and 'Change Palette' in the tool bar. If
you do not get any signal, abort a quisition with the 'Abort A quisition' button (red
ir le), sele t 'Setup A quisition' and hange the 'Shutter Time'. You may also have to hange the 'Gain' in the 'Setup Gate Control' Dialog Box.
• Now you an turn the handle on the spe trograph to fo us on a desirable line.
Make an A quisition • Use the handle on the spe trograph to adjust the line you want to investigate to the
enter of the Display Window, i.e. pixel no. 512.
• To run program, sele t 'Run Program by Filename' in the 'File' menu. In the 'Open'
Dialog Box, sele t the le with the ontents you want to run (F:/Tidsspektroskopi/SimpleWave/AndorS ript/SimpleWave.pgm).
• Make sure the pulser onne tor is onne ted to the dete tor head.
• Enter a quisition data. When this is done the a quisition will start automati ally.
• When the a quisition is done, remove the pulser onne tor from the dete tor head.
MATLAB
(MATLAB 7 Version 7.0.0.19920(R14))
• Start MATLAB.
• Make sure you're in the orre t urrent dire tory, e.g.
'F:/Tidsspektroskopi/SimpleWave/MatlabS ripts'.
• In the Command Window, write 'getdata', whi h is the name of the s ript that loads
data.
45 • In the displayed plot, alled Figure 1, you are able to zoom and sele t the region
ontaining the line of interest and whi h you would like to al ulate the intensity for. If the line was entered about pixel no. 512 in Andor, your interval might be about 485 to 540. Remember these start and stop values.
• Now you an lose Figure 1.
• In the Command Window, now write 'showdata', whi h is the name of the s ript that
displays your time-resolved spe tra.
• The s ript rst ask for the start- and stop values you sele ted from the previous gure. • It then ask for the bre numbers you would like to look at. If you'd like to see all
bres, just press enter.
• A new gure is displayed. It looks better if you maximize it.
• Down in the left orner, a string array with a quisition data is shown. Probably this
doesn't look very ni e. In the Figure Tool bar, sele t the 'Edit Plot' button. Now you an move both the spe trum and the data array around. Tip: put the spe trum slightly to the right and the data array to the left.
• You are able to zoom in the spe trum.
• To save gure, hoose 'Save' in the 'File' menu.
• To plot the diagram with a printer, sele t 'Print Setup' in the 'File' menu. Choose
'Lands ape' and li k 'OK'. Then hoose 'Print' in the 'File' menu.
• If you want to save all sour e data, just opy them to a safe pla e. Tip: also save
the log.dat le. It is not ne essary to save all the other .dat les, unless you plan to generated further spe tra in MATLAB of the same data.
Appendix B Andor The Andor measurement sequen e s ript Simple Wave used for this experiment. rem rem rem rem rem
Created by Johannes Lindén & Sven Huldt Jan, Feb, Mar 2005 Modified for ARM-gater 2005-05-17 SH Modified further 2005-05-24 SH Modified further 2005-06-02 JL
ls() root$ = "F:\Tidsspektroskopi\SimpleWave\saved_data\" q = kill(root$;"log.dat") rem ******** SetTemperature ******** b = -20 path$ = root$;"log.dat" write(path$, "SetTemperature: ";b) SetTemperature(b) rem ******************************** rem ****** SetA quisitionMode ****** b = 1 :rem 1 = signel s an path$ = root$;"log.dat" write(path$, "SetA quisitionMode: ";b;" (1 = signel s an)") SetA quisitionMode(b) rem ******************************** rem ******** SetTriggerMode ******** b = 0 :rem 0 = Internal, 1 = External path$ = root$;"log.dat" write(path$, "SetTriggerMode: ";b;" (0 = Internal, 1 = External)") SetTriggerMode(b) rem ******************************** rem ********** SetGateMode ********* b = 0 :rem 0 = Gate AND fire, 1 = Fire only, 2 = Gate only path$ = root$;"log.dat" write(path$,"SetGateMode: ";b;" (0 = Gate AND fire, 1 = Fire only, 2 = Gate only)") SetGateMode(b) rem ******************************** rem ********** SetShutter ********* b = 2 :rem (1,0) = Permanently open & TTL low, (2,0) = Fully Auto path$ = root$;"log.dat" write(path$,"SetShutter: ";b;" (1 = Permanently open & TTL low, 2 = Fully Auto)") SetShutter(2,0) rem ********************************
46
47
rem ******************** Type of fluores ent tube ********************** p = read(root$;"lamp.dat", rlamp$) path$ = root$;"lamp.dat" if p == 0 then input ("Type of fluores ent tube (e.g. gas mixture): (still '";rlamp$;"' press enter)", lamp$) if (len(lamp$)>0) then q = kill(root$;"lamp.dat") if (q < 0)then print("Failed to delete!") endif write(path$, lamp$ ) else lamp$ = rlamp$ endif else input ("Type of fluores ent tube (e.g. gas mixture): ", lamp$) write(path$, lamp$ ) endif print ("Type of Fluores ent tube: ";lamp$) path$ = root$;"log.dat" write(path$, "Type of fluores ent tube: ";lamp$) rem ******************************************************************** rem ******************************** Who ******************************* p = read(root$;"who.dat", rwho$) path$ = root$;"who.dat" if p == 0 then input ("Enter Laborant: (still '";rwho$;"' press enter)", who$) if (len(who$)>0) then q = kill(root$;"who.dat") if (q < 0)then print("Failed to delete!") endif write(path$, who$ ) else who$ = rwho$ endif else input ("Enter Laborant: ", who$) write(path$, who$ ) endif print("Laborant: ";who$) path$ = root$;"log.dat" write(path$, "Laborant: ";who$) rem ******************************************************************** rem *************************** Phase Type ***************************** p = read(root$;"pt.dat", rpt$) path$ = root$;"pt.dat" if p == 0 then input ("Enter Phase Type (e.g. sine, square or tir): (still '";rpt$;"' press enter)", pt$) if (len(pt$)>0) then q = kill(root$;"pt.dat") if (q < 0) then print("Failed to delete!") endif write(path$, pt$ ) else pt$ = rpt$ endif else input ("Enter Phase Type (e.g. sine, square or tri): ", pt$) write(path$, pt$ ) endif print("Phase Type: ";pt$) path$ = root$;"log.dat" write(path$, "Phase Type: ";pt$)
48
APPENDIX B. ANDOR
rem ******************************************************************** rem ********************* Wavelength on spe trograph ******************* p = read(root$;"spe wl.dat", rspe wl$) path$ = root$;"spe wl.dat" if p == 0 then input ("Wavelength on spe trograph: (still '";rspe wl$;"' press enter)", spe wl$) if (len(spe wl$)>0) then q = kill(root$;"spe wl.dat") if (q < 0)then print("Failed to delete!") endif write(path$, spe wl$ ) else spe wl$ = rspe wl$ endif else input ("Wavelength on spe trograph: ", spe wl$) write(path$, spe wl$ ) endif print("Wavelenght on spe trograph: ";spe wl$) path$ = root$;"log.dat" write(path$, "Wavelenght on spe trograph: ";spe wl$) rem ******************************************************************** rem *************************** Line *********************************** p = read(root$;"line.dat", rline$) path$ = root$;"line.dat" if p == 0 then input ("Line (e.g. Hg I 4360 or 'unknown'): (still '";rline$;"' press enter)", line$) if (len(line$)>0) then q = kill(root$;"line.dat") if (q < 0)then print("Failed to delete!") endif write(path$, line$ ) else line$ = rline$ endif else input ("Line (e.g. Hg I 4360Å or 'unknown'): ", line$) write(path$, line$ ) endif print("Line: ";line$) path$ = root$;"log.dat" write(path$, "Line: ";line$) rem ******************************************************************** rem **************************** Gain ********************************** p = read(root$;"gain.dat", rgain$) path$ = root$;"gain.dat" if p == 0 then input ("Enter Gain (0-255): (still '";rgain$;"' press enter)", gain$) if (len(gain$)>0) then q = kill(root$;"gain.dat") if (q < 0)then print("Failed to delete!") endif write(path$, gain$ ) else gain$ = rgain$ endif else input ("Enter Gain: ", gain$) write(path$, gain$ ) endif Print("Gain: ";gain$) path$ = root$;"log.dat"
49 write(path$,"Gain: ";gain$) gain = val(gain$) SetGain(gain) rem ******************************************************************** rem ***************************** Frequen y **************************** p = read(root$;"freq.dat", rfreq$) path$ = root$;"freq.dat" if p == 0 then input ("Enter ex itation frequen y (Hz): (still '";rfreq$;"' press enter)", freq$) if (len(freq$)>0) then q = kill(root$;"freq.dat") if (q < 0)then print("Failed to delete!") endif write(path$, freq$ ) else freq$ = rfreq$ endif else input ("Enter Frequen y (Hz): ", freq$) write(path$, freq$ ) endif print("Frequen y: ";freq$) path$ = root$;"log.dat" write(path$, "Frequen y: ";freq$) freq = val(freq$) rem ******************************************************************** T = 10000000/freq print("Period time (unit 100ns): ";T) rem *************************Start time********************************** print("Starttime zero orresponds to rising edge of trigger pulse!") p = read(root$;"delstart.dat", rdelstart$) path$ = root$;"delstart.dat" if p == 0 then input ("Enter starttime (unit 100ns): (still '";rdelstart$;"' press enter)", delstart$) if (len(delstart$)>0) then q = kill(root$;"delstart.dat") if (q < 0)then print("Failed to delete!") endif write(path$, delstart$ ) else delstart$ = rdelstart$ endif else input ("Enter starttime (unit 100ns): ", delstart$) write(path$, delstart$) endif print("Start time (unit 100ns): ";delstart$) delstart = val(delstart$) rem ******************************************************************** rem ************************ Resolution ******************************** p = read(root$;"res.dat", rres$) path$ = root$;"res.dat" if p == 0 then input ("Enter Resolution: (still '";rres$;"' press enter)", res$) if (len(res$)>0) then q = kill(root$;"res.dat") if (q < 0)then print("Failed to delete!") endif write(path$, res$ ) else res$ = rres$
50
APPENDIX B. ANDOR
endif else input ("Enter Resolution: ", res$) write(path$, res$ ) endif Print("Resolution: ";res$) path$ = root$;"log.dat" write(path$,"Resolution: ";res$) nw=val(res$) rem ******************************************************************** stp = T/nw print("Seperation Time for Time Windows (unit: 100ns): ";stp) rem ************************ Window Width ****************************** p = read(root$;"ww.dat", rww$) path$ = root$;"ww.dat" if p == 0 then input ("Time Window Width (unit 100ns): (still '";rww$;"' press enter)", ww$) if (len(ww$)>0) then q = kill(root$;"ww.dat") if (q < 0)then print("Failed to delete!") endif write(path$, ww$ ) else ww$ = rww$ endif else input ("Time Window Width (unit: 100ns): ", ww$) write(path$, ww$ ) endif print("Window Width: ";ww$) ww = val(ww$) path$ = root$;"log.dat" write(path$,"Time Window Width (unit: 100ns): ";ww) rem ******************************************************************** rem ************************ # of Exposures **************************** p = read(root$;"nexp.dat", rnexp$) path$ = root$;"nexp.dat" if p == 0 then input ("Number of exposures: (still '";rnexp$;"' press enter)", nexp$) if (len(nexp$)>0) then q = kill(root$;"nexp.dat") if (q < 0)then print("Failed to delete!") endif write(path$, nexp$ ) else nexp$ = rnexp$ endif else input ("Number of exposures: ", nexp$) write(path$, nexp$ ) endif print("Number of exposures: ";nexp$) nexp = val(nexp$) path$ = root$;"log.dat" write(path$,"Number of windows: ";nexp) rem ******************************************************************** ppar = 1 print("Number of periods to skip between pulses: ";ppar)
51 baud(3,38400) :rem Set up om1 to 4800 baud. handshake(1,0) :rem Handshaking turned off. newline(2) rem ppar pulses to skip
omwrite (3,"sp";ppar) print("To gater: sp ";ppar)
omread (3,a$)
omwrite (3,"rp")
omread (3,a$) print("The gater returns (ppar):";a$) rem nexp exposures
omwrite (3,"sn";nexp) print("To gater: sn ";nexp)
omread (3,a$)
omwrite (3,"rn")
omread (3,a$) print("The gater returns (nexp):";a$) rem delstart, when to start (delay)
omwrite (3,"sg ";delstart) print("To gater: sg ";delstart)
omread (3,a$)
omwrite (3, "rg")
omread (3,a$) print("The gater returns (delstart): ";a$) rem ww windowwidth
omwrite (3,"sl ";ww$) print("To gater: sl ";ww$)
omread (3,a$)
omwrite (3, "rl")
omread (3,a$) print("The gater returns (ww): ";a$) rem ********************** Shutter Time ******************************** exptime = (3*T + nexp*(ppar + 4)*T) print("Shutter Time in 100ns = ";exptime;) exptime = exptime/10000000 :rem from 100ns-unit to sek-unit print("Shutter Time in se s = ";exptime;) SetExposureTime(exptime) SetShutterTransfertime(0.01) rem ******************************************************************** rem **************************** Comments ********************************** input ("Comments:", om$) if (len( om$) > 0) then q = kill(root$;" omment.dat") path$ = root$;" omment.dat" write(path$, om$ ) else p = read(root$;" omment.dat", r om$) if p == 0 then k = key("Use old omments: '";r om$;"'? If yes, press 'y'. If no, press enter") if k != 'y' then q = kill(root$;" omment.dat") print(q) if (q < 0) then print("Failed to delete Comment!") endif endif endif endif
52
k = 0 path$ = root$;"log.dat" write(path$,"Comments: "; om$) rem ************************************************************************ rem ********************** Ba kground ********************************** p = read(root$;"bg.dat", rbg$ ) if p == 0 then rbg = val(rbg$) while ( k != 'n' ) and ( k != 'y' ) and ( k != 13) if rbg == 121 then k = key("Ba kground orre tion? y/n: (still 'yes' press enter)") endif if rbg == 110 then k = key("Ba kground orre tion? y/n: (still 'no' press enter)") endif wend if k == 13 then k = rbg endif else while ( k != 'n' ) and ( k != 'y' ) k = key("Ba kground orre tion? y/n:") wend endif if (k == 121) then q = kill(root$;"bg.dat") if (q < 0) then print("Failed to delete!") endif path$ = root$;"bg.dat" write(path$, 121) :rem save y ba kground print ("Ba kground: Yes") print("Hit any key to take ba kground.") key() SetA quisitionType(1) :rem 0 = signal, 1 = Ba kground run() SetA quisitionType(0) :rem 0 = signal, 1 = Ba kground path$ = root$;"log.dat" write(path$,"Ba kground orre tion: Yes") endif if (k == 110) then q = kill(root$;"bg.dat") if (q < 0) then print("Failed to delete!") endif path$ = root$;"bg.dat" write(path$, 110) :rem save n ba kground print ("Ba kground: No") rem SetA quisitionType(0) :rem 0 = signal, 1 = Ba kground rem SetDataType(1) :rem 1 = Counts, 2 = Counts(Bg orre ted) path$ = root$;"log.dat" write(path$,"Ba kground orre tion: No") endif rem ************************************************************************ print("Hit any key to start loop.") key() rem *************************** Date & Time **************************** d$ = date$() q = kill(root$;"date.dat") path$ = root$;"date.dat" write(path$, d$ ) :rem save date print("Todays date is: ";d$)
APPENDIX B. ANDOR
53 path$ = root$;"log.dat" write(path$,"Date: ";d$) t$ = time$() t$ = left$(t$,5) q = kill(root$;"time.dat") path$ = root$;"time.dat" write(path$, t$ ) :rem save time print("The time is: ";t$) path$ = root$;"log.dat" write(path$,"Time: ";t$) rem ******************************************************************** if (k == 121) then SetDataType(2) :rem Counts (Ba kground orre ted) endif if (k == 110) then SetDataType(1) :rem Counts endif d = delstart for b = 1 to nw + 1 step 1 print("********************************") print("This is loop ";b;" of ";nw + 1;" in total.") print("Delay is now: ";d) print("This is now sent to the MGP: sg";str$(d))
omwrite (3,"sg";str$(d);)
omread (3,a$) :rem "*" rem print("The MGP returns 1(*):";a$)
omwrite (3,"rg")
omread (3,a$) print("The MGP onfirms:";a$)
omwrite (3,"go") run()
hangedisplay(#0,2)
omread (3,a$) :rem "dly-bort" rem print("The MGP returns 3(*):";a$)
omread (3,a$) :rem "=" rem print("The MGP returns 4(*):";a$)
omread (3,a$) :rem "#" print("The 'GO' ommand was sent.") fnamn$ = root$;"X";spe wl$;"_";str$(d);".as " print ("Now saving file as ";fnamn$) saveas iiXY (#0,fnamn$,2) d = d + stp next print("End of loop")
Appendix C MATLAB In this appendix the MATLAB s ripts are shown. These are getdata.m, showdata.m,
findzerolevel.m and Zmatrix.m.
getdata.m This s ript reads in data saved away by the Andor s ript. % getdata.m % Created by Johannes Lindén % June 2005
lear all
l root = 'F:\Tidsspektroskopi\SimpleWave\saved_data\'; files = dir ([root, '*.as '℄); nof = length(files); % Sort filename by gate time for i = 1:nof filename = files(i).name; for j = 1:100 if (filename(j) == '_')break; end end
end
for k=j:100 if (filename(k) == '.')break; end end filorder(i,1) = i; filorder(i,2) = str2num(filename(j+1:k-1));
filorder = sortrows(filorder,2); n = 7; filename = files(1).name; filename = filename(1:(length(filename)-4)); % remove '.as ' file(:,:,n) = load(['F:\Tidsspektroskopi\SimpleWave\saved_data\',filename,'.as '℄,' -as ii'); for i = 1:nof filename = files(filorder(i,1)).name;
54
55 allfiles(:, :, i) = load([root, filename℄, '-as ii'); end % Sort done [dim1 dim2 dim3℄ = size(allfiles); allfiles = allfiles(:,2:end,:); % remove fist olumns whi h onsists of index 1 to 1024. [dim1 dim2 dim3℄ = size(allfiles); for i = (1:dim1); sumallfiles(i) = sum(sum(allfiles(i,1:dim2,1:dim3))); end figure(1) plot(sumallfiles); grid on; zoom on; title('Sum of all times and all position') disp('Zoom in the plot and sele t start and stop value and run "showdata"')
showdata.m This s ript displays the time-resolved spe tra. % Created by Johannes Lindén % June 2005 start = input('Insert start value: '); stop = input('Insert stop value: '); for i = 1:dim2 X = allfiles(:,i,:); X = shiftdim(X,2); X = X'; medel = round(mean(mean(allfiles(:,i,:)))); y = findzerolevel(X,medel); for j = 1:dim3 Y = allfiles(:,i,j); Y = Y - y; I(j,i) = sum(Y(start:stop)); end end
% loops through ea h fiber % takes out a 2D matrix from 'allfiles' % adjust the dimensions % % % % % %
evaluate meanvalue for all time windows of fiber 'i' and rounds of
all the fun tion 'findzerolevel.m' whi h returns an y value loops trough ea h time window pi k out time window j in fiber i subtra t the y value evaluate the sum/intensity in the desirable interval
root = 'F:\Tidsspektroskopi\SimpleWave\saved_data\'; s = (['Lamp: ',fread(fopen([root,'lamp.dat'℄),'uint8=> har')'℄); % read lamp data s = s(1:length(s)-2); % remove the two last signs (spa e and enter) S{1,1} = s; % adds text in ell-array S s = (['Line: ',fread(fopen([root,'line.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{3,1} = s; s = (['Phase Type: ',fread(fopen([root,'pt.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{5,1} = s; s = (['Frequen y: ',fread(fopen([root,'freq.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{7,1} = s; s = (['Resolution: ',fread(fopen([root,'res.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{9,1} = s; s = (['Window width: ',fread(fopen([root,'ww.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{11,1} = s; s = (['# of Exp. :',fread(fopen([root,'nexp.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{13,1} = s; s = (['Gain: ',fread(fopen([root,'gain.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{15,1} = s;
56
APPENDIX C. MATLAB
s = (['Spe : ',fread(fopen([root,'spe wl.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{17,1} = s; s = (['Laborant: ',fread(fopen([root,'who.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{19,1} = s; s = (['Date: ',fread(fopen([root,'date.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{21,1} = s; s = (['Time: ',fread(fopen([root,'time.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); S{23,1} = s; s = (fread(fopen('F:\Tidsspektroskopi\SimpleWave\saved_data\bg.dat'),'uint8=> har')'); s = s(1:length(s)-2); s = str2num(s); if s == 121 s = 'Ba kground: Yes'; end if s == 110 s = 'Ba kground: No'; end S{25,1}=s; f lose('all'); F = (fread(fopen([root,'freq.dat'℄),'uint8=> har')); freq = str2num(F'); T = round(1000000/freq); t = linspa e(0,T,size(I,1)); fibres = input... ('Sele t whi h fibres you would like to plot, fibres = str2num(fibres); fibres = sort(fibres); if length(fibres) == 1 figure(10) grid on; zoom on; hold on; swit h fibres
ase {1} plot(t,I(:,1),'Color', [0 0 1℄); %
ase {2} plot(t,I(:,2),'Color', [0 .5 0℄); %
ase {3} plot(t,I(:,3),'Color', [1 0 0℄); %
ase {4} plot(t,I(:,4),'Color', [0 .7 .7℄); %
ase {5} plot(t,I(:,5),'Color', [.7 0 .7℄); %
ase {6} plot(t,I(:,6),'Color', [.7 .7 0℄); % otherwise disp('Invalid input') end legA = (['Fibre ',num2str(fibres(1))℄); legend(legA) elseif length(fibres) == 2 figure(10) grid on; zoom on; hold on; for i = 1:2 swit h fibres(i)
ase {1} plot(t,I(:,1),'Color', [0 0 1℄);
ase {2} plot(t,I(:,2),'Color', [0 .5 0℄);
ase {3} plot(t,I(:,3),'Color', [1 0 0℄);
e.g "1:3" or "1 3 6" (If all, press enter): ','s');
blue green red
yan magenta yellow
57
ase {4} plot(t,I(:,4),'Color', [0 .7 .7℄);
ase {5} plot(t,I(:,5),'Color', [.7 0 .7℄);
ase {6} plot(t,I(:,6),'Color', [.7 .7 0℄); otherwise disp('Invalid input')
end end legA = (['Fibre ',num2str(fibres(1))℄); legB = (['Fibre ',num2str(fibres(2))℄); legend(legA, legB) elseif length(fibres) == 3 figure(10) grid on; zoom on; hold on; for i = 1:3 swit h fibres(i)
ase {1} plot(t,I(:,1),'Color', [0 0 1℄);
ase {2} plot(t,I(:,2),'Color', [0 .5 0℄);
ase {3} plot(t,I(:,3),'Color', [1 0 0℄);
ase {4} plot(t,I(:,4),'Color', [0 .7 .7℄);
ase {5} plot(t,I(:,5),'Color', [.7 0 .7℄);
ase {6} plot(t,I(:,6),'Color', [.7 .7 0℄); otherwise disp('Invalid input') end end legA = (['Fibre ',num2str(fibres(1))℄); legB = (['Fibre ',num2str(fibres(2))℄); legC = (['Fibre ',num2str(fibres(3))℄); legend(legA, legB, legC) elseif length(fibres) == 4 figure(10) grid on; zoom on; hold on; for i = 1:4 swit h fibres(i)
ase {1} plot(t,I(:,1),'Color', [0 0 1℄);
ase {2} plot(t,I(:,2),'Color', [0 .5 0℄);
ase {3} plot(t,I(:,3),'Color', [1 0 0℄);
ase {4} plot(t,I(:,4),'Color', [0 .7 .7℄);
ase {5} plot(t,I(:,5),'Color', [.7 0 .7℄);
ase {6} plot(t,I(:,6),'Color', [.7 .7 0℄); otherwise disp('Invalid input') end end legA = (['Fibre ',num2str(fibres(1))℄); legB = (['Fibre ',num2str(fibres(2))℄); legC = (['Fibre ',num2str(fibres(3))℄); legD = (['Fibre ',num2str(fibres(4))℄); legend(legA, legB, legC, legD) elseif length(fibres) == 5 figure(10) grid on; zoom on; hold on; for i = 1:5
58
APPENDIX C. MATLAB
swit h fibres(i)
ase {1} plot(t,I(:,1),'Color',
ase {2} plot(t,I(:,2),'Color',
ase {3} plot(t,I(:,3),'Color',
ase {4} plot(t,I(:,4),'Color',
ase {5} plot(t,I(:,5),'Color',
ase {6} plot(t,I(:,6),'Color', otherwise disp('Invalid input') end
[0 0 1℄); [0 .5 0℄); [1 0 0℄); [0 .7 .7℄); [.7 0 .7℄); [.7 .7 0℄);
end legA = (['Fibre ',num2str(fibres(1))℄); legB = (['Fibre ',num2str(fibres(2))℄); legC = (['Fibre ',num2str(fibres(3))℄); legD = (['Fibre ',num2str(fibres(4))℄); legE = (['Fibre ',num2str(fibres(5))℄); legend(legA, legB, legC, legD, legE) elseif length(fibres) == 0 figure(10) grid on; zoom on; hold on; plot(t,I) legend('Fibre 1','Fibre 2','Fibre 3','Fibre 4','Fibre 5','Fibre 6'); else disp('Invalid input') end s = (['Line: ',fread(fopen([root,'line.dat'℄),'uint8=> har')'℄); s = s(1:length(s)-2); title(s); xlabel('One period (mi rose .)') ylabel('Rel. intensity') text(0,0,S) fid = fopen('F:\Tidsspektroskopi\SimpleWave\saved_data\ omment.dat'); if fid >=0
om = (['Comment: ', har(fread(fid))'℄); gtext( om) end f lose('all');
findzerolevel.m This fun tion is alled by showdata.m, to be able to al ulate the intensity. fun tion [y℄ = findzerolevel(X,medel) %Compute the value whi h is subtra ted from all the time windows for a %spe ifi fibre M = X - medel; [m,n℄ = size(M); z=zeros(m,n); for k = 1:m for l = 1:n if M(k,l) > 0 z(k,1) = 0; else z(k,l) = 1;
59
end
end
end
X = X.*z; y = sum(sum(X))/sum(sum(z)); %Compute mean value
Zmatrix.m This s ript generates the type of 3D-plot seen in Figure 3.6 %Zmatrix %Created by Johannes Lindén %Jan 2005 %Makes a 3D plot within desirable interval of desirable fibre, ones the %getdata-s ript has been exe uted and have generated the allfiles-matrix start = input('Insert start value: '); stop = input('Insert stop value: '); nf = input('In matrix form, insert whi h fibres you want to look at (e.g [1:6℄): '); for f = nf Z = allfiles(start:stop,f,:); Z = shiftdim(Z,2); Z = Z'; figure(f); mesh(Z); title(['Fibre ',num2str(f)℄) rotate3d on end
Bibliography Boley, F. I., Plasma - Laboratory and Cosmi , D. Van Nostrand Company, INC., 1966 Kitsinelis, S. el al, Relative enhan ement of near-UV emission from pulsed low-pressure mer ury dis harge lamp, using a rare gas mixture, J. Phys. D: Appl. Phys. 37 1630-1638, 2004 Lister, G.G et al, The physi s of dis harge lamps, The Ameri an Physi s So iety, 2004 Lopez, J. et al, Time-resolved opti al emission spe tros opy of pulsed DC magnetron sputtering plasmas, J. Phys. D: Appl. Phys. 38 1769-1780, 2005 Marr, G., Plasma Spe tros opy, Elsevier Publishing Company, 1968 Mentel, J., Elektris he Entladungen 1, Ruhr-Universität Bo hum, http://www.aeeo.ruhr-
uni-bo hum.de
Thorne, A., Litzén U., Johansson S., Spe trophysi s, Springer-Verlag Berlin Heidelberg New York, 1999 Proud, J. M., Luessen, L. H., Radiative Pro esses in Dis harge Plasmas, Plenum Press, New York, 1986 Waymouth, J. F., Ele tri Dis harge Lamps, MIT, Cambridge, 1971
60
