Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Light emission from water irradiated with high energy electrons Shaede, Eric Albert 1967

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-UBC_1968_A6_7 S325.pdf [ 10.53MB ]
JSON: 831-1.0062437.json
JSON-LD: 831-1.0062437-ld.json
RDF/XML (Pretty): 831-1.0062437-rdf.xml
RDF/JSON: 831-1.0062437-rdf.json
Turtle: 831-1.0062437-turtle.txt
N-Triples: 831-1.0062437-rdf-ntriples.txt
Original Record: 831-1.0062437-source.json
Full Text

Full Text

LIGHT EMISSION FROM WATEB IRRADIATED WITH HIGH ENERGY ELECTRONS by ERIC ALBERT SHAEDE B»So., University of British Columbia, 1966 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE In the Department of Chemistry We aooept this thesis as conforming to the st^dard. THE UNIVERSITY OF BRITISH COLUMBIA October, 1967 I In presenting this thesis in partial fulfilment of the require-ments for an advanced degree at the University of British Colum-bia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representa-tives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Chemistry The University of British Columbia Vancouver 8, Canada Ootober 11, 1957. Ii ABSTRACT Luminescence has been observed from water Irradiated with an intense pulse of high energy eleotrons. The angular dependence, electron energy dependence, visible speotrum, lifetime and yield of the light emission have been determined. In addition, the effect of additives on the emission waB studied. The emission spectrum of water was found to be identical to that of methanol, oyclohexane and benzene. All of these re-sults lead to the conclusion that no light emission other than Cerenkov radiation was present in the visible region of the speotrum. The yield of Cerenkov radiation was found to be GhV(4-5000K> ~ 6 * ill TABLE OF CONTENTS PAGES INTRODUCTION 1. Radiation Chemistry (a) The Interaction of Ionizing Radiation with matter (i) Electrons (ii) Electromagnetic Radiation 3 - 5 (b) General Radiation Chemistry of the Liquid Phase 5 - 7 (i) Chemical Effeots of the Badlation 5 - 6 (ii) Linear Energy Transfer 6-7 (ill) Absorbed Dose 7 (Iv) G-values 7 (0) Radiation Chemistry of Liquid Water 8 - 9 2. Cerenkov Radiation 3. Previous Investigations of Light Emission from Irradiated Water 14-16 4. Scope of the Present Investigation 16-18 EXPERIMENTAL 1. Electron Accelerator 19 2. Eleotron Beam Current Measurements 20-21 3. Dosimetry 21-22 4. Spectral Analysis Systems 2 3 - 2 9 (a) Photomultipller-Interference Filter Wedge Spectrometer 2 3 - 2 7 (b) Grating Spectrograph 2 7 - 2 9 5« Densitometry ?Q lv PAGES 6. 180° Camera 29-30 7. Irradiation Cell and Water Plow System 30-31 8. Actinometry 31-32 9 . Spectrofluorimetry 32 10. Materials 32-33 (a) Water 32 (b) Scintillator 32 (o) Photographic Materials 33 •(d) Other Materials 33 11. Electrical Noise 33 RESULTS AND DISCUSSIONS 1. Preliminary Investigations 34 2. The Angular Dependence of the Light Emission 35-36 3 . Variation of the Emission Intensity with Electron Energy 37-38 4. The Emission Lifetime 39 5« Actinometry 40 6. The Emission Spectrum 41 (a) Photomultipller-Interference Filter Wedge Spectro-meter 41 (b) Grating Spectrograph 43-44 7. The Effect of Additives 45-46 8. Comparison of Water with other Liquids 47-49 9. Calculation of a G-value 50-51 10. Conclusions 52 11. Suggestions for Further Study 53-5^ REFERENCES ILLUSTRATIONS V LIST OF ILLUSTRATIONS Figure 1: Diagram showing the oonstruotive interference of Cerenkov radiation by Huygens construction. Figure 2* Field Emission Electron Accelerator photograph. Figure 3* (a) Electron beam pulse shape (b) Light emission from water pulse shape (c) Light emission from scintillator pulse shape. Figure Faraday cup and aperture; photograph and diagram. Figure 5t Electron beam current as a funotion of accelerator oharglng voltage. Figure 6* Photomultiplier-interference filter wedge spectro-meter, (a) photograph (b) diagram Figure 7* Experimental setup diagram showing apparatus layout. Figure 8t The spectral sensitivity of the photomultiplier-interf erence filter wedge spectrometer. Figure 9: The grating spectrograph; photograph and diagram. Figure 10$ Spectral sensitivity of the grating spectrograph with HP4 film. Figure llj 100° Camera; photograph and diagram. Figure 12t Stainless Bteel Irradiation cell; photograph and diagram. Figure I3t Angular dependence of the light emission from (a) water, (b) quartz and (o) scintillator, ^ D as a function of Q , VI Figure 14: The peak intensity of the light emission from (a) water and (b) scintillator as a function of charging voltage of the accelerator. Figure 151 The peak intensity of the light emission from (a) water and (b) scintillator divided by the produot of electron beam peak current and energy as a function of charging voltage of the accelerator. (Int./Exl vs oharging voltage). Figure 16* The peak Intensity of the light emission from water divided by the peak electron beam ourrent as a function of 1(1-VQZx? )• Figure I71 The emission speotrum of water as determined with the photomultipller-spectrometer. Figure 18* The emission spectrum of the scintillator as deter-mined by (a) the photomultiplier-spectrometer (b) the spectro-photofluorometer. Figure 19s The emission spectrum of water as determined with the grating spectrograph. Figure 20* The densitometer traoings of the emission bands of (a) water (b) methanol (0) oyclohexane (d) benzene as determined with the grating spectrograph. Vli I wish to thank Dr. D.C. Walker for his assistance both in the laboratory and in the writing of this thesis. His oomments and criticism were of extreme value. I am Indebted to Mr. Dick Espejo of Field Emission Corporation, who provided technical assistance and information on the operation of the accelerator. I am grateful to Dr. A. Bree for use of his microdensltometer, Dr. E.A. Ogryslo for the loan of his grating spectrograph and Dr. G. Porter for the use of his cali-brated quartz-iodine lamp. INTRODUCTION 1. Radiation Chemistry (a) The Interaction of Ionizing Radiation with Matter "Some knowledge of the processes by whloh radiation inter-acts with matter is essential to an understanding of radiation chemioal phenomena, slnoe the ohemical effects are a direct consequence of the absorption of energy from the radiation",* (i) Eleotrons The interaction of eleotrons with matter ooours by three important prooesses, (1) radiation emission (Bremsstrahlung), (11) inelastlo collisions and (ill) elastlo collisions, the latter essentially resulting only in a change of direction of motion of the eleotron. The relative importance of these prooesses depends upon the eleotron energy. As a high speed charged partiole pasBes In the vicinity of a nuoleus it may be decelerated by the electric field and will therefore radiate eleotromagnetio energy (Bremsstrahlung) in order that the system may conserve both energy and momentum. The rate at which this energy loss ooours, -dE/dx, is propor-tional to Z2Z?/m2, where z,Z are the charges on the partiole and nuoleus respectively and m is the mass of the particle. For eleotrons, Bremsstrahlung emission is negligible below 100 lceV. and beoomes predominate In the energy range 10-100 Mev. Energy loss oan also ooour by coulomb interactions with 2 the electrons of the stopping material. Interactions of this type, or inelastic collisions, produce ionization and excitation in the medium and is the dominant form of energy loss by eleotrons having energies less than 1 MeV. The third process of interaction of electrons with matter, elastlo scattering, is a result of the deflection of the elec-tron by the coulomb potentials of the atomio nucleii of the medium. Elastio scattering is important for low energy eleo-trons and high atomio number materials. Thus eleotrons lose their energy and are defected as they pass through a material. The rate of energy loss and the total initial energy consequently determine the range or penetration distanoe of the eleotron in a given material. For monoenergetic eleotrons, a graph showing the number of electrons at a certain distance within the bombarded material as a function of distanoe is nearly linear with a negative slope and finishing in a small tall. The extrapolated or prac-tical range, Hp, is found by extrapolating the near linear portion of the curve. The maximum range, Ho, is the point where the tail of the ourve merges with the background. For nonmonoenergetic beams, or^-rays, the ourve does not have a linear region and only a maximum range, B0, can be deter-mined. An empirloal formula can be used to relate the energy of the eleotronB to their range In aluminum (and most other light elements slnoe the range varies only slightly with atomio number). For energies in the range 0*01 - 2.5 MeV. the maximum 3 range of ^ -rays or the extropolated range for monoenergetlc p eleotrons is given by* 1.265-0.095^1nE Range (mg/omz) = 412 E (1) where E is the maximum energy of the ^-rays or the energy of the monoenergetlc electrons. (ii) Electromagnetic Radiation In contrast to charged particles, which have a definite maximum range in a medium, X or ^-rayB obey absorption laws common to other electromagnetic radiations. Thus X or ^-rays are incompletely absorbed by a finite thickness of absorbing material. The reduction in intensity, dl, of the incident beam of X or y-xays on passing through an increment of matter of thickness, dx, is given by di = Io fx dx ( 1 1 ) where is the linear absorption coefficient of the material and Io is the inoident Intensity of the radiation. The total linear absorption coefficient, jLL , is a sum of the partial coefficients representing the various processes involved in the absorption. The three most Important absorption processes are (i) the photoelectric process, (ii) the Compton effect and (iii) pair production. The relative importance of these pro-cesses is governed by the energy of the photon and the eleotron density of the stopping material. The photoeleotrio effect is the interaction of the electro-magnetic radiation with an atom or molecule whioh results in the oomplete absorption of the photon and simultaneous ejeotion of an electron with kinetic energy, Ee, given by2 Ee = hV - E0 (iii) where Eo is the binding energy of the eleotron and hjX is the energy of the photon. This process is predominant for low photon energies ( <0.5 MeV) and high atomio number materials. The interaction of a photon with an eleotron of the medium by the Compton effect results in a scattering of the photon and transfer of some energy to the eleotron. The energy of the reooil eleotron, Ee, is given by2 Ee = hi/ - hi/ (iv) where h]/ and hj/'are the incident and scattered photon energies respectively. Ee varies from zero to a oertain maximum, and its value depends on the angular relationship of the collision and the scattered photon. The maximum energy of the Comption electron, E max, Is given by^ E max = Eo _ _ (v) 1 + 0.25/Eq where E0 Is the energy of the incident photon. This process is dominant at medium photon energies (1 - lo MeV) and in high eleotron density materials. Pair production is the result of the annihilation of a photon in the region of an atomio nucleus with the concomitant production of an "electron" pair - a positive and a negative eleotron. This prooess has a threshold energy of 1.02 MeV. whioh is required to produce the two electron rest masses. The klnetio energies of the "eleotrons", Ep and Ee, are given by2 Ep + Ee » h}/ - 2m0o2 = hj/ - 1.02 MeV (vi) where hi/ is the photon energy. This prooess beoomes predomi-nate at high photon energies ( > 10 MeV). Henoe when X or "^-rays are absorbed in a medium, high energy eleotrons are generated whioh then dissipate their energy by the prooesses disoussed in seotlon (I). (b) General Radiation Chemistry of the Liquid Phase flV Chemloal Effects of the Radiation The ohemioal effeots following the absorption of radia-tion are a result of the ionizations and eioitations oauBed by the high speed primary eleotron (or Comptbn eleotron in the oase of X or ^-ray irradiations) and its secondary eleot-rons. The path of the primary eleotron is refered to as a track and it is generally linear at high energies but defleo-tions beoome more common as the electron energy deoreaaes. Seoondary ionizations along the traok result in two types of secondary electrons, the low energy seoondary ( < ~ 100 eV.) and the high energy secondary ( > « 100 eV.). The low energy seoondary eleotrons will undergo wide deflections and form a region of tertiary ionization and exoitation, known as a spur, whioh may, to a first approximation, be regarded as spherical. The high energy secondary electron will form a short traok of its own, known as a §-ray, with a number of spurs along it. Eaoh spur oontalns a number of ezoited molecules, positive ions, and eleotrons. Following their formation, whioh takes about 10 Beoonds, these speoies may undergo various trans-formations in the phyBiochemical stage of the radiolysis. These processes, whioh inolude ion-molecule reaotions, elec-tron oapture, dissooiatlon of exoited moleoules, energy transfer and solvation of eleotrons, oocur in a time of about 10"10 seoonds. The chemical stage of the radiolysis occurs during times longer than 10"10 seoonds and oonsists of the formation of moleoular products by radical-radial, radical-ion or ion-ion reactions within the spurs, the formation of products from radioal-solute or radioal-solvent reaotions following the dif-fusion of radicals and moleoular produots from the spurs and the de-exoitation of eleotronioally exoited molecules by fluor-escence, phosphorescence or quenching. (ii) Linear Energy Transfer (L.E.T.) The Bpacial distribution of the spurs formed by the radia-tion determines the yields of some of the produots of the radiolysis. The rate of energy transfer to the medium by the radiation, namely, the linear energy transfer (L.E.T.), deter-mines the spaoial orientation of the spurs. L.E.T. is usually measured in eV./ft and varies with the type of radiation. It is small ( ~ 0.02 eV/8 ) for Co6° y-rays and high energy elec-trons, medium 0.2 eV/fl ) for low energy eleotrons and fi-partioles, and large ( ~ lo eV/fl ) for heavy nuolear partioles Buoh as a-partioles and Li nuclei. L.E.T. also increases as the energy of the partiole decreases down the traok. The different ohemistry resulting from different L.E.T. radiations originates from the different separation of the spurs down the track, since intraspur reaotions are competing with diffu-sion of radicals from the spurs. For high L.E.T. traoks the spurs are nearly overlapping so that molecular product forma-tion is dominant, whereas in 'y-radiolysis for instance, most radioals completely escape from the spurs. (iii) Absorbed Dose In order to get an absolute measure of the yield of any species formed by the radiation, lt is necessary to know how much energy is deposited in the medium by the radiation. The amount of energy deposited is usually known as the absorbed dose and has many units, the most oommon of whioh is the NradN. One rad is defined to be equivalent to the deposition of 100 ergB gm"1 s(i.e. 6.24 x 10*3 eV. gm"1).^ There are a variety of methods used to measure dose (dosimetry) and they vary from ohemioal reactions with known yields to electronic devices capable of estimating the number of particles.^ (iv) G-Values The yield of a speoies formed or destroyed In a radiation-chemical process is known as the "G-value". It is defined as "the number of molecules, ions, atoms, free radicals, etc. which are formed (or disappear) when the system has absorbed 100 eV. of ionizing radiation energy".^ G-values are normally in the range of 0-10 for primary processes, although G-values greater than ten may ooour for ohain reaotions. 8 (o) Radiation Chemistry of Liquid Water^ The radiation chemistry of liquid water and aqueous solu-tions has been the subject of numerous investigations during the past deoade. "The process confidently believed to occur upon irradiation of liquid water are the following ionizing H2O - V V W -radiatlon H2Q+ + H20 H2O* H20+ + e" or H20» H^O+aq + OH -> H + OH energylosses mm polarization to solvent e~aq + HgO thermal of H20 -> H + OH"aq -»e aq (1) (2) (3) W (5) where H20* represents an unspeolfied excited statef.7 These reactions (2) - (5) represent the physioohemioal stage of the radiolysis and lead to the formation of the produots of the radlolysi ® by the chemical stage whioh consists of reactions such as OH + OH ^ H 2 0 2 ( 6 ) H + H >H 2 ( ? ) e"aq + e*"aq >H2 + 20Haq (8) H + O H - >H20 ( 9) H, OH, e~aq + S produots (10) where S represents an unspecified solute present in the system. Reactions (6) - (9) may ooour within the spurs to form the molecular products and reaction (10) occurs after diffusion of the speoies from the spurs when a solute (S) is present. The yields of the various radicals and products of the radio-lysis are dependent upon many faotors suoh as the nature of the radiation (i.e. Its L.E.T.) and the nature of the solution. For high energy eleotron irradiation of pure water at pH7, the following yields are obtained*2 G e - a q 2 . 3 Gjj = 0.6 Gntr = 2.2 (JM GH202 = 0.71 G H 2 * 0.42 G-h2O = 3.6 The excited water moleoules formed in the radiolysis by reaction (1) are generally in an unspecified eleotronically excited state and they are usually ignored since it is possible to adequately explain the radiolysis of liquid water without including any contribution from exoited molecules.2 The as-sumption, most often made, Is that the exoited molecules either return to the ground state by a non-radiative prooess or else dissociate to H and OH radicals whioh have little excess energy and they merely recombine causing no net ohemioal change9 2. Cerenkov Radiation Cerenkov radiation is a phenomena associated with charged partioles moving through a medium at a speed greater than the phase velocity of light in that medium. The phenomena takes the form of electromagnetic radiation emitted by the medium in the visible and ultraviolet regions of the spectrum. It oan be considered as analagous to the sonlo shook wave produoed 10 by a projectile traveling at greater than the velocity of sound or to the more familiar case of the bow wave of a ship moving through water when its speed is greater than that of the surface waves on the water. Cerenkov radiation was first observed by Mallet? in 1926 but it was not until 193^» when Cerenkov*® studied the effect, apparently unaware of Mallets1 work, that the radiation was characterized. Following Cerenkovs' work in 1934, Frank and Tamm^ developed the theoretical basis for the radiation in 1937 and this was later modified and expanded. A simple treatment of the theory of Cerenkov radiation may be given as follows.^ Assume that as a charged particle travels through a medium, an electromagnetic wave is emitted from each point along the particles' track, due to a polariza-tion of the moleoules in the region of the track and then subsequent relaxation of the polarization as the particle passes. These waves will be able to constructively interfere, as shown by the Huygens oonstruotion In/Fig* 1 if the wavelets from Fig.1 11 points A,B,C, on the particles' path arrive in phase at the conical wavefront, shown in cross section as DEF. This is possible if, while the light wavefront passes from A to F, the partiole passes from A to E. Considering the velocity of the partiole to be v, the refractive index of the medium n and the velocity of light in free space C, then the phase velocity of light in the medium Is C/n, The condition for constructive interference then follows as AF = AE i.e. A£ «= 1 C/n v AE v n (vii) C AF = Cos£ = (vlii) AE • p n where ^ is the relatlvistlo velooity v/o. Hence the "Ceren-kov relation" has been derived. Cos Q m (lx) /3n It follows that, since oos Q must be < 1, this is only pos-sible if 1/ygn $ 1, i.e. if v £ o (x) n Thus the threshold oondition for Cerenkov radiation is that the velooity of the partiole must be greater than the phase velooity of light in the medium. Two other properties are evident from relation (lx). There Is, for an ultrarelativlstio partiole ( 1 ) , a maximum angle for emission given by 0max = Cos"1 (1) (n) (xl) and furthermore, as the eleotrons' velooity approaohes the other 12 extreme, namely fj = 1/n, the radiation direotlon coincides with the particles path. The other property of Cerenkov is that the radiation is in the infra-red, visible and near ultraviolet regions of the spectrum where n > 1. Emission in the X-ray region is impossible since n < 1 due to dispersion of the medium and equation (lx) cannot be satisfied. A further property of Cerenkov radiation whloh comes out of the oonstruotlve intereferenoe oondition is its polarization. In order that the condition be satisfied, it is neoessary to require that the electrio veotor E of the light be everywhere perpendloular to the surfaoe of the Cerenkov oone arid the magnetio veotor H tangential to this surfaoe. Hence the light should be linearly polarized. The duration of the Cerenkov light emission pulse is, to a first approximation, determined by the velocity of the partiole and the length of Its track. Henoe a pulse duration of the order of 10"10 seoonds would be observed for a single eleotron traveling a distance of 3 oentimeters. The radiation yield and spectral distribution of Cerenkov emission can be theoretically determined by a solution of the wave equations and Maxwells' equations for the problem. The original treatment of the problem by Frank and Tamm, as given by Jelley8, makes the simplifying assumptions that (1) the medium is a oontinuum and microscopic structure Is Ignored, (11) dispersion oan be ignored, (ill) the medium is a perfect isotroplo dieleotrlo, (iv) the electron oan be oonsldered 13 to move at a constant velocity and (v) the medium is unbounded with the track length infinite. Their solution to the problem 8 glveB the result dW = e2 41 4J' riP?'™*0* <xll> O .A fj'tf where W 1B the total energy radiated by the electron through the surface of a cylinder of length 1, whose axis ooinoides with the line or motion of the electron. CO is the frequency of the radiation, Q » the relativistlc velocity and n, the refractive index„ This integral may be solved by the introduc-tion of the appropriate limits, to give an equation of the form0 where N is the number of photons, within the spectral range defined by \|, and X2 , emitted by a single electron, 1 is the traok length and OC is a fine structure oonstant equal to I / I 3 7 . The speotral distribution of the radiation may be ex-fi presses in several ways, two 0f whioh are a2N 4 ± (xiv) did A ^ \2 (number of photons per unit path per unit wavelength interval) a2w o4 J. dld\ "\3 (energy per unit path per unit wavelength interval) Application of these formulae ((ix) - (xv)) to the oase of a 500 keV. eleotron ( = O .87 . ) , penetrating a depth of 1 millimetre into water, where n = 1.3^» shows that Qmax = 30° 14 and. the number or photons emitted, between 4000 and 6000& , by a single electron is about 10. A more sophisticated and exaot treatment of the theory of Cerenkov radiation, in which the effeots of finite track length, slowing down of the particle, dispersion and discon-tinuity of the medium are inoluded, gives only very slight modification to the formulae derived from the original theory, 3. Previous Investigations of Light Emission from Irradiated Water At least ten investigations have been made into the phenomena of luminescence of irradiated water. In all but one of these reports, no appreciable light emission other than Cerenkov radiation was noted. The first investigations of the phenomena were made by Mallet9 in I926-I929 and Cerenkov10 in 1934-1930. Mallet observed the emission from water irradiated with radium -rays and made an attempt to determine its spectrum, concluding only that the emission spectrum was continuous. Cerenkov made a more comprehensive study of the light and determined Its spectrum, angular distribution and intensity. Since known quenching agents had little effect on the intensity of the emission, he oonoluded that the light emission was not due to normal de-exoitation prooesses. By observing the influence of a magnetio field on the radiation, he also showed that the emission was due to the secondary electrons produced in the 15 medium by the ^-rays. In the period following Cerenkovs' work, seven more papers1^""^ were published which supported the original dis-covery. The authors used a variety of techniques, including spectrographs and photomultipllers, to determine the emission spectra and photon yields from the Irradiations of water with a variety of radiation souroes, direct current electron ac-celerators, ^-rays, f2~rays and d-partioles. In 1963» Sitharamarao and Dunoan20, studied the light emission from Co^° *y~irradiated water, using sensitive cadmium sulfide orystal detector. They report observing four broad bands in the visible and ultraviolet region of the spectrum and alBO the characteristic Cerenkov radiation produced by the Compton eleotrons. They assigned the bands, on the basis of little evidence, to various processes of the radiolysis. In 1966, Czapski and Katakls , reinvestigated the system, using as a radiation source the ^-particle from tritium in the form of T20. The maximum energy of this ^-partiole is below the threshold energy for Cerenkov radiation. They report observing an extremely low intensity light emission with a photon yield G^^ < 10""5, which is at least four orders of magnitude less than that reported by Sitharamarao and Dunoan, They were unable to measure the spectrum aoourately due to the low intensity of the emission and they did not make any assignment as to its origin. 16 Scope of the Present Investigation Partly In view of the apparent conflict between the two most recent Investigations, lt was deoided to reinvestigate the system using as a radiation source an intense pulse of high energy electrons. This source had the following advantages over radiation sources used previously for similar investigationst (i) it was pulsed and consequently emission lifetime measure-ments would be possible, (ii) the radiation Intensity during the pulse was higher than other souroes by a faotor of 10^ (Sitharamarao and Duncan, who reported observing emission, were using a radiation intensity several orders of magnitude larger than that used by Czapski and Eatakis), (ill) the elec-tron energy could be varied from 0.5 MeV. to below the Cerenkov threshold and (iv) the high energy electrons responsible for Cerenkov radiation were unidirectional and thus the isotropy (and possibly the polarization) of the light emission could be studied. The aim of the investigation was therefore to determine If there was any light emission, other than the expected Cerenkov radiation, and if so, to determine its emis-sion lifetime, spectrum, yield and origin. A consideration of the events occurlng in the system during and after the radiation pulse leads to some predictions as to the nature of the light emission that one might expect to see. When a thermalized eleotron, formed in the ionization 22 prooess, becomes hydrated, to form the hydrated electron , 17 It muBt lose about 2 eV. of energy since the average trap depth is believed to be 2-3 eV. Since this process ooours in a time of the order of 10"11 seconds2, the dielectric re-laxation time of water, the energy might well be liberated in the form of photon emission and thereby give an emission speo-trum for the formation of the hydrated eleotron complementary to its already well established absorption spectrum. Another possible source of light emission might be from excited water moleoules. Although the first singlet exoited state of water is a dissociative level, a theoretical pre-diction of a low lying triplet energy level has been m a d e 2 ^ . Sinoe the ionization potential of water is about 12 eV. and for every 100 eV of energy deposited in the system only about 3 ion pairs are formed, it follows that more than half of the energy deposited in the medium is used in excitation of water molecules apparently with no resulting chemloal change. Hence a possible source of light emission is from triplet exoited water molecules. Other possible sources of light emission are exoited OH radloals, which are known to emit in the near ultraviolet In the gas phase2**, and possibly radical-radical reactions of the type fil* + h 2 - >B3* + (11) B3* —> H3 + hp (12) where Hi* and B y represent an unspecified eleotronioally 18 exoited radical or molecule. Indeed one might speoulate that, "because of the apparent very inefficient use of energy in the radiation chemistry of water, it would be surprising if none was utilized in photon emission. 19 EXPERIMENTAL 1. Electron Accelerator A Febetron pulsed, electron accelerator, shown in Fig. 2, manufactured by Field Emission Corporation, McMinnville Oregon, was used as the radiation source for this investigation. The accelerator, model 701-2660 pulser with model 5235 eleotron tube, produced an extremely intense pulse of 0.52 MeV. eleotrons. The peak beam current, at the tube window, was about 1000 amperes and the beam pulse shape, shown in Fig. 3(a), was roughly triangular with a half width of about 20 nanoseoonds. Both the pulse shape and peak beam current were extremely reproducible from pulse to pulse with a root mean square varia-tion of about The energy of the output beam electrons was variable from near zero to 0.52 MeV and it depended on the D.C. charging conditions of the pulser, which operated on the Marx Surge Circuit principle. The corresponding beam current also varied with charging voltage since the impedance of the tube was nearly oonstant. The maximum radiation dose available was of the order of lo6 rads, with the corresponding dose rate of about 10*3 rads/seo averaged over the pulse. A mounting flange on the accelerator allowed the irradia-tion oell to be positioned at a distance of about 4 oentimeters from the eleotron tube window. 20 2. Eleotron Beam Current Measurements The electron beam current was measured UBlng an apertured Faraday oup. The Faraday cup and aperture are shown in Fig. It oonsisted of two concentrio aluminum cylinders, open at one end, coupled across a T & M Research Produots model GB-1-05 current viewing resistor which had an lmpedanoe of 0.0507 ohmB* The output of the resistor was coupled to a length of doubly shielded EG 58/U ooaxial cable which terminated in 50 ohms at the vertical input of the model 82 plug-in amplifier of a Tektronix model 585A osoilloscope on whioh the voltage pulse developed across the current viewing resistor was displayed. The aperture was an aluminum cylinder maohlned so that the Faraday oup oould be placed in an equivalent position to the Irradiation cell on the aooelerator. In this way the apertured Faraday cup measured the beam ourrent transmitted into the Irradiation oell and gave an estimate of the radiation dose. The appropriate electron window material used in the cell was also included on the aperture to take into account back scat-tering and attenuation of the beam by it. The sensitivity of the resistor was about 20 amps/volt and the voltage range of the amplifier was 0.002 V. to 400 V. so that ourrents in the range from 0.04 to 8000 amperes were measurable with this tech-nique. The eleotron beam current-time pulse shape was monitored on the osoillosoope by soanning the time base at the appropriate 21 sweepspeed and photographing the oscilloscope traoes using a Polaroid camera and high speed film. The typical pulse shape is shown in Pig. 3(a)• The peak beam ourrent measured for an 0.002 inch stainless steel electron window and full oharging voltage was typioally 170*5 amperes and was reproduoible, from pulse to pulse, to within The variation of beam ourrent with charging voltage was determined by this technique and the results are shown in Fig. 5. The outoff of current at 12 KV. oharging was the result of complete absorption of the eleotrons by the stainless steel eleotron window on the aperture and the titanium window of the eleotron tube. Integration of the ourrent-time pulses gave the total number of electrons per pulse. Typical results were, for 170 ampere peak ourrent at full oharging voltage, 2.0 electrons per pulse. 3. Dosimetry The absorbed radiation dose, or total amount of energy deposited in the irradiation cell, was determined approximately by use of a calorimetric technique. An aluminum disk of ap-propriate diameter and of suffioient thickness to stop all of the eleotrons (~ 0.032 Inch) was plaoed in a position equiva-lent to the sample in the irradiation cell and thermally insulated from the surroundings. A ohromel-alumel thermocouple Junotion was attached to the disk and the temperature rise upon irradia-tion of the disk with a single pulse was determined by measuring 22 the change In potential of the thermocouple junction with a miorovoltmeter. Knowing the weight of the aluminum disk, its specific heat and the temperature change (calculated from the change in potential of the Junction and its known sensitivity2^) the total energy deposited in the disk oould be oaloulated. The uncertainty in the value obtained in this manner is primarily due to back scatter of the eleotrons from the aluminum disk and uncertainty in the calibration of the voltmeter. These combined errors are estimated to be less than 20/6. Typical results for full oharging potential and 0.002* - stainless steel window material in the oell were, 0.85 * 0.17 x 1 0 e V . per pulse. (ThiB value is confirmed by oalorimetrlo and chemical dosimetry done by D.A. Head on the same accelerator). Other methods of dosimetry available weres (1) calcula-tion of the energy deposited by combining the total number of eleotrons, as measured with the Faraday cup, with the average electron energy, estimated from a consideration of the mean energy losses in the thin stainless steel oell window. (Using the oalorimetrlo value of total energy and the Faraday oup measurements, the average electron energy was calculated to be 0.42 * 0.08 x 106 V., which seemed reasonable) (2) ohemlcal dosimetry, such as the hydrogen yield from irradiated cyolo-hexane or (3) measuring the optical density change induced in a dyed cellophane and obtaining the dose from a calibration ourve.27 23 4. Spectral Analysis Systems (a) Photomultlpller-Interferenoe Filter Wedge Spectrometer A photomultlpller-interferenoe filter wedge speotrometer was designed so that the emission lifetime and the spectral characteristics of the light emission could be determined simultaneously. The speotrometer is shown in Fig. 6 (a) and 6 (b). Fig. 6 (a) is a photograph of the speotrometer and Fig. 6 (b) is a plan view of it showing its mode of operation. The interference filter wedge used was a VEEIL S 200 continuous line filter and the photomultiplier was an B.C.A. Victor 1P28. Variation of the position of the wedge filter with respeot to the slits was accomplished with the aid of the control rod, which was marked with small divisions so that the exact position of the filter was known. This variation changed the band of wavelengths seen by the photomultiplier and by successively scanning the whole filter a complete speotrum oould be obtained from 4000 to 7000 fi . The dispersion of the filter was about 25 A per millimeter. Light emitted from the irradiation oell was directed at the spectrometer by means of a front silvered plane mirror and two cylindrical light collimators as shown in Fig. 7. The short initial collimator was nickel plated on the inner surfaoe so that the maximum amount of light possible was col-lected. The second collimator, at right angles to the first, was painted flat blaok BO that only a roughly parallel beam 24 of light reached, the slits of the spectrometer. The colli-mators and the speotrometer were enolosed in lead boxes, as shown in Fig. 7, in order to eliminate X-ray effeots on the photomultipller and also to attenuate the primary X-ray beam of aooelerator. The photomultipller load resistance had to be 50 ohms in order to matoh impedance wlkh the ooaxial cable and avoid the use of a cathode follower. (The interdynode oapacitanoe of the photomultipller limits the size of the load resistance to less than 500 ohms if 100 MHz. frequency response is re-quired). The EG58/U coaxial cable leading from the photo-multipller to the osollloscope was doubly shielded and oonneoted to the input of the vertical amplifier of the Tektronix 585A osollloscope. The anode ourrent pulse, developed aoross the photomultipller load resistance, was then displayed on the osollloscope as a negative voltage pulse and it was photographed with a Polaroid camera. The maximum anode ourrent permissible, pQ before saturation effeots oaused non-linearity of response of the photomultipller, was about 2 mllliamperes, which cor-responded to a maximum voltage pulse aoross the 50 ohm load of 100 mV.. Since the maximum sensitivity of the osollloscope amplifier was 10 mV./cm., only a very limited range of light intensity was measurable. However, by interposition of neutral density filters in the light beam, the effective intensity range oould be expanded considerably. Further ampllfloation of the small voltage pulses by means of an auxllary amplifier 25 was not possible because of the high frequency involved. It was also necessary, when operating the photomulti-plier with large anode ourrents, to provide a large dynode resistor chain current in order to prevent non-linearities due to fluctuations of the dynode voltages during the pulBe. As a result, the 1P28 was operated at 500 V. and 10mA.D.C.. (Note: The photomultiplier could not be operated satisfactorily at larger voltages (and thus higher sensitivity) due to a large random noise level oaused by the small load resistance.) A Fluke stabilized D.C. power supply was used to maintain these conditions. Wavelength calibration of the detection system was ob-tained by use of a mercury vapour lamp and a He-Ne laser. The anode ourrent of the photomultiplier was measured as a function of the position of the wedge oontrol rod for both light souroes, using a small slit opening. The He-Ne laser was used as a wavelength marker and the mercury emission lines were used to oheck the linearity of the dispersion of the filter wedge. The device was found to be linear over the range 4000 to 7000 A . Calibration of the spectral sensitivity of the system was performed with the use of a calibrated quartz-iodine lamp placed at an equivalent position to the irradiation oell. The anode ourrent was again measured as a function of the position of the wedge oontrol rod. From a knowledge of the wavelength calibration of the control rod and the emission 26 spectrum of the calibrated lamp,.the spectral sensitivity of the detection system was obtained. It is shown in Fig. b. The procedure for determining an unknown emission spectrum with the spectrometer was as follows. The sample was irradi-ated in the cell, which was positioned In front of the light collimator, and the anode current pulse of the photomultiplier was measured using the oscilloscope. This was repeated for each position of the control rod, using a single pulse and fresh sample of liquid for each wavelength band. The peak anode current, which was proportional to the intensity of the light, was then corrected for the variable spectral sensitivity of the detection system and the relative emission spectrum was thus obtained. It was not possible to determine an absolute emission yield with thiB arrangement beoause of difficulties in estimating the fraction of the total light emitted that reached the photomultiplier. From the shape of the anode current pulses, the lifetime —R of the.emission oould be determined if it was longer than 10 seconds. Lifetimes shorter than this were not measureable since _ o the fall time of the eleotron pulse was 10 seconds. The pulse to pulse reproducibility of the anode current pulses of the photomultiplier for identioal operating condi-tions of the speotrometer was very good. A variation of the order of - was observed and this was likely due to the 27 variations in the output eleotron pulse of the aooelerator. (b) Grating Spectrograph A large aperture, low resolution, grating spectrograph was also used to determine the emission spectra. This spectro-graph is shown in Pig. 9« It was constructed following the design of BBBS and Kessler2^ and consisted of a slit and lens arrangement to colleot the Incident light and foous it on a Bausch and Lomb grating, ruled with 600 lines per millimeter, and a Konlca F.P. 35 mm* camera with a 1.4 wide angle lens to record the spectra. Wavelength calibration was performed by use of a mercury vapour lamp and the absolute spectral sen-sitivity of the system was determined using the calibrated quartz-Iodine lamp. The spectrograph was found to be linear in dispersion over the range from 4000 to 6500A with a dis-persion of about 150 A /mm. when using a second order diffrac-tion angle. The film used to record the emission spectra was the fine grained, high speed, panchromatic Ilford HP4. It was developed in fresh Acufine developer for 5 minutes at 72 P. in a spiral tank with gentle agitation, washed for 30 seconds in water, and fixed for 5 minutes in Kodac Rapid Fixer and Hardener. The emission spectra of the irradiated liquids and the plastic scintillator were determined by positioning the ir-radiation cell in front of the slits and irradiating the sample with a sufficient number of pulses to give a measureable 28 density on the film. The film was protected from the primary X-ray beam with a sheet of lead between the camera and the accelerator. Wavelengths were marked on the film by exposing the centre portion of the film to a mercury vapour lamp. The true emission spectrum was then obtained by the following prooedure whioh corrects for the non-uniform speotral response and the non-linear exposure-density relationship for the film. A series of exposures for the standard quartz-iodine lamp at various exposure times were made and from the densitometer tracings of the films, the characteristic curves, showing exposure versus density, were plotted for various wavelengths. The unknown spectrum was then measured on the densitometer and the appropriate relative exposure was deter-mined from the characteristic curves. This relative exposure was then correoted for the speotral distribution of the quartz-iodine lamp and the emission speotrum of the sample obtained. The spectral sensitivity of the spectrograph was determined in a similar manner and it is shown in Fig. 10. An estimate of the total emitted photon yield from irrad-iated water was evaluated by comparing the wavelength inte-grated densities for the water emission and the standard lamp in equivalent positions and relating the relative densities to a relative number of photons In the spectral region Involved. Taking into account the geometries involved and knowing the absolute intensity of the quartz-iodine lamp, an estimate of the number of photons emitted by the water was found and 29 this oonverted to a G-value using the known radiation dose. 5. Densitometry A Joyce double beam, recording microdensitometer was used to measure the density of the films from the spectrograph and the 180° camera. This Instrument measured the density of the film, relative to the background, directly in density units and laterally scanned the film at the same time, recording the density on chart paper. The density calibration of the instrument was checked by use of neutral density filters and it was found to be linear over the density range from 0 to 1.6. 6. 18QC Camera A 180° camera was designed and constructed to enable mea-surement of the angular distribution of the light emission. Shown in Fig. 11, it consisted of a hemioyclindrloal piece of plexiglass enclosed on the other three sides by a reotangular box with a oollimated entry port (a 1/8H hole in a large aluminum cylinder) for the electron beam at the centre of the half cy-linder. The inner surface of the collimator was covered with a thin piece of aluminum foil so that the vessel was water tight and a filling port was included so that solutions could be ohanged. Provision was made for attaohlng a photographic film to the ourved surfaoe of the cylindrical window. The ourved surfaoe was painted flat blaok on the inside along its edges, leaving a clear 1M wide band around the centre. In this way, a background measurement of the light contribution from Cerenkov 30 radiation from the plexiglass itself and the X-ray effects on the film were determined. (The action of the X-ray beam on the plexiglass generated sufficient light to partially darken the film). The remainder of the box portion of the camera was also painted flat blaok on both surfaoes in order to reduce the effeots of scattered and stray light. The film, Ilford EP4, was exposed to the light generated by several pulses and then developed. A densitometer traoe of the centre portion of the film and an adjaoent portion then gave, by difference, the angular dependence of the light emis-sion. The angular dependence was determined for water, a plastio scintillator and a high purity quartz glass. 7. Irradiation Cell and Water Flow System Some preliminary experiments were performed using an irradiation oell made from aluminum and an electron window of aluminum foil. However, this oell was found to corrode, with the formation of white deposits of aluminum hydroxide, on prolonged oontaot with water. Stainless steel was then tried as a oell and window material and it was found to be satisfactory with no visible corrosion even after 48 hours of oontaot with water. The stainless steel oell is shown in Fig. 12.. It con-sisted of a ring of stainless steel, appropriately maohlned, with a high purity quartz optical window oemented on one side and provision for a thin stainless steel foil electron window 31 on the other side. The electron window was held in place by a teflon ring on the cell and a rubber "0" ring on the aluminum mounting flange. The oell also had filling and draining ports, with Kovar glass-metal seals, to allow solutions to be flowed through it. The water flow system consisted of a two litre flask fitted with a fritted glass bubbler, (to allow degassing of the solu-tion by the passage of helium through itj, a pressure release valve and a glass conneoting line to the irradiation oell. The oonneoting line was attached to the entrance port of the cell by means of a ball and sooket joint. After degassing the solution, by bubbling helium through it for at least a half hour, the flask was pressurised and the water forced into the cell. The flow through the cell was regulated by means of a stopcook on the exit port of the cell. The solution, in the cell, was generally changed after each pulse by opening the stopcook and allowing the solution to flow through the oell for a few seconds. This was done so that impurities formed by the radiation would not build up in the solution. 8. Aotlnometrv An attempt was made to estimate the total number of photons emitted by the sample using a chemloal aotlnometer. Potassium ferrioxylate solutions were used following the method of P a r k e r ^ 0 ' I t was found that no change in optioal density of the developed solutions oould be deteoted following expo-32 sure of the actinometer to the light emission from either water or a plastio scintillator. 9. Speotrofluorlmetry The emission spectrum of the plastic scintillator was obtained with an Amlnoo-Bowman Speotrophotofluorometer. This was done so that a check could be made on the calibration of the light deteotlon Instruments used in this investigation. The spectra obtained from the grating spectrograph and the interference wedge spectrometer for the plastic scintillator were compared with that obtained from the speotrofluorometer. All three spectra were found to be very similar. 10. Materials (a) Water Doubly distilled water was used in all experiments. The second distillation waB made from aoldio permanganate solu-tion. (b) Scintillator The plastio scintillator, NE 102, manufactured by Nu-clear Enterprises Ltd., was used as a check of the calibration of the spectrograph and the spectrometer. The scintillator was in the rorm of a thin plastio sheet and it was out Into oiroular disks which fit inside the irradiation oell. 33 (o) Photographio Materials Acuflne developer and Kodao Eapld Fixer and Hardener were used. The solutions were made using distilled water. (d) Other Materials Analytical grade KOH and ^SO^ were used to prepare acidic and basic solutions. The Benzene, Methanol and Cyolohexane were reagent grade and the compressed gases, He and O2, were regular commercial grade. 11. Electrical Noise Due to the intense electrioal and magnetic fields generated by the Febetron pulser and the electron beam, it was necessary to install a good high frequency grounding system conneoted to all the electronic equipment In use. It was also necessary to filter the 110 V. Input power lines to the oscilloscope and photomultlpller power supply with a radiofrequenoy line filter. All ooaxlal cables used were doubly shielded using copper braid surrounded by thlok copper pipe. The resulting noise level on the oscilloscope traces was less than 5 mV.. 34 RESULTS AMD DISCUSSIONS 1. Preliminary Investigations A few preliminary observations of the light emission from water irradiated with a pulse of electrons gave an Indication of its general characteristics. Visual observation of the emission, by reflecting the light beam with a mirror to an observer behind the radiation shields, indicated that the light emission was short lived (lifetime less than several seconds) and that it had a bluish-white spectral character. An attempt was made to obtain a low resolution spectrum using interference filter wedge and a photographic film. A plexiglass rod light guide was used to transmit the light beam from the irradiation cell to the slits In front of the inter-ference wedge. (The wedge and film oould not be placed directly in front of the cell because the X-ray beam from the accelerator caused appreciable darkening of the film.) This experiment failed to obtain a spectrum since a blank run, where the light from the cell was prevented from entering the light guide, revealed that a considerable portion of the darkening of the film was due to light produced in the light guide Itself. This light was probably Cerenkov radiation from the plexiglass produced by the X-ray beam generated as Bremsstrahlung when the eleotrons were stopped in the water and the eleotron tube window. Similar experiments whioh involved using a photo-35 multiplier tube instead of the photographio film were also invalidated by the balnk run. Since visual observations confirmed the presence of light emission from the water, a more elaborate experimental setup was designed and constructed so that an accurate assesment of the emission could be made. The results of the experiments performed with this apparatus are reoorded in the following sections. 2. The Angular Dependence of the Light Emission In Fig. 13 the angular dependence of the light emission from water, quartz glass and a plastic scintillator Is shown. This data was obtained from measurements of the density of films from the 180° camera as a function of angle using the densitometer. The graphs show the density difference, AD, (between the portion of the film exposed to the light emission and the adjacent background density) at a given angle Q plotted as a function Q , where Q is the angle between the dlreotlon of motion of the eleogron beam and the emitted light beam. In the case of the plastic scintillator, the sample was a hemioyllndrloal pleoe of polished plastic which was fixed over the entrance port for the electrons in the 180° camera whereas for quartz an optical window was used. In both cases the oamera was filled with water to reduce the X-ray effeot on the film and diminish reflection and refraction of the light. As can be seen from the ourves on Fig. 13» water appears 36 to emit light preferentially in the forward direotion where-as the scintillator shows an emission which is isotropic, as would be expected for fluoresoence. The experiment with quartz was performed in the hope that high purity quartz would show only a Cerenkov radiation emission and thereby be a valuable oomparison for the water experiment. However It appears that the emission from quartz has an angular dependence unlike that expected for just Cerenkov radiation. Kawabata and Okabe- 2^ have observed a long lived luminescence in Irradiated quartz, which, ooupled with the expected Cerenkov radiation, might aocount for the angular dependence found in this work. The relative magnitude of A D for the water, quartz and scintillator have little significance because the background densities were not the same. Furthermore, in the water and quartz experiments, several pulses were needed to give sufficient exposures. From these experiments it is evident that the luminescence from water has an angular distribution different from normal fluorescence. The experimentally observed angular dependence is, however, in reasonable accord with it being almost entirely Cerenkov radiation. For water, the maximum angle for Cerenkov radiation for 0.52 MeV. electrons I s a n d , a s t h e electrons slow down the angle of emission becomes smaller. But, In addition, elastio scattering of the high energy eleotrons will oause a broadening of the angular distribution. A complete analysis of the expected Cerenkov distribution is impossible 37 without a detailed knowledge of the primary interaction pro-cesses of 0.52 MeV. eleotrons. These two effeots might indeed give rise to a distribution similar to the observed one. 3. Variation of the Emission Intensity with Eleotron Energy The intensity of the light emission from water and from the plastio scintillator was measured as a funotion of the energy of the incident eleotrons using the interference filter wedge speotrometer. The energy of the eleotrons was varied by varying the charging voltage of the accelerator. The results of this Investigation are shown in Pig. 14 where the peak light intensity (in arbitrary units) is plotted as a funotion of the charging voltage of the accelerator. The corresponding maximum energy of the inoident electrons (data supplied by Field Emission Corporation and corrected for energy loss in the stainless steel window) is also shown on the or-dinate. The curves are normalized to the same intensity at 27.5 KV. charging voltage, but in actual fact the scintillator intensity was considerably greater than that of water. The data was obtained for a narrow wavelength band centred at 5060A . As can be seen by comparing the curve of Fig. 14 for the scintillator with the curve of beam current as a funotion of charging voltage given in Fig. 5, the scintillator emission intensity and beam current have a similar dependence on charging voltage. Water, on the other hand, appears to show a more oomplex behaviour, with an apparent outoff of emission intensity 3» at about 240 KV electron energy. These features of the emis-sions are shown more clearly if one plots the peak light inten-sity divided by the product of the electron energy and peak beam current (i.e. something which is proportional to a G value for photon emission) as a function of electron energy. This is shown in Fig.. 15. It is evident from this graph that the emission yield for water increases as the electron energy increases and also has a zero value at eleotron energies of about 240 KV. In contrast, the emission yield for the scintil-lator decreases quite rapidly and then levels off as the elec-tron energy increases. The radiation yield for an excitation energy transfer process, such as one might envisage being responsible for fluorescence of the scintillator, should be Independent of the electron energy. However, since one is dealing here with extremely high radiation intensities, second order deactivation processes may be the cause of the observed decrease in photon yield. These results are consistent with the angular dependence results in that they tend to support the view that the emission from water is mainly Cerenkov radiation. The low-energy cut-off of the photon yield at 240 KV is in agreement with the ex-pected value of 2b0 KV for Cerenkov radiation in water (i.e. when {$= 1 ) and the increase in photon yield with Increase in electron energy is also as predicted for Cerenkov radiation. In their theoretical treatment of Cerenkov radiation, Frank and Tamm11, establish the relationship 39 where N is the number of photons emitted in spectral range between X| and X 2 by a single electron, a is a fine structure constant equal to 1/137. 1 is the distance the electron travels while it is radiating, Q is the relatlvlstlc velocity and n is the refractive index. Thus for a given spectral range, there should be a linear relationship between N and 1 • (l-l/pZr\Z )• F1S» 1 6 shows t h e P e a k light intenlsty for water at•5060 X divided by the peak beam current (I.e. something proportional to the number of photons per incident high energy electron) as a function of 1-(1-1^2 n2 ). 1 was calculated from the difference in energy between the energy of the incident electron and the energy of the Cerenkov cutoff and using the linear energy transfer value for the high energy electron in water. f$ was calculated from the electron energy taking into acoount its relatlvistic mass. The linearity of this plot may-be Interpreted as either a corroboration of the notion that the light observed is Cerenkov radiation or as experimental verification of the theory of Frank and Tamm. The Emission Lifetime Typical photomultiplier anode current pulses for the light emission from irradiated water and scintillator are shown in Figs. 3(b) and 3(c) respectively. These figures are copies of the photographs taken of the actual oscilloscope traces. Comparison of Figs. 3(b) and 3(o) with 3(a), the time 40 dependence of the electron pulse, Indicates that the emission lifetime from water is ^ 10"a seconds and from the scintil-lator ™ 10-® seconds. The exact shape of the water emission is also noteworthy, In that the secondary peak which is pre-sent in the current pulse, is absent or very small in the light emission in comparison to the main peak. Both of these observations oan be explained on the basis of Cerenkov radia-tion. The lifetime of Cerenkov radiation Is determined, to a first approximation, by the time required to stop the elec-trons and it is therefore of the order of 10"!1 seconds. This Is to be compared with fluorescence lifetimes which may be 10~9 seconds or longer. Secondly, if the energy of the eleo-trons In the secondary peak is substantially less than that of the primary peak, as seems likely because Its intensity has a much stronger dependence on changing voltage than the pri-mary peak, then for Cerenkov radiation (but not for fluores-cence) the intensity of the light emitted by these electrons (per electron) would be expected to be very much less than that emitted by the primary eleotrons. 5. Aotlnometry The failure of the actinometer to measure any light emission from the water is not surprising if one is dealing only with Cerenkov radiation. Since a pulse contains only about 2.xl0*3 electrons and the theory of Frank and Tamm pre-dicts only about 10 photons per 0.52 MeV electron in the range of 3000 5000 A (the strong absorption band of the ferrioxylate 41 actinometer) more than 50 pulses would have been required to exceed the limit of detection of this actinometer as given by P a r k e r ^ 0 ' T h e failure of the actinometer to register a change with the scintillator is however surprising. Perhaps there Is an Important Intensity effect due to second -order processes under the enormous light intensities (<v lo 2 2 photons/ cm.2/sec.) involved In this experiment. 6. The Emission Spectrum (a) Photomultiplier - Interference Filter Wedge Spectrometer The emission spectrum of pulse irradiated water as deter-mined on the photomultiplier - Interference filter wedge spectro-meter Is shown in Fig. 17. This graph Is a plot of the relative number of photons emitted in a given wavelength band as a function of wavelength. The data was obtained from the peak of the anode voltage pulses measured for various positions across the interference filter wedge and corrected for the spectral response of the entire optical arrangement and detec-tion apparatus. The uncertainty in the relative number of photons, shown by the appropriate error bars, is almost entirely due to the uncertainty in the calibration of the spectral response of the spectrometer. The dotted line is a plot of c/ \ versus X (where c Is an arbitrary constant chosen so that a large portion of the experimental points fall on the line) and represents the expected relative variation of Cerenkov radiation Intensity as a funotion of wavelength. 42 It appears from Fig. 17 that the observed emission between 4500 -6500 A is consistent with it being Cerenkov radiation. The deviation of the emission spectrum from the c / \ curve at short wavelength (4000 - 4500 A ) may be due to one of the following reasons: (I) it may be a true emission with intensity greater than Cerenkov radiation; (il) the interference filter may have been transmitting higher order, lower wavelength light at these wavelengths (i.e. 3rd order 2666 8 at 4000 8 since the filter is second order in interference); or (iii) the calibration of the spectral response of the instrument had a large error at these wavelengths. Point (ill) can be checked experimentally. If there was a large error in the calibration in this region, then the apparatus would not be able to reproduce a Imown emission spectrum. The calibration was checked by observing the emis-sion spectrum of the scintillator and treating the data in the same manner as for water. The resulting spectrum is shown in Fig. 18 along with the emission spectrum obtained using the spectrophotofluorometer. Both spectra show the same general shape and exactly the same wavelength of maximum emission. Since the scintillator emission has Its maximum in the suspect region, the calibration of the instrument must be at least as accurate as estimated, otherwise the interference filter wedge spectrometer data for the scintillator would have also deviated from the expected values. Regarding point (ii), if the filter was transmitting 43 ultraviolet light in the region from 4000 to 4500 X by a higher o r d e r interference, the deviation of the spectrum could be explained, since the yield for Cerenkov radiation in the ultra-violet is much greater than in the visible region. However, according to the manufacturers specifications, the interference filter was cemented to a progressively coloured glass to eliminate higher order transmissions. Since the region of 4000 - 4500& occurs at the very end of the filter, it may not have been p r o p e r l y compensated, but this is very unlikely. It appears that the most likely explanation of the devia-tion is that it is due to a true emission. Whether this emis-sion results from a "chemical" radiative process, such as fluorescence, or a physical process, such as transition radia-tion, cannot be decided without further study, although as will be indicated later the same deviation is found for benzene, methanol and cyclohexane, which suggests that it is not "chemical" in nature. The most significant fact about the general appearance of the spectrum is the c/^2 variation which it follows from 4500 - 6500A , in agreement with that expected from Cerenkov radiation. (Notes The spectrum between 6500 - 7000fl could-not be determined with this instrument because of the insen-sitivlty of the photomultiplier In this region.) (b) Grating Spectrograph The emission spectrum was also obtained using the grating spectrograph in order to determine if there was any long lived, 44 low intensity emission which would be detectable by a film but not by the photomultiplier technique. The emission spectrum of water was determined by correcting the measured density on the film from the spectrograph for the non-uniform spectral response of the film. The resultant spectrum, plotted as rela-tive number of photons (in arbitrary units) versus wavelength, is shown in Fig. 19. Because of the relative insensitivity of the film to low energy light, the spectrum extends only to 6000 8 . The dotted line is again a c/\2 versus X curve. Uncertainties in the slopes of the characteristic curves used in the data treatment procedure are the main sources of error. These uncertainties are fairly large due to the fact that the data for the characteristic curves was not obtained using the same piece of film as was used to record the water emission. Although Identical development procedures were used for both films, the background densities were not the same due to the exposure of the film, in the water experiments, to the X-ray beam. (The slope of the characteristic curve is dependent on the development time.-^) As can be seen from Fig. 19, the spectrum obtained for water by this technique, has the same features as the spectrum obtained with the spectrometer (shown in Fig. 1?). Indeed even the deviation at short wavelengths Is present. Since the spectrum was obtained using a second order diffraction angle, the possibility of higher order diffraction occurs here also. However, since the light had to pass through two glass 45 lenses all of the ultraviolet portion of the light would have . been absorbed in the lenses. The possibility of mis-calibra-tion of the instrument in this region is- again ruled out on the grounds that the emission spectrum of the scintillator was found to be very similar to the spectra obtained by the other two methods. (The only difference in the spectra was an emission band in the region of 5900 - 6400 ft which was attributed to a phosphorescence emission which the other two methods would not detect.) Hence it appears that the devia-tion from c/X2 at short wavelengths must again be attributed to a true emission. 7. The Effect of Additives The effect of the addition of Impurities on the water emission spectrum was investigated using the g r a t i n g spectro-graph. It was found that the impurities formed by the radia-tion had no effect on the emission spectrum of water. (H2,02 and H 20 2 are formed by the radiation pulse in the water.) In fact, the density-wavelength band on the film for an ex-posure to the light emission from ten radiation pulses, where fresh water was used for each pulse, was exactly identical to that for an exposure to ten pulses where the same sample of water was used for all ten pulses. It was also found that the addition of O.IN. KOH, O.IN. H2S0/+ or 10-3 M. oxygen had rather little effect on either 46 the emission intensity or the spectrum. Actually the KOH had no effect on either the spectrum or intensity, whereas H2SO^ and oxygen increased the emission yield slightly but the spec-tral distribution remained unchanged. The increase in the case of H ^ O ^ was about 10% while for oxygen only about 2% change was noted. These results indicate that (i) radicals or ions which are reactive towards these additives are not involved in the emission process. The addition of acid would convert all of the hydrated electrons to hydrogen atoms by reaction (11) e-aq + H+ > H (11) while the addition of base would convert all of the hydrogen atoms to hydrated electrons by reaction (12). H + OH" >e~aq (12) The presence or hydrogen, oxygen and hydrogen peroxide would also affect the processes occuring by reactions such as (13) ~ (16). 0 2 + H — »H0 2 (13) 0 2 + e~aq > 02~ (14) H 2O 2 + OH > H 2O + HO 2 (15) H 2 + OH » H 20 + H (16) (ii) There are no normal radiative de-excitation processes from electronically excited water molecules that oxygen at 10"" 3 M. can quench. And (iii) the emission does not result from oxygen itself, or some reaction involving oxygen. De-aeration by bubbling with helium would not completely remove the atmospheric oxygen from the solutions and thus it might 47 have been contributing to the emission; but since a.probable 10^ fold increase In oxygen concentration did not affect the photon yield such processes can apparently be ignored. On the other hand, these results are in complete agree-ment with the emission being composed almost entirely of Ceren-kov radiation, since the intensity and spectrum of Cerenkov radiation depends only on the physical nature of the medium (i.e. its refractive index). The slight Increase in photon yield for the acidic solution might be accounted for by a difference in refractive index between pure water and O.EJ. HgSO^. However, since the absolute difference between the two refractive indices is" only about l%t which would thereby increase the yield by about 2%, a more likely source of the difference is experimental error which is estimated to be of the order of 10J6. (The pulse to pulse reproducibility of the accelerator was about and since the results were obtained for a multiple pulse experiment, an error as large as 10% would not be inconceivable.) 8. Comparison or Water with Other Liquids In order to check the early indication that water appeared to have no emission other than Cerenkov radiation (neglecting for the moment the possibility of some emission in the region of 4000 - 4500 A ) the emission spectra of several other liquids were investigated using the grating spectrograph. A series of experiments were carried out in which the irradiation cell 4-tS and the spectrograph were positioned in identical locations and the experimental conditions were the same for all of the liquids. An equal number of pulses was given to each sample and the spectra were recorded on the same piece of film so that a detailed comparison of them could be made. The liquids ohosen for the comparison were water, methanol cyclohexane and benzene. These were chosen on aooount of their differences in structure, sensitivity to ionizing radiations and refractive index. The results are shown in Fig. 20, where the densitometer tracings of the four spectra are reproduced. It Is evident thatr (I) water and methanol show identical emission bands, both in absolute Intensity and spectrum. (11) Cyclohexane and benzene gave emission bands which were of greater intensity but identical in shape to the water. (The apparent difference In the shape of the emission bands for water and benzene at the red end of the spectrum Is due to the fact that the log (exposure)-density relation for a film is not linear In the region of small densities and in particular a very small change In exposure can give a comparatively large density change. When water was irradiated with more pulses than the benzene, a spectrum of equal density to the benzene spectrum was obtained for all wavelengths), (ill) The order of increasing density of the bands at any given wavelength was water = methanol< oyolohexane < benzene. These results can only sensibly be Interpreted on the 49 basis that the emission in all four cases is almost entirely Cerenkov radiation. It Is extremely unlikely that four liquids with such varying chemical constitutions would give identical fluorescent or phosphorescent emission. The increase in density on going from water and methanol to cyclohexane and benzene can be aocounted for by the differences In the refractive indices of these liquids since Cerenkov radiation has a yield which is proportional (to a first approximation) to the in-verse of the square of the refractive index. Water and methanol have nearly Identical refractive lndicies, while the refractive index of cyclohexane Is larger than that of water but smaller than that of benzene. Absolute correlation between the emis-sion yields for the four liquids and their refractive indioes was not possible since an absolute exposure (I.e. # of photons) was not known and could not be accurately estimated. Qualita-tively the agreement was good. A further significance of these results Is that the apparent emission, other than Cerenkov, in the region of 4000 - 4500 A can not be attributed to a chemical process since it is highly unlikely that the same process would occur In all four liquids. It is possible, however, that the effect may be due to an emission from a physical process of the radiolysis. Possible sources of light emission, due to physical processes, other than Cerenkov radiation are visible Bremsstrahlung and Transition radiation0. Bremsstrahlung In the visible region of the electromagnetic spectrum is extremely weak in intensity 50 8 and Jelley quotes a relative intensity for Bremsstrahlung as about 106 times less than Cerenkov radiation. Transition radiation (radiation produced by electrons at the boundary of two media having different refractive indices) is also known to be of extremely low intensity with a relative intensity of fl-at least four orders of magnitude less than Cerenkov . Thus unless some radiation intensity effect caused an increase in the yield of either of these sources of light, their contribu-tion to the emission should have been negligible. 9• Calculation of a G-value An estimation of the total photon yield was made in order to confirm the findings of the preceding sections. This was done using the calibrated quartz-iodine lamp, which had a known intensity, and the grating spectrograph. An exposure of the emission from the standard lamp at a known distance from the slits of the spectrograph was made for a fixed length of time. Under identical conditions, and using the same piece of film, the light emission from a number of pulses of water was recorded. A comparison of the densities of the water emission band and the standard lamp emission band was made for the wavelength region 4000 - 5000 A in order to estimate the relative number of photons. This was done by comparing the relative densities of the two bands with a number of other exposures made on the same piece of film for a progressive series of water experi-ments where 2, 4, 6, 8 ••• pulses had been glvfen. In this 51 manner the relative number of photons was estimated to be 2.5 * 0.5 times in favour of the q u a r t z -iodine lamp which had been exposed for 0.8 milliseconds and for the emission from 12 pulses of water. The photon yield was then calculated as follows: - The integrated intensity of the standard lamp over the range from 4000 - 5000 A was 130 flwatts cm."2 at 40 cm. Since 2.5 - 0.5 times as many photons were emitted by the standard lamp as compared to 12 pulses from water, one pulse of water was therefore equivalent to 1/30 of the number of photons from the standard lamp. - The total amount of energy emitted by the lamp per cm.2 at 40 cm. in 0.8 milliseconds was 130 x 10"6 watts cm."2 x 8 x 10"4 seconds = 1.04 x 10"7 joules cm.-2. 2 - Thus the total amount of energy emitted per cm. at 40 cm. by the water for 1 pulse of electrons was 0.04 x 10-7/30 = 3.5 x 10"9 joules cm.-2. - Assuming the light was emitted isotropically, the total energy emitted by the water was than 4TT x 402 x 3-5 * 10'9 joules = 4.4 x 1014 eV. since the solid angle formed by 1 cm.2 area at 40 cm. is 1/402 steradians and isotropic radiation is emitted equally over a solid angle of 4<TT steradians. _ s l nce the average energy of the photons in the region 4000 - 5000 * is 2.8 eV., the photon yield is then 4.4 x 10^/ 2.8 = 1.6 x 1014 photons per pulse. - Using the known radiation dose of 0.9 x 1019 eV. per 52 pulse gives a G value of ^ ( 4 - 5 0 0 0 A) = 1.8 x 10 . This discussion has made the assumption that the radiation was emitted isotropically. Cerenkov radiation however would be emitted preferentially in the forward direction, mainly In a cone of <v 65°, and when this is taken into account a G value of about Gh|/(4.5000AV" 6 X 1 0~ 4 i S o b t a i n e d ' No error estimations have been Included In the calculation since they are difficult to estimate. The largest errors are Introduced by the angular distribution and the relative number of photons estimation. A conservative estimate of the uncert-ainty in the G value would be 60% although the number of photons calculation is probably uncertain by only about 30%. The photon yield calculated in this work agrees reasonably well with the yield expected for Cerenkov radiation. From formula (xii) it was calculated that about 4 photons per elec-tron would be emitted in the spectral range 4000 - 5000 % . Using the measured number of electrons 2 x 1013 per pulse, the calculated photon yield Is 0.8 x 1 0 ^ photons per pulse, which compares favourably with the estimated actual photon yield of 1.6 x lO1^ photons. 10. Conclusion In conclusion, the main•finding of this investigation has been that at least 90% of the light emission from water irradiated with an Intense pulse of high energy electrons can be attributed to Cerenkov radiation. The yield of the Cerenkov radiation was found to be Gh^(4-5000ft) * 6 x 10 53 This conclusion was drawn on the basis of; (i) the angu-lar dependence which Indicated that the light was predominately in the forward direction; (ii) the variation of the emission intensity with electron energy and the cutoff of emission at KV electron energy; (iii) the spectrum which showed the expected c / v s X variation; (iv) the effect of additives; (v) the comparison of the emission to emission from other liquids; and (vi) the photon yield which agreed with the expected Cerenkov yield. It is evident from this work that any light emission from water would have a yield of G h j / (visible) x 10"-5 and thereby indicate that it was the result of an unimportant radia-tion-chemical process. The conclusion is also In good agreement with the re-sults of previous work using different radiation sources with 20 the exception of the work of Sitharamareo and Duncan which it appears to contradict. Since they give only a very sketchy outline of their procedure for determining the emission spectra and photon yield, It is not possible to give any explanation of the conflict between this work and theirs. 11. Suggestions for Further Study The fact that water apparently emitts relatively little light, other than Cerenkov radiation, suggests that a very efficient means of energy degradation must exist in the liquid. Some suggestions for further study would therefore include an attempt to scavenge some of the energy by adding an efficient 54 energy absorber, such as anthracene, to the water; observing the fluorescence from ice or glassy water and possibly looking for fluorescence from water vapour, where Cerenkov radiation would be nonexistent because of the small refractive index. 55 REFERENCES 1. J.W.T. Spinks and R.J. Woods, An Introduction to Radia-tion Chemistry. John Wiley & Sons, New York, 1964, p.39. 2. J.W.T. Spinks and R.J. Woods,, op. oit., pp. 15-76. 3. I.V. Vershchinskli and A.K. PiKaev, Introduction to Radiation Chemistry. English Translation from Israel Program for Scientific Translations, Jerusalem, 1964, pp. 6-16. 4. G.J. Hine and G.L. Brownell, ed., Radiation Dosimetry. Academic Press, New York, 1956, pp. 153-528. 5. I.V. Vershchinskli and A.K. Pikaev, op. cit., p. 12. 6. A.O. Allen, The Radiation Chemistry of Water and Aqueous Solutions. D. Van Nostrand, New York, 1961. 7. E. Collinson, Annual Report of Progress in Chemistry (Chemical Society, London) 62, 79-105, 1965. 8. J.V. Jelley, Cerenkov Radiation. Pergamon Press, London, 1958. 9. L. Mallet, Compt. Rend. Acad. Sci. (Paris) 182, 274, 1926. 10. P.A. Cerenkov, Dokl. Akad. Nai*k, SSSR, 2, 451, 1934. 11. I.M. Frank and I. Tamm, Dokl. Akad. Nauk, SSSR, 14 (3), 109, 1937. 12. R.E. Jennings, Science Progress, ^0, 364, 1962. 13. G.B. Collins and V.G. Reillng, Physical Review, ^4, 499» 1938. 14. M.A. Greenfield, A. Norman, A.H. Drowdy and P.M. Dratz, J. Opt. Soc. Am., 41 (1), 42, 1953. 15. J.A. Rich, R.E. Slovacek and F.J. Studer, J. Opt. Soc. Am., 42 (9),750, 1953. 16. E.H. Belcher, Proc. Roy. Soc., A216. 90, 1953. 17. E.W.T. Richards, Proc. Phys. Soo., A6£, 922, 1953. 18. L.0. Brown and N. Miller, Trans Faraday Soc., £1, 1623, 1955. 56 19. E.W.T. Richards, AERE Harwell, C/R 1901, 195b. 20. D.N. Sitharamarao and J.F. Duncan, J. Phys. Chem., 67. 212b, 1963. 21. G. Czapski and D. Katakis, J. Phys. Chem., 2£» 637, 19&6. 22. D.C. Walker, Quart. Rev., 21 (1), 7 9 , 19&7. 2 3 . D.C. Walker, Private Communication. 24. 0. Oldenberg and F.F. Rleche, J. Chem. Phys., 485, 1939. 25. C.D. Hodgman, ed., Handbook of Chemistry and Physics. The Chemical Rubber Publishing Co., Cleveland, Ohio, 44th edition, 1962, p. 2b9«. 26. D.A. Head, Private Communication. 2 7 . E.J. Henley, Nucleonics, 12 ( 9 ) , b2, 195^. 2b. R.C.A. Phototubes and Photocells - Technical Manual PT-60. R.C.A. Victor Company, Ltd., Montreal, 19&3* 2 9 . A.M. Bass and E.G. Kessler, J. Opt. Soc. Am., 42, 1223, 1959. 30. C.A. Parker, Proc. Roy. Soc., A220. 104, 1953. 31. C.G. Hatchard and C.A. Parker, Proc. Roy. Soc., A235. 518, 195b. 32. K. Kawabata and S. Okabe, Osaka Furitso Hoshawen Chuo Kinkujo Saki, Japan, b, 71, 1965. 33. T.H. James, ed., The Theory of the Photographic Process, 3rd edition, The Macmillan Company, New York, 19&6. Fig. 2 \r r H0RIZ0NTAL=50n%ecydlv VERT ICAL : 100 amp/div. V (a) n H0RIZ0NTAL=50nsecVdiv VERTICAL =0.02 V./div. (b) H0RIZ0NTAL=50nsec/div. VERTICAL =0.05 V./div, II (c) Fig. 3 n c l l FARADAY CUP APERTURE Fig. 4 Fig. 5 (3) Photomultiplier power supply Photo multiplier CONTROL CONSOLE Water f low system ( WAVELENGTH (X) Fig.9 100 — 3| (O o R 8 0 E L A T V 601--E S E N S | 40j— I V I T V 2C — C> ( 4000 R e l a t i v e s e n s i t i v i t y of the spectrograph to an equal no. of photons at a l l -wave lengths i a o u o 1 p q-5000 WAVELENGTH (ft) 6000 Fig.11 Fig.12 - 8 0 ° - 6 0 ° - 4 0 ° -20' 9 n o 20 4CJ Fig.13 Fig. 14 6 CHARGING VOLTAGE (KV.) Fig.15 -m-i 1 0.05 0.10 _J 0.15 0.20 0.25 I .(l-l/£2n2) (mm.) Fig.16 T | (Q 3 4 0 0 0 4 5 0 0 5000 WAVELENGTH(ft) T j (Q 00 4 0 0 0 4 5 0 0 W A V E L E N G T H {%) 5000 5500 WAVELENGTH (&) 6000 5 0 0 0 WAVELENGTH (ft) •ADDENDUM The fact that the graph in FigJ6 appears to be a curve rather than a straight line is most likely due to an error in the expression used for the number of photons emitted by an electron. The expression derived in the theory of Frank 11 and Tamm , ; ' was obtained using the assumption that the electrons' velocity, is constant during its passage, through the medium. However, in actual fact. (3 does change markedly as the electrono penetrates into the water. This can be taken into account by using an expression for the linear energy .transfer to the medium by the electrons to relate (3 to 1 in the following manner: 2. dl = - ft dE K ' , , . where dE is the increment of energy loss.and K is. a constant. dE is related to d (3> "by an expression of the form: dE = „ o. where m0c is the rest energy of the electron. Hence the exp-ression for dl becomes: 4 1 = - P ^ T ^ V 2 a 1 3 Including this in the expression for the number of photons emitted per unit path; ( t - i X ' - ^ results in the .folloy/ing'expression':.' dN = AfClLll! J\(3> d(3 v-i-^rv v v v/here A is a constant equal to -Ik c2/ n2K V* 0 Integration of this expression betv/een the limits ; dN from 0 to N and d ^ from 1/n to ^  , gives an expression of the form: N= A -V Z w ( h .6-% A plot of N vs jiiLH-liX, •kZ^ O-fi) then gives a line which is o-c^y2-much more nearly linear than Figo 16. LIST OF CORRECTIONS Line Correction 5th last Z2Z2/m2 to read z2Z2/m2 1st extropolated to read extrapolated / 2/ P 3rd mg/cm to read mg/cm 3rd process to read processes polarization thermal. of E^O 6th last . expresses to read expressed 10th using sensitive to read using a . sensitive 5th last ; Kodac to read Kodak 1st this converted to read this was converted 2nd Kodac to read Kodak 2nd balnk to read blank 6th methanol.cyclohexane to read methanol , cyclohexane 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items