THE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING THE ENERGY DEPENDENCE OF X-RAY DAMAGEIN AN ORGANIC MERCURY COMPOUND ^,, ) Marvin Co Atkins A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in The University of Michi.gan 1961 February 1961 IP-493

Doctoral Committee: Professor Henry J. Gomberg, Chairman Assistant Professor Adon Gordus Associate Professor Terry Kammash Professor William Kerr Associate Professor Robert Ko Ritt -ii

ACKNOWLEDGMENTS The author owes a great measure of gratitude to many people. First it is a pleasure to acknowledge the help of the committee members, especially Professors Gomberg and Gordus, who followed the experimental work most closely. This thesis reflects many ideas which grew out of frequent and lengthy discussions with the author1s colleagues, W. R. Clendinning and A. H. Emmonso The assistance in the laboratory of Messrso Nasser Mahootian and Peter Paraskevoudakis was invaluableo Miss Phyllis Peters demonstrated in her help with preparation of the manuscript that she is a fine typist as well as a scientist. Many other members of the Phoenix Laboratory staff rendered expert assistance. The College of Engineering Industry Program prepared the final draft. Most important of all, the author thanks his wife and son for their patience and forbearance o The research reported in this thesis was supported by the U.S. Atomic Energy Commission under Contract AT(ll-1)-684 and by the Michigan Memorial-Phoenix Project. The author's studied were sponsored by the Air Force Institute of Technology. iii

FOREWORD An extensive development of irradiation techniques and x-ray calorimetry formed an integral part of the research reported hereo These developments were carried out in collaboration with W. R. Clendinning, and are reported in a separate publication entitled "Method of Monochromatic Irradiations and Absolute Dosimetry with Soft X-Rays" (Reference 9)~ The results of this joint effort are summarized very briefly in this dissertation, Reporting in this dissertation and in the appendices may seem at times to include somewhat more detail than is absolutely necessary. The author considers, however, that the purpose of this reporting is not only to substantiate the findings, but also to assist and encourage future investigators iv

TABLE COJF ONT1 P'age FA D'i * O C t * * * a * > a ~ a f a a a a d k a a a 4 i 4 a a J r*~ a i * * a a a # a * a,~ *4 * *~ a. * o 1 a iY;:..iFOEWq iv LIST OF fABtESI G** *. *4 4 4 a v ii LIST OF FIGURBS^ -^^. o < f. ei* o b4 ^ ** *****^ e~K *#*# <4~<~<> #* vi0 *I Tili I. GUE. *..... 1 II. THE THEORY OF THE EECTS OF IONIZING RADIATIO....* *.... 4 A4 Production of Ionizations and Eieitxtions o n......... 4 1] Initial Absorption of an Icident Photon * **** 4 2* Fluorescence Yield and the Auger Efect **...**** 12 5* Intzeractionis of Noving Eleetons4,4*,.***.*..** * 15 B- Sigfin O the Initial Photon AbsoQpi*. *: 19 Oe IsT-ependent Beaeso s*,* * *.... 2 l Tfypes of LETDependent Reactions..*,**., * *.. 28 2P Are There K-Edge Effects Due to LET Dependence?. 50 DB Literature Szrvey on Energy Dependence of Radiation EffeIiteg:t, I, 4a s^^^^ o~ 40.,..^. & 4 a ^ aa, s 3 2 III, EER WITH AX ORGATIC CERGURY CMPOUDI,..... * 40 A# Int+. ut***to,** 40 1 Selection of a System for Measurement of Energy Depe:ndencep,.* * * **.* *. * * 4 4 *0 * * f* 40 24 Literature Survey on Radiation Damage in Organoo metallies, 4444454.fi*+***..*.*...44.444 4.*44 42 B* Prela~niny Experm en ta ork.*4*: a..* * * 4 C. tebhngiques for Measurement of Small Quantities of Mercry Ptelplt * * * ** * ~,.. o*, **.. 46 ^ Statist!is of Measiwrement of the Ra4iation Yield.*.. 51 E* Ixad4ationss with Cobalt6 *,..,:bi~..,..., *., 55 YF Irradiations with XRays,*..,. ~.*,**...**J,** * 59 1* Use qf Xonochromatie Raliatioa+*,.****,*#*.***:. 59 2 Preodution o of Monobrcti Be ams*,........ 60 53 Application to Merciory pox ds~ i'.o..*. ** o*i**ia.o 62

TA3LE OF CONTENTS CONT'D P&ge 4, Techniques of Sample Irradiation..o*............ 68 51* The X-Ray Calorimeter*..***.... *****.***o~~o 72 6, Dosimetry Used in These Experiments.**..*.. *.*.. 74 G* Dose and Dose-Rate Dependence with X-Rays,*.,...*.. 77 H. Energy Dependence in the Soft -X-Ray Region*....,..... 77 IV. SUARY AND CONCLUSIONS**.............. ****o,.* 86 APPENDICES APPENDIX A* PROPERTIES, SYNTHESIS, AND PURIFICATION OF THE MERCURY DERIVATIVES OF CINNAMIC ACID, **~,*~,., ^ 93 APPENDIX B, PROCEDURES FOR TEE DETERMINATION OF SMALL QUALTITIES OF MERCURY WITH DITEIZONE....* o,*.o, 105 APPENDIX C, IHE DARKENING OF aMe4-M e4e-C CRYSTALS BY RADIATION,-**.*.A*>*.*v*. o **<**, A**, o. A**, 115 APPENDIX D USE OF ORGANOMETALLICS AS HIGH DOSE DOSIMEERS.. 131 APPENDIX E4 TABLES OF DATA*.. -o*.....,*.....*.... 135 BwEFECES* #*.*,*.* - *<.*****, * *+*~*-*.**.. 143 vi

LIST OF TABLES Table Page I The Most Prominent Emission Lines*,.................. 14 II Comparison of Experiments with Catalase and a-Me-Et-OAc....... ~. ~....... * ~.......... 90 III Emmons' Experiments on Catalase Using Bragg Spectrometer Radiation.......*,....... *.~**. *...... 91 IV Crystal Irradiation History,...................... 128 V Crystal Irradiation History....*... * * * o ~. * 128 VI Low-Dose Statistics Runs1.,.......... *o.......... 136 VII High-Dose Statistics Runs *.*.......... O 136 VIII Cobalt-60 Irradiations..*......................... 17 IX Dose and Dose Rate Dependence with X-Rays.......... 138 X Irradiation of a-Me-Et-OAc with Filtered K-Spectra... 139 XI Irradiation of a-Me-Et-OAc with Unfiltered K-Spectra. *,, *............................... * 140 XII Irradiations with UIxfiltered LtSpectra,......... *., 141 XIII Diphenyl Mercury Irradiations..... o...... 1...... 142 vii

LIST OF FIGURES Figure Page 1 Axes for the Energy Dependence Curve....... 1........ 1 2 Energy Level Diagram of an Atom Showing Strongest X-Ray Emission Lines (Source of Data: Sandstrom)..... 7 3 Total Mass Attenuation Coefficient of C4HL8O Hg from 10-100 key CQntribution of Mercury is Shown (Source: NBS 583 and Supplement).............. 8 4 Total Mass Attenuation Coefficient of C14H805Hg from 8-20 kevy 4 ***............................... 8 5 K and L Absorption Edges of the Elements in Kev*.* 9 (Source I Data: Fine and Hendee) 6Absorption Coefficient Ratios at the K-Absorption Edges..*................... I o 10 7 Absorption Coefficient Ratios at the L-Absorption Edges4*......,...... ~... *. ~.. ~.. * 11. * 8 K and Total L- Fluorescence Yields of the Elements (After Burhop, Figure 8 and Table VII),..;........ 16 9 Range of Electrons and Rate of Primary Ion Production Per Micron Path (Source of Data: Lea, Tables 10 and 11) 18 10 Instantaneous Linear Energy Transfer of Electrons in Water (Source of Data: Lea, Table 10)4*.**..... 20 11 Number of Ions Produced Per Primary Ionization in Water (Source of Data: Lea, Table 15 ), * X.o. * * 21 12 Low-Dose Statistics Testo........*.+ *.....o 53 13 High-Dose Statistics Test. o. J...... *oo..*..***** 54 14 Decomposition of a-Me-Et-OAc by Cobalt-60 Radiationo,. 56 15 Conversion from Weight of Mercury Liberated to Percent Liberated in a Uniformly Irradiated Sample*....*. * 58 16 Schematic of Fluorescence Irradiation Assembly*,..,. 61 17 Typical Fluorescent Radiators *.*.............., 63 Left to Right: Pt, Sb, Ge, Se, Bi, Os viii

LIST OF FIGURES CONT'D Figure Page 18 X-Ray Machine with Calorimeter in Position.., *. ee... 64 19 Assembly for Irradiation of Large Samples*....... 65 20 Total Mass Attenuation Coefficient of C14H1805Hg from 8-20 kv**a*.*.*ros*************-..* *..... 66 21 Plastic Sample Planchet and X-Ray Samples**....... * * * 70 22 Exterior of the Calorimeter Box+*....,,,,,, r,, 70 23 Internal Assembly of the Calorimeter..*....*........ 73 24 SPG-1 Counting Assembly with Lead Pinhole............ 76 25 Dose and Dose Rate Dopendence with X-Rayso.*......... 78 26 Energy Dependence with Filtered K-Spectra*.t........ 79 27 Energy Dependence with Unfiltered K-Spectrao......... 82 28 Irradiations of a-Me-Et-OAc With Unfiltered L-Spectra, 84 29 Energy Distribution of the L-Spectra................. 85 30 Calibration Curves for Measurement of Mercury with Dithizone,....,.,.,.....,......e...........o..,, 111 31 Probability-Chart Plot of Dithizone Optical Density Values from Nine Independent Determinations of Identical Aliquots (Background O.D, has been subtractded ) O,.. * * * & 0.. * * * * * * * 113 32 Isometric of Crystal Sample Mount (All dimensions in inchesp material is aluminum, screens are 4-40, brass)41........................ 117 33 Beckman DV Cells, Cell Holder, and the Crystal Sample Mounts:..... e *................ * 0 a.. 118 34 Optical Density of a Typical a-Me-Me-Cl Crystal Before and After Irradiations *,..*s. * *.. *.*.... * 122 35 Change in Optical Density of a-Me-Me-Cl After 5 Minute Irradiation (Same crystal as shown in Figure 33)*.,, 123 ix

LIST OF FIGURES CONT'D Figures Page 36 Change in Optical Density of Two a-Me-Me~-C1 Crystals by Irradiation (Exposure field: 200,000r, Gold L-spectrumi) #... *............ o 124 57 Continued Darkening of an aMe-Me-Cl Crystal after Irradiation (Exposure field: 179,000r, Zirconium K-spectr-Lumn). ~ o..... 126 38 Optical Density at Two Wavelengths of an a-Me-Mes-Cl Crystal Through Irradiation, Baking, and Ageing (Exposure field: 700OO000r, Thallium L-spectrumo o o., 127 39 Diphenyl Mercury Irradiations with Cobalt-60O..*.,* 134 x

I INTRODUCTION This thesis concerns an investigation into the extent to which the decomposition of an organic mercury compound by x-rays depends on the photon energy of the x-rayso An extensive development of techniques for irradiating with monoenergetic x-rays and for measuring the energy flux of the monoenergetic beam forms an integral part of the research. The purpose is to contribute to the extremely sparse literature on the energy dependence of effects of ionizing radiation. "Energy dependence" may be defined as the shape or slope of a curve on the axes of Figure 1 which is found for a given experimental system. Effect produced per unit energy absorbed Photon energy Figure lo Axes for the Energy Dependence Curveo The phrase "per unit energy absorbed" on the ordinate must be emphasized. In any experimental system which is thin to incident radiation, it would be expected that the amount of damage per unit energy incident would change with photon energy as the absorption coefficient changed. To eliminate any possible ambiguity it might also be well at the outset to differentiate "energy dependence" as used here from another important quantity which we shall call "total dose dependence." In graphing the latter quantity, photon energy on the abscissa of Figure 1 is replaced by total energy absorbed during the experiment.,1

-2The curve on the axes of Figure 1, which we call the energy dependence curve, is called the "action spectrum" by workers in the field of ultraviolet radiation effects. The amount of work of this type reported with the ultraviolet spectrum greatly exceeds the amount with the x-ray spectrum. In graphing action spectra it is common to plot the ordinate as effect per photon absorbed. We have chosen to adhere to the terminology which is most common in the literature concerning ionizing radiations. This thesis reports measurements of the energydependence of the liberation of free mercury from an organic mercury compound in the energy region of the L-absorption edges of mercury. There are at least three important reasons for measuring energy dependences and for making the measurements in this energy region: 1. There have been very few experiments on the energy dependence of ionizing radiation effects reported in the literature. More complete information may materially improve the still very incomplete understanding of radiation effects. 2. Not only have there been relatively few experiments on energy dependence, but also the techniques for making such experiments have not been well developed. Very few of the measurements that have been reported are really satisfying. 3. Until many more measurements on energy dependence have been made, there always remains the possibility of finding some new energy dependence not included in present

-3theories It might be expected that if such new phenomena did occur they might be in the low energy x-ray region, where absorption cross-sections of all the elements are changing rapidly and some elements have large discontinuities in absorption. The next chapter will review the current theories of radiation effectso Most of the discussion of energy-dependence experiments will be reserved for the succeeding chaptero

II THE THEORY OF THE EFFECTS OF IONIZING RADIATION The theoretical correlation of radiation-effects data may be broken into two subdivisionso The treatment of the first sequence of processes that occur in a damage event is purely physical. This sequence begins with the introduction of an ionizing photon or particle into an experimental sample and ends with the production of ionizations and excitationso The treatment of the second sequence of processes is largely chemical and covers the reactions of ions, raddtials, excited molecules, etc,, ending with the system in a new stable conditiono It may be safely said that the physical processes involve a relatively few phenomena most of which occur in all classes of irradiated systems and which are believed to be quite well understood. Knowledge of the chemical processes is not nearly so complete, howevero Some classes of systems have been studied very intensively, while very little work has been done on otherso The following discussion of these various processes will adhere as closely as possible to the chronological order in which they occur. The purpose of this section of the thesis is not, of course, to describe the interactions of radiation and matter comprehensively, but to provide abasis for interpretation of radiation damage experimentso Only the features of the basic phenomena which are of importance to later discussions and are not obvious from elementary texts will be emphasized. A. Production of Ionizations and Excitations lo Initial Absorption of an Incident Photon A sample of material exposed to x-rays or gamma rays may absorb energy from the beam by the three well known processes~ photoelectric,

-5pair production, and scattering, including Compton, Thomson, and coherento Since pair production does not occur with photons below 1.06 Mev, it is outside the scope of work reported here and will not be discussed further. The facts concerning Compton scattering which are of significance in our work may be summarized briefly as follows: lo The energy absorbed from the radiation field by first scattering processes is given by the product of an absorption coefficient and the average energy imparted to the secondary electron. This product has been tabulated extensively (39) Second scatterings are much more difficult to handle. 2. Compton scattering is not an important energy-loss mechanism below 15 kev in water or 60 kev in lead or mercury. 3. Photons of a given energy produce Compton electrons with a spectrum of energieso The range of possible electron energies, however, varies greatly with photon energyo All Compton electrons ejected by soft x-rays have very low energies. Compton scattering of soft x-rays results in very little energy absorption. Thomson and coherent scatter involve very little change in the photon energy, although, since the photon is deflected, there must in principle be some momentum transferredo The energy deposited in the medium is usually considered to be zero. However, it is of interest to note that it appears that the possibility of chemical effects of Thomson or coherent scattering on the scattering molecule has not been discussed in the literature

-6Photoelectric interactions account for most of the absorption of energy from soft x-rayso The entire energy of the incident photon is given to an orbital electron and the atom with which it is associated. There is no scattered photono The electron is ejected with kinetic energy equal to the difference between the photon energy and the binding energy of the electrono The recoil energy imparted to the atom is negligible for most purposeso A general energy-level diagram of the inner electron shells of an atom is given in Figure 2. The plot of the absorption coefficient of a mercury compound in Figure 3 clearly shows the absorption edges corresponding to these energy levels. While both electrons in the K-shell have the same binding energy, the L-shell has sub-shells of different energieso The same is true of higher shellso The L sub-shells are denoted in x-ray work as LI, Lii, and LIjI Each of these sub-shells has its own absorption edge9 as seen in Figures - fnd 4. As shown in Figure 5, each absorption edge moves to higher energy with increasing Z of the absorbero Because the inner electron shells are not affected by the valence shells, the points lie on a smooth curve ("Moseley's Law")o The values used in plotting these curves, like most energies of edges and emission lines used in this thesis, are from the very convenient compilation of Fine and Hendee (14) The magnitude of the discontinuity at the absorption edges will be of interest to us. Theoretical and experimental values of the K-edge discontinuities are plotted as functions of Z in Figure 6. Experimental values of the three L-edge jumps are plotted in Figure 7~ It will be noted

-7NAME OF SUBSHELL/ / / / ////////// / CONTINUUM OF UNBOUND STATES N (SEVEN SUB-SHELLS) M I~m= 1J^::::=:E= - - - - -::::EE TYPICAL RELATIVE INTENSITIES WITHIN EACH SERIES Ko 100 F u 2ELaeg LLev L. D g o a A S wKa2 50 LaL La,6 La p K,8 30 L-[ L — I - - La, 100 La2 II L,61 52 L/2 20 LY, 7 Kap Ka, K/2 Kf31 Figure 2. Energy Level Diagram of an Atom Showing Strongest X-Ray Emission Lines. (Source of Data: Sandstrom)

0.6 5 15 20 30 40 50 60 80 100 w ~ v< U. 0 0 Ci4 Hs A n C o HgC1 -— f Hg CONTRIBUTION (Source: NBS 583 and Supplement) 0.6 ---------------- I - \ --— I I II 10 15 20 30 40 50 60 80 o100 PHOTON ENERGY (KEV) from 10-100 Key. Contribution of Mercury is Shown (Source: NBS 585 and Supplement) PHOTON ENERGY (KEV) 10o8 o 9 10 11 12 13 14 15 16 17 18 19 20 sooL 90 i| 80- \Sf L8KE L'm'LI 70 LIk 760 I - \ -40 C14H1805Hg \\ Hg CONTRIBUTION Figure 4. Total Mass Attenuation Coefficient of C1~H~OcHg from 8-20 Kev.

200 100 w K- EDGE w w z 0 O- 10 0 co co i -L/ 1/ 111 LEDGE 20 30 40 50 60 70 80 90 Z Figure 5. K and LIII Absorption Edges of the Elements in Kev. (Source of Data: Fine and Hendee)

010~-0 9 - A 0 A 8 7 ~ A A 0 6 5 w w 4 * ALLEN REPORTED BY COMPTON AND ALLISON 3 A REPORTED BY McGINNIES 2 - I I I I 0 10 20 30 40 50 60 70 80 90 100 z Figure 6. Absorption Coefficient Ratios at the K-Absorption Edges.

-*115 * - LI EDGE A - Ln EDGE * - Lm EDGE ALL DATA FROM COMPTON AND ALLISON 03 I- * A 2 S*A AA AAA A I * * o ~ I I I I I 40 50 60 70 80 90 100 Z Figure 7. Absorption Coefficient Ratios at the L-Absorption Edges.

12 that there is a gradual drop in the size of the discontinuities with increasing Z The LIiI jump is nearly twice as large as the LI and LiI discontinuities, as would be expected from the numbers of electrons in these sub -shells It is apparent that there may be a major difference in number and energy distribution of secondary electrons across an absorption edgeo At energies just less than the K-absorption edger for example. most of the photoelectric absorptions will take place in the L-shell and will eject relatively high energy photoelectronso At energies just above the K-edge most of the photoelectrons will have very low energyo 2. Fluorescence Yield and the Auger Effect When a photoelectron is ejected, the atom. is left in a highly excited state, the excitation energy being the binding energy of the electron before ejectiono The atom may return to the ground state through either of two competing processes, or more precisely, through a large number of possible combinations of these processes The two processes are the emission of fluorescent x-rays and the emission of Auger electronso A vacancy in an inner electron shell may be filled. by a transition of an electron from a higher level accompanied by emission of a photon whose energy is the difference in binding energy of the two levelso This phenomenon is called "x-ray fluorescence", analogous to optical fluorescenceo The series of emission lines thus produced by photon or electron bombardment of a given element is called the "characteristic x-ray spectrum" of the elemento Since x-ray fluorescence is of great importance as a source of monochromatic radiation in this work, it w.ll be discussed here in some detailo

-13Let us say that an atom from which a K electron has been removed is in the K state, one from which an L electron has been removed is in III the LIII state, and so on, and denote by EK the binding energy of a K electron, etc. If a K vacancy is filled by a transition from the LIII level, the atom is said to undergo the transition K - LIII, and the photon emitted has energy EK - EL The x-ray lines of an element fall into distinct groups widely separated in energy. Each group comprises lines arising from vacancies in a particular shell. Since the L and higher levels have sub-shells with different absorption edges as shown in Figure 3, for each edge there is a group of emission energies any of which can be emitted if there is a vacancy in that shell. Although the x-ray spectra are simple, the traditional nomenclature is unfortunately complicated. The names assigned to the x-ray energy levels have already been given in Figure 2. The most prominent K and L emission lines are shown in Table I, Transitions between all levels do not occur, but are governed by certain selection rules. The figure shows that the systems K,, Kp, La and L, are close doublets. In the region 10 - 20 kev they are not resolvable in the first order with the spectrometer equipment in the Phoenix Laboratory and will be considered as single lines in this thesis. The Auger effect is a radiationless transition. The energy released by an electron transition to a lower orbit, instead of being emitted as a photon, can be transferred to another electron which is ejected. For example, the following change of states can take place, K -X LT LIIT

_14with emission of an electron of energy E = EK ELI - ELLII It is apparent that in this case it would be impossible to say which of the L electrons was ejectedo TABLE I TIE MOST PROMINENT EMISSION LINES Approx. Relo Energy in Energy in Line Transition Intensity Br (kev) Hg (kev) Ka1 LII - K 100 11.923 68.894 KC'v LI, <K 50 11.877 70.821 Kp MiI I K 113290 80.258 Kp MXIi <K J 13 ~465 82.526 Lai MV = LII 100 o1480 9,896 (Lo2 MIV ~ LI 11 1.480 9o987 LPi MIV - LI 52 1.526 11823 IL. NV Nv LI- 20 11o923'2 NIV LI, 7 13828 -^*-, N~y -^ 7 - 13o828 In visualizing -the energetics of the reaction, it is convenient to think of a fluorescent transition K -- L, with the photon ejecting another L electron. For this reason the process is sometimes called the "internal photoelectric effect " Burhop (4) however, points out that this cannot be an adequate explanationo Many Auger electrons are observed which would have to be ejected by photons arising from forbidden transitionso

-15a The relative probabilities of fluorescence and Auger transition may be determined. from the "fluorescence yield". This is defined as the fraction of ionized atoms in a particular state which will emit a fluorescent x-ray. A summary of measurements and calculations of fluorescence yields up to 1952 has been given by Burhopo (4) A more recent set of values is that due to Rooso(31) Accurate measurements are difficult to make, and disagreement between various experimenters has been wideo In general the fluorescence yield increases with increasing Z, and for a given element the K yield is much higher than. the L yield. A graph of the K and L-fluorescence yields of the elements taken from Burhop0s choice of the best data is shown in Figure 80 Figure 8 indicate that for an atom ionized in an inner level to return to the ground state entirely through x-ray emission is very unlikelyo Even if a K vacancy is filled. by an L electron with emission of a K, photon, it is much more probable that the L vacancy will result in an Auger transitiono For every element, with. the trivial exception of the lightest ones, there is an electron level for which the fluorescence yield is less than 0olo Thus, we may certainly say that most ionizations of x-ray levels result in at least some Auger electronso 3 Interactions of Moving Electrons We have shown that every mechanism for the absorption of photons results in the production of energetic electrons. To speak of the energy of the photon as being absorbedat the se of the t initial photon event is certainly inaccurate. In every case most of the energy must be carried

-161.0 0.9 0.8 0.7 0QK _J.6 - w 0.5 _h) n,, o 0.4 0.3 TOTAL L 0.2 0.1 0 10 20 30 40 50 60 70 80 90 100 Z Figure 8. K and Total L-Fluorescence Yields of the Elements. (After Burhop, Figure 8 and Table VII)

-17away and eventually deposited at other locations by one or more of the following media: 1. A scattered photon or fluorescent x-ray, 2o A photoelectron or Compton electrono 3. A shower of Auger electronso The scattered photon or fluorescent x-ray may of course be absorbed by processes identical to those for the incident photonso The fate of the Compton, photo, and Auger electrons will form the subject of this section. As a fast electron moves through a material, it loses energy almost entirely by production of ionizations and excitationso For electron energies below several hundred kilovolts, other processes are completely insignificant As the electron loses energy, its rate of energy loss becomes greater, or the ionization events occur closer together, until almost the end of its range. For our purposes the details of the physical processes of ionization and excitation are less interesting than the distribution of ion pairs along electron tracks and the energy dissipated per ion pair. Lea(25) has published a number of tables of various quantities concerning electron slowing - down which are of interest in radiation effects interpretation. These tables have been computed from the theory of Bethe,() and are stated by Lea to be in generally good agreement with the rather mea.er experimental data available. While they may not be absolutely accurate, they are at least self-consistent and of proper order of magnitude. Several graphs have been drawn from these tables to illustrate the phenomena involved in electron stoppingo In Figure 9 are plotted the range and the rate of production of primary ions per micron of path in tissue of density 1 gm/cm3. "Primary

-1810 - 1000 PRIMARY ION PRODUCTION / z 0 0.1 1' 10 ) 100 z - IO —' o - RANGE S a. w z C o 0.1 10- =E C_ 0.01I 0.1 I 10 100 ELECTRON ENERGY (KEV) Figure 9. Range of Electrons and Rate of Primary Ion Production per Micron Path. (Source of Data: Lea, Tables 10 and 11.)

-19ions" are understood to be ions produced directly by the fast electron, usually called the "primary electron", regardless of the energy imparted to the "secondary electron". The average rate of energy loss of the primary electron per unit path is plotted in Figure 10. This important quantity is usually called the "linear energy transfer (LET)". Its reciprocal, the "stopping power", is also frequently usedo A secondary electron having an energy of 100 ev or more is often called a "6-ray"o The maximum energy of a so-called 6-ray is understood to be half the energy of the primary electron before collisiono Figure 11 shows the ratio of ionizations produced by the 6-rays per primary ionization as a function of initial energy of the primary electrono The energy dissipated per ion pair produced is usually considered to be 30-35 ev in all systemso This value has been found to hold in a wide variety of substances which can be handled in the gaseous or vapor stateo Its measurement in condensed phases would be very difficult, and the same value is frequently assumed to holdo It should be noted that 30-35 ev is considerable greater than the energy required to form an ion pairo The difference is the energy which goes into excitations. Bo Significance of the Initial Photon Absorption We have seen that primary and secondary electrons form a mechanism for distributing the energy of absorbed photons among a number of moleculesof the absorbing mediumo It is now necessary to assess the relative amounts of damage caused by the electrons and by the initial photon absorption. We may start by assuming that a photoelectric absorption in an atom will disrupt the molecule in which the atom is located. Since in most

-20100 10 I- -J L It Il I 1 I!! 0.I 0.1 I 10 100 ELECTRON ENERY (KEV) Figure 10. Instantaneous Linear Energy Transfer of Electrons in Water (Source of Data: Lea, Table 10).

1.0 z 0 0 r 0.6 Q /: 0.2 0 0 0 0. 0.1 I 0 100 ELECTRON ENERGY (KEV) Figure 11. Number of Ions Produced Per Primary Ionization in Water. (Source of Data: Lea, Table 15).

-22chemical systems a single ionization has a high probability of causing a chemical change, the energy available when an Auger shower takes place within a molecule must make the probability of a change close to unity. The estimate we need of the number of molecules changed by electrons can be made very simply from reported values of the number of molecules changed per unit energy absorbed. These measurements are usually reported in terms of a "G value" which is defined as the number of molecular events of the type of interest produced per 100 ev absorbed, from ionizing radiation. To be completely specific, one should state the photon energy or spectrum with which the value was determined Now the G values for decomposition* of a wide variety of pure organic compounds are found to be in the range 1-10 (Reference 40) Let us assume some typical compound with a G(-M) value of 30 Let us further assume that a photon from a cobalt-60 source is intercepted in the materialo The interaction will almost certainly be of the Compton type, and the average energy deposited will be about 600,000 eVo Then the number of molecules changed will be 600~000 x 3 18, 000. 100 It, is apparent that all but the most minute fraction of damage events are caused by ionizing electronso Whether we assume the molecule in which the Compton scatter occurred to have been destroyed is seen to be irrelevanto We can now take the case of a 10 kev x-ray and a photoelectric absorption followed by radiationless transition. Whether the same G value will hold for these low energy photons is open to questiono Indeed one of * Frequently denoted by G (-M)

-23the primary aims of the project of which this thesis is a part is to throw more light on this question. Let us assume to make a calculation that the G value is also 3 at these low photon energieso Then the number of molecules changed per photon absorbed is 10,000 x 3 = 300~ 100 So again the overwhelming majority of damage events must be due to fast electronso The initial photon event would seem to be insignificant as a cause of damageo A discussion of whether the molecule in which the photoelectric absorption takes place would be destroyed by an absorption in one shell and not by an absorption in another shell is rather meaningless in this situationo We surely would never detect the differenceo In the common occurrence of "indirect" damage in aqueous solution, where the solute molecules are changed not by ionizing radiations but by reaction with radiolysis products of water, the insignificance of the primary photon event is if anything even more impressiveo Ro L. Platzman(30) has proposed that in the case of a system of very low G, where the number of molecules changed per photon (quantum yield) is of order unity or smaller, the location of the initial photon absorption might be important. Carrying his suggestion further, we must assume either that a K-photoelectric event has a much higher probability than another kind of disrupting that molecule, or else that the heavy atom constitutes a "sensitive site", or both. In the sensitive-site hypothesis, an initial event in the heavy atom has aha high probability of causing the measured change, while an initial event elsewhere in the molecule has a low probability of doing the same. We must also assume that the electrons from

-24either event have about the same probability of disrupting other moleculeso Then,we should see a distinct energy dependence across the K-edge of the heavy elemento The ratio of damage produced per unit energy absorbed above and below the K-edge can be readily estimatedo Assume that a K-absorption at a sensitive site will definitely destroy that moleculeo Let the following symbols be defined. F = fraction of photons absorbed in the K-shell of the sensitive site above the K-edge P = number of molecules destroyed by secondary electrons per photon, absorbed anywhere Q = total number of molecules destroyed per photon (quantum yield) Then Q(above edge) F + P Q(below edge) P, and the ratio of the damage is Q(above) l + F Q (below) P It is seen, that for the damage ratio to be even. as large as 2, P must be less than i1 or G(,-M)(10 kev) less than, 001, since F can never be greater than about 0o8o If the heavy element is present in small concentration, so F is small. then the G value must be extraordinarily small indeed for this effect to be noticeable. It is further apparent from the equations just given that the difference between the quantum yields above and below the edge must always be less than. Lo To the writer s knowl edge, this reasonable hypothesis has not been testeo by specitfic experiments. The difficulties of measuring such low quantum yields are rather formidable.,

-25For a K-edge energy dependence to be observed in our "typical" compound with a G value of 3 for cobalt-60, the G value would have to decrease by at least a factor of 100 at 10 kevo In summary, in a material in which the number of molecules changed per photon absorbed is much greater than one, there is no a priori reason to predict a non-zero slope in the energy dependence curve or a discontinuity at an absorption edge of a constituent, unless it is connected with the energy distribution of the electronso In making this statement we are of course assuming that there are no new, presently unknown phenomena. A quick calculation shows that any such phenomenon which caused a large change in total damage in a high-yield system would have to be responsible itself for a very extraordinary amount of damage. Let us assume a molecular system with the following characteristics: Total compound 1.1 Absorption 1.0 -- I X-Ray Coefficient I j Absorption |I Constituent X 0.1 I 0.01! 9.9 10.0 101l hv(Kev) 60 Molecules I Energy dependchanged per ence of damage photon absorbed 30 9,9 10.0 10,1 hv (Kev) * Over this small energy band the difference between damage per photon and per unit energy will be neglected for simplicity.

-26It may be noted this is a rather low-yield system, with G(-M) at 9~9 Kev of 0030 Above the K-edge of element X, about 10% of the total absorption is in X, while below the edge only 1% is in X. We will again assume that every photon absorption disrupts the molecule in which it occurs, and also assume that the change in electron distribution across the edge is insignificant The results follow immediately: 909 Key 10.1 Kev Molecules changed per photon absorbed 30 60 Molecules changed by electrons 29 29 Molecules changed by effect due to primary photon interaction 1 31 Molecules changed due to absorption in atoms other than X 0o99 0 90 Molecules changed due to L-shell absorption in X 0O01 0o01 Molecules changed as a result of K-shell absorption in X 0 30 The relative absorption in the K- and L-shells of X is.10ol Therefore the relative effectiveness of K- and L-shell absorptions in X for disrupting the compound. is K. 30 300 L (Oo0l) (10) From our hypothesis9 a factor of 2 increase in damage across the edge means that a K-absorption in X has to change 30 other molecules by some unknown mechanism The predominance of damage by electrons has led to the general conclusion that damage per unit energy absorbed. is independent of the photon

b27energy except in systems where the different electron energy distributions caused by photons of various energies can cause different amounts of damage. This conclusion has been succinctly phrased many times in the literature: "The chemical action of x-rays is exerted almost entirely through secondary fast electrons oo All ionizing radiations turn out to produce the large majority of activations through the same mechanism. Therefore all ionizing radiations produce, on the whole, strikingly similar effectso o All ionizing radiations have a potency (energy) well in excess of that required to produce any chemical activation, and the potency of each ionizing radiation is a negligible factor." Uo Fano.(13) "The atom from which the photoelectron or Compton electron is ejected is of course ionized; there is no reason, however, to attribute any special virtues to an ionization caused by absorption of an x-ray quantum, and there are very many more ionizations caused by the impact of recoil electrons or photoelectrons. Thus if an effect, e og breakage of a chromosome, occurs at a particular point in a cell there is no reason to attribute this to absorption of an x-ray quantum at that pointo" D.Eo Lea (25) "The chemical effects of high-energy radiation on. matter are due almost exclusively to electronso Differences in the effects are due primarily to differences in geometry of secondary electronso" MoBo Burton et al (5)

-28Co LET-Dependent Reactions 1 Types of LET-Dependent Reactions It appears from the foregoing analysis that if a radiationinduced reaction is conclusively shown to be dependent on photon energy, then either the dependence is explained by the energy distribution of the secondary electrons or is not explained by currently accepted theories of radiation effectso A series of experiments giving results which are not interpretable on the basis of present theory should not, of course, be summarily rejectedo Rather, there should be full understanding that such results are unexpected, and a new model should be soughto A number of reactions, including many important ones, have been shown to be dependent on secondary electron distributiono These are generally classified as "LET-dependent reactions" (The term "LET" was defined in II.oAo3) The study of LET-dependent effects has been facilitated by the wide range of LET values which. are fairly easily available from the minimum LET of about 05 kev/p for fast electrons to hundreds of kev per micron for alpha particleso In considering the effects of soft x-rays we are concerned with the LET values of electrons of energy less than 50 kevo Reference to Figure 10 shows that the LET of electrons increases rapidly as the electron energy drops below 50 kev, and very sharply below 10 kevo Thus it might be expected that LET-dependent reactions, if they exist at all, would be found with low energy x-rayso Indeed, they have beeno LET values used here are instantaneous rates of energy loss at the given energy, not track-average LET's for particles starting out at the given energyo The distinction is frequently very important but often not clearly drawno

-29Most of the LET-dependent effects which have been studied can be broken down into three classes: 1. Systems which show decreasing damage with increasing LETo ao Systems in which interaction between ions and excitons in adjacent clusters lowers the radiation yield if adjacent clusters are close enough togethero This is similar to the frequently observed phenomenon of decreasing damage per unit energy absorbed at high dose rateso The latter effect is due to track-track interactions; the former to intra-track interactionso Example: Many indirect or scavenging reactions, such as the ferrous-ferric dosimetero bo Systems in which there is local "saturation" of the effect for high LETo In other words, the destruction is already complete in the neighborhood of the track, so additional energy dissipation per unit track length simply wastes the additional energyo Example: Some chromosome breakages at high LETo 20 Systems which show increasing damage with increasing LETo These are systems in which a large amount of energy must be concentrated in a small region for an effect to be produced., Example: Some chromosome breakages at lower LETo Discussion of some experiments will be reserved, for the next section. The reader is referred to the paper by Fano(l3) and the book by Lea(25) for additional information

-3020 Are There K-Edge Effects Due to LET-Dependence? It appears that there may be an energy dependence across the Kedge of a heavy element that can be explained by difference in LET of the secondary electrons produced above and below the edge, and that might exist in systems showing any of the three classes of LET-dependenceo This change in effect produced per unit energy absorbed is basically due to production at energies just below an edge of very energetic photoelectrons, and to their absence just above the edge These photoelectrons can in some circumstances dissipate'a large fraction of the absorbed energy at a low LET rate, The Auger electrons produced in the cascade that usually follows ejection of a photoelectron (the photoelectron would necessarily be from the L-shell if the photon energy were just below the K-edge) would dissipate their energy at much higher LET. The data in Figures 9, 10 and 11 are pertinent in consideration of this effecto The strength of this effect might be greatly lessened by the Auger shower that frequently follows a K-shell absorption, assuming that we are talking about effects around the K-edgee If the K absorption is followed by a K =a LL Auger transition, the L-electron will be ejected with energy about the same as that of an L-photoelectron ejected just below the K-edge. Thus there would be little or no difference in effect. A detailed calculation of this effect would be very difficult because of incompleteness of knowledge of the Auger effect, particularly the number and energy distribution of electrons ejected from the outer shells. We may, however, describe the characteristics of a hypothetical system which should exhibit this K-edge effect most strongly~

-311o The system must, to begin with, be LET-dependent in the range 0o5 - 35 kev/.o It may be of any of the three classes. The required LET-dependence might be inferred from a known dose-rate dependenceo 2o The system must contain a "heavy" element -- one with a K-edge above about 5 kev (or Z greater than 23). 3~ It will be helpful if the element is heavy enough so the fluorescence yield is high, say Z greater than 4-0 This will greatly reduce the obscuring effect of K ->LL transitionso Escape of the fluorescence must be taken into account in computing energy absorptiono 4. The sample should be thick to secondary electrons but thin to the K-fluorescence radiationo Reabsorption of the Kfluorescence will obscure the effect by producing copious L-photoelectrons 5~ The system should contain a large quantity of the heavy elemento The larger the fraction of total absorption that occurs in the heavy element, the more pronounced the effecto The experiments of Seeman on AgBr photographic films, which are discussed more fully in the next section,may concern a K-edge effect of this typeo It seems probable that the film is a LET-dependent system of Class lb since a few ionizations may render a grain developableo The film fits the characteristics of a system which might show the effect under discussion as well as any other system which comes to mindo Seeman found a large decrease in darkening per unit energy absorbed for energies just above the Ag K-edge, and a smaller jump at the Br K-edgeo Unfortunately

-352 the fluorescence escape had not been taken. into account in computing the energy absorptionr and at least part of the reported. discontinuity is probably due to escape of fluorescence radiation.o Do Literature Survey on Energy Dependence of Radiation Effects The amount of reported research on energy dependence forms a very small part of the voluminous literature on radiation effectso The literature contains only a handful of x-ray energy depend.ence curves in. which a great deal of confidence can be placed., One of the major barriers to definitive work in this area has been the unavailability of monochromatic beams of su:fficient intensity Many of the earlier energy dependence experiments used the total tube output, filtered to remove very low energies 9 with t;he output spectrum altered by adjusting the tlube voltage to different valueso This is a relatively crude technique Any di.scontinuities of damage as a function on energy would. re-!maidrn undetected, and indeed no energy-dependence curve could be drawn. Most of the remaining energyd.ependence experiment;s have used as 7"monochromatic" beams the characteristic spectra emitted, from various x-ray tube targetso The tube emission was filtered to remove the Kr emission lines and, as well as possible9 -the continuous spectrum on which the characteristic lines are superimposedo There are two major li-mitations on this techniqueo One is that only a few elements are suitable for use as x-ray tube targetso Therefore9 the number of energies available is smallo The other is the impossibility of filtering the continuous x-ray spectrum so that its corntribution is only a small fraction of the energy

-335 flux of the characteristic line that is being used, Any "monochromatic" radiation produced in this way will be heavily contaminated, with a continuous backgroundo The other limitation on earlier experiments, intimately connected with the first, is dosimetryo Most experiments have used some device which measured roentgenso This can give a good measure of the energy flux if the radiation spectrum is accurately known. With soft x-rays, a roentgen measurement to determine energy flux is completely meaningless unless the beam is absolutely monochromatico Suppose, for example, that a beam of pure 40 kev radiation gave a certain reading in roentgens. If the beam were altered so 1% of the energy flux were due to 5 kev photons, the roentgen reading would increase by a factor of sixl These difficulties are discussed further in Reference 9o These considerations indicate that many of the earlier experiments on energy dependence must be viewed skepticallyo The research group in which the author worked is better prepared for the production and measurement of monochromatic beams than any other group of which we knowO Some effort might profitably be devoted to checking some of the published experiments which indicated interesting resultso Most of the data on energy dependence to be found in the literature concern biological systemso In his recent thesis Ao H. Emmonsk12) has discussed a number of these experiments, with particular reference to those on enzymeso It would be pointless to list those articles again hereo Instead, mention will be made only of such. additional pertinent information as has been foundo

-34Fano(13) has discussed briefly the problem of energy dependence from a very basic viewpoint. He emphasizes that more information is badly needed and concludes that there is probably an energy dependence in many systems below 50 kev and particularly below 10 kev because of the rapidly changing LET of the secondary electrons. Lea and Catcheside have measured the energy dependence of the (6,25,26) inactivation of microspores of Tradescantia ) They found a sharp rise in effectiveness of radiation per roentgen exposure dose as photon energy decreased from 8 kev to 3 kevo When the energy was decreased further to 1l5 kev, the effectiveness fell almost to zero. They developed a semi-empirical model to explain these effects. The basic assumption was that a chromosome would not be broken unless an electron produced a certain number of ion pairs in traversing the chromosome. In other words the LET would have to be larger than a certain minimum value. This explains the increasing efficiency with decreasing photon energy. On the other hand, photons lower than some energy would produce electrons so weak they would be stopped before the required number of ion pairs was formed, although the initial LET was very high. This is an example of what we have called a LET dependent effect of class 2. There has been a relatively large amount of work on the LETdependence of indirect reactions in aqueous solution, These are reactions in which most of the damage to the solute is caused by reaction of solute with the radiolysis products of water. These reactions are of great practical importance in vivo, in study of biochemicals in vitro, and in such systems as the ferrous-ferric dosimeter. We cannot begin here to make an adequate survey of this large field and shall not attempt to since the work

-35is not directly applicable to this thesis. Since other areas of work in our project do involve aqueous solutions, however, some typical results will be briefly summarized. In general it is found that ionizing particles of the same LET produce the same effect per unit energy deposited regardless of the nature of the particleo Such differences as have been reported are often ascribed to the difficulties in accurately correlating dosimetry of different particles. It has been found that there is a strong LET-dependence of most indirect reactions between track-average LET values of 3 and 100 kev/(, and particularly between 6 and 25 kev/t. In terms of monochromatic x-rays to produce photoelectrons in water with these LET's, the values are from 10 down to 0ol kev for some dependence and from 5 to 1 kev for strong dependence. It is in this region that radical recombination in water becomes very significant and the relative proportion of the radiolysis products of water is altered. For example, in a review article on this subject, Gray(20) reports that the inactivation per unit energy absorbed of the enzyme carboxypeptidase changes by about a factor of 3 in the critical region, with higher LETYs (lower photon energies) being much less effective. A similar drop is observed for oxidation of ferrous sulfate in the Fricke dosimeter. On the other hand, the reduction of eerie ion in the ceric-cerous dosimeter is increased with increasing LET because the reduction is due to reaction with a different product in water. For additional information and references see the review by Gray and other reviews by Miller,(29) Dainton,(ll) Fano,(l3) and Lea.(25)

-36The work reported in the literature that is most closely comparable to the work we are doing is that of Seeman on photographic films (36) Seeman irradiated two commercial x-ray films at 11 different energies between 6o0 and 78o4 kevo His monochromatic beams consisted of the filtered K-spectra of fluorescent radiatorso For dosimetry, the beam strength was measured in roentgens with an. ionization chambero The roentgen values were converted to erg/cm2 sec by an appropriate formula~ Seeman reported that from 6-18 kev the film blackening per unit energy absorbed. decreases sharply with increasing photon energyo From 18-78 kev the effect per unit energy absorbed increaseso Superimposed on these trends are a pronounced drop in effectiveness just above the silver K-edge and a much smaller drop above the bromine K-edgeo Unfortunately the energy absorption was computed, strictly from collision cross-sections without taking fluorescence escape into accounto Energy contained in the fluorescence escaping from the films should not be counted as absorbed in drawing a curve on the axes of Figure lo The fluorescence escape from these thin films might have been significant, and, could explain all the trends except for the decreasing efficiency with increasing energy below 18 kevo On the other hand. a question can also be raised about possible contamination of his beams with L-spectra at the higher energies used. The L-spectra of tungsten and platinum are quite penetrating, and the filtration used appears to have been inadequate to block them. If there were L-shell radiation present, the dosimetry system would have overestimated the energy flux, and the energy dependence at high energies might be even more pronounced than the plots indicatee

-37It might be that the answers to these questions would show that the reported curves are essentially correct and that the energy dependence can be explained by the distribution of electron energies at the various photon energies used. The fact that such crucial questions can be asked about an experiment that was on the whole carefully done points up the great care that must be exercised in work of this type. In defense of Seeman's article, it should be said that the curve of effect produced per roentgen incident was his primary objective and that fluorescence escape does not enter into this determination. Seeman's work was of great help in pointing us in the direction of using x-ray fluorescence as a monochromatic radiation source. We now come to a discussion of the recent work by Garsou and Emmons in our group which bears most directly on the research reported in this thesis. Garsou s experiments(18) concerned polystyrene films containing various bromine and iodine substituted hydrocarbons together with a dye-forming reagent. Absorption of radiation produced a pronounced color change in the films with extremely small doses. All irradiations used a Bragg crystal spectrometer as a radiation source. Two methods of irradiating the films were used. One consisted of cutting the film into small strips, each strip being irradiated on the spectrometer at an energy given by the Bragg energy-angle relationship. The other method involved using a long strip of film with the spectrometer crystal and collimator being swept back and forth to expose each section of the film at a different photon energy. In spite of the low flux available with the Bragg spectrometer, Garsou was able to make these experiments because of the extraordinary

-38sensitivity of his films. He was handicapped, however, by the small sample size that could be used and by the poor energy resolution it was necessary to go to in order to get high enough flux even for these filmso His experiments indicated that there was a drop in effect per unit energy absorbed at energies just above the K-edge of bromine and iodine, and that there might be an increased effect at energies corresponding to the K-emission lines of these elementso This last effect in particu-lar would be a most surprising phenomenon, entirely at variance with accepted ideaso It may be stated, however, that this effect has not conclusively been shown to existo The jump at the K-edge is qualitatively the same as that found by Seeman in silver bromide films and may also be due to fluorescence escape. Ao Ho Emmons has studied the energy dependence of the deactivation of catalase, an enzyme containing a very small amount of iron, at photon energies in the vicinity of the iron K-edgeo(12) The x-ray work consisted of three separate series of irradiationso The first series consisted of irradiating small amounts of a very dilute solution of catalase on the Bragg spectrometero The high. percentage decomposition of catalase in very dilute solution combined with the ability to measure the concentration accurately yielded another very sensitive system. The kinetics of the decomposition are similar to those of the ferrous-ferric dosimeter Emmons found that the inactivation per unit energy absorbed is much greater at energies just above the iron Kedge In order to obtain higher dose rates, Emmons performed a second series of irradiations of dilute catalase solutions with fluorescence

radiationo The unfiltered K-spectra of iron, nickel, and. manganese were usedo The nickel spectrum lies above the iron K-edge and the spectra of iron and manganese below it. The results showed that the nickel radiation, above the iron K-edge, was a factor of lo7 more effective than manganese radiationo The iron radiation was by far the most effective of all, reminiscent of the emission line effect suggested by Garsouo All these effects in dilute solution are particularly hard to understand -when it is realized that the maximum fraction of radiation absorbed in catalase was 3 x 10-5 and the maximum absorbed in iron about 10 -6 The third experiment consisted of irradiating dry catalase with fluorescence radiation from iron, nickel, manganese, and chromium. The chromium K-spectrum, like the spectra of manganese and iron, lies below the iron edge. The results were somewhat different from those found with solutionso Samples irradiated with nickel and chromium radiations showed damage from the bIeiginning of irradiation, while those exposed with iron and manganese required a sizeable activation dose before any damage was noted. After this activation dose was reached, however, the.destruction was much more rapid. No significant difference was found between samples exposed to iron -and manganeseo These effects are probably related to the very imperfectly understood chemistry of catalaseo The work of Garsou and Emmons has recently been reviewed in more detail in a report written by H, J, Gomberg(l19)

III. EXPERIMENTS WITH AN ORGANIC MERCURY COMPOUND A. Introduction 1. Selection of a System for Measurement of Energy Dependence At the time work was begun on this research reported in this thesis, two other research problems were under way in the AEC Resonance in Radiation Effects program. The experiments of J. Garsou on films containing organic halides(l8) were nearing completion, and A. H. Emmons had recently started his work on the enzyme catalase.(12) From the experimental standpoint, it was desired to study an entirely new system in order to obtain basic information on a completely different radiation effect. At the same time it was realized that the types of experimental systems that could be studied with the available sources of monochromatic radiation and existing methods for their use were very limited. In addition, it was anticipated that a major effort might be necessary toward development of techniques for performing the irradiations. With these factors in mind, the following criteria were established for selection of a radiation effect whose energy dependence was to be studied: 1. The system must be sensitive enough so that measurements could be made of damage produced in some reasonable time by the radiation doses available. 2. The system must include an element with an absorption edge in the energy region where monochromatic beams are available. 3. The system should not include a solvent which might lead to indirect effects obscuring energy dependent effects. -40

-4140 If more than one compound is to be involved, the reactions should be simple and well known to allow fuller interpretation of dat;a 50 Precise measurement of absolute yields is not required, but they should be measurable at least to ord.er of magnitude as this is important for interpretationo 60 Since long irradiation times might be required, the system should be suitably free from outside influenceso The system selected was measurement of the liberation of free mercury by radiation from an organic mercury compoundo This reaction was originally reported by Go Lo Clark and co-workers (8) The reaction was not studied extensively by Clark, but was simply catalogued as part of a broad, survey of radiation damage. To the writer's knowledge, this thesis reports the first study of this or similar reactions since ClarkVs paper in 1930o The information contained in the original paper may be summarized as followso io Decomposition by x-rays was noted in. the compound a-acetoxymecu - P-methoxy hydrocinnamic ethyl ester No decomposition was found in several other aromatic mercury compounds tested, 2. "The yellowish powder rapidly changed color in a few seconds upon exposure to x-rayso" This exposure was made with the direct white beam from an x-ray tube. (It may be noted that this compound is completely colorless when pure ) 3. Decomposition yield was measured by irradiating an ethyl acetate solution of the compound with a white x-ray beam. The mercury liberated was measuredL by picking it up on a weighed gold, foil dropped, in the solutiono

[424o The numerical results given permit rough calculation of the yield as about 60 atoms of mercury liberated per 100 ev absorbed. (It has been found in this study that this estimate of the yield is over a factor of 10 too higho Whether the discrepancy is due to dosimetry or impure preparation cannot be determined ) A study was begun of the energy dependence of the decomposition of a-acetoxymercuri - P-methoxy hydrocinnamic ethyl ester, and the results reported in this thesis have been obtained. Decomposition has been measured in the pure material in crystalline formo This chemical system satisfies the criteria outlined aboveo It involves the direct measurement of the decomposition of a pure material containing a heavy element and irradiated in the dry state. The absolute value of the radiation yield is determined. This is the first direct measurement of the decomposition product of a pure material undertaken as a part of the Resonance in. Radiation Effects programo 20 Literature Survey on Radiation Damage in Organometallics In addition to the literature search on energy depend.ent radiation effects,'it was of interest to survey the literature for previous work on radiation effects in organic mercurials, and organometallics in general. It was desired both, to find what interpretation had been placed. on experiments and to see what experimental methods had been usedo The literature search was virtually fruitless on both countso The overwhelming interest in radiation chemistry during the last fifteen years has been in aqueous solutionso Broadway(2) has published a brief survey on radiation damage in organometallicso The only information given concerns

-43engineering properties of some organometallic polymers. No yields of reaction products are stated. None of the references cited by Broadway yielded any applicable information. Heal(22) has published a review on the chemical effects of radiation in solids. He discusses the paucity of information on effects in organic solids, except for data on engineering properties, and does not give any references on organometallics. Heal's comments on the interpretation of radiation effects in organic solids are, however, of interest here. He states, "A polyatomic molecule can be excited in so many different ways that it is usually quite beyond the power of the present day theory to give a detailed description of the excitation produced by ionizing radiation." He goes on to say that the most fruitful approach in this area has been the "statistical" one. We quote from his review as follows: "The foundation of this method is the following known facto Each 30 ev of energy produces, on the average, one ionization and a couple of excitations; irrespective of the type of radiation and the substance (apart from minor differences).. It is expected that each such event (other than low - level excitations) would cause the rupture of one bondo Since the energy available in most such events considerably exceeds all bond energies, all bonds are expected to stand a roughly equal chance of being brokeno Thus the'ideal' behavior for a complex organic molecule on irradiation would be completely random bond fracture, with about 20-30 ev consumed for each bond brokeno Much of the work in this field has aimed at finding out how closely real substances correspond to this ideal, and explaining why deviations from it occur." B. Preliminary Experimental Work It was decided to study the energy dependence of the release of mercury by radiation from the compound a-acetoxymercuri - P-methoxy hydrocinnamic ethyl estero The properties of this and some related compounds are discussed at length in Appendix Ao This compound, which will be referred

_44to hereafter as a-Me-Et-OAc, is not available commerc.ially and was synthesized. in the laboratory. A considerable amount of work has been devoted to the synthesis and is reported in Appendix Ao The initial preparations yielded a pure white product which it was found could readily be brought out of solution in large, clear crystalso Exposure of the crystals in the full beam of the AEG-50 x-ray tube operating at 50 pkv, 50 ma confirmed Clark's observation that the material does i'ndeed change color in a few secondso The colorless crystals rapidly take on a grayish-brown color and will become very dark in a few minuteso Solutions of a-Me-Et-OAc prepared according to Clark's formula as well as much more concentrated solutions were irradiated at an exposure dose rate of about 5 x 105 rad/hr in the Phoenix Memorial Laboratory cobalt-60 source. 1The cobalt-60 source allows simultaneous irradiation of a large number of samples at a very high dose rate compared to the dose rates available from monochromatic x-rays. It has been invaluable in the preliminary workL associated with the projecto It was observed that after a few hours of irradiation a finely divided, dark gray precipitate sett-led. to the bottom of the tube containing the solution It was also found, that a similar quan.ttity of the powder, irradiated, dry and. then dissolved. would give a similar precipitateo An attempt was made to repeat Clark's method, of measuring the amount of mercury liberated by dropping a weighed. gold foil into the solutiono Bits of 1 mil gold foil measuring about 1 x 2 mm were usedo Despite the fact that the amalgamation o of 0m0n of mercury was sufficient to cause the gold to turn white, no success whatever was achieved in picking up quantities of mercury of the order of magnitude that Clark reportedo Not until enough

-45 precipitate was obtained so that some of the gray powder coalesced into bright droplets was there any amalgamationo This required some 100 mg of precipitate, and amalgamation was still incompleteo The following variations were persued2 lo Thorough cleaning of the gold with boiling HClo 2. Acidification of the organic solution with acetic acido 3~ Centrifuging of the solution with the goldo 40 Placing the gold in the solution before irradiation. 5. Decanting off the solvent and replacing with water or dilute acid. 6. Allowing precipitate to fall on gold and then dryingo 7~ Use of zinc instead of goldo 80 Cleaning the precipitate with boiling stannous chloride solution. These attempts were unsuccessful and. the author still has no idea how this technique can be made workableo At the same time this work was going on, samples of solid Q-Me-Et-OAc were being irradiated with monochromatic beams, At that time our only equipment for obtaining monochromatic x-rays was the Bragg spectrometer. It was found that even after 100 hours irradiation on the spectrometer there was no precipitate of mercury nor any visible change whatever in the crystals. The radiation input was entirely inadequate. It was decided to measure the light transmission of large crystals of the compound to find whether there might be a peak in the absorption spectrum that altered significantly after small doses of radiation. It had been observed that crystals showing an easily visible color change gave no

646visible mercury precipitate when dissolvedo It was hoped that the sensitivity of the crystal darkening coupled with the accuracy and wavelength selectivity of spectrophotometric measurements might allow measurements of some change in the crystals with. very small radiation doseso It was planned to correlate this change with, the liberation of mercury if possibleo To this end, apparatus was constructed. to permit measurement of optical density of the crystals in the Beckman DU spectrophotometer. It was found that no change could be detected after irradiations of the order of 100 hours on the spectrometero We then began developing the use of x-ray fluorescence as a monochromatic radiation source~ Optical density changes could, be measured, after a few hours irradiation with fluorescence beams. However, the magnitude of these changes was found, to be very dependent on other influences, particularly temperature With the further development of more in-tense monochromatic radiation beams it became possible to return to direct measurement of liberated mercury, The e work on. crystals was dropped because direct measurement seemed, to fit better the goal.s of this research effort However, it is believed that the results obtained with crystals, and particularly the techniques employed, are of some interest, and they are discussed further in Appendix Co C Techniques for Measurement of Small Quantities of Mercury Precipitate The first measurements of the yield of mercury from irradiation. of dry a-Me-Et-OAc were made by vaporizing onto a weighed gold foil the precipitate obtained when. the compound was dissolvedo The dissolved. compound was centrifuged in 15 ml tubes an.d the precipitate washed several timeso The solvent was then allowed to evaporate to drynesso Since the amount of

[47precipitate was so small (less than 1 mgo in most cases), removal to another vessel was impossible. A cold finger arrangement was devised, to hold a piece of gold foil in place within 1 inch of the bottom of the centrifuge tube while the tip of the tube was flamed to vaporize the mercuryo The foil covered, the end of the cold finger and sealed fairly tightly against the the walls of the centrifuge tubeo Foil of 00002 inch thickness was used. to keep the weight low. This is the thinnest foil that has enough mechanical strength to be usableo The cold finger around. which the foil was wrapped kept the foil below the boiling point of mercury. The mercury amalgamated with the foil instantlyO The foil could be weighed with very good accuracy on a Cahn Electrobalance However, after a number of tests it was decided that this method was not satisfactory for the work at hand.. Tests made by placing a known quantity of metallic mercury in the centrifuge tube showed that about 25% of the time the measured quantity was much too low, indicating that some mercury vapor had escaped without amalgamatingo Once the precipitate from a sample had, been vaporized and lost, there was of course nothing to do but start over. The problems of adapting to micro work the vaporization. method which is standard for gram quantities of mercury, (42) seemed severe. The final measurements reported in this thesis were made by dissolving the mercury in acid and determining the mercury present in an aliquot of solution by a colorimetric measurement with dithizone. Use of dithizone is a fairly involved procedure, and discussion of the techniques is reserved, for Appendix B. We shall describe in this section only the laboratory procedure for obtaining the solution containing mercuric iono It may simply be

-48stated that a 10o aliquot of a solution containing 1 mg mercuric ion has been determined, with a standard deviation of less than 2% in nine runs The dithizone method provides the great advantage that another aliquot of the same sample can always be run if desiredo This becomes very significant when a single sample takes two days or more of x-ray machine time Most of the samples irradiated were of one of two typeso Those exposed to x-rays were flat plastic planchets filled with the finely powdered UO-Me-Et-OAc These will be discussed later in more detail Those irradiated in the cobalt source consisted of the same powder placed in 3 ml Pyrex test tubes and corked. The following procedure was developed and used for all samples Reagents and Apparatus Acid digestion. mixture consisting of 130 ml concentrated nitric acid and 110 ml. concentrated sulfuric acid carefully mixed. Three polyethylene wash bottles containing respectively absolute alcohol, benzene, and double distilled watero 12 ml heavy duty Pyrex test tubeso Suction flask connected to lab vacuum systems with long, finetipped glass nozzle attached to plastic tube connected into top of flasko This was always used for removing the supernatant liquid from the precipitate after centrifugingo Centrifuge (An International Model 1-SB was used in these experiments ) Procedure Lo Have centrifuge tubes clean) dry, and dust-freeo Remove sample to centrifuge tubeo If a test tube sample^ dump powder into

-49centrifuge tube and. wash out test tube into centrifuge tube with alcohol. If a planchet sample, carefully remove powder to glassine powder paper with nickel spatulao Then slide powder into centrifuge tube, Wash off powder adhering to paper into tube with alcoholo Do not wash off plancheto It should be unnecessary and may get dirt into sampleo 2o Fill tube to about 2 inches from. top with alcoholo Place tube into water at 50~Co Stir contents of tube with nickel spatula until dissolvedo Rinse off spatula into tube and fill to about 1/2 inch from top 3~ Centrifuge for 2 hours at 3100 rpm. 4, Carefully remove the liquid with the suction tube down to 1/8 - 1/4 inch above precipitateo With a 1 mg precipitate, this leaves about 0,25 ml liquid. 50 Fill with alcohol to about 2 inches from topo Stir with spatula. Rinse off spatula into tube with alcohol and. fill to marko Stopper and. centrifuge 1 hr at 3100 rpm6, Repeat steps 4 and 5~ 7, Repeat steps 4 and 5, except use benzene instead, of alcohol. 80 Repeat steps 4 and. 5, Use alcohol. Before stirring place tube into beaker of water at 70 ~C 9, Repeat steps 4 and. 5, using double distilled water instead of alcohol, Squirt water directly into bottom of tube to insure mixing with the remaining alcohol, Do not stir, as precipitate is likely to stick to spatula due to surface tension of watero Place tube into beaker of water at 100 ~C.

-50that has just been removed, from heato Leave for 5-10 minuteso Stopper and centrifugeo 10o Remove water with suction tube down to 1/8 - 1/4 inch above precipitate. Add 5 ml of the acid digestion mixtureo Allow to sit several hours. 11o Boil very carefully over microburner until nitric acid is all driven off 12o Cool and pour through small funnel into 100 ml volumetric flask. Rinse tube into flask with double distilled water several, times and dilute to marko The sample is now ready for determination of mercury by the dithizone test described in Appendix Bo Remarks lo A brief recapitulation of the procedure may be in order: ao Dissolve sample in alcoholo Centrifugeo bo Wash with alcoholo Centrifugeo Co Wash with alcoholo Centrifugeo d. Wash with benzene. Centrifuge. e Wash. with alcohol Centrifugeo f. Wash with watero Centrifugeo go Add. acid and. boilo Dilute to 100 mil 2o As a general precaution, the steps down through removal of the supernatant at the end of (b ) above were carried through as rapidly as possible After that there is no danger, for example, in letting the samples sit overnighto The precipitate must not

-51be allowed to dry out at any time There can be noticeable evaporation of finely divided mercury in a few hourso 3~ When. the precipitate is washed with water it is recommended. that it not be stirredo The water should be squirted into the bottom of the tube, however, to insure the alcohol is mixed with ito Otherwise there may be a violent reaction when acid is addedo 4o When the supernatant is removed, about 025 ml remains of the original 10 ml or moreo This provides a dilution factor of 40 in the next stepo With five additions of organic solvent, the fraction of soluble compound remaining is'then 40-~5 or about 1080o The wash with water is used principally to provide a bridge between organic solvents and. acid., 50 The usual precautions with such. a toxic material as mercury must of course be observed, D. Statistics of Measurement of the Radiation Yield. The major problem encountered, in. the chemical work has been reproducibility of radiation yield measurementso The problem has been. difficult because of the very low yields that must be measured in samples exposed to monochromatic x-ray beams even wi.th the use of fluorescence rad.iation A factor of 10 higher x-ray output would have simplified the situation enormouslyo In a typical monochromatic irradiation lasting two days the yield of free mercury is only 0.05 - 0.1o% of th.eamount of mercury presento In absolute numbers this is 002 O- 04 mg mercury from the sampleo

-52We shall not be concerned here with errors in the dithizone measurements. These errors are discussed in Appendix B and are small compared with the overall errors in measuring low-yield samples. Efforts toward improving accuracy of measurement of these small radiation yields have been directed primarily along two lines: 1o The process of separating the liberated free mercury from the undamaged compound and soluble damage products. The procedure described in the preceding section is the result of experience and is rigidly adhered to for best results. 2. Purification of a-Me-Et-OAc to insure reproducibility of large radiation yields and to reduce to a minimum the "background yield" of unirradiated sampleso The purification process described in. Appendix A has reduced the amount of mercury centrifuged out of an unirradiated sample to 0.04 - 0.06 mg per gram of compound. Efforts to reduce this background still further have not been successful. To pin down the statistics of yield measurements, groups of nine samples in. test tubes have been irradiated to two different total doses in the cobalt-60 source. The samples weighed about 0.5 gm each. The doses were selected so that the mercury yields would represent approximately upper and lower limits of yields obtained in x-ray runs. The results are plotted on probability charts in Figures 12 and 13. Distributions are seen to be approximately Gaussian in each caseo The standard deviations of the measurements were 14% for a mean yield of 0.070 mg per 1/2 gram sample and 4t% for a mean yield of 0.70 mg. These values include the statistical deviation of the

-53z X U.O 00 20 I < - 0> \ nI 3E oJ 0 o 0 O d o 05dWS 0 w 0 an 2 0 C -, ~ ": d ^<~~~~0 00 3\IdWVS'VIO o9'0 83d ow _o ON oQ 0o O r 0. 0 0 000 c d d 0 31dl4IVS' VO 9'O 813d'gl~l -131A

-540) Ot \z I4 \\~ ~ " " e~~~~LL \ ~ ^ No w! Om 0 Z o0 o \ Q M 4 —,4l: 0 p ) C]Q~~~~~ ~ o 0o o o o o ~ 33dWVS'"9St0 QC C;W3 Z Co 0r-i 0 cr, t u0.H C |^ ~^ h <~ (0 (0 (0 d d d d d d 6 d 31dlNVS'H9 9O:3d'9i 0131A

-55dithizone measurements. A constant background found by averaging several unirradiated samples from the same batch of material was subtracted from the values before plotting, so the charts indicate what one would actually measure as the radiation yield. The data are presented in tabular form in Tables VI and VII in Appendix E. E Irradiations with Cobalt-60 A number of samples of O-Me-Et-OAc have been irradiated in the cobalt-60 source to check linearity of damage with total doseo The results are plotted in Figure 14o Data are tabulated in Table VIII, Appendix Eo The samples exposed at a dose rate of about 5.1 x 107erg/gm-hr were irradiated at a standard position on the outside of the cobalt source cover. All these samples were irradiated at one location, one at a time, to eliminate the effect of any variation of dose rate from point to point around the sourceo The points indicating a dose rate of 6.8 x 106 erg/gm-hr represent the maximum, minimum, and mean readings of the nine samples in the "high dose statistics runo' These samples were located 30 cm from the center of the source. The deviation of the mean of this cluster from the line probably indicates an error in positioningo An error of 1 cm in location at this distance results in dose rate change of 7o% The dose rates in rep were obtained from the work by Serment and Emmons of the Phoenix Laboratory staff (37) The conversion to absolute energy units was made by the following calculation: 1 rep = 97 erg/gm water Energy absorption cross-section of water at 1.25 Mev = 3.01 cm2/gm Energy absorption cross-section of C14H1805Hg at 1.25 Mev = 3.07 cm2/gm.

-560 0 0 0 o 0 -P rd 0 w _ 0 \5-~~~~~~~~~~~~~~~~~~~~ I~~~~~ I C 0^ \ 0. S K \, O S o 0 0 I) \ l -* J 0 ~z c 0 0 w O O O - 0 0 0u Q Q b)0 0 G 0 - O f (WO/ll) a3..V83(1 Adno83w

-57(Energy absorption rate in C14H1805Hg) (Dose rate in rep) x 307 x 97 The energy absorption cross-sections were computed from the tables by Storm, Gilbert, and Israel (39) The yield figures in Figure 14 are plotted as milligrams of mercury liberated per gram of compoundo These figures may be converted to percent of mercury present which is liberated by the relation Percent liberation mg Hg liberated -3 Mol Wt Comp x 10 Percent liberation = o x 100% x —..- x gm compound Mol Wt Hg mg Hg liberated x 0~232 gm compound A convenient graph of this relation is given in Figure 15o The yield versus total dose is seen to be linear over the more than three decades where tests were madeo The compound appears to follow a simple decomposition proportional to total dose unaffected by decomposition'-products at least up to 8% destruction. From this linear relationship the G value for liberation of mercury by cobalt-60 radiation may be computed readilyo In the modern terminology this will be G(Hg) as distinguished from G(-M), the G value for destruction of the original compoundo Any decomposition resulting in the formation of a soluble mercury compound is of course undetected by the measurement procedure used. 1 mg Hg is liberated. by 2~32 x 108 erg 1 mg Hg (6.023 x 1023) (10-3) = 2~92 x 1018 atoms 200 6

-588_ o 0 2 o 0o 1111 I I I I1Jill I I I I I i' _ 0^ 8^0?j n3 o urd o H 0 k^ bO0 CH — \\ - 4E 4-) 0.l _QO~O; ~~lV38l' IN U)d ~~~~Qo ~ 0~~~~~~ a31VU38 lN300 Q3/V~g3I1l 9N30I3d

-59232 x 108 erg 232 x 108 02018 2o 32 x 108 erg 6 x 10- = lo45 x 1020 ev lo45 x 1018 x 100 ev lo6 x 1l0-12 G- (Hg)(Co-60) 292 x 1018 = 202 1.o45 x 1016 This is seen to be close to the "normal" G value for direct effects discussed by Healo The difference between 2o02 and the nominal value of 3 if there is one destruction per ionization may be due to mercury going into soluble compounds. The yield is shown to be independent of dose rate over a factor of 10o This range tested, surprisingly enough, includes the maximum local destruction rates used in the x-ray experimentso The yields from x-ray experiments are low because this maximum dose rate exists only over a very small amount of material because of the extremely short penentration of soft x-rayso With cobalt-60 radiation. on the other hand, the dose rate is for practical purposes uniform throughout the samples used. This analysis is carried further in. the next sectiono Fo Irradiations with X-Rays 1o Use of Monochromatic Radiation The techniques used in this work for production and dosimetry of monochromatic x-rays were developed jointly by WO Ro Clend.inning and the author. These techniques are described and discussed in detail in a separate publication in technical report format entitled, Methods of Monochromatic Irradiations and Absolute Dosimetry with Soft X-Rayso (Reference 9) Since inclusion of this material would at least double the length of this thesis, only the barest outline of the methods will be given here in order to complete the description of the experimento

-60o 2o Production of Monochromatic Beams After extensive experimentation and review of previous work, it was decided that use of x-ray fluorescence would. provide the highest radiation intensity consistent with the requirement of very clean, spectrae In this method the "white" beam from an x-ray tube is allowed to fall on a sample of a given element The characteristic x-ray spectrum of the element is emittedo (A more comprehensive description of x-ray fluorescence phenomena is given in Chapter II of this thesis ) It is most convenient to use the K-spectra if a series of irradiations at different energies is desiredo The K, line of each element can be almost completely removed, by inserting in the beam a filter containing an element whose K-edge lies between the KU and K, lines of the emitting elemento The arrangement is shown schematically in Figure 16o At least 95% of the energy striking the target sample is carried by photons having the K. energy of the emittero The spread in energy of this line is only a few electron volts No continuous background is present except for a very small amount of scattering of the x-ray tube backgroundO The sample of the element whose spectrum is desired is called the "radiatoro" The term "fluorescer' is also seen in the literature~ The element need not be pure, but may be. in a compound with low-Z elements. The fluorescence yield of these elements is very small.9 and their low energy fluorescence is entirely absorbed, in a short air patho The radiator must be quite free of contaminants having spectra in the energy region 4 —50 kevo When pure elements were not obtainable, compounds were used and

X - RAY TUBE WHITE BEAM CHARACTERISTIC, MONOCHROMATIC XSPEOTRUMBEAM RADIATOR FILTER SAMPLE Figure 16* Schematic of Fluorescence Irradiation Assembly.

-62were packed in plastic planchets 1-3/4 inches squareo Several typical radiators are shown in Figure 17. Radiators of about 45 elements have been prepared and usedo All x-irradiations discussed in this thesis have been performed on General Electric XRD-5 machines equipped with Machlett AEG-50 tungsten target x-ray tubeso Two of these machines are available in the Phoenix Memorial Laboratory. They have a maximum voltage of 50 pkv and current of 50 ma. A photograph of this equipment is shown in Figure 18. A closeup view of the x-ray tube, radiator, filter frame, and an experimental sample is shown in Figure 19o Although the principle of use of fluorescence as a monochromatic radiation source is simple, its application has presented numerous practical difficultieso Many of the elements are available only in forms with unfortunate chemical or physical properties. Preparation of filters of the required thinness and uniformity has been particularly troublesome These problems and their solution are discussed at length in Reference 90 35 Application to Mercury Compounds Energies of the monochromatic beams used in these experiments were centered about the L-absorption edges of mercuryo The absorption spectrum of a-Me-Et-OAc (C14HI1805Hg) over the range 5-20 kev is depicted in Figure 20. The contribution of mercury to this cross-section is also showno These absorption coefficients were calculated by interpolation of the (27) data in the recent compilation by McGinnies7) of the best experimental value s.

-63Figure 17. Typical Fluorescent Radiators. (Left to Right: Pt, Sb, Ge, Se, Bi, Os.)

-64Od H a o H...... o _ rdP: ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ I 0. ~ ~~~4j~~~~~~~i....................................... k~~~~~~~~~~~~~~~~~~~~~~~~~~................. F9 X~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..... d~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~............... pr;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...........

-65 - Sample X-ray Tub e Radiator Filter Frame Figure 19. Assembly for Irradiation of Large Samples.

-66PHOTON ENERGY (KEV) 10 9 0 11 12 13 14 15 16 17 18 19 20 I I l l l I 90 80 \\ -~'aO LII 7- \ 60 w 0 30L -- C 0 z I\\\ w 40 C14Hie05Hg __ —- Hg CONTRIBUTION 30 Figure 20. Total Mass Attenuation Coefficient of C14H1805Hg from 8-20 key.

-67The locations of the L-absorption edges as given by Fine and (l4) Hendee( are LI 12.285 kev Li 14.212 key LI -14o841 key Monochromatic irradiations have been limited to the region of the L-edges because the x-ray machines available for this work could not reach the region of the K-edge at 835106 kev. However, this should not at all preclude comparison of the results with those obtained by Emmons, who irradiated around the K-edge of iron, and Garsou, who irradiated around the K-edges of bromine and iodineO It appears that there are only three physical factors concerned with the absorption edges which could enter into any model which explains unexpected effects observed across the absorption edges: 1. The depth of the absorbing shell in kevo The K-edge energies of the elements just mentioned are Fe - 7o111 key Br - 130475 kev I - 33o164 kev Thus the L-edges of mercu:ry are seen to bracket the bromine edge and lie between the K-edges of iron and iodine 2. The magnitude of the absorption coefficient jump. The ratio of K-shell absorption to L-shell absorption in iron, bromine and iodine is about 8 at energies just above the K-edge. In mercury the ratio of L-shell

-68absorption to M-shell absorption at energies just above the LI edge is 43~ Intuitively it seems most improbable that a difference of a factor of 2 in this quantity could eliminate an effect at one absorption edge if it existed at another edgeo 3. Difference in secondary electron distribution across the edge Just below an edge an energetic photoelectron is produced. while above the edge much of the energy is dissipated in an Auger showero This holds true for the Ledges as for the K-edge. The fluorescence yields for the elements in question are as followso Hg L-fluorescence~ Oo36('Refo 4) Fe K-fluorescence~ 0o308 (Refo 31) Br K-fluorescence~ O056 (Refo 4) I K-fluorescence 0o88 (Ref. 4)'Thus it would seem that any difference between the energy dependence across the absorption edge in a-Me-Et-OAc from that found. in catalase and in the plastic fil.ms must be attributeod to the chemical systems rather than to physical factors associated witbh the absorption edgeso 4.c Techniques of Sample Irradiation Design of samples for soft x-ray work must take into careful consideration the small penetration of the x-rays and the characteristics of the x-ray machine It can be seen, from Figure 20 that the mean free path. in a-Me-Et-OAc at 15 kev (just above the Lj edge) is only 0o014 gm/cm2. Thus only a thin layer of material on the surface will

-69receive any appreciable x-ray doseo It is also apparent that any container for the sample can have only a very thin film between the a-Me-Et-OAc and the x-ray source without diminishing the dose significantly The samples have been contained and irradiated on the x-ray machine in the plastic planchets shown in Figure 21o The outer dimensions of the planchets are 1-3/4 x 1-3/4 x 1/4 inch. The depression which contains the powder is 1-1/4 x 1-1/4 x 1/32 inch. The material can be packed in these planchets with a surface density of up to 0.1 gm/cm2 depending on the fineness of the powder and the pressure used in compactingo In the preparation of a sample, the clean planchet is weighed first and then weighed again after the a-Me-Et-OAc is packed ino The surface density is assumed uniform since a small error in this quantity can cause only an insignificant error in the final resulto After the powder is packed and weighed, the sample is wrapped in 0,00025 inch Mylar which is taped together on the back side. The a-Me-Et-OAc powder is distinctly darker after irradiationo This is the same color change that was measured in the large crystals. With powders, the apparent color change depends on the size of the fine crystals. If the material is very fine, the eye sees primarily surface reflections from the crystals, and the decrease in transmission is not so obvious. It has been determined, however, that there is no difference in the yield of mercury. These samples are "almost totally absorbing". Each sample absorbs at least 95 of the incident x-radiation, or in other words, has a thickness of at least 3 mean-free pathso Some compromise has to be

Figure 21, Plastic Sample Planchet and X-Ray Samples. Figure: 22.ii~ Exterior!of theCalorimeteriBox Figure 22. Exterior of the Calorimeter Box.

-71made between the desirability of absorbing 99~99% of the radiation and thus minimizing the escape correction and the desirability of using a small amount of material to reduce the mercury background correction and prevent wastage of the material which has to be prepared in the laboratory. If the sample is 95% absorbing, the correction for incomplete absorption can still be made with high accuracyo Because of the very strong x-ray absorption in a-Me-Et-OAc, the percent decomposition at the surface of the sample is much higher than the average decomposition of the sampleo A calculation was made to determine the magnitude of the surface decompositiono Let us consider a parallel beam of monochromatic x-rays with energy flux I erg/cm2 sec incident on a sample of thickness To Now the absorbed dose rate per unit area in dx at x is Ixtdx = Ioe-Xtdx where p. is the linear absorption coefficient for the photon energy of the beam. The total dose deposited per unit area per unit time is, of course, T T Io f e -_X[dx -Io [e- LX] Io(l - e 4T) x=O where e- LT is always less than 0,5 in these samples and will be neglected for this calculation. Thus the fraction of the total dose per unit area which is absorbed in dx at x is just F(xdx) = e-Xpdx Let us find the fraction of the total dose which is absorbed in the first 1/100 mean free path at the surfaceo The dose rate can certainly

-72be assumed uniform through this thickness. We have F(O,Ax) = pAx = i A(p), (01p) - o01 P ^ We can now take the case of an x-ray sample irradiated. to a very high dose so that 1 mg Hg is liberatedo We assume that the photon energy was 12.3 kev, where the compound has the maximum cross-section in this regiono The absorption coefficient is 75 cm2/gm, so 1 mfp = 0o0132 gm/cm2o The area of the sample is about 10 cm2, so Ool mg Hg was produced per cm2o From our calculation then, (Oo01)(0ol) mg Hg was liberated from (0o01(0o0132 gm) of compound, or a yield at the surface of 10' Img gm, or 7~55 mg/gmo This corresponds to a maximum of 1o32 x 10 lo8% decomposition at the surface, which. is well within the linear region previously establishedo 5o The X-Ray Calorimeter In order to obtain an absolute measurement of the x-ray flux at different photon energies, a soft x-ray calorimeter was designed, built, and used. This equipment is described in detail in Reference 9~ The calorimeter measures the energy flux of the x-ray beam by intercepting the beam with a 0002 inch gold foil and measuring the rate of temperature rise of the foilo Views of the exterior of the calorimeter and the internal assembly are shown in Figures 22 and 23~ The gold receiver plate is suspended by silk threads in a hole cut in a plexiglas plateo The plexiglas plate is in turn suspended by

-73Receiver Plate Plastic Plane Lead Wire Supporting Rod Heat Sink Stupakoff Seal End Plate Figure 235 Internal Assembly of the Calorimeter

-74threads inside the brass calorimeter box The box has a mylar window large enough to allow the full beam to pass The inside of the box is evacuated and continuously pumpedo The temperature rise of the gold plate is measured by a thermistor connected into a sensitive Wheatstone bridge circuito The sensitivity is calibrated through use of an electrical heater wire wrapped around the gold plateo Since the calorimeter was used within a few inches of the hot x-ray tube, the major problems encountered were in insulating the receiver plate from external temperature fluctuationso These problems were eventually surmounted, and the results obtained with the calorimeter were very satisfyingo Energy fluxes of 20 microwatts in the entire fluorescent beam were measured very reliably, and fluxes down to 5 microwatts were measured with some confidenceo 60 Dosimetry Used in These Experiments The primary standard for measurement of radiation dose has been the soft x-ray calorimetero Since the calorimeter is not suitable for routine day-by-day measurements, a General Electric SPG-1 argon-filled proportional counter tube has been used as a secondary standard and carefully calibrated against the calorimetero The calibration procedure and final calibration curve for this tube are discussed in detail in Reference 9o Before and after each irradiation the radiation beam is monitored with the SPG-l in its normal position on the goniometer and the large protractor at 0o00~o A lead sheet pierced by a pinhole is placed in front of the SPG-1 window to reduce the count rate enough to eliminate

-75coincidence losses. The count rate is always measured at the x-ray tube current that is to be used for the irradiation, normally 40 ma. The counting set-up with the lead pinhole in place is shown in Figure 24o The lead pinhole is a permanent piece of apparatus, and the entire counting set-up is completely standardized and reproducibleo It has been found that the counting equipment can be removed from the machine and replaced with the same count rate being obtained within the limits of statistical error. In order to check on operation of the SPG-1, an Fe55 x-ray emitting source is counted every time the counter is usedo There is at times a slight shift from day to day in the count rate obtained This is ascribed to drift in the high voltage setting, and the SPG-1 counts are corrected to a standard Fe55 rate of 4000 cpSo Uniformity of the radiation beam over the sample has been checked with x-ray films. The dose rate has been found to be uniform within the error of measuremento Therefore, the absolute value of the energy content of the beam incident on the samples has been calculated by multiplying the energy flux determined with the calorimeter, described in Appendix F, by the area of the sampleso The absorbed energy is then calculated using the known absorption coefficient of the material at the given photon energy and the weight per unit area of the sampleo This procedure is believed to give the absolute value of energy absorbed by the sample within the limits of error of the calorimetero

c7 il-_ —::i-:::-:::::i._i5-Biiiii:ii-i ii-:i: a:::: " 16i Iil: I: i.iii:il I -iiP ii:i i:: -"F — iil:i::_: - i::::::_-iii::li::i-i —i —l:- I:i ii::-iiil-:i::l:- ii_:::i: —i!i i —l-ii:_i-iiiii —ii-li -ii-ii_;iiiiii-iii:i:iiii:i-iiiiiiii:li,_: -::::_:- -:::_-i-i: i-:,:ii:iiiii:i::-::-:-:::_:-_iiii-i: —i:-i -i-:i_:i i:-iiii Figure 24, SPG-1 Cowrting Assembly with Leaa Pinhole,

-77G. Dose and Dose Rate Dependence with X-Rays Although the cobalt-60 experiments are very indicative, several tests of dependence of the yield on dose and dose rate were made with fluorescent x-rays to ascertain that there were no unexpected effects. The results are shown in Figure 25. The data are given in Table IX, Appendix Eo As expected, the- yield is found to be linear with total dose and independent of dose rate over the range tested. H. Energy Dependence in the Soft X-Ray Region In the final phase of the work reported in this thesis, the energy dependence of the liberation of mercury from a-Me-Et-OAc was determined in the energy region of the mercury L-absorption edges. Samples were irradiated with filtered K spectra of twelve elements having K, lines between 9 and 18 kev. After filtration at least 95% of the energy in the beam resided in photons whose energy lay within a few ev of the nominal energy. the K. line of the radiatoro Nearly all the remaining energy content of the beam consisted of Kg photons from the radiator which were not stopped by the filter and lines from the tungsten L-spectrum which were scattered by the radiatoro The data are presented graphically in Figure 26 and in tabular form in Table X, Appendix E. The yield from each sample is shown in milligrams of mercury released per erg absorbed. In addition to the nominal energy, the element used in the fluorescent radiator for each energy is shown on the horizontal axiso At least two samples were irradiated at each energy. Values of all measured quantities and the ensuing calculations are given in Appendix Eo

-78C: CD \ o Ua 0 0 0E U q) q0 q0Oc -') W ]' - x \~^ ~ ---- < i. N. -. \ -~0 Wn I o rad jr a) Q fl eO \ ) IO o o d o 0 CD P 0) 0 ~ ~~~ ~ ~ ~ 0 0

-790 Z* 0 -J). —- C -r ~d I 4) I - I 1 o) rd 0.P'4 V) lv: ~ _ 0 H ~- 0 _J Q.~~~~~~. b0 (D *~~~C 0 0 * 0 <

-80The precision of the data was limited somewhat by the long irradiation times required. Each point in Figure 26 represents about two days of x-ray machine time. It was not possible within the limitations of available equipment to achieve radiation yields large enough to give very high precisiono The radiation yield was found to be independent of photon energy in the region of the mercury L-absorption edges within the limits of experimental precisiono The mean radiation yield from all the points in Figure 26 is 3~68 x 10-9 mg/ergo The standard deviation from the mean is 0.769 or 21% of the mean. If there were a difference in the radiation yield across the mercury L-edges, -we might expect the largest discontinuity to occur at the LII edgeo From Figure 20, above the LIjI edge a minimum of 70% of the radiation is absorbed in the Lshell, while below the edge of course no absorption takes place in that shell. The mean radiation yield at the five energies below the LII edge is 3~75 x 10-9o We can safely say that within experimental error there is no difference in the radiation yield across the L edges. The measured radiation yield of 3~68 x 10-9 mg/erg can be readily compared with the G value found with cobalt-60 radiationo A yield of 3068 x 109 mg of mercury is equivalent to (3667 x 10l12 gm)(6o023 x 1023 atom/mole) 1 10 x 10 atomso 200 6 gm/mole and. 1 erg = 6o25 x 1011 ev = 6~25 x 109 x 100 eVo Therefore G(g) (98 kev) 110 1x 7610 G(Hg)(9-18 key) 6=2 x 109 = 176

-81This is to be compared with G(Hg)(Co60) 2=.02 Since the dosimetry methods used for the two radiations are very different and a number of rather imprecise measurements and assumptions are involved in correlation of dosimetry, this is considered to be good agreement between the two G values. No significance is to be attached to the 13% difference between the valueso On the contrary, the close agreement indicated the possibility of using this chemical system, or a similar one, as a "transfer standard" for dosimetry of radiations of different energieso In an attempt to establish with somewhat more precision whether or not there is a discontinuity across the LIII edge, samples were irradiated with several different unfiltered K-spectra. The results are plotted in Figure 27~ Data are given in Table XI Appendix Eo The unfiltered K-radiations contain both Kd and KB photons in the ratio of about 51 (32) In Figure 27 the energies of both KM and Ki photons are marked on the horizontal axis. The experimental points are plotted at the?i energies, Calculation of an average or effective energy would not be useful for our purposeo Indication of the actual incident spectrum is to be preferred. Yields of duplicate samples irradiated with five different spectra are shown in Figure 27. With the Ga spectrum no photons were absorbed in the mercury L-shello With the Se and Br spectra only 10 - 15% of the photons were absorbed in the L-shell, while with the Kr and Zr spectra over 70% were absorbed in the L-shell, Confirming the data shown in Figure 26, no dependence of radiation yield on the site of initial photon absorption is foundo

-82CD N " o - a_.0 Z- 1 — -t_ __1-1 —-------- --- i - 0 0 0" N- ). CH CH.p a),00 a) om _ 0 m - (0 0~ c ________ _ _______________________________ _9)_'-_ 0 (1OI A9 Adi.Lnrl) (~je/~uu) a-3iA NOli.LaVI

-87The last series of data to be presented concerns the indication by Emmons and by Garsou that incident photons of the same energy as the emission line energies of the heavy element present in a substance may be extraordinarily effective in producing damageo To determine whether such an effect existed in a-Me-Et-OAc, samples were irradiated with the L-spectrum of mercuryo As discussed at length in other sections of this thesis, the L-spectrum consists of five prominent lineso Since each line is rather weak compared to the Ka line, filtration to separate single lines is not feasible in experiments with this chemical systemo The intensity incident on the sample would be entirely too lowo Instead the samples were irradiated with the entire L-spectrumo To provide a direct comparison, additional samples were irradiated with L-spectra of several elements above and below mercury in atomic number The results are plotted in Figure 28o The abscissa is the atomic number of the fluorescent radiator, since again an average energy would not have any significance Data are tabulated in Table XII Appendix E. In Figure 29 the energies and relative intensities of the five principal L-lines of the radiators used are showno The data presented in Figure 28 are the oldest data presented here, and the spread is fairly large However, it is strongly indicated that there is no exceptional damage produced by photons of the emission line energies of the sampleo There is certainly no effect here comparable to that found by Emmons and by Garsou.

m84rd HI ro - 0'0 OZ U o 0 ) -H oj OD -AN 00 0'o (C ( 01 A9 Aldl.9ln) (5Je/6w) 0131A NOIVIlOVy 6""

-85(D CH I 0) 0 w - o H OJ -. JQI 0..0~ h 00 (L F I < 0-~~~~~~ --- - -- -.- --- -.* —. —. - ---- ---- --- ---- ---- ---- --- 0

IVo SUMMARY AND CONCLUSIONS Samples of a-acetoxymercuri-p-methoxy hydrocinnamic ethyl ester have been irradiated with monochromatic x-ray beams at a number of different energies between 9 and 18 kevo This energy region includes the L-absorption edges of mercuryo The release of metallic mercury during irradiation has been determinedo A curve of the energy dependence of the radiation damage between 9 and 18 kev, of the type defined in the Introduction, has been presented in Figure 26 It has been determined that there is no dependence of the radiation yieldo on total dose or dose rate which would vitiate the resultso In addition, similar samples have been irradiated with unfiltered L-spectra of six consecutive elements including mercuryo With these spectra, each consisting of five distinct energies, it is not possible to determine an energy dependence curve of exactly the type defined in the Introductiono However, Figures 28 and 29 taken together give the information desired for our purposeso The literature survey indicates that this is probably the first experiment undertaken on energy dependence of yield of a parti. cular radiolysis product of a pure compound. in the soft x-ray regiono It was found that within experimental error the radiation yield is independent of photon energy between 9 and 18 kevo There is no change in damage per unlit energy absorbed across the L-absorption edges of mercuryo The L-spectrum of mercury is no more effective than other L-spectra in this region. In addition it has been found, that the amount of mercury liberated per un it energy absorbed is the same, within -86

-87experimental uncertainty, for soft x-rays as for cobalt-60 radiation. These results bear out the conclusions discussed in the second chapter that, except in some special circumstances, the damage per unit energy absorbed should be independent of the location of the initial photon absorption. These results are distinctly different from those obtained with catalase by Emmons and with the halogenated hydrocarbon-dye system by Garsou. If we assume that all the experimental work is valid., the a-Me-Et-OAc system has an energy dependence which is in agreement with accepted theories, while the catalase and halide-dye systems seem to give data which are definitely unexpected. We cannot well explain the differences between these results without having some idea why the "resonances" were found. in catalase and, in the halide-dye system. To date no model has been proposed which convincingly explains these resonance effectso A comparison between a-Me-Et-OAc and the halide-dye system is particularly difficult because of differences in irradiation methods. Garsou s experiments, while carefully done, were performed at an early stage in the development of experimental techniques by this groupo A re-examination of the halidedye system with more advanced methods is probably indicated. The a-Me-Et-OAc data should be more directly comparable with Emmons' data obtained on catalase with fluorescence radiation. A searching examination has failed to uncover any essential difference between the experiments which might lead to correct results in one case and incorrect results in the other. A comparison ofanumber of different aspects of the experiments is given in Table IIo

-88For completeness, Emmons two series of experiments on catalase solution using the Bragg spectrometer as a radiation source have been compared similarly in Table III: Each series of Emmons' experiments shows that the damage rate of his system depends in some non-trivial way on photon energyo It will be seen in the bottom, row of each table that analysis shows that each. series of Emmon.s results conflicts with accepted. theorieso This statement should probably be explained further It was shown in Chapter II that there exist, at least conceptually, systems in which absorption at a sensitive site can cause a change in the quantum yield across an absorption edgeo It was also shown, however, that for this effect to be observed the quantum yield. must be of the order of magnitude of the fraction of radiation absorbed, at the sensitive site above the edge, or smallero It was further shown that in this effect the numerical difference in the quantum yield across the ed~ge cannot exceed unity. In discussion of LET-dependent reactions, it was shown that a difference in LET across an edge could not be -I -ib-le unless the element with the edge absorbed a large fraction of the incident photons Let us examine now whether any of Emmons9 experiments fit the sensitive site hypothesiso One is immediately attracted to this idea since the enzymatic activity of catalase is thought from other experiments to depend on the action of the few iron sites in the giant moleculeo In Emmons' experiments which we have called I an.d IV, however, the measured quantum yl.elds are very much too high to fit this hypothesis,

-89In his experiments III and IV, moreover, the change in quantum yield across the edge was much too largeo It should be pointed out that these disagreements are only quantitative, and that refinement of the experiments might show the original quantum yield values were in erroro The correction, however, would have to be by several orders of magnitude to make the results fit the theoryo We-lmay now examine whether Emmons' results can be explained by LET-dependence. This also is an attractive hypothesis, since the catalase solutions are probably LET-dependent. Inspection of the numbers for the fraction of photons absorbed in iron, however, shows immediately that the change in LET across the iron edge is extremely small and could hardly be conceived to be significanto Emmons' experiment II indicated the iron fluorescent spectrum to be exceptionally effective at deactivating catalase solutiono It is stated in the table that this result disagrees with theory because there is no hypothesis supported by other experiments to explain this phenomenono Two major differences between the catalase and a-Me-Et-OAc systems are in molecular weight and fraction of the radiation which is absorbed in the heavy elemento The significance of these differences in explaining the results of the experiments remains obscureo An explanation for the resonance behavior of catalase, if the earlier experiments are confirmed, may be intimately connected with enzyme chemistry. This is obviously outside the scope of this papero

W90TABLE II COMPARISON OF FLUORESCENCE EXPERIMENTS WITH CATALASE AND a-Me-Et-OAc Ref: Emmons(12) Emmons I Emmons II Atkins Catalase (Dry) Catalase (Solution) a-Me-Et-OAc Radiation Source X-ray Fluorescence X-ray Fluorescence X-ray Fluoresecence Radiation Spectra 4 K-spectra, 3 K-spectra, 12 K-spectra, Used unfiltered unfiltered filtered 6 K-spectra, unfiltered Material irradiated Purified enzyme Purified enzyme Pure compound in aqueous solution Molecular weight 225,000 225,000 in 2 x 10-7 467 M solution Heavy element Fe Fe Hg Fraction absorbed 0.04/0.005 1.3 x 10-6/1.6 x 10-7 0.994/0.970 in heavy element above/below edge Dose rate dependence Probably none Strong None Stability to Unstable Unstable Relatively very environment stable G (x-ray) 1 (Dose rate 1 G (Co60) for Cr and Ni dependent) Molecules damaged 100 - 200 0.02 - 0.2 200 - 400 per photon (x-rays) Damage criterion Loss of enzymatic Loss of enzymatic Release of activity activity metallic mercury Results Ni and Cr give direct Fe factor of 6.6 No energy damage curves, Fe and more effective dependence Mn give induction than Mn periods followed by steeper damage rate Compatibility with Disagrees Disagrees Agrees accepted theory Quantum yield too No apparent reason high for Fe radiation to be more effective

-91TABLE III EMMONS' EXPERIMENTS ON CATALASE USING BRAGG SPECTROMETER RADIATION Ref: Emmons (12) Emmons III Emmons IV Catalase (Solution) Catalase (Solution) Radiation Source Bragg Spectrometer Bragg Spectrometer Radiation Spectra used About 18 energies 6.9 and 7.3 kev from 6-12 kev Spectra stated to be purer than in column at left Material irradiated Purified enzyme in Purified enzyme in aqueous solution aqueous solution Molecular weight 225,000 in 2 x 10-7 M 225,000 in 2 x 10-7 M solution solution Heavy element Fe Fe Fraction absorbed in 1.3 x 10-6/1.6 x 10-7 1.3 x 10-6/1.6 x 10-7 heavy element above/ below edge Dose rate dependence Strong Strong Stability to Unstable Unstable environment G (x-ray) (Dose rate dependent) (Dose rate dependent) G (Co60) Molecules damaged per 0 - 51 2 - 7 photon (x-rays) Damage criterion Loss of enzymatic Loss of enzymatic activity activity Results Damage curve follows Three times as much shape of Fe damage above K-edge absorption Compatibility with Disagrees Disagrees accepted theory Change in quantum Quantum yield too large yield too large. compared to percent absorption in heavy element. Change in quantum yield too large.

-92We shall conclude by listing the accomplishments of the research program reported in this thesis~ lo Significant additions have been made to the scant data available on radiolysis of organometallicso 20 Analytical techniques have been developed by use of which much more data can be obtainedo 3~ A potentially very useful dosimeter has been suggested. (Appendix D)o 4. A start has been made on study of colloidal metals in organometallic crystalso (Appendix C) 50 Techniques for production and use of monochromatic x-ray beams have been extendedo (Reference 9). 60 A calorimeter has been built and used in collaboration with another researcher for absolute dosimetry of soft x-rays (Reference 9)~ 70 Data have been obtained on the energy dependence of the radiation yield of a single product from a pure compoundo 80 The fact that results have been obtained which are in agreement with accepted theory lends more credence to experiments using similar techniques which produce other results o

APPENDIX A PROPERTIES, SYNTHESIS, AND PURIFICATION OF THE MERCURY DERIVATIVES OF CINNAMIC ACID The preparation of a number of mercury derivatives of cinnamic acid was first described by Schrauth, Schoeller, and. Struensee. (33 34,35) The only work in English describing these compounds which has come to the attention of the writer is Whitmore's Organic Compounds of Mercury.(42) It should be noted that the information in. Whitmore is merely excerpted from the original articles and contains several important mistranslations. A search of Chemical Abstracts for the last twenty years has not revealed any reference to the compounds discussed in this appendix. Physical Properties The compounds to be discussed have the structural formula H I -Hg - R" C] C | 0 H R where the radicals R and R' and typically Me (CH3) or Et (C2H5) R" is typically OAc (O C=O CH3) or Cl, Br, or I. For convenience, such a compound will be denoted here by the shorthand name a-R-R'-R". The compound with which the energy dependence data were taken is then H 0 |I gHg- 0 C- CH3 fn )-^\ | C_ OC2H5 O- H CH3 -95

Short name: a-Me-Et-OAc Full name: a-acetoxymercuri-p-methoxy hydrocinnamic ethyl ester. The purified compounds can be prepared in the form of colorless, brittle, needle-like crystals. The ease of forming sizeable crystals varies greatly among the compoundso The a-Me-Me-Cl can be prepared rather easily with a crystal size of 2 x 10 mm, while a-Me-Et-OAc tends to form longer, thinner crystals, 1 x 15 mm being common. The a-Me-Me-OAc, on the other hand, has always formed very tiny needles. Details of the crystallization procedure will be given later. The compounds are all very soluble in ethyl acetate, methanol, ethanol, and acetoneo The halides are somewhat less soluble than the acetateso The compounds are only slightly soluble in 30 - 40~ boiling petroleum ether, but somewhat more soluble in higher boiling petroleum ethers and diethyl ether. They are slightly soluble in hot water, but tend to form an amorphous state with water with some decomposition. Melting points of some of the compounds as given by Schrauth, Schoeller, and Struensee are as follows: o-Me-ie-M-Ac: 139~C a-MeeMe-Cl: 132.5 C a-Me MeMeBr 110o5~C a-Me Me-Io 100 C a-Me-Et-OAc: 107~C Preparation of the Cinnamic Alkyl Esters Since the desired compounds are not standard commercial items, they have been synthesized in the laboratory. Synthesis procedures have

-95in general followed those of Schrauth, Schoeller, and Streusee(33,34,35) Several modifications have been devised, however, to simplify production and to provide a very pure product. The laboratory procedures found to be the most practical will be presented here along with the descriptive chemical equations and notes on our laboratory experienceso One of the starting materials for synthesis of an U-R-RI-R" compound is the cinnamic R1 estero Since these esters are not commercially available either, they have been prepared from cinnamic acid and the R1 alcohol by the conventional method: H H iI i,0. _..C = C - C-. OH + R'OH Cinnamic acid RI alcohol H H I I >0 --- ______ (^ \C = C-C —-ORI + H20 Heat Cinnamic RI ester water Sulfuric acid is used to remove the water produced and prevent the back reaction. The following procedure has been found very efficient for producing cinnamic ethyl ester: 1o Mix in 1000 ml round bottom flask with standard taper neck: 460 ml absolute ethanol 120 gm cinnamic acid (Eastman No. 75) 20 ml H2S04 Boiling stones

-962o Boil under reflux for 12 hours or more. Use Glas-Col heating mantle and reflux condenser with ground-joint connection to flask, 30 Remove condenser and boil off excess of alcohol until rate of boiling diminishes and temperature starts to rise 40 Cool and add 5 parts'water, Add Na2C03 solution until neutralized 50 Using 1000 ml separatory funnel, wash with water 4-5 times o 6S Filter through No. 50 paper into glass - stoppered bottle, Yield is 110-1220 gmo Noteso 1. The methyl ester may be prepared similarly, using methanol in place of ethanol. In washing the methyl ester, it is necessary to use hot water and keep the funnel warm with a heat lamp since the melting point is 36~C, 20 The esters are clear and almost colorlesso They have a pungent but pleasant odor which allows their detection in extremely small quantities, Their properties are listed in the Handbook of Chemistry and Physics, 3 When the starting materials are mixed. the cinnamic acid will not all di.ssolve until the mixture is nearly boiling. 40 After several hours boiling, the mixture may become cloudy because cinnamic acid is less soluble in the ester than in the alcoholo This is i.nsignificanto

-97Preparation of a-Me-Et-OAc In general, the compound G-R-RV-OAc is prepared according to the equation H H 0 I | o0 n r H -C C -C —- ORI + ROH + Hg 0 \0 - C'- nTT Cinnamic RI ester R alcohol Mercuric acetate H 0 Hg - 0 - C CH3 | i l \C- OH 0 U- I R a -RR -OAc Acetic Acid The details of the preparation vary somewhat with the differen compounds. The preparation. of the impure a-Me-Et-OAc is very simpleo Two solutions are made~ Solution A: 25o2 ml cinnamic ethyl ester 100 ml methanol Solution B~ 37 gm mercuric acetate 150 ml distilled water 5 ml glacial acetic acid to prevent hydrolysis The two solutions are mixed thoroughly in a dis:h with a tight-fitting lido (A glass refrigerator dish 3 x 4 x 3 inch contains one batch very necelyo) The reaction is quite slowo The a-Me-Et-OAc usually begins crystallizing out in 3-4 days, and crystallization is complete in about 12 days. If-no crystals have formed in 5-6 days, a seed from a previous preparation will start crystallization immediatelyo

-98Purification of a-Me-Et-OAc While the preparation of the impure compound is simple, attaining the desired purity with, a good yield of material, has involved a number of problems When the reaction is complete after 12 days, the dish contains a thin layer of cinnamic ethyl ester on the bottom, and above this a layer containing water, alcohol,, acetic acid, and. mercuric acetateo The crystals of U-Me-Et-OAc are in both layers It has been found to simplify purification considerably, with. a waste of no more than 10-20% of the yield, if crystals from the aqueous phase are removed with a spatula without disturbing the ester phaseo The material left in the dish may then be discardedo The crystals are contaminated by acetic acid, mercuric acetate, and cinnamic ethyl ester. The first two may be easily washed off with cold water, but removal of the ester is a major problemo Schrauth, Schoeller, and Struensee recommend fractional. crystallization from ethyl acetateo The material is dissolved in a minimum amount of ethyl acetate near its boiling pointo The solution is then allowed to cool slowly, and the crystals of pure material are filtered from the liquid which still contains the estero It is desirable to get as large crystals as possible so a minimum amount of liquid will remain on them after filtrationo We have not found this technique at all satisfactory for quantity productiono Proper control to obtain large crystals is tedious and timeconsumingo Even under the best conditions, the material will still smell strongly of ester after the first crystallizationo Several measures can be taken to improve the prod.uct~ a o Repeated crystallizations.

-99bo Use of a considerable excess of ethyl acetate. Co Washing the crystals on a filter bed with ethyl acetateo Any of these procedures will cut the yield drastically, since a-Me-Et-OAc is quite soluble in ethyl acetate even at room temperatureo Also, repeated solution in boiling ethyl acetate appears to produce some decomposition, We have solved the purification problem by turning to another classical procedure -- precipitation with petroleum ethero Like many other compounds, a-Me-Et-OAc is soluble in ethyl acetate, alcohol, acetone, etco, but practically insoluble in 30-40~ boiling petroleum ethero It is also insoluble in a mixture of petroleum ether with one of these other solvents over a wide range of proportionso The cinnamic ethyl ester, on the other hand, is completely soluble in petroleum ethero Our successful purification procedure has consisted of preparing a concentrated solution of a-Me-Et-OAc in ethyl acetate at room temperature and adding petroleum ether until a precipitate formso The precipitate is then filtered as rapidly as possible on a coarse sintered-glass funnelo This process is repeated at least five timeso The product is a fluffy, snow-white powder with no detectable odor. The melting point agrees with that reported by Schrauth, et al, and the spread is 0O5~ or less. It has been found that despite the sharp boiling point, the material prepared in this way sometimes gives an appreciable precipitate when a solution is centrifugedo This precipitate consists of some metallic mercury and a waxy material which is insoluble in everything we have triedo It gives a "background" of precipitated mercury which in some batches is comparable with the radiation yield for small doseso Even. worse, the

-100background varies considerable between samples from the same batch, This problem has been brought under good control by centrifuging the material in solution as the last step in the purification. processo The supernatant is poured off and the product again precipitated with petroleum eithero When the a-Me-Et-OAc prepared in this way is carried through. the procedure described. in Chapter III for measuring the radiation yield. it gives a blank reaction of about 50 micrograms of mercury per gram of compoundo In summary, the procedure used for purifying a-Me-Et-OAc is as follows o 1. Remove crystals from aqueous phase of dish with. spatula. 2. Wash on glass filter 2 or 3 times. 3. Dissolve in ethyl acetate at room temperature and precipitate with 30-40~ boiling petroleum ether 4o Repeat solution and precipitation, 4 more times or until dry precipitate is odorless. 50 Dissolve in ethyl acetate and centrifuge in 50 ml tubes for 1 hour at 3000 rpm. 6. Precipitate with petroleum ether and filtero Dryo 7o The material. must be kept as d.ust-free as possibleo It is never allowed to contact papero Preparation and Purification of -Me-Me-OAc The preparation of a-Me-Me-OAc differs somewhat from that of the Me-Et compound: 1o Prepare the following solutiono 20 gm mercuric acetate 120 ml warm. methanol

-101Trace of water 5 ml acetic acid 2. Stir until dissolved and pour into disho 3. Add 12 ml cinnamic methyl ester and mix thoroughly. In this case crystallization is complete in 3-4 days. Purification is the same as that for a-Me-Et-OAco Preparation of Chlorides The compounds a-Me-Me-Cl and a-Me-Et-Cl are prepared from the corresponding acetates by adding a concentrated aqueous solution of a slight excess of sodium chloride to a methanol solution of the acetate. The chloride precipitates: H i| Hg - 0 - C CH3 / _C -C C + NaCl I | | \ - OCH3 0 H / CH3 a-Me -Me -OAc H H| gCl C__C - OC 0o + CH3COONa I ---. -- OCH3 0 H Sodium acetate CH3 a-Me -Me -Cl a-chloromercuri-B-methoxy chydrocinnamic methyl ester

-102A methanol solution of the purified acetate may be used, or the chloride may be salted out of the mother liquor remaining from the initial preparation of the acetate. In this latter case, a sizeable amount of material is recovered that would otherwise be wasted, and precipitation of the chloride serves as one stage of purificationo Subsequent purification is as for the acetateso Growing Large Crystals In the work on color changes in crystals described in Appendix C, it was necessary to prepare crystals of a-Me-Me-Cl about 2 x 10 mm or larger in size. A procedure was worked out for growing satisfactory numbers of crystals using only very simple equipment~ 1 Equipment required.: 4 liter stainless steel beaker for water bath 50 ml Pyrex suction filtration for holding small flasks 1200 watt hot plate Ring stand or other arrangement for holding small flask in water bath Fine sintered-glass funnel 2o Prepare solution saturated at about 70 of O-Me-Me-Cl in ethyl acetateo 3~ Using suction, filter into 50 ml flask. 4. Wash compound which dries in funnel through with minimum amount of ethyl acetate 5. Place flask in water bath at 70~ with only neck of flask above watero

-10360 Allow ethyl acetate to evaporate until solution is saturated. This may be completed quickly by blowing clean air into the flasko 7o Raise temperature to boiling point, about 83~~ Make sure all material is dissolved. 80 Place cork in flask, cover with towel, and allow to cool slowlyo 9~ At about 70~ drop in a seed crystalo Recover and allow to coolo 10. When bath reaches room temperature, remove large crystals from flask with forceps 11o Wash off each crystal by dipping quickly in pure ethyl acetate and place on filter paper to dryo 12. Usually only a small fraction of the chloride has come out in large crystalso The flask may simply be reheated to 830 and allowed to cool againo 13o After several crystallizations, the material in the flask has usually suffered some decomposition and become dirty enough so large crystals will not form, It must then be filtered again. Notes: lo A saturated solution at 70~ has about the right concentration for best results. If the solution is more concentrated, crystallization occurs too rapidlyo

-1042o Taking the bath below room temperature is of little helpo 3. The solutions should of course be kept as free from dirt and dust as possibleo

APPENDIX B PROCEDURES FOR THE DETERMINATION OF SMALL QUANTITIES OF MERCURY WITH DITHIZONE Use of Dithizone as an Analytical Reagent Diphenylthiocarbazone (dithizone) is one of the most widely used of the organic complexing reagents used in quantitative analysis. It has been used for gravimetric, titrimetric, and spectrophotometric determinations of many different metallic ions. The spectrophotometric procedure is the method of choice for small quantities. It can be used to measure 1 Y of mercury to 5% accuracy, and of course its range can be extended to as large quantities as desired. The classic articles on use of dithizone as an analytical reagent were written by H. Fischer and his associates.(l5l16) Dithizone is a violet-black solid, very soluble in chloroform and carbon tetrachloride. It has an intense green color in these solvents. When a solution of dithizone in chloroform or carbon tetrachloride is shaken with an aqueous solution of a salt of a heavy metal, a complex salt is formed. This salt is much more soluble in the organic than in the aqueous phase and imparts to it a color ranging from violet through red and orange to yellow, depending on the metal present. The most helpful reference on use of dithizone has been the long article in Welcher's compendium, Organic Analytical Reagents.(41) Much of the foregoing information has been extracted from this source, which also gives over 200 listings in a bibliography on dithizone. The * 1 = 10-6 gram. -105

_106article on mercury in the manual by Fister(17) has also been of great helpo Other useful references have been Mellan(28) and. Lang and Nelson (24) Procedure for Determination of Mercury The following pages describe the procedure used in this work for the determination of small amounts of mercuryo Preliminary Notes lo The mercury is assumed to be mercuric ion in dilute acid solution, prepared as described in Section IIIoC. 2. Dithizone is decomposed to some. extent by light and should never be exposed to very strong lighto The mercury dithizonate is quite light sensitiveo Room light must be reduced to a very subdued level before dithizone is added to mercury solutions, and the dithizonate should be handled and transferred to the DU as rapidly as possibleo 3. The mercury dithizonate absorption peak at 490mp. in chloroform is being measured, 4~ Hydroxylamine hydrochloride is used as a precaution against reaction of dithizone with oxidizing agents which may be presento 50 All glassware must be washed thoroughly, rinsed with d.emineralized water, and dried before useo Maintenance of a separate set of glassware for this work which has never been used for anything else is a good precautiono It is also desirable to mark some glassware to be used only with dithizone and never with mercury solutionso

-1076. Calibration and test procedures are described for 2 "scales" covering the ranges 0-50 and 0-125 7 Hg in the separatory funnel. Application to other ranges will be obvious. 7. Any run that gives an OoD. of 1.25 or greater should be discarded. The sample should be re-run on a higher scale or with a smaller aliquot of mercury. At higher OoD. there will be depletion of dithizone and a color change due to keto-enol transformation. 8. Since the DU accommodates four cells, it is efficient to run three samples with a control at the same timeo Therefore, the procedure is written in terms of multiples of three samples. 9. Since chloroform is somewhat toxic, breathing of large amounts of vapor should be avoided. Apparatus and Chemicals 125 ml Pyrex separatory funnels, pear-shaped. Pro-pipette bulb, used for all pipettingo Volumetric flasks and pipettes as necessaryo Diphenylthiocarbazone (Dithizone), Eastman Kodak No. 3092. Chloroform, reagent, to meet A.CoS. specifications for suitability with dithizoneo Hydrochloric acid, reagent. Hydroxylamine hydrochloride, reagent. Mercuric chloride, reagento Double distilled watero

-108Preparation of Reagents Hydrochloric Acid9 0~25 No Add 22 ml concentrated hydrochloric acid to 1000 ml double distilled water and mixo Shake with 10 ml, 0o002% dithizone and allow organic phase to settleo Hydroxylamine HCl1 20% Aqueouso Dissolve 20 gm hydroxylamine HC1 in double distilled water and dilute to 100 mlo Shake with a few ml 0.002% dithizone and allow organic phase to settleo Store in brown glass-stoppered. bottleo Dithizone, 0o05%o In 100 ml volumetric flask place 50 mg dithizoneo Dissolve in chloroform and dilute to marko Preserve in brown tightly-capped bottle in refrigeratoro Stability~ about one month. Dithizone, 0o002%o In. 250 ml volumetric flask place 10 ml dithizone 0o05%o Dilute to mark with chloroformo Prepare fresh, dailyo Keep in stoppered volumetric flask. Store in dark cupborad when not in useo Mercury Stock Standard (1ml = 1 mg mercury)o In 100 ml volumetric flask place 13504 mg crystalline mercuric chlorideo Dissolve in 0025 N HC1 and dilute to mark with same Use to prepare dilute standards as required. Procedure for Running Unknown lo Set up three (or multiples of three) separatory funnelso Use Dow-Corning silicone stopcock grease On stopcockso Greasing of stoppers is unnecessaryo 20 Place 25 ml of 0~25 N hydrochloric acid in each funnel

-1093o Place 1 ml of 20% hydroxylamine HC1 in each funnel. 4. Place aliquot of unknown mercury solution in funnel estimated to give 0-50y of mercury if 10 ml dithizone is to be used, or 20-1257 if 25 ml dithizone is used. 5. Shake funnel to mix, 60 Reduce room light to low level. It is not necessary to make the room so dark it is difficult to seeo 7. Into the first three funnels only place 10 or 25 ml 0.002% dithizone as appropriate. 8. Fill Cell No. 1 of the Beckman holder with 0.002% dithizone and capo 9. Shake each funnel vigorously for 1 minuteo Invert funnel and open stopcock to relieve pressure at intervals. 10o After three funnels have been shaken, pick up each in turn, run several ml of chloroform phase into waste container, and fill up a cell and cap. Normally the chloroform phase will have completely separated by the time it is needed. 11o When the cells are full, place the holder in the DU cell compartment. The room lights may now be turned on again. 12. After the O.D. of each sample is measured, the liquid may be emptied from the cells and the next three funnels run similarly. A suction funnel attached to a vacuum line, with a very thin glass tube attached to the flask by plastic hose, is a great help in completely emptying the cells without getting anything on the outsides. Normally it is not necessary to rinse the cells or remove them from the holder if they are to be used again immediately.

-110Calibration Procedure for 0-125 y Hg 1o Prepare dilute Hg standard, In a 100 ml volumetric flask place 2 ml of the mercury stock standard (1 ml = 20y). Dilute to mark with distilled watero 2o Prepare six separatory funnels as for running unknown samples. In the funnels place respectively 0, 1, 3, 4, 5, and 6 ml of the dilute standardo 3~ Using 25 ml of 0o002% dithizone in each funnel, carry through the remainder of the procedure as described aboveo Calibration Procedure for 0-50 7 Hg 1o Prepare dilute Hg standard In. a 100 ml of volumetric flask place 1 ml of the mercury stock standard. (1 ml = 10y). Dilute to mark with distilled watero 2o Prepare six separatory funnels as for running unknown sampleso In the funnels place respectively O, 1. 2, 3, 4, and 5 ml of the dilute standard. 3~ Using 10 ml of the 0.002% dithizone in each funnel, carry through the remainder of the procedure as described aboveo Calibration and Statistics The two calibration curves are shown in Figure 30o These curves have been used in evaluating all radiation yieldso In order to determine the error in the dithizone measurement itself, a series of nine runs was made with. samples drawn from a mercuric chloride solution of known concentration. Data on this experiment are as follows

-111(zia -iv O0) OH X 0 0 0 0 0 0 ~. I N 0 H1200 0 C) NJ 0\-~~~~ Cd.. ^ w 0 ^.0 \ zX- C o H InH \ \~ 0 0 0 0 0 0 N 0 G N (ZIGia ] - H H

-112Test solution: mercuric chloride in double distilled water, 1 ml = 17 Hg++ Aliquot taken: 10 ml in each funnel Dithizone used: 10 ml 0o002% solution in each funnel The following O.Do values were obtained, shown as recorded and as corrected for the blank O.Do value of.014. Deviations from the mean, x = 263, are also shown. OoDo Recorded Corrected OoDo (x) x - x o269 o255 -o008 ~273 o259 o004 o274 o260 -o003 o275 o261 -o002 o277 0263 o000 o280 o266 o003 o280 o266.003 o281.267 o004 o283 o269 o0o6 Then the standard deviation in the dithizone measurement itself is (X)2 ( XX0) 0 o004 - lo5% 9 OoD. values of these nine readings are plotted on an arithmetic probability chart in Figure 31L The points are seen to fall on a normal probability curve quite nicelyo The standard deviation as read from the probability chart is also 1.5%o It should be emphasized that these statistics apply only to the determination of a single mercury sample and do not include errors occurring before the mercury solution is preparedo

-1130> 01 0) o_ -P 0) H 04OCD _10 c0m Xo X H 0 0) *HH HN00 H o H r0- 0-0-P? P 0 CI * H0II \~~Q0 -0! W H-l 0 0< * P- Q 0 0 CD NQ ~ ~ ^ 0 0 CQ CQ cm cm D CQ cm c< H O~~~~~~~~~~ (Y - - b0? o~~~~~~~~*^~ OCO (0^' CM OOO <0 ^*c^~~~~k h- <D <D (0<0<0l010tf>l CM CM CM CM CM (M CM CM CM~~~~~~~~~~~~~~~~~~i Q~~~~~~~~~~~~~~~~~~~~~~~r 0~~~~~~~~~~~~~~~~~~h

-114. There are only two steps in the dithizone measurement where errors directly affect the accuracy of the measurement: lo Pipetting mercury solution into the funnelo 2. Pipetting 0002% dithizone solution into the funnelo Errors in other operations will have only a small effect on the accuracy

APPENDIX C THE DARKENING OF a-Me-Me-Cl CRYSTALS BY RADIATION Clark(7) noticed that crystals of the compound a-Me-Et-OAc turn dark in a very short time under an intense x-ray beam. After a series of experiments in the early stages of the work reported in this thesis indicated that it would not be practical to measure radiation yields directly using the Bragg spectrometer as a source of radiation, an attempt was made to develop the color change of the crystals as a practical method of determining radiation yieldso When the techniques of obtaining relatively high monochromatic dose rates with x-ray fluorescence were perfected, it became possible to measure radiation yields directlyo It was apparent that direct measurement of yields would better meet the goals of this thesis problem, and no further work was done on the crystals. It is believed, however, that the data obtained on the crystals, as well as the techniques used, are of sufficient interest in themselves to warrant mention here. Most of the data were taken with a-Me-Me-Cl, since it proved to be the easiest member of its family to prepare in crystals large enough for convenient optical density measurements. In the very limited data taken with a-Me-Et-OAc, no significant differences were noted. Therefore only a-Me-Me-Cl will be discussed here. Measurement Techniques Clark merely reported that "the... powder changed color rapidly in a few seconds upon exposure to x-rays." This observation was verified, The clear, colorless crystals acquire a distinct brown or grayish-brown -115

.ll6color in less than one minute in the direct beam from our x-ray machineso The first step in using this color change as an indicator of radiation damage was to make these observations quantitativeo A Beckman DU spectrophotometer was used for making the measurements reported here. The DU is designed to measure oDo. of liquids contained in glass cellso The manufacturer offers no accessories suitable for measurements on crystalso Several devices for adapting the DU to measure OoDo of solids have been described in the literature, but none of them satisfied all the following requirements: lo The device should, of course, provide accurate and reproducible measurementso 2o To minimize the problem of growing large perfect crystals, the device should incorporate a slit limiting the area of the light beam in the DU' 3. To avoid difficulties in repositioning a crystal in exactly the same place for measurement at different times, the crystal should be permanently mounted on the light beam defining slit and left on that slit during irradiationo 4o The slit assembly on which the crystal is permanently mounted should not be so large as to restrict unnecessarily the number of samples which could be irradiated at one timeo A sample mount was designed which fills all these requirements and has been very satisfactory in useo A drawing of one of these mounts is shown in Figure 329 -and a photograph in Figure 33~ The mount is exactly the shape and size of the standard glass cell for the DU, so it fits

-117-.480- t.235 /4.~ /.235 ~' 3 210 DD~ ~~I6,'3 32 9 F-T TIC F1igur6 32. IsometricofCry 4'16 Fiur 3 I 32 - _ 4 136 16 THICK'1 6 Figure 32. Isometric of Crystal Sample Mount.

-118Figure 335 Beckman DV Cells, Cell Rolder, and the Crystal Sample Mount.

-119into the standard cell holdero The mount was designed to be easy to fabricate in the shop. The body of it is formed simply from a bar by three mill cuts. A number of the sample mounts were madeo The slit on the mount is adjustable. An opening about 1 mm wide was found very satisfactory with these crystals, however, and all the mounts were set to the same width. A sample mount with no crystal on it and with its slit set at a suitable width was used as a standard for O0Do measurementso Use of these mounts required a slight modification to the DU itself. The square glass cells are wide enough to pass the entire light beam from the monochromatoro The slit in the crystal mount, however, passes only a small part of the beam. The intensity of the beam is very non-uniform, and the standard click-stops of the sample carriage do not position the mounts accurately enough to give the same reading every time, To surmount this problem, a cell holder carriage was modified so the cell holder could be moved back and forth in the beam until the transmission of each was a maximum The indicated OoDo was then recordedo This measurement was fairly tedious to perform, but the results were very goodo Because of difficulty in locating the exact peak of transmission for standard and, sample when the carriage was moved back and forth, the spread in successive readings of a sample was always larger than in successive readings of a glass cello Also, a mount would occasionally be inserted into the cell holder askew resulting in a large error in OoD. To reduce the former source of error and eliminate the latter, O~D. measurements were made according to a program of which the following is a

-120typical example: O.Do of Sample No 120 when prepared: 400 mi o286, o286, o287 * 290, o287, o290 o288, o287, o288 Successive values on the same line indicate rezeroing of the standard and remeasurement of the unknowno An asterisk indicates complete removal of the standard and unknown from the machine and reinsertiono If the values on one line differed from the other lines by more than a few percent, the mount was assumed to be askew. That line of data was thrown out and another line taken. The standard deviation of the set of readings shown above is.0016 or 0o55%o The OoDo of the crystals varies with wavelength so much over the range 200-2000 mpi that it is not possible to measure O.Do over the entire interval using one reference standardo The DU cannot accurately compare a standard and sample differing in OoDo by more than 2 (or in transmission by a factor of 100)o This problem is readily solved in this case, however, by use of two slits adjusted to different openings as standards Radiation-Induced Changes in the Optical Absorption Spectrum A semi-quantitative determination of the changes in the absorption spectrum of a-Me-Me-Cl upon irradiation can be made relatively easilyo Several crystal samples were prepared, and the optical densities were measured at a number of wave-lengths. The samples were then irradiated for short times in. the full beam of an x-ray tubeo The optical absorption

-121spectrum of a typical crystal before and. after irradiation is shown in Figure 34. The change in O.Do as a function of wave-length in plotted in Figure 35~ It should be emphasized that the OoDo values are referred to an arbitrary standard and that spectra of different crystals observed were only qualitatively similaro Radiation produces pronounced movement of the "absorption front" to longer wave-lengthse The greatest change in OoDo is at about 280mp in the ultraviolet, although the location of this peak varies with different crystals and radiation doseso Quantitative Determination of Radiation-Induced Changes In an effort to pin down the observed variations, the changes in OoDo for a number of crystals were studied more intensively in the region 280-700m|o, It was found that the changes in shape of the absorption spectra varied from crystal to crystal. This is illustrated in Figure 360 The change in OoDo of one crystal showed the same behavior as that in Figure 35, while the change in the other crystal was distinctly different. These variations are not to be ascribed to OoDo measurement errors, which are relatively very smallo Measurements were concentrated on the wave-lengths 290 and 400mjo The 290mij value was chosen because the change in OoDo with radiation is maximum near that point The 400mi value was selected to find whether OoD. changes might be more reproducible at a wave-length farther removed from the maximum changeo It was realized that a major factor contributing to irreproducibility was the continuing change in OoDo after irradiationo The continued

-122. 0 Us1^. 0 r ~~~~~~~~~~~~~0 / ~ 0~,.-*zI 0 ^ ~ f <~~~~It Q =) C ID0 / cc <~~~r w.< | - OD u a o ~qE3 ~ ] 0 v 0~~~~~~ Qg Q. I s u^ it'0 -. /"C o o q ir~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~r. ir I, e I I u~ ~~~~~~~~o ~ ~ ~ ~ ~ c 2Z Io uJ n it o ed L~~~~~~~~~~~~~~~~~~~~~~~I /~~~~~~~~~~~~~~~~~~C'/- r i~ ^ Cd O r- cd Q: cv -------' —-*- J-r4 6J " o O Q)~~~~~~C > H0 U)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Q d ~ ~-e M ~~~~o O CV ~~~~o~

,1230 O O0 O E o 0) H *H 0 I- 0 0) o: 00 ~rl Cd,clr. oo -s68P d rdH n-4 oA 4 0 w oo CL'(3'0 9 f I- r-l 0~~~'cr -u ^

-1R40.8 0.7 0 SAMPLE NO. 106 E3 SAMPLE NO. 105 0.6 EXPOSURE FIELD: 200,000r GOLD LINES 0.6 0.5 0 0.4 0.3 0.2 0.1 II I I I I I 275 300 350 400 450 500 650 675 WAVE-LENGTH (mup) A Figure 36. Change in Optical Density of Two a-Me-Me-Cl Crystals by Irradiation.

-125darkening of a crystal, over a period of 40 days at room temperature after irradiation is shown in Figure 37. This effect would obviously have to be carefully controlled to give good agreement when trying to compare radiation effects under different conditions. Baking the crystals for a period of time was tried to see if elevated temperatures would accelerate this post-irradiation darkeningo The results of accelerated ageing of a crystal at 800C.are shown graphically in Figure 38. The optical density comes to a maximum in a few hours. The drop in OoDo at 400mj. after continued baking was consistently observed, although not always after only 2 hourso This accelerated ageing did not improve the reproducibility, however, but on the contrary impaired it The change in OoD. during baking varied much more among different crystals than the change during irradiation The last treatment tested offered much more hopeo This consisted of irradiating a crystal after a preliminary baking at 800~C, or a preliminary irradiation, or botho Table IV shows the changes of O.Do in three crystals through successive processes of baking, irradiation, baking, and irradiationo Table V shows the history of three crystals starting with several irradiations before bakingo There seems to be a definite improvement in reproducibility following the initial processingo It will be noted in Table V that the first irradiation produced far more change than succeeding oneso The salutary effect of initial processing may be due to bringing initial strains and flaws closer to equilibrium Most of the crystals had some flaws within the area where light transmission was measured.

2.0 1.8 1.6 O.D. 9~ 1.4 c.2I o I I I....... ~ I 0 10 20 30 40 50 DAYS AFTER IRRADIATION Figure 37. Continued Darkening of an a-Me-Me-C1 Crystal After Irradiation.

2.0 1.0 1.8 / -0.9 1.6 / 0.8 290 / 1.4- 0.7 1.2 / 0.6 =. /, E. /. 0.5 E 0 0 d 0.8 / 400 m, 0.4 d dCj d 0.6 0.3 0.4 -0.2 0.2 -0.1 0RD' 0?RRAD- 8 DAYS AT ROOM TEMP 2HRS 2HRS I ATED 80~C 80~C 18HRS Figure 38. Optical Density at Two Wavelengths of an a-Me-Me-Cl Crystal Through Irradiation, Baking, and Ageing.

-128TABLE IV CRYSTAL IRRADIATION HISTORY Irradiations with unfiltered Ti spectrum, about 39,000 r/hr with Victoreen r-metero Baking was at 80~C The operations were carried out in the order shown, from left to righto The numbers represent observed OoDo changes due to that operationo Wave-length Sample Bake Irradi. Bake Irrad. (mni) No. 12 hrso 4 hrs. 4 hrs. 6 hrs. 400 130 o139 o012 o149.030 131 o055.027 0o68.032 132.024.025.092 o038 290 130.467.080 o283 o108 131 o135 o090.145.115 132 o010 o107.142 119 TABLE V CRYSTAL IRRADIATION HISTORY Irradiations with unfiltered Tl spectrum, about 399000 r/hr with Victoreen r-meter. Baking was at 80~C The operations were carried out in the order shown, from left to righto The numbers represent observed OoDo changes due to that operation. Wave-length Sample Irrad. Irrado Irrado Irrado Irrado Bake Irrad. (mni) No. 6 hrs. 5~8 hrso 6 hrso 5.7 hrs. 6 hrs. 4 hrs, 6 hrs. 400 127.072.027.033.027 o04 o.177.023 128 o057 o025 o041 o017.039 o120.021 129 o063.033.039.017.033.087.035 290 127 o149 o107 oII6.129 o090 o720 o 080 128 - o 111 o 140 124 114 o562 o074 129 o156.117 o142.122,111.652 o072

-129A Possible Explanation for These Phenomena A literature search in this area uncovered only one publication which seemed to be directly applicable Matejec(26) has described his experiments on darkening of silver halide crystals by heatingo It was proved by microscopic examination that the darkening was caused by "colloidal" silver particles in the crystalso The phenomena described in this appendix show several points of similarity to the phenomena in silver halide crystals~ lo Matejec found that freshly prepared. silver halide crystals show an "absorption front" at short wavelengths, similar to the "front" found hereo 2o Heating the crystals shifts the absorption front to longer wave-lengths and makes it somewhat flattero Compare with Figure 360 3. After prolonged heating at lower temperatures the OoDo decreases again This was shown to be due to agglomeration of the collodial silver particles into larger particles which present a smaller total crosssection to the light beamo The similarity of Matejec's observations to the ones reported here make it very plausible that the darkening of a-Me-Me-C1 crystals is caused by colloidal mercury liberated by radiationo It is known that the darkening of a-Me-Me-Cl is not due to any electron trapping effectso If an irradiated crystal is melted, the apparent color is unchanged. During several hours at a temperature just above the melting point the the dark coloration settles toward the bottom of the melto Alternatively

-130it has been suggested to the author that these effects may be due to formation of a colored polymer during irradiationo It is of course recognized that the experiments reported in this appendix are not definitiveo They are presented for their heuristic value. To the writer's knowledge9 observations similar to those of Matejec in silver halides have not been previously reported in organometallic crystals. This may be a fruitful fi.eld for further studyo

APPENDIX D USE OF ORGANOMETALLICS AS HIGH DOSE DOSIMETERS A pressing need has been felt for some time, particularly by those working with the engineering properties of materials after irradiation, for a gamma dosimeter capable of measuring high total absorbed doses. By "high" we mean at least 1010 erg/gm (108 rads), and preferably 1011 erg/gmo A number of groups have actively sought such a dosimeter that could, for example, be placed in container of oil undergoing static irradiationo See as an example the report by Kircher et al (23) The requirements for such a dosimeter are that it be completely self-contained during irradiation, small, dose rate independent, and unaffected by ageing over a period of months or by temperatures of, say 120~Fo It should be reasonably equivalent to water in its energy absorption coefficient over the range of photon energies encountered, and its response per unit energy absorbed should be independent of photon energyo To be of widest application it should be reasonably cheap and easy to prepare and reado The system of measurement of the radiation decomposition of organometallics described in this thesis appears to offer considerable promise of providing such a dosimetero Four organometallic compounds in addition to the a-R-R -R" family have been irradiated in the cobalt-60 source and carried through the solution and centrifuging processo These compounds were diphenyl mercury, tetraphenyl tin, triphenyl tin chloride, and triphenylarsineo The first three named gave appreciable precipitates after a dose of 109 erg/gm, and gave very little precipitate from an unirradiated sampleo Thus, it seems that this technique will probably work for a wide variety of organometallicso -131

-l52The problem of equating the energy absorption coefficient of the dosimeter material to that of water or carbon does not seem insurmountable. Compounds of aluminum, magnesium, or boron might be usable with the detection system employed here. A compound of one of the elements from titanium through germanium might give an adequate energyindependence in the dosimetric sense. particularly if the metal were present in fairly low concentrationo As an example of a commercially available compound which could be used as a dosimeter if its high absorption of low energy radiation were not a problem, diphenyl mercury (Eastman Kodak No0 3868) was irradiated to several different doseso The dose ranged between 109 and 4 x 1010 erg/gm. absorbed from cobalt-60 radiationo The irradiated diphenyl mercury was carried through a solution-centrifuging procedure identical to that for a-Me-Et-OAc except that benzene was used for the first four stages since the compound is not sufficiently soluble in alcoholo Alcohol was used for the fifth wash and water for the lasto The data are shown in Figure 39 and tabulated in Table XIII, Appendix Eo The yield of mercury is seen to be linear with total dose over this range9 which includes the lower part of the dose region in which dosimetry continues to be a problem. There is no reason to assume this compound is not usable to even higher doses~ Only the time required for the necessary irradiations limited the doses used in this experimente The curve of Figure 39 gives a value of G(Hg)(Co-60) = 1olo In summary9 we may list the characteristics of a good organometallic dosimeter materialo lo Stable to expected environmental conditions except radiation~

51335 2o G (metal) (Co~6O) less than 2, so the dosimeter will not saturate at too low a doseo 3o Soluble in a common solvent so the solution-centrifuging analysis procedure can be usedo 4o Yielding a reproducible precipitate linearly with total dose but independent of dose rate 5o Possessing an energy absorption coefficient similar to water or carbon for photon energies expectedo 60 Preferably commercially available. though this is not vital if the dosimeter has outstanding characteristicso

-134200 100 E E 0~ Q / DOSE RATE= < | / 5.0 x Io erg/gm -hr I g 10 10 10 Figure 39. Diphenyl Mercury Irradiations with 9 An I0 I0 10 ABSORBED DOSE (erg / gm) Figure 39. Diphenyl Mercury Irradiations with Cobalt-6o.

APPENDIX E TABLES OF EXPERIMENTAL DATA -1355

-1^6TABI% VI LOW-DOSE STATISTICS RUNS Background = 0.06 mg/gm (1) (2) (3) () (5) (6) (7) Sample Weight Irrad. Yield Gross Net Net Rank No. (gm) Time (mg) Yield Yield Yield (min) (mg/gm) (tag/gm) 1/2 gm ~/ ~- 06 Sample 75 rin 278.50,05 75.093.121. 0605 2 279.5113 75.122.238.173 o0865 9 280.5019 75.103.206.141 50705 281.4920 75.104.212,147.0735 7 283.5071 75.098.193.128.o64o 3 284.5086 75 o086.169.104.0520 1 285.5059 107.5.133.263.198.0695 4 293.4884 75.107.220.155.0775 8 294.5135 75.107.209 144.0720 TABLE VII HIGH-DOSE STATISTICS RUNS Background'= 0.06 tng/gm (1) (2) i "(32) (4) (5) (6) Sample Weight Yield Gross Net Net Rank (gmin) (mg) Yield Yield Yield (mg/gm) (mg/gm) 1/2 gm 0/0 ~~.06 SamPle (mg) 297.0535.731 1.46 1.4o.700 5 298,.4942.701 1.42 1,36.680 3 299.5144.720 1.40 1.34.670 2 300.4970.735 1,48 1.42.710 6 301.4960.739 1.49 1.43.715 7 302.5014.765 1.52 1.46.730 8 303.4973.786 1.58 1.52.760 9 307.5079.730 1.44 1.38.690 4 308.5304.690 1,30 1.24.620 1

-137U.) 0 CQ -P 43 - O O O o - n _: O OO - O H U) oD OOHO(-O\O LAHO\J o 0 0 -H P t r *Z1 HLC \ T l 4 JOO JOHOO\ o 0 4 H > 0 N n o O 4 4 0 0 \ Q - P4 0 ^ C O Ua,~ ~~~~~~~~. r,\ o oo r q_+ o H l A O o o 0.rd D P.0 r O bO 4 L- -0\O —LrO-I r0 - O H OH r C qO U)l C3\0) -.' X H-Lr\'C\ 0 O r-_ 0 H H - * 0 0O.,.},'.9 r1-l w 0 4lf 4 0 0 * o -I CHO - _co ~o 44 r -' J C- I *rd H H H H HHr- H-H Hb 4- 00 0 O O <C L / %C N 00L 0 O0 a * 0\ * O..._. m 0. "-C (DL3'~ L CX:U) O 0000 0. L'O O O O t C L Cb 0 r12 X L r- H O C, pC A - H r 1H 0 0 O 0 O H C co ^^ O~r-I^HlAOJLAH OJOOOOJ Oq OJ i1 t 1 O O Or 00 0 -0: A00' 00 000 0Q +Q St rq 0 0 OQ Q 0 0O- ) r4 rC4 C rH 0 0(3. ~H H J OH4 -= H.. 0 00000000000 0 0 -p * a -OOOO~~~~~~~ I P4 p _* C ~ U) - o ~J 0 U),L4 b 0 L O L F- \ —- \o. 0 %Lr\L 0%\ 0%\,-! 0 r' ~ - - - cTJ Le:: b-O - - - -LN- I 000000000000 0 U, o -s1) *H Q -t d f i LC\LALL(NLrA OOOOLA0 0 o 0. ) n 00000000000 00 0 0) U,-^ b C <kl~ < 0Ct< C~ 404 0 0\ 0k C C C 0J d p` -C, q 0 P 3 0 r4^~r'0?~ bD; lClP p p p qpqpQ O S r% 4)I-0 HHHHC~oJC~icM~r\00 00 ^ ~ ~ ~ ~ ~ ~ ~ ~ ~ N

158 TABIE IX DOSE AND DOSE RATE DEPENDENCE WITH X-RAYS Irradiations with Mercury L-spectrum Sample Batch Absorbd Irrad. Absrbd Yield Net No, Dose - Time Dose (mg) Yield Rate (hr) (erg) (mg) _~ ~. ((erg/sec) 211 II-A 0 0 0 o0l4 0 209 II-A 612 20 4 4 x 107.175.161 210 II-A 612 12 2o64 x 107.080 o 66 213 II-A 612 20 4,4 x 107.172.158 214 II-A 612 12 2.64 x 107.125.111 220 II-A 612 48.2 10o6 x 107.390.376 221 II-A 306 42.85 4.71 x 107.171.157 229 II-A 612 6 1.32 x 107 o0585 o044 230 II-A 153 53-75 2.96 x 107 o108.094

-139C0\0 7\\ C0\0\C\ 0\0\CY\ C\ \ 0\ 0\ 0\ O \ 0\\ O 0\ 0\ 0\ 0\ \ o''o'o'o'o'o 1 I o'o'o b'o1' o bb'o 00000000 0000000000000,^ 1HH H HHHHHH H HHHH HHH H HHHH -~- ~ ~ ~ O~ o~ o~ ~ o~ OkO rqcO u~ O~ O ~1 O!-~ ~1 o~1 ~1 oJ oJ~ o0 co foE Sm< H}.~'-~... m m~ o o o o ~...-~ m o.... -, m o..... o... CQ r_ -i r CC r l CO \ L Un C r\ CC _ \ m n ri a\ CU C H-Q-) o' - c c..co C0 cOcccccOOO=j- c 0o o00 -"o u.-x 1..HLr'x,cru-\ rd o UN-j- o H bDH I \ C 0 0 o LrCM o o o 00 o oo o o\ o n \o 0\0 4 ) a) S. * x -- ON n \ H M H H o ON \ O n rC CY ) H n MO H O\Lr\ r\c CQ rd Hq H - Cc\ co LC\ O0 L\n 00 - Lr\ 0) O Hs Un D O\ -O Lr H _-O O ) 0 -0 L'1 0O<U L\ \D1 H O C OC\ L O O \O cr)f c O O - O COO\ Cj Cr C H?-I~HC ( n mc nc ^^ 4Lr\( a Ci ri r Ln 4 \,O Lf\ Lr\-:I- rA Coo\ Cd j C' h0 CY\ ON NX N \L 0 0 0[ r\L \ Y L\ 0 \ U N co S _,.,. *,,*. a. a..........* * ~ t A 0 ON Lr- - pt- t- t- b- t- - M- tc- t- l - M - -b — \ -t- c - t- -b-b t-iL- A-t-i 0 CO CO O0 0 O 0 OC o 0 O 00 00000000000 0 ~ cO L O N 0 \D < L\ c0 c a C OO CU Q0\ (O O o a 0 O\L0t O O CO)r-Lr0oo CO M C \ \ C) Lo \ Lr 0i\rU-\ UN H H CM CM CM LCI\4 CM L4 L) 4\ t tD LC) L\ \ C M H 0 O A rl 0 0 0 0 n L\ 1 —\ 0 C 4C M \r\r 0- \ \C cO 0\ \c fi LC\ L tc Ccc 0c h4 k- I) n-f — -,. O O cO o0 0 O -- t —=- f 0 0 0 o M O 0 o O 0 c cO oo o cO M o cO LD L H Lr, LrP 4- u E-1 H o >< H a C) 4-0 0 u ON ON O \Cr\ ON 0\ \o o n o \ C \ON\0\ ON C\o n o O O\ O F^ mo <u ~i( osC roc c oo o o o o o -o t oo o o o o o o o o co 0o 0 L O o 000 0. *...0 0.. a. 0...0. 00 Lf\..\ B uo N C)\1 c 0\ \(xl\F\ c coo \ H A- I, Q ) l *p o,9 cOOOO OOoHHHHHOOH SD a s X LOdOOOOOOOO IxJ ^-^ OJ su *0. o*.t.^.~ i,;.^.o.......o... h C CM - ~ o C~ cO I\ CW \D O CU If cO c. Q- I O C r r o o o cO o o (N C k-c 0 CI U cO 4 c O cN o ONiN L ri ^ O 4 O4O4 OO'OO 0\- k\ —-0- 0 c -00c N (' -NH CM h^ Ui $i i'sma ssti'Zm an ^^i5i''ai^ H a: ---- t * ^ Lrrl h h 4(Q Q Ol 4 X- 0 ir\ Q ainchOhO@ ir h X 0 0\(7 O l^ 0h ^^^ ^ipLfz r\rti\\D \^ ^I>lm coocHOi-mjrnn W; c ( ( ( ( o < ( o (i j aj j (j( a a c jp ( f r o CQ!

C O\ O o O O O\ 0 O 1 - B 0 I I B 11 B o 0 0 0 0 0 0 0 0o 0 >b d H3 H r-4 r-,-1 r — r-4 rd r — A r-d i — 0 rH H H H 0Hg-% B N N N N X N N X N *C-., -I 00 CM L.\ C' \D O 0 O (\j (Y C 00 CM 00 (Y') C\j CY) (\J HHCM 0 (D00 00 \0) 00.C 0/ p^ <^ @V^ 0\ — 0 0B, —I L 0\ 0 0 O 0 0 "* c'q CY \ID CoY c) o' Lrx ~-~. ~. - 0CO MO CO CO c o — q- LC\ cO CO co ~ -~ 0 to _:I- MO MO L-= H — t \ CiDf) C00) _0- 00 0 0 0 0 0 0 0 0 0 0 0 p o ~0 0 0 0 0 0 0 0 0 t) Qrd ~ 0 L — Q 0 Lr'x 00 U'X LrX rH 0 PH ON ON O C..O 0o O 0 00 00 CO. C\J -.- LC - t- l -- 0- - S- (P r 0 0 0 0 0 0 0 0 0 0 cxO 0 0 0 0 0 0 0 0 0 0 \MO O 0 C\J O O 0O O 00 OO'o (X0 0 0 0 0 0 0 0 co 0 0 00000 r-0CM C\ r-q L co B-E vs-^ r( rIOj r-4 r-I r-I id ^ L1L L \' L00 L — ~ z ~ C a) M CM -- MO Hr- 00 _- Lm 0\ O P4 o o%^H E -00 o 00x ON o00 00 VO bi +3 LC\0 Cy-) t- L0 0 C9 CC) 00 C\m \MO,O-O M O MO MO LO.O.-. 0 0 0 0 0 0 0 0 0 0 o o J 0 M 0 \ ON D ON M: O O \D \O. O El Jd rd 0'O,.I c cj 0.1 04\ 0 0 " 0 c C o'J -P CM CM CM CM C' ON 0 O r H 0CC B^rM t 0 0 00 00 ON m M 0 M ozP..- mx mx L: zp ^ m -- LPH 0 0 0^0 c o 0\ r\ \0 cO b- D ~* o- -j\ _:I — L(lB ^-^ (D^)fq00 0 0 00 00 00 0 0 0 00 CO

-14l, TABLE XII IRRADIATIONS WITH UNFILTERED L-SPECTRA Sample Radiator Absrbd Yield Yield No. Dose (mg) Per (erg) Unit Energy (mg/erg) 153 Pt 11.7 x 107.48 4.1 x 10-9 154 11.7 x 107.48 4.1 x 10-9 146 Au 11o9 x 107.385 3.24 x 10-9 157 11.9 x 107.638 5,36 x 10-9 147 Tl 12.7 x 107 50 3.95 x 10-9 149 12.7 x 107.38 2.99 x 10-9 148 Pb 11.9 x 107.31 2.60 x 10-9 174 23.8 x 107.68 2.86 x 10-9 155 Bi 13.4 x 107.575 4.30 x 10-9 156 13.4 x 107.545 4.06 x 10-9 151 Hg 10.5 x 107.44 4.20 x 10-9 152 20.6 x 10.56 2.72 x 10-9 158 10.5 x 107.309 2.94 x 10-9 166 21.0 x 107 o73 3.48 x 10-9 167 10.5 x 107.299 2.84 x 10-9 201 31.4 x 107.153 4.87 x 10-9 204 23.6 x 107.664 2.82 x 10-9

-l42TABLE XIII -DIPHENYL MERCURY IRRADIATIONS All irradiations in cobalt-60 source. Absorbed does rate = 5.0 x 107 erg/gm-hro Sample Weight Irrad Absrbd Yield Yield No. (gm) Time Energy (mg) (mg/gm) (hr) (erg/gm) 212.6645 0 0.008.012 207.9736 4809 244 x 107 5.49 5.65 309.5052 43.3 216 x 107 2.65 5.25 310.5100 23.8 119 x 107 1o415 2.78 311.4950 834.15 4160 x l07 5Q08J 103,312.5116 12.4 62 x l07.524 1.02 313.4938 6.0 30 x 107.218.443 314.5035 3.0 15 x 107.088.175 315.5196.535 2.-68 )x l07 o005.00965 316.o4972 212.6 1060 x 107 14o175 29.7

REFERENCES 1, Bethe, H. "Quantum Mechanics of One- and Two-Electron Problems." Handbuch der Physik, 24, 273, (1933). 2o Broadway, N. Jo The Effect of Nuclear Radiation on Metallo-Organic Compounds and on Polyethylene. Tech. Memo. No. 3, Battelle Radiation Effects Information Center, Columbus, Ohio, (1958). 35 Brode, W. R. Chemical Spectroscopy. New York: John Wiley and Sons, (1943). 4. Burhop, E. Ho S. The Auger Effect. Cambridge University Press, (1952). 5. Burton, M. B., Hamill, W. H., and Magee, J. L. A Consideration of Elementary Processes in Radiation Chemistry. Second Geneva Conference Paper A/CONF 15/P/916, (1955), 60 Catcheside, J. D. and Lea, D. E. "The Effect of Ionization Distribution on Chromosome Breakage by X-Rays." J. Genetics, 45, 186-196, (19453). 7. Clark, G. L. Applied X-Rays. New York: McGraw-Hill Book Co, (1955), 8. Clark, G. L., Pickett, L. W,, and Johnson, E. D. "New Studies on the Chemical Effects of X-Rays." Radiology, 15, 245, (1930)o 9o Clendinning, W. Ro and Atkins, M. C. Methods of Monochromatic Irradiations and Absolute Dosimetry with Soft X-Rays, Resonance in Radiation Effects Report No. 3, The University of Michigan, Ann Arbor, (1960). 10. Compton, A, Ho and Allison, So Ko X-Rays in Theory and Experiment. New York: Van Nostrand Company, Inc., (1935) 11. Dainton, F. So "Chemical Effects of Radiation." Radiation Research, Supplement 1, 1-25, (1959)o 12, Emmons, A. Ho Resonance Radiation Effects of Low Energy X-Rays on Catalase. Ph.D. Thesis, The University of Michigan, Ann Arbor, (1959). 135 Fano, U. "Principles of Radiological Physics." Radiation Biology, A. Hollaender, Ed., VolO I, 1, New York: McGraw-Hill Book Co., (1954). 14. Fine, S. and Hendee, C. F. "X-Ray Critical Absorption and Emission Energies in Kevy" Nucleonics, 13, No. 3, 36, (March, 1955). 15. Fischer, Ho Wiss. Veroffentlich. Siemens-Konzern, 4, 158-70, (1925)e -145

-14416o Fischer, H. and Leopoldi, Go Z. anal. Chemo, 103, 241-257, (1935)o 17. Fister, H. J. Manual of Standardized Procedures for Spectrophotometric Chemistry, Standard Scientific Supply Corp., New York, (1950)o 18, Garsou, J. Contribution a l'etude de l'efficacite des rayons x monochromatiques sur quelques systems dihalogenures organiques solides et liquides. PhoDo Thesis, University of Liege, Belgium, (1959) 19. Gomberg, H. J. Technical Report No. 1, Resonance in Radiation Effects, University of Michigan, Ann Arbor, (1960)o 20. Gray, Lo Ho "Probleme des Relations Entre le Rendement Radiochemique et les Aspects Physiques de la Qualitie des Radiations." J. Chim. Physique, 48, 172, (1951)o 21, Hardwick, T. J, "The Effect of the Energy of the Ionizing Electron in Irradiated Aqueous Systems." Disc. Faraday Soc,, No 12, 203-211, (1952). 22. Heal, Ho G. "The Chemical Effects of Ionizing Radiation in Solidso" Atomics, 6, 205-208 and 241-247, (1955)o 23. Kircher, J. F., et alo Investigation of Effects of Radiation Applicable as Gamma Radiation Dosimeterso Wright Air Development Center Technical Report 58-158, (1958)o 24. Lang, Eo P. and Nelson, Ko Wo "Report on Mercuryo" J. Assoc. Official Ago Chem,, 25, 399, (1942)o 25o Lea, D, E Actions of Radiations on Living Cells. Cambridge University Press, (1955). 26, Lea, Do Eo and Catcheside, Jo Do "The Mechanism of the Induction by Radiation of Chromosome Aberrations in Tradescantia." Jo Genetics, 44, 216-245, (1942)o 27. McGinnies, Rosemary To X-Ray Attenuation Coefficients from 10 Key to 100 Mevo Supplement to NBS Circular 583, National Bureau of Standards, Washington, (1959). 28. Mellan, I. Organic Reagents in Inorganic Analysis, Blakiston, Philadelphia, (1951)7 29. Miller, No "Radical Yield Measurements in Irradiated Aqueous Solutionso" Radiation Research, 9, 633-646, (1958) 300 Platzman, R, L. "On the Primary Processes in Radiation Chemistry and Biologyo" Symposium on Radiobiology, Jo Jo Nickson, ed., New York~ John Wiley and Sons, (1952)o

-145310 Roos, Co E. "A New Method for Measuring Auger Transitions." Physo Rev,, 93, 401 (1954)o 32, Sandstrom, A. Eo "Experimental Methods of X-Ray Spectroscopyo" Handbuch der Physik, Volo XXX, 78-245, Springer, Berlin, (1957) 330 Schrauth, W, Schoeller, Wo, and Struensee, Ro "Uber einige komplexe Quecksilberverbindungen des Zimtsaure Methyln und der Zimtsaureo" Ber,, 43, 695, (1910) 34. Schrauth, W,, Schoeller, W.,. and Struensee, Ro "Uber einige komplexe Quecksilberverbindungen der Zimtsaureester und der Zimtsaureo" Ber,, 44, 1048, (1911). 35. Schrauth, W,, Schoeller, W., and Struensee, Ro "Uber Atherverbindungen der P-Phenyl-hydracrylsaure." Ber,, 44, 1432, (1911) 36, Seeman, Ho E. "Spectral Sensitivity of Two Commercial X-Ray Films Between 0.2 and 2.5 Angstromso" Revo Scio Insto, 21, 314, (1950) 37. Serment, V. and Emmons, Ao Ho Data and Graphs on the Dose Rates In and Around the 5000 Curie Co-60 Source. unpublished. report, Phoenix Memorial Laboratory, University of Michigan, Ann Arbor, (1958)o 38, Siegbahn, Ko Beta- and Gamma-Ray Spectroscopyo Interscience, New York, (1955)o 39~ Storm, Eo, Gilbert, E,, and Israel, Ho Gamma-Ray Absorption Coefficients. LA-2237, Los Alamos Scientific Laboratory, (1958). 40. Tolbert, B. Mo and Krinks, Mo Ho Abstract of Paper, Bioenergetics Symposium, Brookhaven National Laboratory, (1959). 41. Welcher, Fo J, Organic Analytical Reagents, Volo III, 463-550, Van Nostrand, New York, (1947).

UNIVERSITY OF MICHIGAN 3 9015 02499 5337