W“ “W W .1 ABSTRACT THE L—AUGER SPECTRUM OF IN113 AS DERIVED FROM THE ELECTRON CAPTURE DECAY OF EN113 by Raymond J. Krisciokaitis The low energy range of the M.S.U. 77": iron— free beta-ray spectrometer has been extended down to l kev by means of electrostatic post-focusing acceleration of the electron beam. It has been shown that accurate relative intensity measurements of electron spectra are possible at least down to these energies. The spectrometer data acquisition system has been automated to facilitate more reliable and efficient data taking particularly as this applies to weak low energy Auger spectra. 1 , , 13 as derived from The L-Auger spectrum of In the electron capture decay of Sn113 QV2 to H kev) has been measured and the major Auger lines and groups identified. Order of magnitude of the relative intensi- ties of these lines has also been determined. The effective incremental charge AZ has been determined for the most prominent and important transitions. AZ = 0.55; 0.67 for transitions of the type Ll,2,3-Ml,2,3Ml,2,3; Ll,2,3-MM,SMH,S’ respectively. It has been shown that the spectrum exibited intermediate coupling features and that shifts of Auger lines due to double vacancy effects probably exist. u. THE L-AUGER SPECTRUM OF IN-ll3 AS DERIVED FROM THE ELECTRON CAPTURE DECAY OF SN-ll3 By Raymond J? Krisciokaitis A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Physics and Astronomy 1965 all ACKNOWLEDGEMENT I would like to express my deepest appreciation to Professor Sherwood K. Haynes for providing me the Opportunity to carry out this investigation and for giving me invaluable guidance throughout the course of the problem. Professor William H. Kelly has been a kind and, helpful adviser during the period of Professor Sherwood K. Haynes absence. Many thanks to him. Professor Hugh McManus has given me important perspective concerning some theoretical aspects of the Auger process. I express my gratitude for this. I am grateful also to the men in the electronics and machine shops for their aid during the construction phases of the problem. Other peOple to whom I wish to express my grati- tude are Robert Francoise and George Flemming, who eagerly assisted in various ways in the preparation of the thesis, Jerry Lerner, who was in charge of the iso- topic mass separator at Argonne and Argonne National Laboratory for the use of the separator. ii :l ACKNOWLEDGMENT. LIST OF TABLES. LIST OF FIGURE CHAPTER I. CHAPTER II. CHAPTER III. CHAPTER IV. REFERENCES. TABLE OF CONTEIITS S o o o o o o o o o o o o o 0 INTRODUCTION. . . . . . . . . . . . AUTOMATIZATION OF THE DATA .AC QUISITIOIQ’ SYSTEM. 0 o o o o o o o A. Introductory Remarks. . . . . . B. Design and Operation. . . . . . POSTFOCUSING ACCELERATION FOR A 7“!“ SPECTROMETER. . . . . . . . A. General Design Considerations B. The Acceleration Cell and Sup- porting Electronics . . . . . . C. Operation and Performance . . . THE L- AUGER SPECTRUM OF IN113 AS DERIVED FROM HE ELECTRON CAPTURE DECAY OF SNll. . . . . . . . A. Introductory Remarks. . . . . The Auger Effect. . . . . Source Preparation. . . . . . Measurement . . . . . . . . . The Spectrum and Discussion . . *IitIJUOUJ Conclusion. . . . . . . . . . . iii Page ii iv 0\ l7 l7 23 29 38 38 A2 #5 1+8 1+9 64 11' LIST OF TABLES Table Page I. Estimated Relative Vacancy Distribution fOI‘ the K, 111, L2, L3 Shells. o o o o o 52 II. Energies and Relative Intensities of the L-Auger Lines . . . . . . . . . . . . 58 iv LIST OF FIGURES Figure Page 1. (A) Schematic Representation of Automated Data Acquisition System (B) Power Flow Diagram . . . . . . . . . . 10 2. Details of Fig. l (B) and A = 20 Program. . 1% 3. Noise vs. Post Acceleration Voltage from Giroux and Geoffrion, ref. 25 . . . . . . 21 4. Postfocusing Acceleration Cell and Voltage Connections . . . . . . . . . . . . . . . 24 5. Noise Curves Showing Retardation Effect . . 31 6. Noise vs Retardation Voltage With Fixed Accelerating Potential. . . . . . . . . . 32 7. Noise vs Accelerator Voltage With Several Focusing Fields . . . . . . . . . . . . . 33 8. Sections of L and K Auger Groups of In113 Demonstrating Accelerator Performance . . 37 9. Decay Scheme of Sn113 from Schmorak, Emery, and Scharff-Goldhaber, ref. 12. . . . . . MO 10. L-Auger Spectrum of In113 as Derived From Sn 13 . . . . . . . . . . . . . . . . . . 57 11' CHAPTER 1 INTRODUCTION Contemporary experimental investigations of nuclear structure and nuclear forces consist of a vast array of techniques and methods represented by a variety of instruments. It is generally found that no one tech- nique or method will monopolize an area of investigation, for it will either be displaced by a new and an improved method, with greater probing power, or will be forced to refine itself, and thus continue in its usefulness by complementing its competitor. The underlying premise for success of such a pattern of development is that the scientist must respond to the challenge and stimulus provided by new discoveries, both in the theoretical and experimental areas. Low energy nuclear physics, more specifically nuclear spectroscopy, has made a quantum Jump during the past decade, both in terms of instrumental improvements (e.g., perfection of high resolution electron spectrome- ters and solid state detectors), and advances in theory (e.g., critical study of the conservation law of parity in weak interactions by Lee and Yang). This is attested' in the recent edition of d-fl-J-Ray Spectroscopy by K. Siegbahnl which doubles the amount of material in the previous edition. 2 One of the more recent deve10pments has been the refinement of the theory of orbital electron capture by 2,3,H nuclei as formulated by Bahcall. Orbital capture theory has been presented by Marshak,5 Brysk and Rose,6 0diot and Daundel,7 and reviewed by Bouchez and Depommier,8 and Robinson and Fink.9 The latter have pointed out a systematic discrepancy between measured L/K electron capture ratios and the predictions of Brysk 6 2’3’h which include and Rose. Bahcall's calculations, atomic variables in the initial and final states and take account of exchange and imperfect overlap effects, tend to bring theory and experiment closer together. Besides stimulating the need for more precise measurements of electron capture ratios using the accepted techniques,8’9 the refined theoretical results warrant the investigation of the usefulness of modern precision spectrometers as tools for such refined measurements. The M.S.U. iron—free double-focusing 7/72 beta- ray spectrometerlo seemed to afford an excellent Oppor- tunity for this kind of study. It is a precision instru- ment capable of very high resolution. A detailed account of its characteristics may be found in reference 10. The global problem, as envisaged at the start, was to be the measurement of the relative orbital capture probability ratio L/K, and possibly Ll/Lz/L The key to 3. the measurement was the study of the L-Auger and K-Auger electrons (I. Bergstrom and C. Nordling ref. 1) created 3 during the deexcitation of the atom, after it has been ionized by the orbital capture process. Noteworthy is the fact, that the Auger process itself has not been thoroughly investigated. The exception is the K-LL Auger group. The other groups, such as the K-LY, K-XY and the L-XY, have been investigated relatively little, both ex- perimentally and theoretically. Thus in effect, the Auger process for the element in question must be studied experimentally and more fully understood and interpreted before Auger spectra can be used in an analysis to obtain the electron relative orbital capture probabilities. The choice of the specific capture process,ll Snll3(1l9d)-££L+>- In113m(1.7h), which in turn decays to 113 with the emmission of a 392 kev the ground state of In gamma and conversion electrons,* was made for the follow- ing reasons. 1. The study of the L-Auger spectrum emanating from a medium Z atom had never been attempted previously and our attempt to study such a spectrum involving the diffi- cult problem of detection of low energy electrons was a matter of interest by itself. 2. Additionally, Z = #9 is a good place to search for energy shifts of L—Auger lines due to double *A weak electron capture branch (2%) also feeds 113 a high lying (6H8 kev) In state. THE! 1, vacancies following the K—LL transitions.* The under- standing of this effect would be important for any Z if the relative orbital capture ratio such as Ll/L2/L3 is ever to be measured. 3. The radioactive substance was easy to produce due to a large neutron capture cross section12 for Snll2, which could be readily obtained possessing an isotopic 113 purity of 72%. Also, Sn could be easily separated 112 and other isotopic contaminants by from the stable Sn means of an electromagnetic mass separator, thus ful- filling the need for a high specific activity source. E. It was found during earlier experiments in this laboratory that separation of trace amounts of In113 from 113 was possible by means of a vacuum evaporation tech- Sn nique using the much higher volatility of Indium to an advantage. This gave us the capability of observing the Auger process separately, due to the vacancies created by the conversion process. It is inherent in the study of Auger spectra (generally of low intensity) that the experimenter is faced with lengthy running times. For the sake of ex- pediency the data taking procedure of the spectrometer *The K Auger yield (aK) is substantially higher in this region than for heavy elements (I. Bergstrom and C. Nordling ref. 1). THES 5 was automated. This is treated in CH II. Our choice of the above electron capture process required the extension of the low energy range of the spectrometer down to at least 2 kev by means of post— acceleration of the focused electron beam.* Therefore, a successful post-focusing acceleration cell had to be developed in order to insure reliable electron intensity measurements. It is described in CH III. The requirement of producing extremely thin sources (several atomic monolayers in depth), needed to avoid source absorption losses and consequent spectral distortion, constituted also another problem that had to be solved. Chapter IV shows the results of the composite (capture and conversion) L-Auger spectrum of Snll3-In113 and constitutes the preliminary attack on the problem of measurement of the relative orbital capture probabilities through the examination of the Auger spectrum by means of a modern high precision electron spectrometer. *The 2 kev figure is the endpoint energy of the L-Auger spectrum for elements with Z~50. TH! CHAPTER II AUTOMATIZATION OF THE DATA ACQUISITION SYSTEM A. Introductory Remarks It has been said that necessity is the mother of invention. This chapter is an illustration of such a phenomenon. Peculiar to low transmission instruments such as the ”fispectrometer at Michigan State University, is the fact that data collection or acquisition is usually a very long and tedious process requiring running times often counted in units of months. This is particularly true if one confines himself to the study of Auger spectra, which generally are weak in intensity to start with; the gathering of statistics is therefore very time consuming. Once the spectrometer is set in operation it is best to collect data continuously and therefore mini- mize the rate of equipment breakdown and consequent down— time. Furthermore, when studying short half—life sources, the need for continuous collection of data becomes even more stringent. Typical sampling of momentum is of the order of two to ten minutes per momentum position. This means that, before spectrometer automatization, the mo- mentum step—up or step-down had to be made manually after each counting interval, requiring the Operator not only to change the momentum value by turning the knobs on the 6 el 7 10 but also to reset the counting decade voltage divider, cycle. Needless to say, then, automatization of such a system was desirable, not only from the standpoint of doing good physics, but also from the standpoint of con- venience and the saving of the costs of manual labor. The approach to this problem was one of an in- ventor rather than of an electrical engineer. The heart of the design is a multistation electromechanical rotary switch.* It is powered by a direct current rotary sole- noid** which is coupled to an output shaft through a floating ratchet mechanism. The shaft drives the rotors of a series of wafer switches*** which may be programmed to perform multiple switching functions. A set of four of these rotary switches is mechanically coupled to the lower four decade units. By a programming process uti- lizing the wafers to drive the decade divider (or step up the momentum or magnetic field) systematically at predetermined intervals, one may sweep the momentum region of interest at preset time intervals. A workable model was constructed and Operated for 500 hours during the RaD L Auger run.13 It was later discovered that what appears to be a similar design was already in Operation at Uppsalalh with a magnetic spec- *Ledex Inc., Dayton, Ohio. **Size SS awg.no. 26 (no interrupter). ***l-pole l2-throw W-548 (phen.). El 8 trometer for neutron capture experiments. The present system is modest in terms of the data acquisition and data handling systems that are in vogue at many larger laboratories. The ground work for a more sophisticated arrangement has been laid, however, and the basic circuits of the present system can be easily extended to a more versatile automatic operation of the spectrometer. B. Design and Operation Figure 1A illustrates the basic elements of the automatized system. For automatic operation the timer* is set to count for a desired time interval, and the starting momentum position is set manually by adjusting the decade divider. The intervals at which the momentum region of interest is to be swept are programmed on the program deck. When the scaler count switch is thrown, counter pulses start accumulating until the preset counting time interval elapses. Any time during this first counting interval for the starting momentum po- sition the power switch connecting the d.c. supply (0-36 v) to the pulser may be thrown (switch 84 fig. 2).** *Eagle Signal Corp., MOline, Ill. **Switches SO, 51’ S2, S3, remain normally closed for automatic operation. They were primarily installed for testing isolated Operation of individual solenoids. THES TH! Figure l. (A) (B) Schematic Representation of Automated Data Acquisition System. Power Flow Diagram. .H mmson zamoso 33m $38 Am: A _ _ Zgaoma a 32200 ¢m>o 5.53 6 >._._n=m<..=<>< «530n— mo.ozw._om rmdhom cum; 3.. ‘l 10 . «kudzm two—>5 wom 20.5.5509“ $.40 09529—54 Aq: no 20—h4kzwmwmawm oFSzwIUm _ nompzou _ . .Esm — 5&8 on _ g 5:: meozuqom Egret «we _> .o m 0350) wo9. Ia 13:23...qu nnnnnnnnnnn no. -o nnnnnnnn . u - .. u n - u u 828 _ . . .x v ’N v D! .Q Av . u 3.0.“ .39 38. .v n . 7 Av 8.8 4v #v . u 2... ta u . a.» , coo. . . . . “v . . Av I u m lg ”3.8N T . ‘ I . l “ 3n Av 80.!“ . u .8. M" n 9...! "GP :6 " coronuo _ _ . V 1 F 1.. em 15 operation of the pulser. The pulser will then provide eleven pulses and be turned off (by the last wafer in the program deck) bringing the wafer rotor tabs of the program deck to the initial position.* The pulser will remain off until the next print command is given. In other words, during this process twelve pulses are gener- ated and all of them (or some number less than twelve) may be directed to pulse any of the four driving sole- noids. For example by connecting the 11 and 0 contacts on the second wafer from the left (program deck), the 101 solenoid is permitted to turn two steps (1/6 revo- lution). If it is desired that the 101 solenoid would take, say, six steps then contacts 11, O, l, 2, 3, E, would need to be mutually connected. Thus, information is stored in this relatively crude way. The carry over operation for the A = 20 case is accomplished by mutually connecting contacts 9, 10, ll, of the top wafer in the 101 column. During the last Operation the lower wafer provides power transfer to the 102 column. With the present system the decade divider may be programmed to go unidirectionally in steps of 5, 10, 20, 25, 50, 100, 200, 250, etc. which is sufficient for *The rotary solenoid of the program deck com- pletes one full revolution. 16 the typical problems encountered so far.* More versa- tility could be provided by a better memory, and by installation of bidirectional solenoids. The programming from one mode of Operation to another, which takes now about two minutes, could be eliminated by prewired circuits. A program may also be incorporated which would cycle a given momentum region. An extra wafer would need to be used in the 103 column.** *The A = 5, 10, 25, etc. programs are not shown in this diagram, since very little mental exercise is needed to program them once the fundamental operations of the A = 20 mode are understood. **The third row of wafers in fig. 2 are spares. THE81 CHAPTER III POSTFOCUSING ACCELERATION FOR A 7772 SPECTROMETER A. General Design Considerations Detection of low energy (<5 kev) electrons with constant efficiency or transmission is a frontier region in beta-ray spectrosc0py. Physicists have employed, with various degrees of success, several methods of de- tection of focused beams as they emerge from the spec- trometer.15 For example, it has been found that a de- tector consisting of a scintillator and a photomultiplier tube removed via a light pipe from the spectrometer mag- netic field and cooled to liquid nitrogen temperatures can be used effectively to detect electrons above roughly 20 kev; however, below these energies one is confronted with varying detection efficiencies. The efficiency of the scintillator and photomultiplier combination de- creases with decreasing energy, and careful detection efficiency curves must be run to insure reliable intensi- 16 Electron multipliers may be used, but ty measurements. here again the detection efficiency is by no means constant for electrons of different energies. MOst commonly, gas filled counters are employed. They require a window to keep the counter gas from leaking into the spectrometer vacuum system, i.e., an electron absorbing slab of matter is introduced into the beam path, thus 17 THESI 18 requiring the experimenter to consider its transmission properties. If the window is a very thin film, (say collodion,nr30 ug/cm2), then electrons down to approxi- mately 10 kev may be detected with 100% transmission. However, if the spectrometer is adjusted to focus e- lectrons of lower energy, than absorbtion corrections have to be made.17 If the detection of even lower ener- gies is desired, then one will find that for the above mentioned film the transmission of electrons will stop completely at about 2.5 kev (so-called cutoff energy) where the gas counter becomes effectively useless. Its usefulness may be extended by the use of even thinner films. However, there is a practical limit to this since films below 10 ug/cm2 GVl.l kev cutoff) tend to be diffi- cult to produce, tend to be leaky and destroy their usefulness by creating poor vacuum, and thus introducing another type of unreliability in measuring relative intensities. For example, an electron of 2.6 kev energy will have a probability of 0.03 to be scattered out of the beam while traveling through one meter in a vacuum H 18 of 10' Torr. The author has found that it is possi- ble by means of a support to manufacture windows,210 2 which can maintain a vacuum better than 10-5 Torr ug/cm and which will support counter pressures of 55 mm Hg for several weeks. The support that is used is an etched cOpper grid with approximately 55% transmission (Buckbee Mears Co.). The problem is far from solved, however, :l, 19 since such windows will transmit electrons with constant efficiency only down to 5—6 kev in energy, and anything below that figure must be corrected for absorbtion. Thus the problem reduces itself to making another experiment, namely, the measurement of the absorbtion properties of the thin film, or using theoretical results which are questionable and certainly unreliable for precision spec- trosc0py.19 This problem can be circumvented by the use of an electrostatic accelerating potential, e.g., between the detector and the spectrometer wall. By increasing the energy of the electron beam in this way, one can drive electrons through the window without absorbtion losses, (i.e., push the effective cutoff energy to nega- tive values). It is possible in principle to detect with constant efficiency electrons approaching zero energy. During the last fifteen years several investi- gators have attempted electrostatic acceleration.15’ 18’ 20-29 Both preacceleration and postacceleration have been tried. The success has been only partial. Acceler- ation of the electrons at the source can lead very easily to defocusing problems unless extreme care is taken. This is eSpecially true for precision instruments running at high resolution. Acceleration at the source also leads to a loss of effective resolution. Note that two lines separated by AE at El now differ by AE at E2 = El + Eacc with application of a preaccelerating potential .Il 20 V (Eacc.= eV). Noteworthy is that the electronics gener- ating the preaccelerating potential must be very well regulated, e.g., better than 1:105 for an instrument such as ours. In these respects acceleration between the spectrometer exit port and the counter seemed to show more promise. However, previous attempts with postac— celeration have lead to very serious noise problems when- ever the accelerating potential brought up the electrons to the window cutoff energy. Giroux and Geoffrionzs (fig. 3) have shown the sort of behaviour that was en- countered. If the experimenter stayed below this po- tential, the method was partially fruitful, since e- lectrons below the window cutoff were registered, al- though with variable transmission for different energies. Furthermore, it was possible to obtain a partial trans- mission curve for the window. The noise was identified as coming from outside of the counter window and it was hypothesized that it was due to secondary electrons, and/ or ultraviolet radiation produced in the preliminary stages of discharge. The earlier investigators tended to use relatively high (v20 kev) accelerating potentials and 26 using heavy windows ch.1 mg/cmz). Work done by Achor potentials below 5 kev and films <10 ug/cm2 indicated that spurious counts were not due to electrical dis- charges, but probably were due to essentially zero energy electrons being accelerated through the window. With the above considerations in mind, we .mN .Mmh «COHHHHOmO USN NSCHHO EORH wmprO> COHpmeHmOOww umom .m> mWfiOE .m mhdmgmrm .3; cotwnfiooom 05 «EB écmE .83 2 H58 Eczemo—omm .v .05 enzo> v: < N. o. o w v H _ 0 00 0o. I Ova _ a... P z 22 decided that the most promising and the most practical approach would be to stay below 10 kev in voltage, and thus (possibly) avoid electrical discharges. This also simplified things in the sense that fancy high voltage electronics would not be necessary. The ZTEE'Spectrome- ter lends itself more readily to the use of a postac- celeration technique than, say, a lens spectrometer, since a typical beam focus angle is about 5°, and danger of spectral distortion due to the action of the ac— celerator electric field on the divergent beam can be kept at a minimum. During the initial phases of the accelerator construction it was learned that Geiger et al30 had used a simple version of an accelerator with the Chalk River ‘Zrfli— spectrometer with a 2.H kev acceleration to study the M and N conversion lines Q~3 kev) of Smlsl. This encouraged our decision to confine ourselves to lower accelerating voltages. It was also learned that Mehlhorn and Albridge31 were using a negative retardation potential to experiment with noise reduction. They have published a version of the technique used.32 A retardation provision was there— fore incorporated in our accelerator. Other factors worth mentioning that were con- sidered during the initial stages of construction and design were: (1) elimination of all steep electric field gradients and regions of high field in general; (2) mini- THES 23 mization of high field in the immediate vicinity of the counter window where leaks were expected; (3) use of very regular and smooth surfaces; (H) cleanliness of the surfaces in the acceleration space; and (5) penetration of the electric field bubble from the accelerator field past the resolution defining slit. B. The Acceleration Cell and Supporting Electronics Figure M shows the combination of slit sizes and geometry which was used to run the L-Auger spectrum of In113m (v2 to H kev). The spectrometer baffles were set so that the incident beam had a~5° spread at the focus. The left side of the cell was tightened more firmly than the right to correct for approximately 1 mm deviation of the curving beam at point no. 3. A network of plastic screws was used to sandwich the plates together. Follow- ing is a point by point consideration of the various ele- ments of the cell as they are shown in fig. %. (1,2) Counter Anode And Outside Wall The basic counter design is described in Parker's thesis.33 The anode is a 0.002" stainless steel wire, which is raised to the Operating potential by a high voltage power supply stacked on top of the accelerator power supply by means of an isolation transformer. The counter is run in the Geiger-Muller region using a 67% argon- 33% ethylene mixture at 55 mm Hg with a resultant 12% slope. Teflon tubing is used to isolate the counter from the gas tank. The high voltage connections are made All .mcofluomcmoo mmmuao> Ucm mama cowpmpmawoo< mchSOOMGmom .: oasmwm 58 En. agoumm© ¢ a. state \K\\\\\ . 24w; tam 02.2....8 29.5.58: 395202 0 oooooooooooooo ”HI ................ _ _ “CUTE” .1233. 3.0-0.0 )Solo .................. "DEE >5 g .09 gum .. -— 2.9-00 Gun—(Om >5 2.0 + 3.00. 248 aaz< inu [II II. o i. o .L o o M @ 3.4.x. NEE v ‘ :82 8 .8 $05008 .Ig 3023 mm; M 5260.20 4 .................... ................. ................. ION ................. ................. ................. ................. 4§§N gawk—€50. E83. whammy/A mzozbgoo wan; Q44 jmo 29.354804 96800”. ._.mOn_ 25 via a coaxial cable,* where the shield carries the ac- celerating potential and the inside conductor the counter potential. The counter wall and the window support are at a positive potential with respect to the spectrometer vacuum tank which is at ground potential. (3) Counter Window on COpper Mesh The windows are collodion, produced by the usual technique of dropping a solution of collodion in n-amyl acetate on distilled water. They are then mounted on the window plate and COpper mesh support, which has been attached to the plate using low melting point solder paste** to minimize oxidation of the copper mesh. It is worthy to note that the window plate which supports the mesh and subsequently the organic film is highly polished on both sides, so that molecular attraction between the film and the plate together with the counter pressure pressing the film against the plate seems to give a relatively leak proof window seal. It was found that films of >10 ug/cm2 will yield system pressures of 10-5 Torr or better when mounted in this manner. (H) Intermediate Plate The intermediate plate was introduced after experimenting with single insulating slabs of several thicknesses, and after an informal study of the relation- *Belden 8239 **Force; American Solder and Flux Co., Phila. 26 ship of system vacuum* to noise as observed with an os- cillosc0pe. It was observed early in these investi- gations that by using accelerating potentials of greater than 3 kv, with a single insulator spacer of approxi- mately 1 cm thickness, and with system vacuumwlO'5 Torr, one induced the appearance of a new class of pulses (as distinct from the unmistakably identifiable counter pulses). Their amplitude and frequency seemed to increase with vacuum deterioration, and with the narrow- ing of the insulating slab (i.e., use of stronger e- lectric field); or, the increase of accelerating po- tential. The number of true counter pulses started to increase only when the accelerating potential reached the cutoff energy of the window. It was possible to isolate the former rather dramatically by simply turning off the counter potential. The counter pulses disappeared while the others stayed. Unfortunately during these tests it was difficult to keep the variables under a strict quanti— *"System vacuum" is used here synonymously with the extent of leakage of counter gas into the spectrome- ter main vacuum. The vacuum gauge was located in a draft free spot closer to the source than to the counter and therefore a vacuum check meant always that the vacuum as monitored in this spot was by a factor of one or two better than in the vicinity of the window. al 27 tative control and therefore the author wishes to issue warning concerning the strength of the above statements. It was felt that these residual pulses may have had origin in the accelerating volume, and possibly were due to the formation of positive ions which when ac- celerated to the cathode produced secondary electrons, thus producing a kind of capacitive discharge which could have easily reflected itself through the electronics into the counting circuit. The pulses were observed at the scaler input. Another possible explanation hypothesized was that these curious pulses were due to some sort of surface discharge along the lucite wall, which was both dependent on the density of molecules near its surface and the strength of the electric field. It must be remarked that these pulses at moder- ate accelerating potentials and good vacuum were smaller in amplitude than the counter pulses and could be rejected by raising the sealer treshold. It became impossible to do this, however, oncelvk kv were applied. It was decided then that an attempt should be made to see if any improvement occurs if most of the imparted energy is given immediately after the momentum resolution slit, allowing the region where the escaping gas concen- tration would be strongest to be a region of weaker field comparable in magnitude to what was considered to be a safe value from previous experiments. A happy medium 28 was attempted by slightly narrowing the original Spacer and introducing an even narrower second spacer nearest the window, both being separated by an intermediate plate which would be set electrically at 2/3 of the total applied accelerating voltage. Of course, dimensional provisions had to be made to make sure that the beam could travel unhampered to the window. The spectrometer magnetic field was off during all above tests. (5) Momentum Resolution Defining Slit The momentum resolution defining slit (1 mm) consists of two pieces fit snugly into the cavity. (6) Retardation Slit The retardation slit (1.5 mm) consists of two similar pieces. The roles of the two slits may be easily interchanged, and it would be interesting to see the effect this has on the accelerator performance. The pene- tration of the electric field bubble into the region where the retardation slit is located is estimated to be about 10 volts with 8.0 kv acceleration. The advantage of its present location is that otherwise it would be moved into a deeper part of the accelerator potential and therefore would probably require more retardation voltage. The primary disadvantage to its present location is that it offers the incident beam, which misses the aperture, a negatively charged obstacle. This likely produces secon- dary electrons which are attracted in by the acceler- ating field and contribute to some of the noise 29 which will be discussed in the following section. (7) Spectrometer Exit Port The spectrometer exit port is actually preceded by a vacuum gate which is not shown. The basic material used for the electrodes is yellow brass (67% Cu-33% Zn). The counter material is also brass with a somewhat higher content of zinc. The electrode surfaces were polished thoroughly starting with rough sandpaper and finishing with emery polishing paper, and effort was expended to manufacture a special rounding tool for rounding the internal edges (0.12 cm). Before assembly and attachment to the spectrometer the surfaces were further cleaned with Brasso and washed with benzene. Great caution was of course taken to avoid introduction of foreign materials on the electrodes. Home made r.f. oscillator type high voltage supplies are used, with some special circuitry for the counter supply. Maximum current output is about 25 uamp each. C. Operation and Performance After the accelerator was assembled as depicted in fig. H it was observed that with no acceleration and retardation fields the accelerator hardware had intro- duced no changes in the beam transmission or instrumental resolution, as a matter of fact there appeared a slight improvement in the intensity of the Auger K-L2L3 line (20.1 kev), which was used in making this transmission THEE 30 test. Actually in the preliminary studies a 0.5 mm mo- mentum resolution defining slit was used.* The author considers that the primary weakness of the accelerator in the above form is the danger of production of secondary electrons due to the large number of slits and edges along the beam path. However, the transmission study seemed to indicate that this fear is probably unwarranted. We also observed that when the accelerating field was applied (no retardation), the earlier discussed non counter pulses appeared only after reachingav6 kv ac- celerating voltage and were small enough at 8 kv (maximum range of supply) so that they could be rejected by changing the sealer treshold. The ratio in amplitude of counter to noncounter pulses was about 5:1 for 10-5 Torr. The ratio of their durations was approximately 3:5. The number of counter pulses increased with accelerating voltage (refer to fig. 5), regardless of whether or not a source was present, although the plateau level for the case of the source was higher. It could be raised even further if the ionization gauge degas current was turned on. Fortunately the application of the retardation *The change of the width of the momentum reso— lution defining slit from 0.5 mm to 1.0 mm had no effect on the accelerator performance. The plateau level of the type shown by the top curve of fig. 5 increased in going to the wider slit. El 31 .uommmm cofiumppmpwm mmfizonw mm>nso mmwoz ”>3 moflso> mOF zo_hmN_I ” 0.5—u o_.rwzo ¢0k4mw4woo< o> mm_oz O GAME 9.52042 02 . ifi o ..wo<......0> mOhdmmeoo< u> wm_oz O BSION J T tu/o +L+J ‘2 :1' a); DJ .Hmaucmpom mcflpmamamoom cwxwm Spas mmmufio> coaumchmumm .m> mmfloz .o oaswam .2: 34.53 20.5845: m.| 0.- t- m.| ~.Il _. | L. n .1 n A A o 11 n. N + l' c Aw * r C 5:0 . . 1' 5. m if to. x... N— aeox 22. m. .xomoma 382.3 e. zockmwnfioolu >._ m It; wo<.50> zo_._. wm_oz w. BSION m/o .mpamfim mcfimsoom Hmpm>mm spa: mmmufio> neumpmfiooom m» mmfloz .m E: moan? «3454304 @2689. 258 a o m e m N _ _ b P h b b F _ J 1 u A d a 0 II N Av v ‘ ilhllwlul $ ._. fl \\¢\\\\\\ 1 w \ \x _ \ x \ \ m \ o \\ a i. m 0 \ \ a \ \\ a \ \ \ \ \ x \ \ \ z #o. \\\ \ x a \ \ \\ \ \ nlp \ \. \ \ \ \\ A... N— \ \ \ \ \A \ \ x 1.. v. I \ \\ \\ \\ \\ \ «SETS .xommmq 382.3 \ \ \ 1:! w— \ \ \ d \\ \x x 3459 zofiaomflmm ,3? 5.3 $36 9.1% a \\ \\\ \\ \\ \ \ \\ a \ moses om. o .. m. \ \\ $36 3. o \\*\\. _\\ \ \ . i. ON \\ 836 on x I. .o..\\\ $33 8% < ..- 8.6.... @2680... 44mm>mm It; 34.59 10.25.5004 2, wmfiz ..e~ “”3 BSION ohdmfla 3% potential, fig. 6, (no spectrometer magnetic field) was able to reduce this noise to nearly natural background. The lower curve in fig. 5 shows the retardation effect along the full accelerator range.* With this information on hand, one is tempted to say that indeed the top curve in fig. 5 represents the delivery of low energy electrons from the main vacuum system to the detector. This is, of course, most strongly substantiated by the fact that they may be repelled by the application of the retarding po- tential; and by the results shown in fig. 7. These results show the noise dependence on the spectrometer magnetic field (no source), suggesting the possibility that some of the electrons which have been attracted by the accelerating field from the main vacuum tank are being removed from detection by the action of the mag- netic field. Further, it is interesting to note that the departure point of the retarded and the unretarded curves (say, fig. 5) corresponds very closely to the cutoff energy of the window. This point of departure of the two curves was, for all windows tried, consistent with the cutoff point computed from the estimated window thicknesses.l7 Noteworthy also is the fact that for 2 theIVIZ ug/cm window the shape of the curve corresponds *The accelerator supply delivered only 7 RV maxi- mum at the time these curves were run. 35 very closely to the absorbtion curves by Lane and Zaffarano.l7 However, since fig. 6 says that the energy of these low energy electrons floating in the spectrome- ter tank ranges, say, from O to 130 ev, then the de- parture point of fig. 5 is somewhat cloudy. For that reason and the fact that Lane and Zaffarano give very uncertain figures for their film thicknesses, a strict comparison between Lane and Zaffarano curves and the curve of fig. 5 is not made. The difference between the lower curve in fig. 5 and natural background is probably due to events in the acceleration space itself. Considerable time was spent studying the effects of the post acceleration cell once the spectrometer was allowed to focus an electron beam. It was found that at about 20 kev, as mentioned previously, there seems to be no defocusing, fig. 8. However, there was a noise con- tribution to the line which was roughly dependent on the beam strength, e.g., 52% at 6 kv acceleration, and de- creased as the accelerating potential was lowered. This appears to be far superior to anything previously reported.32 At lower energies there was no simple way to check the accelerator performance without a strong isolated line in the region of interest. By running cunms of the type shown in fig. 8 (left) it was possible to show, however, that defocusing is negligible for this resolution. This was done by a point by point comparison of neighboring shapes both for the case of acceleration F1 36 and no acceleration. Examination of peak to valley ratio of both curves (using a rough absorbtion correction) yielded a accelerator introduced noise figure of about 1% of beam strength for this energy region. The insert shows graphically the growth of the L3-MMM5 peak with ac- celerator voltage. No inconsistencies were found within statistics. Further, the insert demonstrated (although this is not indicated on the graph) that the introduction of 7 kv accelerating voltage did not shift the line position, thus encouraging the belief that no serious artificial energy shifts were introduced in the L-Auger spectrum (~2 to 4 kev). An extrapolation of the good results of performance in this energy region therefore seems appropriate down to at least 1 kev. .mocmEpomamm acumhmamooq mnwpmapmcoaoa MHHCH mo mosomo newsmam was a no mcowuomm n~.8§moL~m 3.4-x. >8. non in. 3...: 598 was! t n 311-..: >8. on L J J O «£553.. g g .69.; 20593»! 8% 595.389 8% . 29.3932. $1295.58”... 33 . gun. mohammqmouq 02.3% 3cm 55mm oumeo m4 9.5mede mwoaxgmuo mzozbwm 3.3.. - 5. >3. o.~ I OOfiN \. sooo. +009 tOOON $88 CHAPTER IV 113 THE L-AUGER SPECTRUM OF IN AS DERIVED FROM TIE ELECTRON CAPTURE DECAY OF SN113 A. lgtroductory Remarks Chapter III represents the most serious exper- mental problem that needed to be solved, namely, the ex- tension of the low energy range of the spectrometer by means of postfocusing acceleration. Another important problem was the production of a very thin source in order to reduce the number of ener- gy degraded electrons due to self—absorbtion by the source. Source absorbtion effects can be serious at the energies that were studied since they introduce long tails to the low energy side of monoenergetic electron lines and increase their width. This effect tends to defeat the high resolution capability of the spectrometer and also serves as a handicap in the analysis of the spectrum. A later section of the chapter is devoted to the description of the source production technique that was required to minimize this effect. The reasons for the choice of the electron capture decay of Sn113 as the subject of study were stated in CH I and are given again here for completeness. First, the L-Auger spectrum of a medium Z atom 38 39 had never been studied previously. This was a matter of interest by itself. Second, Z = 49 is a good place to look for double vacancy effects following K-LL transitions since the K-Auger yield, aK, is a factor of three greater for this Z (ref. 1) than for, say, Z = 80.* The energies of L-Auger spectra decrease with decreasing Z enhancing the experimental difficulties. Z = #9 was thought to be a good compromise of fairly high Auger yield and yet high enough L-Auger energy. 113 Third, Sn was easy to produce due to a large neutron capture crossection,l2 112 and could be easily sepa- rated from Sn by an electromagnetic mass separator. Fourth, previous experience in this laboratory of separating trace amounts of In113 from Sn113 , using volatility differences of the two elements to an advantagg gave us the capability of observing the Auger contri- bution due to the conversion process alone. Figure 9 *aK is the probability that a K vacancy will be filled by means of an Auger transition. The L Auger yields, aL , aL2, aL3, (similarly defined) are also 1 substantially higher for this Z, meaning that the L- Auger spectrum intensity would be similarly higher for the same number of radioactive atoms than for a heavy element (with other factors assumed to be constant). to 9 0/0 . 877which also *Magnaflux Corporation, Chicago, Illinois. 50 shows the energy ranges of the different L—Auger groups. The interpretation of such a spectrum is not an easy problem. By way of illustration, let us consider an ideal set of conditions which would be desirable from the standpoint of an easy study and clear cut interpretation of the L-Auger structure. These conditions could be described as follows: (1) Bxistance of no primary vacancies except in the L she11.* 3 (2) Choice of atoms of large enough Z such that the two vacancies produced in the M shell during a L3—MM Auger transition could be described by pure j-j coupling. (3) Capability of manufacturing a high specific activity source of one or two atomic layers in debth, which would exibit no source thickness effects. (4) Use of a conducting source backing, say nelO ug/cm2, to avoid the classical problem of back- scattered electrons which produce an ambigous *K. Risch69 has demonstrated that it is possible to create vacancies in the L3 subshell of bismuth by means of monochromatic x-rays without ionizing the L1 and L2 shells and in this way study the resulting isolated L3- Auger spectrum. 51 background under the spectrum. (5) Unlimited resolution. Under these circumstances the L—Auger spectrum would consist of the L3—XY groups (X = M,N; Y = M,N) due only to the L3 vacancies (refer to fig.]fi)for the L3 energy ranges). The widths of the lines would be merely the sum of the widths of the atomic states involved in a given transition. The backseattering would probably be negligible. Any overlap of lines would be due only to the overlap of atomic widths. If in the light of the above considerations we examine our spectrum, we find that: (l) The primary vacancy contributions to the L sub- shells are due to, (a) orbital electron capture, (b) conversion process, (0) transfer of holes from the K shell by means of radiative tran- sitions, and (d) transfer of holes from the K shell by means of Auger transitions (K-LL and K-LX). Table I gives a first order approximatyxl of what would be expected for the relative vacancy distributions per disintegration for the three different L subshells. The 2% capture branch to the 6H8 kev state (fig. 9) has been neglected. Thus, we see that the set of primary vacancies which give our spectrum comes about in a complicated way and the spectrum (fig. 10) is now a superposition of the Ll-XY, Lz-XY, and 52 TABLE I. Estimated Relative Vacancy Distribution for the K, L , L2, L3 Shells 1 Process Creating Vacancy K L1 L2 L3 a Electron Capture 0.871 0.1047 0.0022 b Internal Conversion 0.282 0.0410 0.0061 0.0082 c Radiative Transfer 0.2660 0.5240 from K d Auger Transfer 0.0740 0.0915 0.1300 from K e Sum 1.154 0.2197 0.3658 0.6622 (a) Calculated using Rose and Brysk formulas (ref. 6). (b) Calculated using tables of internal conversion by Sliv and Band (ref. 1). (c)(d)The K flourescence yeild was assumed w = 0.83 K (ref. 1). The number of vacancies forming in each of the L subshells per vacancy filled in the K shell was taken from Listengarten (ref. 56). 53 L3—XY groups with very serious overlaps starting at approximately 3130 ev (22500, fig. 10). Further, it is not possible to deduce what the relative intensities of the three major groups might be on the basis of the figures of row (a) Table I, because the partial Auger yields a1, a2, a3 are not known. They are related to the Coster-Kronig yields, fij’ and the partial fluorescence yields, wi, (i,j = 1,2,3,) in the . * following manner, and are known very imprecisely. This is also true for the fluorescence yields although these are known to be 2-3% for Z = 49. The most important C-K transitions for this Z have been found to be of the type Ll- L3Mu,5 (refs. 44 and 50). Their effect would be *rij is defined as the probability that the i shell vacancy will be filled by a transition of the type Li-L.X,Y. a1 is defined as the probability that the i J shell vacancy will be filled by an Auger transition of the type L -XY. is defined as the probability that i wi the i shell vacancy will be filled by an x-ray tran- sition. (2) 54 to create additional vacancies in the L3 sub- shell (in addition to those set forth in Table I). Further, because they cause ionization of the M4,5 shells, the subsequent L3-M4,4M5,5 lines would be shifted due to the fact that the My,5 electrons that are to be ejected are now more tightly bound. This phenomenon has been observed in x—ray studies. Refer to references 70 and 71. In our case this will be neglected since by Table I the most intense L—Auger group is due to the L3 vacancies and the Ll-L3M§,5 transition can be considered negligibly small in terms of intensity contribution, particularly, since the primary Ll vacancy is by a factor of three less than the L3 vacancy. For the case of Z = 49, pure j-j coupling is not a good approximation. Both theoretical deductions by BurhOp and Asaad,50’ 52 as well as experimental measurements of the K-LL group, e.g., by R. L. Graham et al,72 have shown that intermediate coupling features play a strong role for ZAISO. That is to say, in the in- terpretation of the spectrum we can no longer ignore the fact that a j-j coupled line may have associated intermediate coupling satellites. In previous studies of L-Auger spectra of high Z elements a discrete j-j coupling approximation (3) (4) (5) 55 was still warranted. Source thickness effects as exibited by long tails are important handicaps in the analysis, particularly, since they vary over the spectrum becoming more severe towards lower energies. The possibility of discrete energy losses of electrons as they emerge from the source must also be considered. A characteristic bump on the low energy side of a line when particles mumstravel through a thick source has been observed by, e.g., Nordling, et al73 and J. S. Geiger, et a1.30 It has been attributed to quantization of inelastic energy losses of the electron, and has been treated theoretically by Bohm and Pines.71+ Backscattered electrons are estimated to give a background contribution of 5-10% of intensity making it difficult to draw a base line for the spectrum. Even though the spectrometer resolution is excellent (0.12%) the above reasons make it impossible to isolate a "standard" line shape in a clear cut manner. Moreover, at this resolution and energy the natural widths of the lines play an important role. Due to the fact that no measured or calculated values of L widths for Z = 49 exist, they could not be 56 taken into account during a crude line fitting effort. As a guide in analysis extrapolated values were used for the L widths obtained from J. S. Geiger, et a1,30 and Listengarten.41+ Clearly, the interpretation of this spectrum is a difficult problem. Table II gives the measured positions of lines together with the values calculated by means of the following expression, first suggested by Bergstrom and Hill,‘75 E(L-XY)Z = E(L)Z - B(X)Z - E(Y)Z - AEXY where, L = primary vacancy in any of the three sub- shells. X,Y = resulting vacancies in any of the suc- ceeding subshells of major shells. E(L),E(X),E(Y)= binding energies of the respective sub- shells.* where, AZ was allowed to take values 0.55; 0.67 for tran- *Binding energies used were obtained from tables in reference 1 due to S. Hagstrom et a1. .maancm Bohm mm>wpma mm mafiucH mo Esapomam ammsqum .QH shaman E 2 9:23 «$0.20 3480 ‘ W N om m. n. gt Tzz..l T223]. Tzzjll Tlllzzjli .zI TI 22: Ill TI 22... 2.2..“ _ TV 22... T J- 22.4 _ 3.38 38 l 58 TABLE II. Energies and Relative Intensities of the L-Auger Lines Line Trans. Energy Energy Intensity Comments No. Calc. Meas. K-L2L3 (ev) (ev) = 1.00 1. L3—M1Ml 2047 2053 <0.10 2. L3-M1M2 2173 2167 0.14 3. L3—M1M3 2213 2211 0.24 signs of multiplet 4. L3—M2M2 2296 2290 0.30 5, 2320 possibly D.E.L. 6. L3—M2M3 2336 2339 0.85 signs of multiplet 7o 2352 L ‘M M IoCo séteTlfte (?) 8. L3-M3M3 2373 2375 1.14 signs of multiplet 9' L3‘M1M4 242” 2418 0.27 L2-M1M3 2420 10. L3-M1M5 2432 2431 0.48 11, L2—M2M2 2504 2506 0.56 12. L2-M2M3 2543 2542 0.77 13. L3—M2M5 2558 2555 1.53 14. L2-M3M3 2582 2577 0.88 16. L2-M1M5 2640 2639 0.40 L2-M1M4 contr. l7. Ll—MlIVI2 2680 2666 -v0.10 (?) 18. Ll-MlM3 2721 2712 .v0.10 19. LQ-MQMLL 2758 2763 2.09 L2—M2M5 icontr. 20- L2'M3M4 2795 2792 0.92 L3—M4M4 2800 21. L3—M4M5 2808 2808 4.18 12_M3M5 contr. 59 TABLE II. Cont'd Line Trans. Energy Energy Intensity Comments No. Calc. Meas. K-L2L3 (8V4 (8V) = 1.00 22. L3-M5M5 2816 2821 1.90 23. Ll—M2M3 2843 2860 0.52 Ll—M3M3 2882 24. 13’42N3 2940 2939 0.39 L —M1M contr. 25' L3‘M3N3 2978 2985 0.55 26. L3-M2N5 3004 2999 0.30 27. LZ-MMMS 3016 3016 2.51 28° L2‘M5M5 3024 3028 0.32 29. L3—M3N5 3042 3050 0.69 30. Ll-M2M5 3066 3063 0.19 31. Ll—M3M5 3103 3107 <0.10 (?) 32. L2-M2N2 3146 3150 L3-M5N1 3150 3160 L2-M2N5 3212 34. L2-1vi3N4 5 3250 3246 0.26 35. L3-M5N4 5 3262 3266 1.77 36. Ll—MuMB 3316 3316 0.50 37. Ll‘M5M5 3324 3333 (v0.10 38. L -M N 3398 Lg_MgNg,§ 3408 3400 0.10 39. Ll-M2N2 3 3447 3448 <0.10 40. L2-M5N5 3471 3471 0.60 L2-M4 5N4,5 contr. TABLE II. Cont'd Line Trans. Energy Energy Intensity Comments No. Cale. Meas. K—L2L3 (EV) (EV) = 1.0:) 41. Ll‘M2N4,5 3512 3512 <0.10 3 2,3 3,3 43. L3~N2’3N4,5 3610 1V0.10 3 77 L1”M5N2,3 45. Ll-144’5N4’5 3753 ”0.10 L2—N2,3N2’3 3773 40. L2-N3N5 3836 3836 ~0.10 47. LE-N4 5N4 5 3904 <0.10 61 sitions of the type Ll,2,3—M1,2,3Ml,2,3; Ll’2,3-M4,5M#,5, respectively. For transitions of the type Ll,2,3-M1,2,3 M4,5 estimates were made for example by taking an average of Ll-MlM5 (AZ = 0.67) and Ll-M5Ml (AZ = 0.55). AZ = 1.0 was used for all others. The energy calibration in the L—Auger region was based on the location of the K—L2L3 line which was assumed to have an energy of 20.144 kev based on a semiempirical table by Hornfeldt.76 The energies in this table are estimated to be good to about 0.05%. A work function correction of 4 ev was also taken into account, due to the fact that the electron leaving the source will have to overcome a small electric field and loose an amount of energy roughly equal to the W.F. of the vacuum tank surface. It must be remarked that the measured positions of a line were really center of gravity estimates of multiplets due to the fact that for Z = 49 we are removed from the pure j-j coupling limit. A crude line fitting effort, as hazardous as it is due to the previously mentioned reasons, was nevertheless used in a comple- mentary way to aid in the identification and positioning of the lines. The estimated relative intensities are also given in Table II. A computer program assuming a gaussian shape with an exponential low energy tail was tried on some regions of the spectrum. This indicated that by means of parameters that were obtained 62 graphically, together with empirical data such as po- sition of lines, estimated widths and heights, the standard deviation from an optimum fit was 20-30%. Thus, the column in Table II indicating the relative intensi- ties should be construed as possessing errors of at least this order of magnitude and even worse for the very weak lines. After the line fitting trials we feel that discrete energy losses (D.E.L.) were not prominent and may have manifested themselves only as gentle ripples on the low energy sides, particularly, in the low energy region of the spectrum. Subsequently they were ignored. The existence of double vacancy effects which would arise when the L-Auger process follows the K-Auger process was investigated, insofar as the experimental data allowed it. The effect may be illustrated by con— sidering a K-LlL3 transition which leaves the L shell doubly ionized, thus, effecting the screening of the electrons which are about to participate in a Auger tran- sition, say, L3—MMM5. To a first order approximation, if one can take into account the changes in screening correctly, the estimate of the energy shift of a double vacancy satellite line from its parent could be made purely on energetic grounds. The situation, however, is much more complex than in the usual Auger process since the initial state is comprised of two holes which can couple in some particular way and the matrix element 63 would look as follows, 2 : (pmsr/<25/}£Z/a% f / Z; l 31.73 3499/2- Aside from the fact that this would introduce broadening of the double vacancy (multiplet) lines, it was not obvious how the energy shift could be estimated in an easy manner. A straight forward application of Slater's screening rules77 gave a double vacancy shift of 16-18 ev to the low energy side of the most intense L3-M4M5 line. From the standpoint of intensity it was reasonable to expect noticeable double vacancy effects around, e.g., the L3-MRM5 line (see Table I).* However, the search was impeded by the heavy overlaps of lines in that region, together with our inability to define the shape of the low energy side of lines (not to mention our lack of knowledge of their exact position). There seemed to be evidence in the form of unexplained residues in intensity around the low sides of lines, however, the *The analogous process of x-ray satellites had 78 been first reported by Siegbahn and Stenstrom and was studied although not exhaustively in the 1930's. 8 Recently Deodhar79 and Hayasi 0 have given reviews of the origin and theory of x-ray satellite lines. 64 conclusion that the double vacancy effect was seen must remain somewhat speculative. F. Conclusion It has been shown that post—acceleration of magnetic spectrometer electron beams (with a gas counter as a detector) is a practical and successful method of studying very low energy beta-ray spectra. The results of measurement of the L-Auger spec- 113 by means of a7fi3Efprecision spectrometer trum of In indicate the very real feasibility of studying orbital electron capture ratios in this way. A sharper spectrum would be required which from the vantage point of this experiment appears possible. As it is the results of the relative intensities of the L-groups are too uncertain to attempt analysis of Ll/L2/L3, relative orbital capture probabilities. Although the ultimate goal of measuring these probabilities is not attained,* information about the L-Auger spectrum of a medium Z atom has been obtained, and constitutes the first attack of measuring very low energy Auger electrons with a precision spectrometer by means of a post—acceleration technique. From the standpoint of technique, we believe *A full discussion of this aspect of the problem must await the complete measurement and analysis of all 113_3n113 of the electron spectra of the In system. 65 that thinner conducting source backings should be used in order to minimize backscattering. Carbon films would seem to offer the best solution for this. Source thick- ness effects could be further minimized if more stringent measures were taken in preventing the exposure of the source to air and vacuum pump oil. This could be managed by better vacuum systems (refrigerated baffles) both in the vacuum evaporator and the spectrometer itself. Ac- celerator produced sources also offer a new area by means of which high quality sources could be prepared. From the standpoint of physical information gained the effective incremental charge AZ has been determined for the most prominent Auger transitions and the energies tabulated. Order of magnitude of relative intensities of the lines has been also determined. It has been shown that the spectrum exibited intermediate coupling features and that double vacancy effects probably exist. Chm-FLA) [\J 10. ll. l2. 13. 14. 15. 16. l7. 18. REFERENCES Siegbahn, Editor, OL- -I- Ray Spectroscopy North- Holland Publishig Co., 1965). . N. Bahcall, Phys. Rev. Letters, 9, 500 (1962). K. ( J J. N. Bahcall, Phys. Rev., 122, 2683 (1963). J. N. Bahcall, Phys. Rev., 131, 1765 (1963). R. E. Marshak, Phys. Rev., 61, 431 (1942). H. Brysk and M. E. Rose, Revs. Modern Phys ,39, 1169 (1958) S. 0diot and R. Daudel, J. Phys. Radium, 12, 60 (1956). R. Bouchez and P. Depommier, Reports on Progress in PhySiCS: 23: 395 (1960)- B. L. Robinson and R. w. Fink, Revs. Modern Phys., 32, 117 (1960). L. J. Velinsky, Ph. D. Thesis, Michigan State University, (1964). Nuclear Data Cards NRC-60-2-99, 100, 105, 106. M. Schmorak, G. T. Emery and G. Scharf- -Goldhaber, Phys. Rev. , 124,1186 (1961). L. J. Velinsky, et al, to be published. G. Backstrom, A. Backlin, N. E. Holmberg, Nucl. Instr. and Meth. , 16, 199 (1962). C. S. Cook, Nucleonics, 12, No. 2, 43 (1954). M. S. Freedman, private communication. R. 0. Lane and D. J. Zaffarano, Phys. Rev., 24, 960 (1954)- M. S. Freedman, F. T. Porter, F. ‘Wagner and P. P. Day, Phys. Rev. , 108, 836 (1957). O\ O\ 19. 20. 21. 22. 23. 24. 25. 27. 28. 29. 30. 31. 32. 33- 34. 35- 36. 37- 67 H. A. Bethe, Handbuch der Physik (Verlag Julius Springer, Berlin, 1933), Ed. 23, p. 519. C. H. Chang and C. S. Cook, Nucleonics, 19, No. 4, 24 (1952). A. Juillard and M. A. Mousse, J. Phys. Rad., 12, 94 (1958)- H. M. Agnew and H. L. Anderson, Rev. Sci. Instr., 20, 8690 (1949). L. M. Langer and C. S. Cook, Rev. Sci. Instr., 12, 257 (1948) D. K. Butt, Proc. Phys. Soc. Lond., 63A, 986 (1950). G. Giroux and C. Geoffrion, Can. J. Phys., 33,153 (1956) W. T. Achor, Ph. D. Thesis, Vanderbilt University, (1958) W. Mehlhorn, Z. Physik, 160, 247 (1960). P. Erman and Z. Sujkowski, Ark. Fys., 22, 209 (1961). Jose Sant'ana Dionisio, Ann. Phys., 2, 29 (1964). J. S. Geiger, R. L. Graham and Janet S. Merrit, Nuclear Physics, 48, 97 (1963). R. G. Albridge, private communication. W. Mehlhorn and R. G. Albridge, Nuc. Instr. and Meth., 26, 37 (1964). R. A. Parker, M.S. Thesis, Vanderbilt University, (1960). R. C. Greenwood and E. Brannen, Phys. Rev., 122, 1849 (1961). Willie E. Phillips and John I. Hopkins, Phys. Rev., 112: 1315 (1960). S. B. Burson H. A. Grench and L. C. Schmid, Phys. Rev.,]. 13, 1é8 (1959). W. T. Achor, W. E. Phillips, J. I. Hopkins and S. K. Haynes, Phys. Rev., 114,137 (1959). 38. 39- 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53- 54. 55. 68 R. K. Girgis and R. Van Lieshout, Physica, 24, 672 K. S. Bhatki, R. K. Gupka, S. Jha and B. K. Mandan, Il Nuovo Cimento Serie X, 6, 1461 (1957). C. Ianduchi, G. Nardelli, M. T. Russo—Manduchi and G. Zannoni, Il Nuovo Cimento V01. XXXI, N. 6, 1380 (1964). P. Avignon, Ann. Physique, 1, 10 (1956). Y. Deschamps and P. Avignon, Compt. Rend. Ac. Sc. Paris, 236, 478 (1953)- C. D. Broyles, D. A. Thomas and S. K. Haynes, Phys. ReV-3 229 715 (1953). R. G. Jung and M. L. Pool, Bull. Am. Phys. Soc., 1, 172 (1956)- M. A. Listengarten, Izvestia Akad. Nauk SSSR, (Physics) 24, 1050 (1960). (Transactions of the 10th Conference on Nuclear SpectrOSCOpy.) Jose Sant'ana Dionisio, Ann. Phys., 8, 747 (1963). E. H. S. Burhop, The Auger Effect and other Radiationless Transitions (Cambridge University Press, London, 1952). P. Auger,Comptes Rendus, 180, 65 (1925). M. B. Stearns, The Spectroscopy of Mesic Atoms 12 Progress of Nuclear Physics, 6, 108 (1957). D. Coster and R. De L. Kronig, Physica, 2, 13 (1935)- W. N. Asaad, Nuclear Physics, 63, 337 (1965)- E. J. Callan, Rev. Mod. Phys., 35, 524 (1963). W. N. Asaad and E. H. S. BurhOp, Proc. Phys. Soc., 21, 369 (1958)- E. J. Callan, Phys. Rev., 124, 793 (1961). w. N. Asaad, Proc. Roy. Soc., 242, 555 (1959). W. N. Asaad, Nuclear Physics, 44, 399 (1963). 57- 58. 59- 62. 63. 64. 65. 68. 69. 70. 71. 72. 73- 69 M. A. Listengarten, Izvestia Akad. Nauk SSSR, (Physics) 25, 792 (1961). M. A. Listengarten, Izvestia Akad. Nauk SSSR, (Physics) 25, 182 (1962). R. A. Rubenstein, Ph.D. Thesis, University of Illinois, (1955). Z. Sujkowski and O. Melin, Arkiv Fysik, 22, 193 (1961). G. Albouy and M. Valaderes, Comptes Rendus, 250, 2877 (1960). J. C. Nall, Q. L. Baird and S. K. Haynes, Phys. Rev., 118, 1278 (1960). J. Burde and S. G. Cohen, Phys. Rev., 104, 1085 (1956)- K. Risch, Z. Phys., 1 0, 87 (1958). C. Geoffrion and G. Nadeau, Can. J. Phys., 35, 1284 (1957)- R. G. Albridge and J. M. Hollander, Nuclear Physics, 22, 554 (1961). Z. Sujkowski and H. Slatis, Arkiv Fysik, 14, 101 (1958). I. Bergstrom, F. Brown, J. A. Davies, J. S. Geiger, R. L. Graham and R. Kelly, Nuclear Instruments and Methods, 21, 249 (1963)- Jerry Lerner, Argonne National Laboratory, private communication. K. Risch, Z. Phys. 152, 89 (1960). E. G. Ramberg and F. K. Richtmeyer, Phys. Rev., 51, 913 (1937)- F. K. Richtmeyer and E. G. Ramberg, Phys. Rev., 51, 925 (1937). R. L. Graham, I. Bergstrom and F. Brown, Nuclear Physics, 32, 107 (1962). C. Nordling, E. Sokolowski and K. Siegbahn, Phys. Rev., 105, 1676 (1957). 74. 75- 76. 77- 78. 79- 80. 70 D. Bohm and D. Pines, Phys. Rev., 22, 609 (1953). I. Bergstrom and R. D. Hill, Arkiv Fysik, 8, 21 (1954). 0. Hornfeldt, Arkiv Fysik, 23, 235 (1963). A. H. Wapstra, G. J. Nijgh and R. van Lieshout, Nuclear Spectrosc0py Tables (North-Holland Publishing Co., Amsterdam, 1959). N. Siegbahn and W. Stenstrom, Physik. Zeit., 1 , 48, 318 (1916). G. B. Deodhar, Proc. Nat. Acad. Sci. India A, 32, 320 (1962). Takesi Hayasi, Sci. Rep. Tohoku Univ. First Ser. (Japan) 46, No. 2, 82 (1960). HICHIGRN STATE UNIV. LIBRARIES III)IIMIIHIIII)MIAIll)!\INWI)HHIWIINIHII 31293017640297