UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The precise intercomparison of lead isotope ratios Kollar, Francis 1960

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

Item Metadata


831-UBC_1960_A1 K6 P7.pdf [ 5.54MB ]
JSON: 831-1.0085535.json
JSON-LD: 831-1.0085535-ld.json
RDF/XML (Pretty): 831-1.0085535-rdf.xml
RDF/JSON: 831-1.0085535-rdf.json
Turtle: 831-1.0085535-turtle.txt
N-Triples: 831-1.0085535-rdf-ntriples.txt
Original Record: 831-1.0085535-source.json
Full Text

Full Text

T H E P R E C I S E I N T E R C O M P A R I S O N OF L E A D I S O T O P E R A T I O S by F R A N C I S K O L L A R Dipl. Eng. University of Budapest, 1952 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF Doctor of Philosophy in the Department of Physics We accept this thesis as conforming to the required standard The University of Bri t i s h Columbia April, 1960 i In presenting t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree th a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I f u r t h e r agree that permission f o r extensive copying of t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the Head of my Department or by h i s r e p r e s e n t a t i v e s . I t i s understood tha t copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be allowed without my w r i t t e n permission. Department The U n i v e r s i t y of B r i t i s h Columbia, Vancouver 3 , Canada. P U B L I C A T I O N S Kollar, F . Electronics, (Hungarian) Allami Jegyzetkiado, Buda-pest, 1955 (Mimeographed University Textbook). Kollar, F . Seismic Amplifier, (Hungarian) Magyar Tudomanyos Akademia, Muszaki Tudomanyok Oszfaly Kozlemenyei, Vol . 1-2, 1951. Russell, R. D . and F. Kollar. Transistorized power supplied for a mass spectrometer, Can. Jour, of Phys. (To be published May, 1960.) S U B M I T T E D F O R P U B L I C A T I O N : Kollar, K. , R. D . Russell and T . if Ulrych. Precision Intercom-parisons of Lead Isotope Ratios: Broken Hill and Mount Isa, Nature. Faculty of Graduate Studies P R O G R A M M E O F T H E FINAL ORAL E X A M I N A T I O N FOR T H E D E G R E E OF DOCTOR OF PHILOSOPHY of F R A N C I S K O L L A R Dipl. Eng. University of Budapest IN R O O M 301, PHYSICS B U I L D I N G F R I D A Y , A P R I L 29th, 1960, A T 1:30 P . M . COMMITTEE IN CHARGE Dean F. H . S O W A R D : Chairman J. A . JACOBS R. D . R U S S E L L R. W. S T E W A R T G . L . P I C K A R D S. Z . P E C H W. A . B R Y C E K. B. H A R V E Y W. H . M A T H E W S J. F. S Z A B L Y A External Examiner: Dr. R. M . F A R Q U H A R University of Toronto T H E P R E C I S E I N T E R C O M P A R I S O N O F L E A D ISOTOPE RATIOS A B S T R A C T The isotopic constitution of lead bocame important in geophysics as the lead-uranium and lead-thorium methods of absolute geological age determination were established. Interest in the isotopic abund-ances of common lead in lead minerals was a natural development. The interpretations of observed variations in lead isotope ratios has been more simple than had been expected twenty years ago when they were first measured by A . O. Nier, and results have been obtained that are of very great importance to geophysics. Conse-quently this field is now expanding very rapidly. Many interpretations now being made are limited by the avail-able precision of the measurements, which is of the order of several tenths of a per cent. A mass spectrometer laboratory was set up in the Department of Physics at The University of British Columbia, and a mass spec-trometer capable of measuring heavy elements with high precision was designed and constructed. It is a 90 degree sector, 12 inch radius, direction focusing instrument with a copper tube and using a modified Nier-type gas source. It is essentially of orthodox design, but special attention was given to try to eliminate small sources of error. T o establish more stable source conditions, an exceptionally stable filament emission control was constructed and purified lead tetramethyl samples were used. The ion beam was measured with a servo-voltmeter of original design that is capable of a high precision. Readings can be made from a calibrated dial on the voltmeter, eliminating sources of error in the usual chart recorder. Tests have shown that the improvement in precision warrants this step. This mass spectrometer is believed to be the first to make extensive use of transistorized circuits. Error was reduced further by comparing each sample measured with a standard within reasonably short intervals of time (about twenty minutes). Three comparisons were made to loop two samples with the standard, and the looping error was determined and distri-buted around the loop. The analyses obtained in this way were com-pared with existing analyses from other laboratories, and an improve-ment in precision between a factor of five and a factor of ten seems to have been obtained. T o demonstrate the precision obtainable with this mass spec-trometer and with the improvement in operating techniques, analyses were made on the isotopic composition of leads from Broken H i l l and Mount Isa, Australia. It was previously known that the leads are very similar in composition at both localities. From these new measurements it was established that the Broken H i l l and Mount Isa deposits contain lead of distinctly different isotopic com-position. In both deposits the isotope ratios were found less variable than could be inferred from previous measurements. A fine structure was found in the isotope ratios in both localities. These small variations indicate contaminations due to radiogenic leads. Values of the thorium/uranium ratios of the source of contamination were estimated. The precision of analyses made it possible to determine an age difference between the Mount Isa and Broken H i l l deposits of 40 million years. G R A D U A T E S T U D I E S Field of Study: Geophysics Geophysics J: A . Jacobs Isotope Geology R. D . Russell Modern Aspects of Geophysics J. A . Jacobs Electron Dynamics R. E . Burgess Related Studies: Electronic Instrumentation F. K . Bowers Analog Computers W. Dietiker Servomechanisms E . V . Bohn Electrical Engineering Seminar J. F . Szablya i i ABSTRACT The isotopic constitution of lead became important in geophysics as the lead-uranium and lead-thorium methods of absolute geological age determination were established. Interest in the isotopic abundances of common lead in lead minerals was a natural development <, The interpre-tations of observed variations in lead isotope ratios has been more simple than had been expected twenty years-age when the f i r s t measurements of this kind were made by A„ 0„ Nier, and results have been obtained that are of very great importance to geophysics. Consequently this f i e l d i s now expanding very rapidly„ Many interpretations now being made are limited by the avail-able precision of the measurements, which is of the order of several tenths of a per cento A mass spectrometer laboratory-was set up in the Department of Physics at The University of British Columbia, and a mass spectrometer capable of measuring heavy elements with high precision was designed and constructed,, It is a 90 degree sector, 12 inch radius, direction focus-ing instrument with a copper tube and using a modified Nier-type gas source. It is essentially of orthodox design, but special attention was given to try to eliminate small sources of error. To establish more stable source conditions, an exceptionally stable filament emission control was constructed and purified leadtetramethyl samples were used. The ion beam was measured with a servo-voltmeter of original design that i s capable of a high precision. Readings can be made from a calibrated dial on the voltmeter* eliminating sources of error in the usual chart recorder,, Tests have shown that the improvement in-precision warrants i i i this step« This mass spectrometer is believed to be the f i r s t to use extensively transistorized c i r c u i t s . Error was reduced further by comparing each sample measured with a standard within reasonably short intervals of time (about twenty minutes). Three comparisons were made to looptwo samples with the standard, and the looping error was determined and distributed around the loop. The analyses obtained in this way were compared with exist-ing analyses--from other laboratories, and an improvement in precision between a factor of five and a factor of ten seems to have been obtained. To demonstrate the precision obtainable with this mass spectro-meter and with the improvement i n operating techniques, analyses were made on the isotopic composition of leads from Broken H i l l and Mount Isa, Australia. It was previously known that the leads are very similar in composition at-both l o c a l i t i e s . From these new measurements i t was established that the Broken H i l l and Mount Isa deposits contain lead of d i s t i n c t l y different isotopic composition. In both deposits the isotope ratios were found less variable than could be inferred from previous measurements; A,fine-structure was found in the isotope ratios •in both l o c a l i t i e s . These small variations indicate contaminations due to radiogenic leads. Values of the thorium/uranium ratios of the source of contamination were estimated„ The precision of analyses made^possible to determine an age difference between the Mount Isa and Broken H i l l deposits of 40 million years. iv TABLE OF CONTENTS ABSTRACT i i LIST OF ILLUSTRATIONS v i LIST OF TABLES v i i i INTRODUCTION . 1 CHAPTER 1 LEAD ISOTOPE ABUNDANCES The present state of lead abundance interpretations 6 The need for precise isotopic abundance measurements of lead 23 CHAPTER 2 INSTRUMENTATION General considerations 27 The design of a precise mass spectrometer 35 Power supplies 38 Filament emission control 41 The source and collector electrode assembly 42 The measuring system 45 The sample handling and vacuum systems 50 CHAPTER 3 LEAD ISOTOPE MEASUREMENTS WITH HIGH PRECISION Sample preparation 53 Measuring techniques 55 Analytical results 63 V TABLE OF CONTENTS (cont'd) CONCLUSIONS 81 BIBLIOGRAPHY 84 APPENDICES Details of the tube and magnet 88 Magnet current power supply 91 The filament emission control 97 The measuring system 100 v i LIST OF ILLUSTRATIONS Fig. 1 Radiogenic lead isotopes 10 Figo 2 Growth curves and isochrons 10 Fig. 3 Distribution of experimental points for 160 randomly selected samples 14 Figo 4 Examples of anomalous leads 16 Fig. 5 Conformable lead growth curve 19 Figo 6 Conformable lead growth curve 20 Figo 7 Relation of conformable, meteoritic and a suite of anomalous leads 21 Fig. 8 Relation- o'f conformable,,-meteoritic and a suite of anomalous leads 22 Figo 9 Reproducibility of analyses 33 Figo 10 Fi r s t order direction focusing mass analyser 37 Figo11 Mass spectrogram with two peaks representing atomic masses . M and M+£M„ 37 Fig„12 Source assembly 43 Fig.-13 Collector assembly 43 Fig„14 Construction of s l i t s 43 Figo15 Simplified diagram of the ion current ratio-measuring servo-voltmeter 47 Figo16 Pressure scattering correction 59 Fig,17 Representative cross section across the main lode at Broken H i l l 65 Fig.18 Idealized cross section of the Broken H i l l mining area 66 Figo19 Section at Black Star Mine, Mount Isa 67 Figo20 University of British Columbia analyses 77 Fig„21 University of Toronto analyses 78 Figo22 University of British Columbia analyses 79 v i i Fig. 23 University of Toronto analyses 80 Fig„ 24 Outlines of analyser tube and i t s relative position to magnet 89 Fig„ 25 Details of tube construction 89 Fig„ 26 Approximate impedance characteristics of the magnet measured at i t s terminals." 90 Figo 27 Magnet current supply 92 Fig. 28 Filament emission control 98 Fig„ 29 Voltage coefficient of Victoreen resistors 102 Figo 30 Ion current preamplifier 103 Fig, 31 Lead equalizer and attenuator 103 Figo 32 Original circuit of output stage i n Brown, No. 358816 chopper amplifier 104 Fig, 33 Modified Maxwell bridge for velocity damping 104 Fig„ 34 Servo motor-driven potentiometers and attenuators in measuring circuit 105 v i i i LIST' OF TABLES Table 1 Isotopic composition of lead in meteorites by Patterson 12 Table 2 Interlaboratory comparisons of sample T-1003 31 Table 3 Isotopic analyses of galenas from main Broken H i l l lode by Russell, Farquhar and Hawley (1957) 32 Table 4 Printed version of data tape 61 Table 5 Printed version of the output tape from the ALWAC computer Showing the calculated isotope ratios 62 Table 6 Leadfi from Broken H i l l , New South Wales 68 Table 7 Leads from Mount Isa, Queensland 69 Table 8 Replicate analyses of sample No.l 70 Table 9/. Differences i n isotopic composition of Broken H i l l samples 75 Table 10 Differences in isotopic composition of Mount Isa samples 76 Table 11 Shunt coefficients 107 INTRODUCTION Natural isotopic abundance variations of elements occuring in rocks and minerals have provided new and powerful methods in geophysics. Small variations in isotopic composition are often more significant than the natural abundances themselves. These variations reflect the different histories of rocks and of minerals and may give quantitative information about processes and conditions in different geological environments. Considerable progress has been made in this exciting f i e l d of geophysics during the last twenty years as mass spectrometers and associated tech-niques have improved and have supplied more and more satisfactory data. Of the many North American laboratories engaged in studies of isotope abundance variations (five i n Canada and three or four times this number in the United States) a l l use instruments differing only in detail from that used by A. 0. Nier i n his pioneer work around 1940, and instru-ments of this type are commonly referred to as "Nier-type". With only three recent modifications, namely the use of solid sources, electron multipliers and simultaneous collection techniques, versions of the Nier-type instrument have been capable of determining satisfactorily the isotope ratios of interest to geophysics. It is only within the past few years that some of the particular investigations being undertaken have required a precision not easily obtained with most of these instruments. Many of the Canadian contributions to geophysical isotope studies have dealt with the element lead, and have been primarily carried 1 2 -out in the Geophysics Laboratory at The University of Toronto, where the present writer began his work in this f i e l d in 1957 - Since most of his previous research experience was in geophysical electronic instrumentation unrelated to mass spectrometry, he was In the position of being able to examine the problems in this type of instrumentation c r i t i c a l l y and without the natural prejudices of one trained i n i t i a l l y in mass spectrometry. At that time the studies of R. D. Russell and R„ M„ Farquhar were at the point where some aspects of the future work would be limited by the precision of the mass spectrometer, and the- opportunity occurred for the writer to_co-operate with Dr. Russell in a c r i t i c a l examination of possible developments in the instrumentation, A survey of the literature shows remarkably few developments in recent y e a r s i n improving the precision of isotope abundance determinations of heavier elements. For example, -the recently published book "Advances in Mass Spectrometry" (1959), incorporating over 7G0 pages contributed by approximately 100 authors, contains no discussion of electronic ci r c u i t s ; the bibliography of 600 papers on instrument design starting at 1938 l i s t s surprisingly few papers dealing with this aspect. Even in the f i e l d of ion optics, which has been the subject of a great deal of theoretical study, such basic questions as the effect of the fringe f i e l d of the magnet is known only sufficiently for rough estimates to be made of the deviations in focusing conditions from the idealized case. The writer soon came to the conclusion that the less stable com-ponents of the mass spectrometer are the electronic units and that the available precision i n conventional operation is limited to a few tenths of one per cent-mostly as a result of imperfect s t a b i l i t y of the power supplies and inadequate measuring cir c u i t s . It therefore seemed that he 3 could make a suitable contribution to this f i e l d by giving attention to various instrumentation problems which have been largely neglected by mass spectrometrists s and by demonstrating in particular that much improved precision in lead isotope analyses can be obtained. At The University of Toronto the writer made some preliminary designs of anew metal mass spectrometer tube and contributed-to the development of electronic units. When'first he (and later Dr. Russell) came to The University of British Columbia the possibility arose of build-ing a complete mass spectrometer, designed specifically for the measurement of isotope ratios of heavy elements*with greater precision than had been previously achieved. This thesis is concerned with-the development and operation of that mass spectrometer. The first-chapter demonstrates the need for precise'measurement of lead isotope abundances. The second chapter, after showing typical mass spectrometers and their precisions deals with those aspects of the present instrument which represent de-partures from the usual practice. This is followed i n the third chapter by a demonstration of the use of this mass spectrometer for the precise intercompartson of;*lead Isotope abundances of lead from Broken H i l l and Mount Isa, Australia. The results-show effects-of geophysical importance that were missed in previous less precise analyses. The appendices contain short"technical descriptions of important parts of the instrument. Although most of the development work was invested in the electronic units the author" does not want to overload the reader in the same proportion, A great deal of time was spent in designing and assembling many of the more standard parts, because this was a prototype instrument not copied in any part from another mass spectrometer. Descriptions of two of the transistorized supplies have been accepted for publication (Russell and Kollar^ in press). Also such problems as design for ease of maintenance and other purely engineering considerations are omitted from the thesis, which contains only those parts of the writer's work that are novel and have some sc i e n t i f i c interest for geophysics. From an economic aspect, the materials used in this mass spectro-meter cost between one-fifth and one-tenth the price of comparable com-mercial instruments,- The f i r s t spectrograms were obtained 21 months after the decision*was made to build the instrument and only 17 months after the actual construction began. Many people have contributed to this program. Dr. R„ D, Russell made the overall design of-the mass spectrometer as well as specifying the performance required for his research problems. Within this framework the detailed design and construction of most parts of the mass spectrometer were under the immediate supervision of-the writer, who wishes to acknow-ledge that-the success of this project wasvmade possible by the help'and co-operation of many people. The valued assistance of Professor J, A, Jacobs, who served as the writer's supervisor during the1 i n i t i a l stages of his research and the chairman /of -his'Ph.D committee throughout, is sincerely appreciated, T . J , Ulrych constructed'certain of the electronic units, and P. Neukirchner, others, J, S, Stacey designed the high voltage supply with i t s incorporated safety circuits and the elaborate system of supply interconnections that has proved very valuable'in making1 refinements and modifications to the instrument," A, B„ L„ Whittles measured the voltage coefficient of the Victoreen resistors used in the ion current amplifiers and J. Lo Allard wrote vthe computer program used for some of the calcu-lations. The mass spectrometer'construction'depended greatly on the f a c i l i t i e s of the Department of Physics and the help of many of i t s faculty 5 and, staff members. The writer'also wishes to acknowledge the assistance of Professor J, T. Wilson who was responsible for making the i n i t i a l arrangements for him to continue his graduate studies in Canada, and of Dr„ R, MO Farquhar who gave valuable technical advice. This project was financed by-^he National Research Council of Canada, The-Geological Survey of Canada and The California Research Corporation, CHAPTER 1. LEAD ISOTOPE ABUNDANCES The present state of lead abundance interpretations At the beginning of the century Lord Rutherford had proved . that the radioactive elements uranium and thorium decay ultimately into helium and lead. • The work to establish an1- absolute time scale for the earth's history soon began. From the ratio of lead to the radioactive parent the ages of uranium and thorium minerals were calculated. These ages were correct as far as the assumptions of purely radiogenic lead and unaltered samples were justified. .Contamination of the minerals by common leads-caused difficulties and further advances were possible only after research on-the isotopic composition of common leads. Nier (1938) and Nier, Thompson and Murphey (1941) measured the isotopic composition of 25 common lead samples from different localities and from different geological environments. They found a correlation between the1 age'of-the lead ore'deposits and the abundances of lead-206, lead-207, lead-208 relative to lead-204. This was interpreted as resulting from the production in the earth of these isotopes by the radioa-ctive. ' decay of uranlum-238, uranium-235 and thorium-232 (Fig, 1). Geochemical models were suggested which would produce during the earth's history the isotopic abundances observed at the present. These models have been the subject of continuous re-examination by the addition of newly-measured data and by considering the reasonability of the geophysical consequences, 6 7 Instead of giving a f u l l historical account of these studies or describing the less significant modifications of important models, the writer'wants to show some representative ideas to serve as a back-ground for-this project. Interpretations of lead isotope abundances are always based on the following assumptions. (i) There was an early stage in the earth's evolution when a l l the leads were well mixed. The isotopic composition at the end of this stage Is well-d"ef ined; lead of this composition is called primeval lead". That time serves as the origin of our time scale. This time is denoted t 0 ( i i ) Any observed variations in lead isotope ratios are entirely the result of the addition of different amounts of radiogenic lead to the primeval lead. The time Interval between the present and t 0 has been identi-fied with the age of the earth. Common leads, i.e. leads found in minerals free from uranium and thorium, contain an isotope of lead of- atomic mass-204 which is stable and which is not known to be generated by any radioactive decay. Therefore a l l lead-204 existing today is primeval. Because mass spectrometers measure relative abundances i t is suitable to express quantities of the other lead * Lead-204 has been reported to be radioactive with a half l i f e of 1.4 x i o 1 7 years by Riezler and Kauw (1958). This result would not require any modifications of existing lead isotope theories. 8 isotopes relative to lead-204. Alterations of the isotopic compositions of leads other than by radiogenic addition such as by isotopic fraction-ation due to physical-chemical processes, are negligible 1, certainly in the present range of possible^precisions, (This last point w i l l be elaborated later). Based on the assumptions above the relative amounts of the lead isotopes can be given as follows, denoting the pertinent isotope ratic by x, y, and z and the time of isolation of the sample from uranium and thorium by t : t 0 P b206/ P b204 = x = ao + o t J v A e ^ dt t P b207 / p b204 = y = b 0+ JV A' e V t dt (1) .to P b208 / P b204 = z = C q + JW A " e V't dt t where A , A ! and V are the decay constants of U 2 3 8 , U 2 3 5 and T h 2 3 2 . a Q 8 bo and c 0 are the isotope ratios of primeval lead 8 and V equals U 2 3 5 / P b 2 0 4 , W equals Th232/pb204 f and °C equals U 2 3 8/U 2 3 5 a l l extrapolated to the present time ( t = 0 ), The equations (1) are rigorous but very general, because V and W are unspecified functions of both time and place, Gerling (1942), Holmes (1946, 1947, 1949) and Houtermans (1946, 1947) added a new assumption to the two above: 9 ( i i i ) V and W are different at different places within the" earth, each place representing a closed system. During the time interval t 0 to t these systems preserve their uranium/lead and thorium/lead ratios, except for the radioactive decay. In geological time the lead "grows" to an isotope composition determined by t 0 ; t , V and W are therefore characteristic of a particular environment as well as the age of a particular mineral. Thus Xi = a 0 + * V i ( e * t o - e ) 7 i = b 0 + Vi ( e * t o - e** ) (2) H = co ^ w i ( e V' t o ~ e > " * ) The lead ores deposited at time t , i f unaltered afterwards", • supply the information about the variables i n equations (2). It is usual to plot the- measured isotope ratios on graphs with families of growth curves, using y and x co-ordinates. Because £ y /. ( e V t ° - a * ' * ) S x <* • ( e *o - e ^  * ) = R (t) (3) the x and y values l i e on straight lines through a 0' and b 0 with a slope determined by t . Houtermans named these lines "isochrons". Isochrons are therefore the l o c i of leads having the same age. A diagram of this sort is shown on Fig. 2. 10 137-8:1 4-50 xI09years 0713xI09years 139*I09years(half-life) FIG. I RADIOGENIC LEAD ISOTOPES (AFTER COLLINS FREEMAN AND WILSON ) Growth cur ves —-— •06 t»30by. A —— Isochr •ons FIG. 2 206pb/207pb 1 GROWTH CURVES AND ISOCHRONS ( AFTER RUSSELL AND FARQUHAR ) The fact that the primeval lead composition was unknown prompted many attempts to determine it- from measured lead samples, using'similar relations like (1) or (2) or their derivates. Such- calculations by Gerling (1942), Holmes (1947), Houtermans (1947), Bullard and Stanley (1949), Collins, Russell and Farquhar (1953) and Russell and Allan (1955) gave different results for t'0* varying from 2„8 x l o 9 to 5 x 10 9 years, depending on the set of data used or rejected. New information about lead isotope-abundances-in the solar system has been provided by Patterson (1953) and Patterson, Tilton and Inghram (1955) from measurements on meteorites. The extreme d i f f i c u l t i e s in the micro-chemical techniques required most elaborate sample preparations. He published the measured data given in Table 1. Meteorites apparently represent materials belonging to our solar system, but of extra terrestrial origin. They satisfy-the assumptions about' closed systems (as required for the Holmes-Houtermans model), and i f they originated from an i n i t i a l l y homogeneous mixture of primordial matter their early lead' composition was identical with primeval lead. According-to Houtermans' ideas about the isochrons and arguments of Burling (1952) who re-considered the'properties of closed systems containing uranium and thorium, Patterson's points have to be linearly related in y and' x . The set of yTable 1."does"this, (see Figures 7 and 8) as was shown by Patterson (1956), proving both the assumptions involved and the quality of the measurements. Since modern terr e s t r i a l lead also f i t s these lines, the equality-of the age of the" earth and of meteorites is established. The Henbury and Canon Diablo meteorites were of the iron type, the lead being extracted from the t r o i l i t e phases for which V was shown to be approximately 0.0002. The isotopic composition of those two t r o i l i t e s was identified with primeval TABLE 1 Isotopic composition of lead in meteorites by Patterson Meteorite - 206/204 207/204 208/204 Nuevo Laredo, Mexico 50,28 34o86 67,97 Forest City, Iowa 19o27 15o95 39,05 Modoc, Kansas 19o48 15,76 38,21 Henb ury, Aus t r a l i a 9o55 10,38 29,54 Canon Diablo , Arizona 9,46 10,34 29,44 * Three additional meteorite analyses have been published by Starik, Shats and Sobotovich (1958) but these are believed by the writer to be much less reliable. 13 lead in the solar system. From equation (3), setting t = 0 and the slope <Sy/£x to the measured value of 1.678, and using the terrestrial oC equal to 137.8 (Inghram 1946). t 0 = 4 500 ±70 million years as determined by Patterson. The most remarkable thing is the good f i t " o f points to a straight line on the z versus x plot. The slope of this line from (2) with t = 0 is $z W " ( e to - 1 ) « ; ; ( 4) <$x <*V '" ( e * t o - 1 ) 3.72 was found for the ratio W/<*V which equals T h 2 3 2/U 2 3 8. Besides the similarity of this ratio to te r r e s t r i a l values i t s constancy is extremely interesting, and shows that no significant fractionation took place between uranium and thorium when uranium was fractionated relative to lead by a large amount ••(Patterson, 1956). Without going into details about the-geochemical significance of the results given above, i t is reasonable for our present purposes to accept-the -age of the solar system as "the age of the earth and use the average of the Henbury and Canon Diablo abundances as te r r e s t r i a l primeval lead. As the number of measurements increased, i t became clear that certain leads- had isotopic compositions which do not f i t into the growth pattern of the ordinary leads. These leads had large negative • apparent age values and were called anomalous by Holmes, After the accum-ulation of more data, suggestions were made for modified models. They were based i n part on the s t a t i s t i c a l nature of the spread of the observed points, such as is shown in Figure 3. This figure apparently supports the P b ^ / P b 2 0 4 15 necessity of 'relaxing the geologically d i f f i c u l t conditions about the Holmes-Houtermans closed local subsystems, and of assuming uranium/lead ratios which vary during time-. Such-types of-growth mode-Is represent conditions requiring very sophisticated geological mechanisms, and also this type of approach does not seem to help much in understanding the spread of points around any growth curve. On this basis, analyses and their interpretations met varied acceptance. A solution was shown for the problem of anomalous leads by " Russell and a l . (1954), Farquhar and Russell (1957) and Russell, Farquhar and Hawley (1957). They experimentally proved a linear relationship among leads from the same mineralization episode. Fig. 4 shows three examples. By the explanation of the above authors these were ordinary leads which have received-anomalous additions of radiogenic lead from near surface rocks containing uranium and thorium in significant amount. The slope of the straight line interconnecting the measured points depends on the ratio of P b 2 0 7/Pb 2 Q 6 (or P b 2 0 8/Pb 2 0 6 for the z - x plot) of the radiogenic lead addition and by simple relations give age limits for the surface subsystem generating the radiogenic lead. The shift along the straight line depends on the proportionate amount of radiogenic lead added. This mechanism was convincingly proved on both mathemetical and geological bases. It was soon realized that in spite of the existence of a spread of isotope ratios ;about an average growth curve these deviations were not great (less'than -5 per cent) indicating sources of lead ores remarkably uniform in the uranium/lead ratios; a/similar argument applies to the thorium/uranium ratios. The rejection of obviously anomalous leads improved that apparent uniformity. The assumption of uniform source for a l l leads gives the mathematical form of the model as given by Alpher and Herman (1951). 16 17 ...... r The equations are essentially the same as (2) but ' V and W are the same for a l l leads. Russell and a l . (1954), and Wilson, Russell and Farquhar (1956) give parameters for a single growth curve to which a l l the leads' that are not obviously anomalous f i t within t 3 per cent. The unique growth* curve was more attractive for physicists than for geologists. Their assumption-that the leads were derived from very great depth, where large scale uniformityismuch l i k e l y , could not be proved' directly and many arguments were given for and against. The suspicion that most leads were anomalous in different small amounts demanded a good criterion to recognize-a l l 'the anomalous leads. The other possible'approach was to find c r i t e r i a , for recognizing at least some of the ordinary leads. Stanton (1955, 1958) had been investigating sulfide deposits of' the group referred to as "conformable" by King and Thompson. (1953). His conclusions wero- that the galena deposits of this type are derived by deep source volcanism and quickly localized i n off-tfeef -sediments by re-placing iron in iron sulphide deposits. There are many lead deposits of this type and- their method of formation described above would suggest, that these* leads are ordinary (Stanton and Russell', 1959)'. Independent invest-igations on galena samples from Broken H i l l , Australia, which is the .type : conformable deposit, have shown a high degree of uniformity in the lead isotope composition among 17 samples. Such uniformity can ber used as one criterion to establish that a lead is ordinary. V a r i a b i l i t y in composition, trace element' content, and geological features proved that the leads at Broken H i l l in the s a t e l l i t e deposits in veins or fault zones are'anomalous. The linear relationship*of these samples (see Fig. 4) again indicates simple processes (Russell, Farquhar and Hawley, 1957). 18 Stanton and Russell (1959) selected lead analyses for eight different conformable galena deposits. The selection was based purely on geological grounds. The graphs in Fig. 5 and Fig, 6 show that these conformable leads f i t remarkably well to a single growth curve. Russell and Farquhar (in press) have proposed three classifications of leads as: (1) Meteoritic leads (2) Conformable leads (3) Anomalous leads Individually these classes can be explained by simple models*. Conformable leads are derived from deep sources (from the mantle of the earth according to Russell and Farquhar) which are uniform in lead, uranium and thorium. The original primeval lead continuously gets radio-genic additions in this source and i t s isotopie composition is thereafter a unique function of time. The conformable lead deposits are samples of that deep environment. Anomalous leads are derived from conformable leads by the addition of radiogenic leads generated i n surface rocks. The observed scatter of ore leads about a single growth curve in the pre-vious models is assumed to be solely the result of the inclusion of anomalous leads that have been not recognized as such. Because of the single growth curve Houtermans isochrons became meaningless (except for meteoritic leads) but'are replaced by the relationships for anomalous leads. In Figs. 7 and 8 are shown the conformable lead growth curves in • larger scale, with the straight lines determined by Patterson's meteoritic lead abundances, and a! highly anomalous set of samples from Thunder Bay d i s t r i c t , Ontario. 19 20 40 38 36 o Q_ 00 o CVJ CL 34 32 30 3 CD CD leiberg-— athurst—f -• •/-Mt uchansy^ Sullivan Isa X—Brok / ""Yukor en Hill I Tread well y^-Gene J^— Manitouwa va Lake dge •• i 13 15 Pb 2 0 6 /Pb 2 0 4 17 19 FIG. 6 CONFORMABLE LEAD GROWTH CURVE (AFTER STANTON AND RUSSELL ) P b 2 0 6 / P b 2 0 4 FIG. 7 RELATION OF CONFORMABLE, METEORITIC AND A SUITE OF ANOMALOUS LEADS (AFTER R U S S E L L AND FARQUHAR ) 22 23 The need for precise isotopic abundance measurements of lead In the preceding section certain aspects of the interpretations of common lead- isotope ratios were described. This description was necessarily a-brief one and omitted many of the finer points of this subject. However, i t is clear that there are two principal models for interpreting lead abundance ratios. These are the model originally proposed by Gerling and later by Holmes and Houtermans for their calculations of the age,of the earth, and the second is the model originally proposed by Alpher and Herman which is the'model that*has been extended and expanded by Farquhar and Russell. Although these two models are in gross agreement they dif f e r in many of the-finer points'and because the models- bear on a number of aspects of the earth's history,such as the formation of the continental masses and the origin of sulphide ore deposits, i t is of very great sc i e n t i f i c interest t o t r y to distinguish which of these is in better agreement with experimentally obtainable facts. It is generally conceded that isotopic analyses of leads carry an uncertainty of several tenths of one per cent in the best cases and a c r i t i c a l examination of some of the measurements reported in the l i t e r a -ture loadstone to suspect that these analyses often contain uncertainties approaching one per cent, particularly i n the case of lead-204 which has generally proved to" be the most d i f f i c u l t isotope to measure. It i s the writer's thesis-that an adequate appraisal of these two models can only be made i f a "greater precision is obtained in the experimental data i t s e l f . It is reasonable to expect that such a precision can be obtained, as i t is known that in the case of isotopes of the lighter elements, such as for example oxygen and sulphur, precisions better than 0.01 per cent are obtained on a routine basis by adopting specialized mass spectrometer -24 techniques and by using'Tapid-intercomparisons between a sample and a standard. It is to be-expected, however, -that the experimental d i f f i c u l t i e s in extending- these techniques to measure the isotope ratios of lead would be very much more d i f f i c u l t because i t is necessary to work in a mass range an order of magnitude higher. Supposing for a moment that i t is possible to obtain more precise analyses, there come to mind immediately several questions whose answers could throw a great deal of light on lead isotope interpretation problems. The f i r s t of these concerns the nature of the conformable lead curve pro-posed by Stanton and Russell (1959'). The data-presented by these authors and more recently by Russell and Farquhar (in press) suggest that conformable leads f i t a single growth curve within the presently available precision of the results.- -It i s of very great importance to discover whether this apparently good f i t between the measurements and a single growth curve is real or whether i t i s the result of coincidence and also to discover the limits within- which^these particular lead isotope ratios f i t the assumed curve. A second application of immediate interest is In 4the re-examination of a number of suites of anomalous leads. Russell and Farquhar have postulated that such leads from a single l o c a l i t y should be related linearly and have suggested ways in which the linear relationships can be used to find age and- geochemical information about the ore deposits and about the surrounding "geological formations.- In one case, including leads from the Thunder Bay region of Ontario, the linear relationships have been amply demonstrated^but-in other'well known cases'of anomalous leads, such as at Sudbury,-Broken H i l l , and Joplin, Missouri, the total range of variation of the isotopic abundances is so small that i t is extremely d i f f i c u l t to establish without question the linear nature of the relationships. This is 25 another problem that could be at least c l a r i f i e d by a rather small number of isotopic analyses of-lead-providing i t were possible to obtain a pre- ' cision significantly better than that previously available. In addition, i t would be highly desirable to re-examine in more detail the relationships of the meteori-tie-leads- among themselves and to terrestrial leads, but this is a problem of even greater d i f f i c u l t y and is one that is not under present consideration at this laboratory. Another poss i b i l i t y that must be borne in mind in working on these problems-is the fractionation of the lead isotopes in geochemical and geophysical processes. This,has never been considered a very l i k e l y process for lae.d but i t unquestionably does occur on some scale and i t is to be expected that more and more-precise measurements w i l l sooner or later reveal i t . Adamson (1959) used available spectroscopic data to calculate partition function ratios for compounds of lead and from this produced a table of equilibrium constants for typical lead reactions that suggest fractionations of the order of 0.05 percent at 0°C and 0.01 per cent at 327°C unless- multi-stage processes are considered. It would appear that fractionation effects from this cause are not l i k e l y to present a limita-tion to lead isotope studies in the near future. The fractionation of lead isotopes during diffusion processes has also been calculated (Senftle and- Bracken, 1955). The effect due to this cause is predicted to be somewhat larger than that due to equilibrium fractionation processes and this in fact may be observable when a reasonable improvement in pre-cision is obtained. It has been the writer's object to develop a useful technique and to use this technique enough to provide a reasonable test. To do this 26 he has selected certain'samples from Broken H i l l , New South Wales, Australia, and Mount Isa, Queensland, Australia, to serve-as two test cases. Leads from both of' thes-e locations- have been studied before, primarily at the University of Toronto although'a few analyses extend back to the original results of Nier. These are both1-stated to be conformable deposits of the type considered by Stanton and hence are interesting from the point of view of testing the model of Stanton for ore deposit formation. Since both are mining areas of considerable economic importance both are described well in the literature, thus f a c i l i t a t i n g the present investigation. There are two additional reasons for being interested in Broken H i l l samples. One is that a sample from Broken Hill-has been distributed widely to labora-tories around the world as an interlaboratory standard and1 the second is that there are known to be anomalous leads at Broken H i l l which occur in distinct and easily recognizable structures. The known analyses for the samples from Broken H i l l , Australia, and Mount Isa, w i l l serve as a standard against which to compare the results obtained in this research. This is a-suitable standard because these mea-surements, according to Russell, were made with great care and are typical of the best analyses ordinarily obtainable with previously existing techniques. CHAPTER 2 INSTRUMENTATION General considerations Depending on the purpose for which a mass spectrometer is i n -tended, namely, precise mass determination or relative abundance measurement*-, the demands on instrumentation are different in many aspects. Similarly, applications i n physics and chemistry have led to a great diversity in design according to the particular requirements, Inghram and Hayden (1954)' give an impressive l i s t of many kinds of mass spectrometers. It shows the great variety of attempts to use different physical phenomena depending on atomic mass in mass spectrometer designs. Instruments of the type which are most successfully used inisotope abundance work-have been developed from Dempster's magnetic analyser and particularly from Nier's mass spectrometers since 1936. Abundance determinations need only moderate mass separation capabilities from the spectrometer compared to modern standards of pre-cision mass measurement. Therefore, the structure of the mass :spectro--meter can be much less complex since the ion optical aberrations of the simpler designs can be unimportant. The typical arrangement of a mass spectrometer (-Figure 10) consists • of 1 ion source and collector assemblies -and a-magnet. Either-a sector magnet with a 60 or 90 degree angle is used, or else an arrangement where the ionic paths are entirely in a homogeneous magnetic f i e l d (the 180 degree instrument). From the source a ribbon of monoenergetic ions is injected into the magnetic analyser which deflects and refocuses the diverging ion beam at the collector. 27 28 This type of instrument is therefore called a direction focusing mass spectrometer. At the focus the collector s l i t accepts one component of the dispersed ion beam. By varying the magnetic field the total spectrum can be scanned. The-ion currents of the different masses are subsequently measured at the collector by measuring the voltage drop developed- across a high resistor. A high gain d.c'. amplifier with negative feedback supplies-an output signal proportional to the ion current at a conveniently low impedance level. Automatic recording permits the spectrum to be examined repeatedly and the abundances can be determined from the spectrograms with a higher precision than could be obtained by direct measurement. Though this scheme is simple enough, many subtle sources of' error arise when precision measurements are attempted. During the last two decades many mass spectrometers and their associated problems were described in'-publications and were reviewed in handbooks and monographs. At the beginning of his studies the writer experienced some difficulty in judging the degree of perfection necessary for the different elements in a good mass spectrometer. It is clear that for a precise measurement: (i) The analyser must disperse the different mass components sufficiently so that ions from the adjacent masses do not affect the measured intensities. (ii) Discriminations by the source or sample handling system-may not cause unknown variations in the apparent isotopic composition of the sample. ( i i i ) The source should supply an ion beam steady enough so that there are no significant fluctuations in the intensities of the successively measured isotope components. 29 (iv) The measuring system has to perform linearly and has to be capable of a high precision. This l i s t , containing the minimum requirements, actually starts with the least d i f f i c u l t problem. For isotope abundance measurements-there is usually not much d i f f i c u l t y in establishing the necessary resol-ution which is determined largely by the geometry of the system (See Figure 11). Discrimination exists in a l l mass spectrometers causing the measured abundances to be different from those in the original sample. Nier (1950) published absolute isotope abundances of a number of elements by calibrating his instruments with known synthetic isotope mixtures. His absolute measurements were vessentially relative comparisons with a known standard. But a mass spectrometer changes with time although the operating conditions seem identical. The most significant source of variable discrimination is the ion source, and the processes in this part of the mass spectrometer are only partially understood. Surface ionization-'sources,' advantageously used to analyse" microgram quantities of solids,-are especially unstable. In gas sources, where the ions are produced by'electron bombardment, a more steady behaviour can be obtained. However, the intensity fluctuations of the ion beam appear to be one of the important factors limiting the precision when a scanning technique is used. Ingenious solutions have been found to eliminate the effect -of these fluctuations.. To-measure the ratio of two isotopes Strauss (1941) and Nier, Ney and Inghram (1947) used two collectors and compared the two 30 ion currents simultaneously in a bridge arrangement without scanning;' In this way any ion current intensity "variations-appearing i n the same proportion in both components were cancelled. With the application of this method McKihney, et a l . (1950) reported a precision of 0.02 per cent obtained in-measuring differences in O^/o-1-6 isotope ratios. Unfortunately simultaneous collection and measurement do -not lend themselves to the analysis of complex spectra. To reduce the effects of source in s t a b i l i t y when a spectrum Is scanned an alternative method can be used, as described by Gorman, Johns and Hippie (1951) and by Stevens and Inghram (1953). A s l i t or grid intercepts a constant fraction of the total ion current and by a ratio measuring circuit the current of a particular mass at the collector can be compared to the total ion current. This technique is usually referred to as grid-ratio recording. Sources of error i n the'measuring system, although small, may not be negligible. The nonlinearity of the resistor and the d.c- amplifier may cause systematic errors which become important i f conditions of mea-surement are not always identical. Noise of the measuring system limita-the precision of measurement of the isotopes with low abundances. To reduce random errors, measurements are generally repeated five or more times. Certain laboratoriesare reported to record as many as thirty successive pairs of spectra for one complete measurement. An i l l u s t r a t i v e example of the present-precision of lead isotope abundance measurements is* given-in Table 2. Parts of the same galena sample from Broken H i l l , Australia, were distributed among different laboratories :by the Geophysics Laboratory of The University of Toronto. The agreement in relative abundances is good in view of the fact that both the techniques and instruments were different in each case. 31 , TABLE 2 Ihterlaboratory comparisons of sample T-1003 100 times ratio of 204/206 207/206 208/206 6,24 ± .03 96.21 ±,04 222.4 ± .5 (3) • 6.239 ± ,008 96.33 ±,14 •223.8 ± ,4 (3) 6.212 96.09 223.2 (4) 6.238 ±.013 96.28 ±.20 223.1 ± .5 (5) 6.216 ±.015 96.53 ±.08 222i5 ± .2 (6) 6,169 ±.031, 95.87 ±.09 220.2 ±2.2 Instruments and techniques analyser magnet ion source collector spectrum correction (1). 90°, 12 in. solid electron mult. Pbl + square root (2) 60°, 12 in. solid electron mult. Pb + of'the mass (3) 90°, 10 in. gas Faraday cup Pb(CH 3) 3 + C 1 3 and" (*)... 90°, 6 in. gas Faraday cup Pb(CH3-)i3+ hydrogen (5) 180°, '6 in. gas Faraday,cup Pb(CH 3) 3 + loss -(6) 180°, 10 in. gas Faraday cup Pb + Hg 2 0 4 and hydr: Laboratories (1) Atomic Energy Research Establishment, Harwell, England (2) California Institute of Technology, Pasadena, California, U.S.A. (3) Geological Survey of Canada-, Ottawa, Ontario, • Canada • (4) University1 of - Toronto, Toronto, Ontario,'Canada is) University of Toronto, Toronto, Ontario, Canada (6) University of Tokyo, Tokyo,' Japan 32 Table 3 Isotopic analyses of galenas from main Broken'Hill lode by Russell, Farquhar and Hawley (1957), 100 times ratio of Sample No, 204/206 207/206 208/206 T-667 96,64 222.8 T-668 6,186 96,56 222.4 T-669 6 , 23rj 96.56 222.4 T-670 6,22g 96.43 222,0 T-671 6,20q 96,64 222,8 T-672 6,198 96,68 222,9 The tabulated isotope ratios are averages calculated from spectrograms of five consecutive double scans. 2 0 4 / 2 0 6 2 0 7 / 2 0 6 2 0 8 / 2 0 6 &20 INTERLABORATORY 6.25 96.0 96.5 co 0 CvJ OJ 222.0 223.0 6.20 6.25 96.0 96.5 222.0 223.0 TORONTO — 180' FIG. 9 REPRODUCIBILITY OF ANALYSES 34 Only the values of The University of Tokyo appear to contain errors greater than i s considered acceptable. The""degree of' reproducibility' in one laboratory is shown by Table 3, The '-same "instrument 'and' same' technique'was used with different-samples from the main lode and conformable layer at Broken H i l l . Accord-ing to the theories about the conformable''lead deposits, and according to the measurements made at The University of British Columbia with the new instrument (presented Chapter 3), there are only very small varia-tions in isotopic composition of lead throughout this ore. The measure1-ments in Table 3 were normal routine analysis made at The University of Toronto on different days. Figure 9 shows the data from the1 two tables plotted to show the nature of the variations. The author- has had some experience with the 180 degree instrument at Toronto which has been used here for comparison both in the design and in the particular measurements shown. He'could also study a well engineered,' commercial mass spectro-meter (Metropolitan Vickers, MS- 2). Vi s i t s to the mass'spectrometer laboratory at* the Physics Department of McMaster University, Hamilton, Ontario, were 'Stimulating, too. During these studies the writer realized the simple fact that the accuracy of a well working mass spectrometer can be kept to i t s best level only by tedious effort. During the project i t was proved that to be guided only by the published literature creates a very wrong perspective and personal experience and the guidance of an expert mass spectrometrist is essential. Through personal experience and by considering published results, the writer believes that the lead isotope analyses made at The University of Toronto are among the best of the results currently being published. Therefore the' techniques and results of the Toronto laboratory have been 35 used as a standard against which to compare The University of British Columbia measurements. The design of a precise'mass spectrometer Considering the requirements in precision,it was decided that" an improvement by an order of magnitude compared with the Toronto standards would be desirable. This precision would be certainly enough to find answers to many particular problems of lead-isotope abundance variations. The instrument was intended for use in general abundance • measurements although i t was f i r s t equipped for lead analysis. At the start, operation with lead tetramethyl samples and the use of an electron bombardment gas source wefe agreed upon. To build the analyser tube from metal with easily accessible source and collector was visualized as a necessary feature. The f i r s t step i n the design was the selection of reasonable dimensions. It was considered that in spite of the good experiences at Toronto with the 180 degree instrument, i t s advantages do not outweigh the disadvantages of having a very restricted space in the source and collector regions. Furthermore, in the sector instrument the source conditions can be made less dependent on the f i e l d of the analyser magnet which is varied during a measurement. In addition, the small f i e l d i n the collector region permits placing the electrometer tubes in the high vacuum close to the Faraday cup resulting in smaller stray capacity and electrical leakage.- The resolution of the instrument was planned to be higher than typical lead spectrometers'have, so a larger radi-us-than 6 inches seemed desirable. This also suggested the selection of a sector magnet or else an unreasonably large electromagnet would be necessary. 36 Both 60 degree (mostly in'the United States) and 90 degree- (Canada and Great Britain) sector magnets are favoured-. For the same radius, instru-•ments"of -small"er. angles1'have'"longer 'ion -paths- and -therefore-"larger press-•>ure -seatteriogT This can be a significant source • of1 error in lead -1etramethyl aaaiy«efl-;'**'"!Hiu$ , v ia • 90 'degree-•iil"ns^ume-nt,'*-wasm-prererred- to a 60 degree one. -Similarly, increasing the radius also increases the- path length,and the aberrations too are proportional to the radius. For this reason an instrument of larger radius does not give as great improvement in performance as might at f i r s t be expected. Therefore, the selection of instrument size was not a simple one. However, we feel at present that there is not too much to be gained by using a 90 degree instrument larger — than 12 inches for this problem, so a mass spectrometer with a 12 inch radius, 90-degree, f i r s t order direction focusing analyser s t i l l seems to be a good - choice"; Here should be mentioned that higher resolution -could-•also«ibe'-'-'ob'-tiai'ned"-'-itt:-,the^ C'asev--of-•'a-'isma!l-l-er''''radius- instrument by working with narrower s l i t s . The use of s l i t s too narrow exaggerates the d i f f i c u l t i e s of s l i t alinement and reduces the intensity of the beam. If this loss in ion current i s prevented by a better focusing of the ions to the exit s l i t , i n t e n s i t y fluctuations may arise much easier by electric or magnetic f i e l d changes-causing"a- displacement of the beam. By specifying a resolution of 450 the sum of collector and source s l i t widths was closely determined. The dispersion is D ~ 0.048 inch per unit mass around mass 250 so a collector s l i t Wc » 0,020 inch permits an image width Wj_= 0.0066 inch according to relations given i n Figure 11. The source s l i t used has a width Ws = 0.004 inch making an allowance of 0.0026 inch for the R<xz term and for broadening due to the additional miscellaneous aberrations. This increased image width is caused by the 37 M M-K$M — D — Resolution = M D D = dispersion Wc= collector s l i t width W'i = image width = W„-tR c < 2 +;misCo Fig. 11 Mass spectrogram with two peaks representing atomic masses M and M+SM. 38 i n i t i a l energy spread of the ions, fluctuations in accelerating voltage, non-uniformities of the magnetic f i e l d , misalinement of s l i t s , scattering at s l i t edges-;--ion-molecule •col-lisions-and 'by-space- charge-effects to this broadening the observed peak shapes have rather rounded corners and t a i l s asshown by the dotted lines on Figure 11.The resolution which gives- the dispersion in units of the idealized peak width at i t s bottom is not adequate by i t s e l f to show the analytical quality of an instrument. For accurate measurements the peaks should have a f l a t portion on their tops and t a i l s * s u f f i c i e n t l y small. In connection with t a i l i n g Duckworth (1-958, P.60) has given a definition of "abundance sensitivity" as the ratio of the peak ion current at mass M to the ion current of the t a i l at mass M+l. According to him, a value somewhat greater than 10 4 at mass number 100 is the best value obtained with single focusing 5 instruments, giving a limit of 10 i n abundance ratio when detecting a faint •isotope adjacent to an abundant one'. In oup instrument the writer has found the t a i l i n g of peaks to be caused mostly by pressure scattering and has been able at normal operating pressures to obtain an abundance sensitivity of 2 x 10 4 at mass 250. To reduce errors caused by the t a i l s "he: applies' a- correction- to -the•'measured abundances' as shown i n the section describing the measuring techniques (Chapter 3)„ Power supplies Because the magnetic analyser acts as a momentum f i l t e r the homo-geneity i n energy for the ions is essential to get a well defined mass spectrum. The type of ion source used produces ions with an i n i t i a l energy spread in the order of a few tenths of a volt. The use of a high acceler-ating voltage has the advantage that the relative energy spread became small, the focusing properties of the ion optics are better and the number of ions 39 which are lost to the walls of the analyser tube due to la t e r a l velocity components in the ion beam can be reduced« The high voltage supply delivers 5 kilovolts, more than twice the voltage-used with many lead instruments. The s t a b i l i t y requirements for this supply-' were -obtained' by "specifying the permitted displacement of ••lon--'-be,am"--at-'-i}iie-,'-coli'ector---due''™to'-vhi-gh"Vol"ta,ge ' f l u c t u a t i o n s A s t a b i l -i t y better than 1 i n 10 4 was required to obtain a displacement notgreater than caused-bythe 'unavoidable i n i t i a l energy-spread of the ions. A s t a b i l i t y better than this was achieved by designing the supply in a conventional way but with special attention on the voltage reference in the regulator section and by stabilizing the line voltage by a Sola •a.'C.-'!'Stabi»l!ize^  voltage is- continously monitored by a 1 per cent mirror-scale meter and a fine control with a calibrated potentiometer makes small changes, known within 0.1 volt,for resolution checks. This supply was designed and constructed entirely by J. S, Stacey. The magnet current supply shows major departures from the con-"Tentional^-=d«8i"gns?aus«d,;-for'-mass'-spectrometers.- 'This circuit r i s a combination"of'transistor^ vacuum tube- and magnetic amplifiers. It regulates •the current•of the magnet coils by comparing the voltage drop on a manganin-resistor-in series with the - coils with a reference voltage of 5.36 volta'^upplied»'by--'foup--mereu3V''cel"Is»' A transistor difference "amplifier -drives 'power transistor working-as- a current regulator -connected im'-Beries-wi't^ the coils as a load. To make possible a control over a considerable current range a saturable reactor in-series with the primary of the power transformer acts as an auxiliary-regulator by keeping the voltage drop on the power transistor at about 27 volts. The transistor regulator can handle fast voltage 40 changes in the d„c. output from the r e c t i f i e r within a range of about plus or minus 25 volts. Slow variations are regulated by both the saturable reactor and transistor amplifiers so that the regulation obtained is -the product of - the stabilfzation factor of -these' two amplifiers. Because the-maximum collector voltage is specified as 60 volts for the power transistor type used,a clipping circuit protects the transistor during the switch-on transient- period or i n case of a sudden large change in line voltage. The outputJ-of the magnet current supply is 500 m i l i -amperes at 360 volts to excite^the magnet to 5.5 kilogauss, the largest • f i e l d required. With 5 kilovolt ions this f i e l d is sufficient to analyze masses to nearly 300, The magnet current can be adjusted to scan mass spectra by turning a 15 turn helical potentiometer either manually or by a small variable •'-speed reversible motorr The s t a b i l i t y of the current is the order of 1 part in 25,000 for periods of seconds or minutes. This is highly adequate on the same basis as specified i n connection with the high voltage supply. An interesting aspect of the design of this supply was that because of i t s complex feed-back loops i t was necessary to know the "transfer functions -of"the--components - rather'well".An -electromagnet is often hand-l'ed ;as-a Bimple • inductance. This'ls'a great - overs impii--fication and"the"impedance-frequency function of the magnet was experi-mentally determined. The importance of coupled eddy currents in the solid iron-core cannotbe ignored for components of frequency greater than 0.2 cycles per second. The magnet current is monitored on a large 1 per cent precision mirror-scale instrument, which has also a scale proportional to the square of the current. This scale can be used as a mass scale. It is 41 worthwhile mentioning the great overall efficiency of this supply; with the maximum output power' of 180 watts 'the power dissipated by the circuit is only about" "50-"watt,s";"""""Thl-s--'smal,l-:power'"doe'snot-affect- the heat balance Of "Other :unit;s esigntfi'cantly'i'a^ Conventional' mass 'Spectrometer••magnet-supplies dissipate from a few hundred watts to a kilowatt. Filament emission control To measure mass-spectra by the usual practice of repeated scanning over a mass range i t is of paramount importance to keep the ion beam intensity quite constant. By symmetrical scanning the errors due to linear dri f t s can be eliminated, but deviations from a constant d r i f t with periods less than 5 to 10 minutes should be kept small. Using gaseous samples and electron bombardment to produce the ions, these s t a b i l i t y requirements must be f u l f i l l e d by the filament supply. The d i f f i c u l t i e s are that the electron-emitting filament and associated electrodes are-at a high voltage above 'ground and the filament represents a very low impedance load (one ohm). Therefore conventional circuits -have regulated' the- power in-'the a'.c, primary side of this supply. The availability-'Of compact s i l i c o n - r e c t i f i e r s and power transistors have made i t possible to•design this circuit in a different and more adequate way. For-normal operation the 0.001 inch by 0,030 inch f l a t tungsten filament 0.5-inchlong, requires about4 amperes for the emission of 500 microamperes. It works under temperature saturated conditions and there-fore the emission can be controlled by a very small change in filament current. Two power transistors in the Darlington compound arrangement 42 (Shea 1957) having a large current gain work as a regulator in series -with t he f ilament.- A small preamplifier- transi s to r wit h an-- impedanc e-matching emitter-follower imput stage controls the power transistors to adjust the-filament-current and consequently the emission until the •dlfference-'-'between"the'>27' 'Volt'-Zener'diode' reference and voltage- drop due to the"Memii8'Si-on!i*current-i,on«,,-the-viad?jus'-t-a;ble'-"amiss-i-on•••-control resistor •becomes close'to zero, 'A d.c, supply designed for a 6 volt and 5 ampere output was built as a 6 phase half wave r e c t i f i e r , having an r,m,s. rippl of only 4.5 per cent. The ava i l a b i l i t y of three phase power made possibl this compact arrangement making unnecessary the bulky f i l t e r components otherwise -necessary if ' only single phase- power is used. The secondaries of the three phase"transformer are insulated to withstand 5 kilovolts. A small glow-tube stabilized sub-supply provides the electron accelerating voltage, the trap bias and the repeller voltage, as well as the current for the Zener reference diode. The s t a b i l i t y of this unit i s adequate when operated from the unregulated mains. In the case of typical line voltage changes (a few volts) the filament emission varies a few parts in 10,000 causing corre-sponding ion current changes. The source and collector-electrode assembly The-source electrode assembly was designed by R. D. Russell =' and i t is a modified version of the ion source used in the Toronto 180 degree instrument. Its main feature is a pair of half s l i t s between the draw-out 'plate and exit" s l i t , and connected to a negative potential (Figure 12)-,; This-arrangement-was found-to give a small- angular -spread and higher ion beam intensity than the conventional Nier-type source. 4 3 as sample S l i t width inch 0 . 0 4 0 0 o 0 1 2 O0O8O 0 o 0 0 4 i t Case Repeller Electron beam (1 to paper * - 4 kv Draw-out plate -1 kv Focusing and deflector half s l i t s Gl© Exit s l i t F i g o 1 2 Source assembly 4 >-0 o 0 4 0 0 . 0 2 0 1 by 0 . 3 0 0 Resistors and electrometer tubes Mounting glass plate . Collector No, 1 - 1 2 v Supressor OND Collector s l i t ( + 2 5 v) Collector No. 2 GND Fig. 1 3 Collector assembly O Q O O Q O jGuide pin holes Fig. 1 4 Construction of slits 44 In the same time these half s l i t s are used to center the ion beam' by adjusting the voljage difference on the two sides,, The voltages given approximately in Figure 12 are obtained- from a 'bleeder'network built- from deposited carbon resistors and from high-quality carbon potentiometers used for fine adjustment. These bleeders are operated from the 5 kilovolt high voltage supply and from a small regulated supply with -1,5 kilovolt output o A set of 1.5 inch by 2 inch rectangular source electrodes were machined from a 0,020 inch thick Nichrome V sheet, A rectangular opening in each holds the actual s l i t by a press f i t . The'slits are machined •"-• from two pieces i n the manner used by McMaster University (Figure 14). Guide pin holes, with the help of a j i g , make possible the precise alinement of- the- s l i t s i n the assembled package. Between the plates cylindricalv-f-i'-red--'--ateatite-->-S'paee-3?6--dete-Hiiine--the spacing. These short tubular pieces f i t - on the four glass tubes with thin stainless steel bolts working as mounting posts through their centers. The collector assembly (Figure 13) is built from the same type of s l i t s as"the source. The Faraday cup is supported by a glass plate on which the-10^ and 1 0 ^ ohm- resistors and electrometer tubes are also clamped. This arrangementof the electrometer tubes- inside the vacuum prevents the i n s t a b i l i t i e s , otherwise hard to avoid, due to surface leakage on the glass-bulb caused by moisture. The particular 1 construction of the electrometer tubes (Raytheon type CK 5886) makes possible their operation in-magnetic fields since the construction of these tvjbes deter-mines electron"paths'-essentially copianar. A magnetic f i e l d as high as 100 gauss parallel to the electron beam was found to cause no increase in grid current, although i t produced small changes in the static character-i s t i c s . 45 The purpose and operation of the different collector s l i t s w i l l be shown i n connection with the measuring system,, The measuring system Special care was- given to- develop • a measuring system with a higher inherent precision than the typically used arrangements•„ The abundances- are generally calculated from spectrograms recorded1 on 10 or * 11 inch wide chart paper. Sensitivity ranges of the recorder are designed to make l t possible to record the peaks between about half and f u l l deflection. The peak heights, giving the particular-isotope currents, are measured with a transparent, precisely divided scale, so an estimate of 0.1 millimeter can be made. This would give 0.1 per cent as the possible error i n peaks recorded at a half scale deflection. The actual accuracy is worse because the base lines from which the- peak heights should be measured have similar errors. Moreover, the variation of length of plastic scales is quite noticeable, and we have substituted-a scale ruled on glass. Even with this precaution variations are ob-served upon repeated measurements, probably due to the change of the paper dimensions-with'humidity. The accuracy obtainable from the spectrograms is also limited by the imperfect performance of the recorder i t s e l f . The diverse opinions about the precision of these instruments made i t necessary to check i t . The continuous-balancing-recorders are not capable of finer resolution that is determined by the number of convolutions on the slide wire, * Bern University uses a mirror galvanometer with a 1 meter scale rather than a paper chart recorder. Readings are taken directly from the galvan-ometer. 46 even when backlash is assumed to be negligible. Two different custom built Brown chart recorders were used with the present mass spectrometer. In the f i r s t , single-pen recorder the slide wire had 700 turns. This means a resolution of 1 in 350 for half scale deflections. Our second, two-pen recorder is twice as good in this respect but s t i l l not as-good as would be desired, (Neither recorder exhibited the back-lash of a Leads and Northrup recorder we had previously used,) Realizing these limitations of-recorders, often ignored, a design was made for a continuous balancing servo voltmeter incorporat-ing similar elements as used in the conventional chart recorders, but using a newiy'"avai'l!ab-l-e'---l-0--turn---'hel'i--c-al--'-potent-i-ome-ter (Helipot 7600 series) linear with 0.025 -per - cent,-5 *• Figure 15 shows- the simplified" : diagram of this unit. The circuit is designed to operate as- a "grid-ratio" measuring system when collector number 2 monitors a constant fraction of the total ion- current of the spectrum. To overcome the-usual troubles-with this technique, particularly varying secondary emission asthe spectrum moves along the grid during scanning, mechanical' scanning with-the collector s l i t has been proposed by R. D. Russell. The mechanical details of this system which has not been used yet in mass spectrometers, are s t i l l under development, but a l l the necessary electronic parts have been incorporated in the mass spectrometer and tested. Measurements presented in this thesis were carried out with an experimental copper tube of more or less conventional design. The present collector assembly has a s l i t (collector 2 on Figure 13) for the ratio measurement. The improvement in s t a b i l i t y with a stationary beam was satisfactory when both collectors were operating. However, when the spect 47 Pre amp. Brown chopper amp. Rg. (10 1 0 ohms) BVr Preamp. A' -13 v ;Rj: (10 1 1 ohms) lOv ,»iT^  gi.^ R l - ~ ^ ^ V j To chart recorder -v Q = i i R i B i sE 2A Constant input impedance attenuator controlling sensitivity in five calibrated stages, A' = Attenuator ganged with A A x A' = const, B = Constant input impedance attenuator -continuously adjustable. Determines operating range. Fig. 15 Simplified diagram of the ion current ratio-measuring servo-voltmeter. 48 was scanned false corrections made by the auxiliary system were intro-•«.e duced by^varying secondary emission. This indicated that without the moveable s l i t the effectiveness of the ratio recording is dubious/ and further investlgat'ions were made -on the precision that measurements with the single collector can furnish. For the single'collector operation the second amplifier system is not working,so the voltage on the terminals of the precision 'potentiometer of the main system is constant. Experiments on secondary emission effects from the s l i t s i n the collector assembly resulted in the choice of the supressor voltages shown on Figure 13. The reproducibility of repeated analyses have become better as step by step improvements in reducing the noise of the measuring circuit and a uniform operating practice were achieved. The f i r s t stage of the ion current amplifier consists of• the Raytheon CK 5886 subminiature electrometer pentode. It operates as a cathode follower to obtain the low output impedance necessary to'drive the following transistor emitter-follower stages. The gain of this rather unorthodox circuit is only about 0.5, but Its s t a b i l i t y is better than other arrangements -tried. The Brown chopper amplifier, which has a low input impedance (a few hundred ohms), demands the low-output impedance of the preamplifier to avoid serious matching losses. Great care was given to the equalizer networks compensating the phase lags caused by the stray capacity of the collector and inherent in the servo, motor and amplifier system. The-Brown amplifier was modified by increasing i t s velocity feed-back and a>phase lead network was incorporated into the attenuator section following the preamplifier. This attenuator is ganged with the sen s i t i v i t y -changing resistors controlled by five pushbutton switches. In this way the loop gain of the servo system is independent of the measuring range • 49 setting. Direct coupled- amplifier systems typically employed have a gain of 1000, while the loop gain of this electrometer amplifier isestimated to be about 10,000„The superiority of the servo volt-meter is mostly due to the- fact that the motor as an integrating element makes the position error zero when the gain is high enough to overcome f r i c t i o n in the bearings. This improved servo system permits the check of this position error by monitoring the error voltage on the input terminals of the chopper amplifier. This has. to be zero i n the ideal case. During preliminary measurements the error signal was recorded together with the spectrum and,the measured ion currents could be corrected from the error record. With later improve-ments of the system, this precaution became unnecessary. The shaft rotation of Mj_ (Figure 15) is proportional to the ion current with a precision not worse than 1 in 5000, Either a visual observation of shaft position on a precise d i a l , as i s used now, or photo-graphic recording, or a di g i t a l converter driven by the potentiometer • shaft to control a printing or tape punching device have been considered. -The latter p o s s i b i l i t i e s would reduce human errors in the reading of the di a l . Measuring less abundant isotopes needs attention to average the noise background which-has a relative-amplitude-not greater than 2 i n 1000, This noise is actually small; i t is only a few times more than the Johnson noise from the lO^ 1 ohm collector resistor, Nonlinearity of the collector resistor was thought to be an important potential source of systematical error. The 10^- ohm Victoreen resistor which is used is specified by the manufacturer to have an average voltage coefficient of not more than 0.03 per cent per volt between 1 to 100 volts. The voltage on this resistor i s the order of 10 volts with the 50 ion currentof 10~-'-(-> amperes obtained with the mass 253 peak from the lead-208 isotope.Whittles (I960) carried out measurements onthe same resistor now-Mused"-and'has-shownthe voltage coefficient to vary with voltage and ire be- 0.008 per cent per volt or less under 12 volts. Temperature-variations have a great effect on this resistor which-has a temperature" coefficient of -0.1 per cent per degree centigrade. Keeping this resistor inside the vacuum results i n sufficient reduction of the thermal disfr^ibances. The servo voltmeter motor rotates also a second "potentiometer ganged with the f i r s t which el e c t r i c a l l y transposes the shaft rotation to a signal suitable for driving the chart recorder. This instrument records the spectrogram in the conventional form with no possibility of loading the measuring c i r c u i t . Both spectrograms and dial readings marked on the corresponding peak tops during scanning have been used to calculate the isotope abundances. Checks on noise of the electrometer amplifier have shown base • line fluctuations of-not more than 1 x10~^ amperes during periods of the -order'-of' 10"-minutes-. -The- response time of the system, mostly determined by the motor, is less than two seconds for a -full scale deflection. The sample-handling- and vacuum systems This:part of the instrument consists of a 100 cubic -centimeter cylindrical flask connected with the ionization chamber in the source by about two feet of 1 millimeter glass capillary tube; a constriction at the source side of this capillary limits the rate of flow of the gaseous sample. The pressure of the lead tetramethyl sample in the sample line is •si about 10 millimeters of mercury for an ion current of- 10"^ amperes corresponding"to*the lead-208 isotope. Under these conditions the gas -flow through-"the-leak--ts-'Tis-cous;The •capillary" on"the • high' pressure •side r e p r e s e n t molecules is - high enough to overcome back diffusion from the leak entrance (Halsted and Nier, 1950). Vis'cous"sample-flow^provides' uniform-isotopiC'Oomposi-tion- during a measurement'jb'eoause -no depletion 'in lighter isotopes occurs as would be the case under conditions of molecular flow. A mercury ventil separates-the sample flask from the greased stopcocks and tapered joints serving to attach the sample container. A Toepler system adjusts the mercury level i n the flask'and, by changing i t s volume, controls the sample pressure. A small two stage mercury diffusion pump evacuates the sample handling system between analyses. The -main mercury diffusion pump for the analyser tube is a three stage glass unit. The pressure is typically 2x10"^ mm Hg measured on the 41 millimeter diameter pumping lead between the cold trap and copper analyser tube'. Pressures as low as 3 x 10 mm have been obtained'. During analysis the pressure rises to about l x 1Q~& mm. After i n i t i a l troubles with a two inch diameter isolation valve containing neoprene 0 rings i t was replaced with a straight tube section. Now only aluminum gaskets are used at joints and between the stainless steel flanges * It is interesting to note that this valve was described as "helium-tested" by the manufacturer. Presumably the d i f f i c u l t y is the vapour pressure of the neoprene not leakage in the conventional sense, and this would not be detected in an ordinary leak test. 52 and their covers both at the source and collector side. For adjustments on components inside the vacuum envelope the mercury pumpiscooled down, requiring about 15 minutes, the system is f i l l e d with argon, and the bolts of the electrode covers are removed. Following short exposures1 to atmos-pheric pressure,••"operating-pressure 'conditions can- be restored within two or three-"hours;" A- 750 watt-"glass-Insulated flexible heater• strip i s wrapped around the tube for baking. CHAPTER 3 LEAD ISOTOPE MEASUREMENTS WITH HIGH PRECISION Sample preparation Chemical procedures for the preparation of lead-tetramethyl by the reaction of methylmagnesium bromide with lead halides have-been the subject of extensive-studies', and are'described by Collinsy-Freeman and Wilson (1951), Collins, Farquhar and Russell (1954), and others. Suspicions "about- I'sotopic 'discriminations In" this" procedure were cleared by Ducheylard, Lazard and Roth (1953). The last stage of the chemical techniques supplies the lead tetramethyl—dlssolved i n ether. According to the usual practice, the evaporation of most of the ether is followed by a careful vacuum d i s t i l -lation on t-he-'mass ^ spectrometer sample- handling 5 system. Each d i s t i l l a t i o n step is checked<'by 'the'-mass' "spectrometer f o r - satisfactory leadtetramethyl concentration. This manipulation can be made quite well by a skilled operator but the f i n a l concentration is uncertain and the source conditions are therefore never the same among different analyses. Moreover, contam-inating peaks in the interesting mass range have been occasionally observed. With the object of higher uniformity in the sample preparation, experiments were made on methods for a suitable separation between the ether and lead tetramethyl'; - The presently used technique-is described in detail by T. J. Ulrych in his M.Sc. thesis (Ulrych 1960). 53 54 This method f i r s t uses a miniature fractional d i s t i l l i n g apparatus 't"0-"red,uce--"the'-ori-ginal sample, about 10 cubic centimeters-, to ^ or J cubic" centimeter; highly-concentrated in lead-tetramethyl". This fracti-on'-'i'S' further'-purified by-a- vapour phase chromatographic column. The' ether -and <any?' water "present^are-we'll -separated- from -tie lead tetramethyl-so- that -a reasonably pure^sampl-e is- obtained. --To get about 80 milligrams of clean sample 250 milligrams of lead iodide'is-required because of the 50 per cent -yield of the chemical reactions. In the present-application this is a quantity easily obtained from small pieces of galena specimens. No attempt has been made to reduce the required sample size. The purified, samples from the separating column system are trans-"f e rred •lnt®:-'*a temporary ^container -having "a" gre ased-s t op cock and- a s t and ard tapered joint-; After'-stori^ the stop-cock grease was found to be partially dissolved and mixed with the lead tetramethyl* However, no d i f f i c u l t y was caused i n using these samples and no difference in isotopic composition-was observed when fresh and older samples were compared. To avoid the loss of samples through the deterior-ated -grease^in'-th^ are now stored frozen at the bottoiir-of • 'these1 "containers'.--For storage "for more -than-a-' few -days ^-b seal tubes-are used to avoid grease contamination. Under electron bombardment lead tetramethyl dissociates" into a wide spectrum of products including Pb + ions, PbCH3+ ions, Pb(CH3-)2+ + • + ions, PbfCHg)^ ions and a small proportion of Pb(CHjj}4 ions, a l l of which contain lead and therefore are potentially capable of supplying information about the lead isotope ratios when examined in a mass spect-rometer. Of these, only the Pb + ,.ion group and the Pb(CH3)g+ ion group 55 are simple to interpret. Analysis of the lead ion group is attractive because it-requires the minimum resolution in the mass spectrometer. However, portions of its spectrum are overlapped by the mercury spectrum that usually is present as a •small background, and - often by mercury dimethyl peaks-. Moreover j" there-is superimposed on the- Pb+ spectrum a PbH+ spectrum forming about ten per cent of the whole. The proportion of the latter is critically dependent on the mass spectrometer source conditions and therefore the writer preferred not to use this ion group. Since the present mass spectrometer has sufficient resolution the lead trimethyl-ion group was used for a l l analyses. This group is also complicated by the--presence of multiple spectra. The presenoe of carbon-13 in the'raoleoule'results in a secondary spectrum displaced upwards one mass unit, and the finite probability of loss of a single hydrogen atom from the trimethyl ion group results in another secondary spectrum displaced downwards one mass unit. These secondary spectre amount to approximately 3v3 per cent -and 0.80 per cent of the main spectrum, respectively. The first secondary spectrum is not dependent on mass spectrometer source conditions, and the second, which does depend on souroe conditions, is small enough to cause no difficulties. The inter-pretation of the lead tetramethyl spectrum has been discussed by Dibeler and Mohler (1951) whose interpretation differs in detail from that given above. Measuring techniques After a sample is introduced into the gas handling system of the mass spectrometer its pressure is adjusted to about 10 mm Hg. By monitoring-a "convenient"isotope in the lead trimethyl spectrum the 56 controls adjusting the source electrode voltages, namely the repeller, draw out s l i t , focusing and deflector- electrodes are set to give the • maximum ion-current.--Ther electron -accelerating voltage is kept always on i t s hi-ghes-t---value' (about,-"l-60""vol-ts-)-v.-'The ion beam intensity is much lower with'-smalle^ in other mass spectrometers. This is probably due to the construction of the filament- assembly which uses electrostatic focusing and works -without a source magnet. There is i n i t i a l l y a d r i f t upwards in the ion current for a few minutes. The rise of a few per cent is followed by a steady •slow-decay. Usually when the instrument has been under high -vacuum for a long time -the f i r s t sample shows this transient with a longer duration (10 to 15 minutes). It is found necessary to wait out this transient before starting the measurement, otherwise the varying d r i f t causes a poorer r e l i a b i l i t y of the f i r s t one or two pairs. The scanning rate is adjusted-tO""eompl;ete'the'1 recording' of a pair, scanned down-and up-mass, within 9 or 10-minutes. After a measurement the sample is returned to the sample tube by cooling i t with liquid a i r . This can- be made surprisingly quickly; after about go seconds the detectable ion currents are reduced by a factor of approximately 5000, Experiments about memory showed that the stopcock grease is the only significant source of such effects, by releasing dis<-solved lead tetramethyl. Keeping the sample reservoir closed by the mercury ventil for 10 minutes and checking the background spectrum did not indicate that the mercury would contribute to the memory effect. Realizing that intercomparisons between samples within a short time are possible because of the neglibible memory of the instrument, this procedure was tried. This overcame the slow changes in the instrument 57 causing the "generally observed long term variations in i t s performance. Tntercomparisoii' "patterns '•, o^nsl'st±ng",of-,''ait'e'rnat,i'ng'v"samples"after-"measuring two pairs of'spectra from-each- were later• changed to alternating by threes, giving better proportion of measuring to sample handling time. The- time of measurement in this way is about three to three and a half hours from which two hours is the actual time used to record the twelve pairs of spectra. Further reduction of uncertainities i n the measured differences can be achieved by repeating the intercomparison measurements. An economical way of obtaining smaller errors is to arrange measuring patterns in loops containing three intercomparisons of two samples and a standard. The analyses shown in this thesis were made in this way. Each loop contains three intercomparisons; f i r s t sample A is measured with our standard sample, then B, and f i n a l l y the comparison of A with B closes the loop. In this loop, sum of the differences should be zero in the ideal case, here the closure errors are distributed evenly around the loop for each isotope ratio to obtain i t s most probable value. I n i t i a l measurements showed systematic changes in the abundances correlated with pressure. The dependence of t a i l i n g on pressure at the two sides of the three intense ion beams causing a rise i n the base line of the smaller peaks at mass 250 and 254 was apparent. Experiments with different pressures of the mass spectrometer tube i n the IO"** •mm Hg range gave the experimental points on Figure 16. They were obtained from the records by measuring the base line rise under peaks recorded on the two sides of the group of three abundant masses, assuming that the t a i l is significant only at one and two mass distance from a peak. According to the analyser pressure during the measurement, t a i l correction coefficients from this graph were used to correct the peak at mass 252 for t a i l s from i t s adjacent peaks, and the mass 2^ 1 peak for t a i l s from 58 mass 252 and mass 253 peaks. The peak at mass 253 is corrected similarly for scatter- from^masses 252 and 251 - The base line for these peaks is obtained by--measuring i t ^ corresponding'sensitivities of the ion -current 'amplifier. Toimprove the precision of the lead-204 measurement, which comes from the peak height at mass 249, i t is measured from the base line at the two sides of this peak and is corrected for i t s own t a i l using twice the t a i l i n g coefficient applied to adjacent masses. The ion current at masses 250 and 254 which are used only for small corrections i n the reduction of a spectrogram are determined directly from the record. As an example, suppose that the t a i l from a peak i s found to contribute a fraction x of the peak at the adjacent mass ( f i r s t t a i l i n g values in Figure 16) and a fraction • y 'two masses away (second t a i l i n g value). Suppose further that the measured ion intensities at masses 249, 250 .....254 are Aj_, Ag..-. .-.Ag. Then the tail-corrected values used are ^ f 2xAi A 3 - -.yAg A 4 - xAg - xAg Ag - xA 4 - yAg A6 It is the opinion of the writer that this means"of correction is approxi-mately equivalent to making measurements at zero pressure in the analyser tube. The reduction, of the trimethyl group of the lead tetramethyl spectrum requires that six unknown quantities be evaluated. These are X. 59 FIG.I6 PRESSURE SCATTERING CORRECTION 60 the four lead isotope abundances and the proportions of each of the displaced secondary spectra. To obtain this information there are known the proportions of the six ion currents at masse^ 249, 250,..,,254. Since the secondary spectra -are small, and since ^he equations-are" non-linear, i t Is 1 neeessary to solve them by a suitable iterative procedure. Since these calculations cannot be described briefly, and because they -are -now well established, ^ hey w i l l not be detailed here. They are discussed by Ulrych (I960). The computation work, which is tedious, was made with the help of the Alwac III di g i t a l cqmputer of The University of British Columbia. To avoid small discrepancies, some original, hand calculated analyses were a l l recomputed. The computer-program-was written by J. L. Allard. Table 4 shows a data tape printed. Each pair of six measurements is proceeded by a dollar sign. The f i r s t two rows are the servo-voltmeter dial readings in units of 0.1 division from mass 249 to 254. The third row contains the multipliers (x 1000) corresponding to the pushbutton-controlled sensitivity ranges. The correction peaks with masses 250 and* 254 are measured from the • spectrogram in" units of Q-. 1 mm. Their -multipiier in the third row is determined from any peak as -the, ratio of dial reading to corresponding chart deflection. The fourth row has the t a i l correction coefficients- for one -and two mass distances. Theij: values are obtained from the ta-i41ng'-i-correct-ion---ii'iagram'-aecordi-ng--to ttxe pressure. Double dollar signs after the last set indicate that no more data w i l l follow. The last row gives the approximate hydrogen loss and carbon-13 coefficients for the f i r s t step of the iterative calculation. A special feature of the program is that anything typed immediately preceding any dollar sign w i l l be retyped at the corresponding place on the result sheet. The heading 6< Table 1+. Printed version of data tape. The input to the ALWAC computer is from this tape. No. 1 Standard compare with No. 59 Kollar March 31/60 $ 9982 260 16108 16258 12175 1896 9952 260 16032 I6165 12085 1871 1000 61+23 10101 10101 30101 61+23 0.0007 0.00025 $ 9799 255 1583!+ 15990 11979 1850 971+9 252 15699 15821+ 11832 18I+3 1000 6I+2 3 10101 10101 30101 61+2 3 0.0007 0.00025 $ 10255 267 16579 16733 12533 19^8 10262 262 165^9 16681+ 121+80 1936 1000 61+23 10101 10101 30101 61+23 0.0007 0.00025 $ 10160 260 161+26 16586 121+22 1930 10157 260 16371+ 16508 12352 1920 1000 61+23 10101 10101 30101 61+23 0.0007 0.00025 10079 255 16275 161+29 12301+ 1911+ 10067 258 1623!+ 16363 1221+2 1893 1000 61+23 10101 10101 30101 61+23 0.0007 0.00025 $$ 0.008 0.033 Table 5» Printed version of the output tape from the ALWAC computer showing the calculated isotope ratios. No. 1 Standard compare with No. 59 Kollar March 3l/6o 1.3975* .2338* 22.6985* 22.8758* 51.1006* 1.6938* 100.0000 1.3965* .2323* 22.69^3* 22.8739* 51.1107* 1.6922* 100.0000 1.3953* .2308* 22.6966* 22.8720* 51.1111* 1.6942* 100.0000 1.3952* .2290* 22.6912* 22.8720* 51.1168* 1.6958* 100.0000 1.3962* r: .2280* 22.6974* 22.8724* 51.1136* 1.6923* 100.0000 5 • 1.3961* .2308* 22.6956* 22.8732* 51.1106* 1.6937* 100.0000 .0009* .0023* .0029* .0016* .0061* .0015* Std Devs j k dn k up 204 206 207 208 5. .00800 .03300 .00816 .03362 l.1+51+3* 23.^566* 22.6091* 52.^799$ 100.0000 1.0000 16.1289 15.5462 36.0856 prog 3 6.2000 100.0000 96.3872 223-7322 63 of the sheet- is an example of this routine. The printed results are shown in Table 5. The heading is followed by-r-t'he--t,'ali-«correct-ed'-l'0-n.,-current percentages adding up to 100. The average values of the percentages are printed in- the red row, • followed by-the- standard deviation ( ). Under the preliminary-hydrogen loss and carbon-13 coefficients their f i n a l computed values are printed'-'-in---'redv"--:'-Th-i8 i s followed by the reduced percentage lead abundances;•• and-underneath the isotope ratios relative to lead-204 and to lead-206, respectively. Analytical results The f i r s t measurements with the new technique were*made on -sampies •••t,tot^,lw4"previwfflly-'^'ett--8t udi'ed •-,at~-*The"'DhlveTstty- -of-Toronto-and elsewhere^ in-order that "there- could be a reasonable basis for eval-uating the-results-. - A suite of galenas from the mining d i s t r i c t at Broken H i l l y New South-Wales, Australia, was sent directly by Dr. H. F. King, Chief Mine'-Geologist. •••• Although-'more-than-•'thirty-samples- were received, .-only••••six wre^ttsed"''!^ One-sample from- the Black Star mine at Mount Isa, Queensland, Australia, was supplied by Dr R» Thompson of this university, and four more were donated by Dr. R. M. Farquhar of The University of Toronto. The latter samples were identical with those previously studied at Toronto. At Broken H i l l the lead extends over an area of hundreds of square miles. At the main lode near the city of Broken H i l l the lead is found principally in three lodes numbered in order from the surface, but also in a rhodonite zinc lode near the surface and interspersed between the lodes. The lead forms an almost perfectly conformable layer, but some 64 evidence of replacement is seen under the microscope,'•• Figure-17 illustrates a typical section across the main lode. The conformable lead layer l i e s beneath a marker- horizon known as the Hanging Wall Granite, which is a highly- grami-tized-' gneiss ^-"••"•Th±S'--mark6r'<'hori-zon',";,*wh±!ch,"-can-be identified throughout the entire area,- has been highly folded, presumeably after mineralization (Figure 18). Immediately below i t at various places i n the area where i t reaches the surface are lead sulphide deposits similar mineralogically and structurally to1 the main lode and which are, therefore, referred to as Broken H i l l type. Low lying'fractures in folded sediments -younger than-*the'Hanging Wall Granite also contain lead deposits, but with d i s t i n c t l y different gangue minerals. These are the Thackaringatype deposits which have been shown to contain anomalous lead-. Discussions of the lead isotope ratios for various'Broken'Hill samples-are given bys Russell, -FaTojuhar >and'"Hawley" (-'1957) - and by Stanton and Russell- (1959). These resu!1w-'';are'-'"ei!umnaTi'zed'---ifl>'TablieB'-;-6-'and 7. -King has proposed a classification of conformable leads of which the Broken H i l l type are typi c a l . The Mount-Isa leads are also supposed to belong to this c l a s s i -fication. The extent of-the Mount Isa deposits, which are shown in cross section in Figure 19, is much smaller than Broken H i l l . Stanton and Russell suggest that such conformable leads are least l i k e l y to be anomalous and show that according to the Toronto analyses these leads f i t a common growth curve within a few tenths of one per cent and therefore could have been produced in very similar environments. The present analyses"were carried 1 out by comparison with a standard sample. This is the University of British Columbia'No, 1, which is a- s p l i t 'from-Toronto sample No, 1003. The long range s t a b i l i t y of the mass spectrometer can be estimated by comparing the results of the 65 Fig. 17 - Representative cross section across the main lode-at Broken H i l l . (After Russell, Farquhar and Hawley) 66 • Hanging Wall Granite "Broken H i l l lode Thackafihga-type depoiits X Conformable Broken H i l l : type deposits Fig. 18 Idealized cross section of the Broken H i l l mining area showing . general structural relationships between Broken-Hill-type and Thackaringa-type mines. (After Russell, Farquhar and Hawley) 67 Fig. 19 Section at Black Star Mine, Mount Isa, showing the structure of the silver-lead lode and i t s relation to the copper orebody. (After Blanchard and Hall) 68 TABLE 6 Leads from Broken H i l l , New South Wales (After Stanton and Russell) Sample 100 times ratio of •No.. Location 204/206 207/206 208/206 667 No. 3 lens 6.191 96.64 222.8 668 Between No.2 and No.3 lenses 6.186 96.56 222.4 669 No. 2 lens 6.233 96.56 222.4 670 Rhodonitic zinc lode 6.225 96.43 222.0 671 No. 1 lens 6.208 96.64 222.8 672 Siliceous zinc lode 6.198 96.68 222.9 912 L i t t l e Broken H i l l Mine 6.177 96.14 221.8 913 Pinnacles Mine 6.179 96.26 222.0 914 Rupee Mine 6.184 96.39 221.9 915 Centennial Mine 6.184 96.30 222.6 916 Parnell Mine 6.154 96.56 221.9 917 Balaclava Mine 6.188 96.73' 222.-2 918 Rising Sun Mine 6.192 96.73 222.1 , 919 Great Western Mine 6.213 96.55 223.8 920 Globe Mine 6.183 96.73 222.0 921 White Leads Mine 6.211 96.23 220.5 922 Nine Mile Mine 6.179 96.50 221.1 j 69 TABLE 7 Leads from Mount Isa, Queensland (After Stanton and Russell) Sample 100 times ratio of No. Location 204/206 207/206 208/206 928 Level No. 11 , No.2 orebody 6.206 96.08 223.0 929 Level No. 12 , No.5 orebody 6.207 96.29 224.2 930 Level No.9, No. 7 orebody 6.204 96.45 224.4 931 Level No. 9, No. 11 orebody 6.185 96.34 222.3 932 Level No. 9 6.215 96.53 224.9 933 Level No.l, Sub above 7 6.226 96.50 224.7 70 TABLE 8 Replicate analyses of sample No. 1 Date of Analysis 204/206 100 x ratio 207/206 of 208/206 February 17 February 19 February 24 February 25 March 9 March 10 March 14 March 16 March 31 6.198 6.197 6.203 6.205 6.212 6.213 6.205 6.208 6.200 96.60 96.58 96.40 96.53 96.43 96.42 96.49 96.33 96.39 223.77 224.13 223.50 223.90 223.72 223.75 223.96 223.71 223.73 Mean of A l l Analyses with 6.205 Standard Deviations +.006 Mean of Last Five Analyses 6.208 with Standard Deviations ±.005 96.44 ±.09 96.42 ± .06 223.80 ±0.18 223.77 ±0.08 * Samples indicated were measured from the recorded chart 71 various''analyses' of this sample carried out • Over a-' period--of-six weeks These are tabulated in Table 8. The f i r s t four analyses in this table were carried out by measuring the peak heights from a record obtained from a chart-recorder; - The last five-were made- by noting readings on the'servo-vol^tmeter'-dial.- In^^•all-cases'the'reproducibility is better with the servo^voltmeter; ; but'>-this!-is-"pa-rtl-cul-arly-apparent- in the lead-208/leadr-206 ratio where the improvement is about a factor of 2.5. The improvement in the lead-204/lead-206 ratio is not great because noise from the 1011 ohm resistor and' from the electrometer amplifier is the limiting factor. For a l l the isotope ratios the standard deviation of an analysis i s less than ±0.1 per cent, and this is- indicative of' a very stable-mass spectrometer. Since a l l Mount Isa"and Broken H i l l samples were analysed twice, a reproducibility with a standard deviation of 0.07 per "-cent" would-be-expected- even-"-if the comparison'technique had not been used. The-averages of the values in Table 8 were used for the value of the standard when the measured differences were plotted for comparison with the Toronto results. Table 9 shows the deviations measured between the Broken H i l l samples and the standard. These were a l l obtained by measuring the spectra displayed on the Brown chart recorder. The ratios lead-207/ lead-204 and lead-206/lead-204 are plotted in Figure 20, which, shows a distribution of points just ten times smaller (in linear dimensions) than had been'shown by previous measurements. For the Mount Isa samples (Table 10)> the -same ratios-show a significant- spread, but--the spread is s t i l l smaller than had been indicated by the Toronto measurements 4 F i g u r e 21). In the corresponding lead-208/lead-204 versus lead-206/lead-204 plots (Figures 22 and 23) there are significant differences in the case of both 72 Mount Isa -and-Broken'-Hill••;-"-'-Howeverv""the'"spre,acts"varB"-,agat'n'substantially less than -had'^ een-'-p^  from locations over 800 miles apart, give average values that f i t a mean lead-207/lead-204 - lead«206/;lead-204 growth curve within 0.2 per cent of the lead-207/lead-204 ratio and a mean lead-208/lead-204 - lead-206/ lead-204 'growth-•'curve:--withlnua,Tsimilar"percentage of the lead-208/lead-204 ratio. This corresponds to a difference in the uranium/lead ratio of 0.7 per cent and in the thorium/uranium ratio of 0.5 per cent for the two deposits. This can be considered to show that the requirements of an exceptional regularity i n isotopic constitutions of the conformable leads are satisfied much better than could be shown by previous analyses. The solid curve i n each of Figures 20, 21, 22, and 23 shows a portion of the mean growth curves for conformable leads as deduced by Russell and Farquhar(in press). The present results would require modi-fications of these curves of not more than one or two tenths of a per cent, well within the experimental error of the curves themselves. Perhaps the most interesting part is the fine structure in the isotope ratios which- has just become apparent with the increased precision. This fine structure had not been anticipated by the writer. Although one has to be cautious i n interpreting the measured ratios until the techniques have been verified by s t i l l many more analyses, there seems to be'no question but that the leads at Broken H i l l and Mount Isa vary in isotopic composition. In the case of Broken H i l l , the variation shows up principally in the abundance of lead-208, which seems especially significant in view of the repeated measurements on the standard sample which showed the 7 3 lead-208/lead-206 ratio to be the most reproducible. Russell, Farquhar and Hawleyhave shown that the thorium/uranium ratio of the rocks con-tributing radi*0'geni'c*-lead',i;o make-anomaloiis- leads is- proportional to the slope of the--^ bes1;''-stTai'ght'---lltte''vt-hrough''--the'"poi-nts"-on the lead-208/lead-204 -versus lea^2t36/lead-'204" ^ of these writers, -it appears that the conformable-leads at Broken H i l l have received variable amounts of radiogenic lead contamination from crustal sources, and the the thorium/uranium ratio i n these sources was about three times the crustal average. Since the crustal average is usually taken to be about 3 . 7 atoms thorium per atom of uranium, this suggests souree rocks which would now have 11 atoms of thorium per atom of uranium. fh-e structures at Broken H i l l strike north-east to south-west. Samples occuring along a l i n e parallel to the structures through the main lode, are :lsast"ico-ntamimted^ on-either side of this l i n e a r e more contaminated. A similar trend is also apparent from the analyses of Farquhar and Russell, although i t is largely masked by the greater uncertainties in the measurements. The-i?esults for the Mount Isa-samples show qualitatively similar effects. "Iii^tMsi;ease":th'e 'contamination-from- thorium lead i s similar to that found at Broken Hill,- -but- the contamination from uranium lead is larger. It is impossible to-make a good estimate of the thorium/uranium ratio in the^so-uree of "contaminating lead, but i t seems to be about- twice -the crustal-average, -fhe- writer does not have as complete geological descriptions -for'Mount''Isa-4as f or--- Broken Hill.-The- leads occur in a shear zone i n highly folded shales. Although there seems to have been some controversy about this matter, the general opinion seems to favour 74 the idea that the mineralization is subsequent to the folding (Blanchard and Hall, 1937). Blanchard and Hall also take issue'with the contention • of other writers that the lead at Mount Isa is syngenetic. Stanton, in the papers-'cited,-has suggested that these leads, as well as those at Broken Billv'-were beds off active island arcs and replaced iron in iron sulphide lenses. Leads from both l o c a l i t i e s have apparently been derived from source rocks having very similar proportions of uranium, thorium and lead. This makes i t possible to determine' a more reliable difference in age between the deposits. Comparing the least radiogenic compositions found at each location, results i n a n estimated difference*in age of 40 million years, Mount-Isa being -younger than- Broken Hill.-Approximately the same age difference is obtained either through the procedures of Farquhar and Russell or of Houtermans. 75 TABLE 9 Differences in isotopic composition of Broken H i l l samples Sample 100 times ratio of No. 204/206 207/206 208/206 206/204 207/204 208/204 1 0 0 0 0 0 0 30 -.002 -.03 +.006 +.001 +.03. 4 31 +.002 +.06 -\ -.005 -.016 + .06;,^  34 +.003 +.02 -.2 2 -.007 -.004 -,050 44 -.006 -.05 +.015 +.006 +<057. 49 +.002 -.04 - X 7 -.004 -.011 r.03 ? Location of samples 1 Main lode, No. 2 lens, No. 15 level 30 Centennial Mine 31 Pinnacles Mine 34 Main lode, No. 2 lens, No. 17 level 44 L i t t l e Broken H i l l Mine 49 Globe Mine 76 TABLE 10 Differences i n isotopic composition of Mount Isa samples Sample 100 times ratio Of No. 204/206 207/206 208/206 206/204 207/204 208/204 1 0 0 0 0 0 0 59 -.029 -.21 -.3 2 + .074 + .037 60 -.050 -.35 - 4 o + .129 + .069 + .22g 61 -.040 -.32 - 3 4 + .104 + .048 + .17? 62 -.043 -.22 + .116 + .072 + .173 63 -.040 -.21 ~.2g + .104 + .061 + - 1 8 5 Location of samples 1 Broken H i l l reference sample 59 Black Star orebody 60 See No. 928 sample i n TABLE 7 61 See No. 929 sample in TABLE 7 62 See No. 930 sample in TABLE 7 63 See No. 932 sample i n TABLE 7 .20 U n i v e r s i t y of B r i t i s h Columbia analyses .21 U n i v e r s i t y of Toronto analyses 7 9 3 6 3 Pig,22 U n i v e r s i t y of B r i t i s h Columbia analyses CONCLUSIONS Chapters 2 and 3 have described measuring techniques for precise intercomparison of lead isotope ratios, as well as some typical -measurement's;-''- •Measurementswhave been* obtained'that are substantially more precie-e"-thaii-','ted''';previous']:'y'-'been,''possl'ble'.-'''The following f i r e points-are believed to be responsible for the major part of the increased precision, although careful attention to every small detail during the measurements has also been essential-. (1) The filament-emission control has a s t a b i l i t y great enough to obtain sufficiently steady ion currents (typically a few parts in 10,000). ( i i ) The measuring system has a high inherent precision (0.025 per cent/full scale) and the procedure used, reading a servo-driven d i a l , makes i t possible to realize this precision without the added errors of a chart recorder, ( i i i ) The purification of the sample i n the vapour phase chromato-graphic column makes the possibility of contamination in the interesting mass range highly unlikely. This is important because such contamination hap occasionally been observed with conventional techniques, (iv) The greater purity of the samples has made i t possible to work with more reproducible source conditions. Even small variations in source conditions can affect the measured ratios. Moreover, the increased purity results in a lower pressure i n the analyser tube and this i n turn reduces the proportion of scattered ions.-Although the pressure t a i l i n g i s included as a correction i n the calculations i t i s clearly desirable to keep this correction small. 81 82 (v) The direct intercomparison of samples requires that" the-mass spectrometer be stable only for periods of about-thirty minutes. The d r i f t of the instrument can be kept much smaller for such short periods of time. Further developments in a l l of these aspects are considered possible and work is continuing at this laboratory on further improvements in the precision of lead isotope ratio measurements. The actual results obtained for Broken H i l l and Mount Isa I were somewhat unexpected. The fact that the conformable leads at both lo c a l i t i e s vary very l i t t l e in isotopic composition had been expected, and this fact was confirmed. However, an interesting fine structure in the results has been found, which suggests that measurements of this type w i l l have an important future. Obviously interpretations of the- results at-this- early -stage -must be considered tentative. However, the following points seem to be indicated. ( i ) T h e Broken H i l l and Mount Isa deposits contain lead of dis-t i n c t l y different isotopic composition, ( i i ) In both l o c a l i t i e s the isotope ratios are less variable than could be Inferred from previous measurements, ( i i i ) Both deposits have isotope ratios f i t t i n g their average lead-thorium-uranium growth curves within about 0.1 per cent of their values. (iv) At both l o c a l i t i e s the conformable leads have been contaminated by small amounts of radiogenic lead, probably during the pro-cesses preceding mineralization. At Broken H i l l the contamin-ation is from a source with a thorium/uranium ratio about three 83 times the crustal average, while at Mount Isa the contamina-tion came "from - a' •source' which, although s t i l l rich in'thorium, is-much nearer • "the-crustal average." - The radiogenic - 1ead contamination is small-, amounting to less than one-half per cent of the lead, (v) The Mount lea leads are younger than the Broken H i l l conform-able leads;• • If-the hypothesis of-surface contamination as --described in* "fiv}" 'is •'•accepted, •-•••the "difference in age-does not exceed 40 million years. The average age of the two deposits according to Russell and Farquhar (in press) is about 1600 million years. Therefore, the present measurements are capable of resolving an age difference of 8.5 per cent in this case. Further speculations-about these results are possible. For example, Stanton (1955) has suggested that this type of deposit originated off volcanic island arcs. The close similarity in age between the Broken H i l l and Mount Isa leads might be interpreted as indicating the existence about 1600 million years ago of a chain of volcanic islands approximately parallel with the present eastern coast of Australia. Much work remains to be done before speculations of this type can be replaced by a rigor-ous interpretation. 84 BIBLIOGRAPHY Adamson, C.B., The equilibrium process of chemical exchange leading to isotopic fractionation of lead in nature, B.Sc. Thesis, University of Toronto, 1959. Alpher, R.A. and R.C. Herman, The primeval lead isotopic abundances and the age of the earth's crust, Phys. Rev. 84, 1111-1114, 1951. Blanchard,R. and G. Hall, Mount Isa ore-deposition, Econ. Geol. 53, 1042, 1937. Bui lard, E.C. and J.Pi- Stanley,- The age of the earth, Publication of the Finnish Geodetic Institute 36_, 33-40, 1949. Burling, R.L., The determination of geological time, Nucleonics 10, 30-35, 1952. Collins, C.B., R.M.; Farquhar and R.D. Russell, Isotopic constitution of radiogenic leads and the measurement of geological time, Bull. Geol. Soc. Amer. 65_, 1-22, 1954. Collins, C.B., J.R. Freeman and J.T. Wilson, A modification of the isotopic lead method for determination of geological ages, Phys. Rev. 82, 966-967,- 1951. Collins, C.B., R.D. Russell and R.M. Farquhar, The maximum age of the elements and the age of the earth's crust, Can. J". Phys. 31_, 402-418, 1953. Dibeler, V,H., and F.L. Mohlerj —Mass spectra of- some organo-lead and organo-mercury compounds, Jour. Res. Natl. Bur. Standards, Washington, 47_, 337-342, 1951. Ducheylard, G., B. Lazard and E. Roth, Idotopic analysis of lead with the aid of a mass spectrometer. I Preparation of tetra-methyl-lead. J. Chim. Phys. 50, 497-500, 1953. Duckworth, H.E., Mass spectroscopy, Cambridge University Press, 206 pp., 1958. Farquhar, R.M., and R.D. Russell, Anomalous leads from the upper Great Lakes region of Ontario, Trans. Amer. Geophys. Union,38, 552-556, 1957. Gerling, E.K., Age of the earth according to radioactivity data, CR. Acad. Sc. U.S.S.R. 34, 259-261, 1942. Gorman, J.G., E.J. Jones and J.A. Hippie, Analysis of solids with the mass spectrometer, Anal. Chem. 23, 438-440, 1951. 85 Halsted, R.E. and A.O. Nier, Gas flow through the mass spectrometer viscous leak, Rev. Sci. Instr. 21, 1019-1021, 1950. Holmes, A., An estimate of the age of the earth, Nature 157, 680-684, 1946. Holmes, A„'y A revised estimate of the age of-the earth, Nature 159y 127-128, 1947. Holmes, A., Lead isotopes and the age of the earth, Nature 163, 453-456, 1949. Houtermans, F.G., The isotope frequency in natural lead and the age of uranium, Naturwiss, 22., 185-186, (addendum: ibid., 219), 1946. Houtermans, F.G., Time of the formation of uranium, Zeit. Naturforsch'. 2a., 322-328, 1947. Inghram, M.G., Mahlia^^ , National nuclear-energy series, division II, Vol. 14, Chap. 5, 35, McGraw-Hill Co., 1946. Inghram, M.G. and R.G. Hayden, A Handbook on mass spectroscopy, Nat. -Acad. Sci. (U.S.) Nat. Res. Council Nuclear Sci. Ser. Report No. 14, (1954). King, H.F. and B.P. Thompson, Geology of Australian ore-deposits, Chap. 9, The geology of the Broken H i l l d i s t r i c t s , 5th Empire, Min. and Met. Cong. I, 533-577, 1953. ' McKinney, C.R.,. J.M. McCrea, S. Epstein, H.A. Allen and H.C. Urey, Improvements in mass spectrometers for the measurements of small differences of isotope abundance ratios. Rev. Sci. Instr. 21., 671, 1950. Nier, A.O., Variations in the relative abundances of the isotopes of common lead from various sources, J. Amer. Chem. Soc. 60, 1571-1576, 1938. Nier, A.O.^ A redetermination of the relative abundances of the isotopes of C, N, 0, A and K, Phys. Rev. 77, 789-793, 1950. Nier, A.O., E.P. Ney and M.G. Inghram, A null method for the comparison of two ion currents in a mass spectrometer. Rev.'-Sei. Instr. ig.,-101-103, 1947. • Nier,, A.0., R.W. Thompson and B.F. Murphey, The isotopic consti-tution of lead and the measurement of geological time III, Phys. Rev. 60, 112-116, 1941. Patterson, C.C., The isotopic composition of meteoritic, basaltic and oceanie leads and the age of the earth, Proc. First Conf. on Nuclear Processes in Geological Settings, 36-40, 1953. 86 Patterson, C.C., Age of meteorites and the earth, Geochim. et Cosmochim. Acta IP", 230, 1956. -Patterson, C.C., 0. Tilton and M.G. Inghram, Age of the earth, Science 121, 69-75, 1955. Riezler, W. and G. Kauw,"" Natural radioactivity of lead-204 and the question of natural activity of dysprosium-156, Z. Naturforach". 13a,~ 904-905, 1958'. Russell, R.D. and D.W. Allan, The age of the earth from lead isotope abundances, Roy. Astron. Soc, Geophys. Supp. 7_, 80-101, 1955. Russell, R.D. and R.M. Farquhar, Dating galenas by means of their isotopic constitutions II, Geochim et Cosmochim. Acta, in press. Russell, R.D7,' R.M. 'Farquhar-, G.L. Cumming and J.T. Wilson, Dating galenas by means of their isotopic constitutions, Trans. Amer. Geophys. Union 35, 301-309, 1954. Russell, R.D., R.M. Farquhar and J.E. Hawley, Isotopic analyses of leads- from Broken H i l l , Australia, with spectrographs analyses, Trans. Amer^ Geophys.Union 38, 557-565,"1957. Russell, R.D. and F. Kollar, Transistorized power supplies for a mass spectrometer, Canadian Journal of Physics, in press. Senftle, F.E. and J.T. Bracken, Theoretical effect of diffusion on isotopic abundance ratios in rocks and associated fluids, Geochim. et Cosmochim* Acta. 7_, 61-76, 1954. Shea, R.F., (ed.) Transistor circuit engineering, John Wiley and Sons Inc. 468 pp. (especially page 130) 1957. Stanton, R.L., The genetic relationship between limestone, volcanic rocks and certain-ore deposits, Austr. J. Sci. 17_, 173-175, 1955. Stanton, R.L., Abundances of copper, zinc and lead in some sulphide deposits, Jour. Geol.', 1958. Stanton, R.L. and R.D. Russell, Anomalous leads and the emplacement • • of lead sulfide ores, Econ. Geology 54, 588-607, 1959. Starik, I.E., M.M. Shats and E.V. Sobotovitch, On the age of meteor-ites, Dokl. Akad, Nauk SSSR, 123r 424-426, 1958. Stevens, CM. and M.G. Inghram, Ratio recording in isotopic,analysis, Rev. S c i . Instr. 24, 987-989, 1953. Strauss, H.A., A new mass spectrograph and the isotopic constitution of nickel, Phys. Rev. 58, 942-948, 1941I 87 Ulrych, T.J., The preparation of lead tetramethyl for mass spectro-meter analysis, M.Sc. Thesis, Department of Physics, University of" British Columbiay 1960. Waldron, J.D., (ed.) Advances in mass spectrometry, Pergamon Press, 704 pp., 1959. Whittles, A;B.E.-, ; Voltage -coefficient of Victoreen &igh-meg resistorsv vRev."Sci.' Instr. 31_, ;208-209, i960. Wilson, J.T., R.D. Russell and R.M. Farquhar, Radioactivity and age of minerals, Handbuch der Physik, 47_, 288-363, 1956. 88 APPENDICES Details of the tube and magnet (Figures 24, 25 and 26) The tube was constructed of copper, joined' by common plumbing fittings (wrought copper, not cast). The flanges were machined from No.. 303 stainless steel. These parts were silver-soldered together, after which the tube was cleaned as i f i t were to be electroplated'. The electrode covers were fabricated from stainless steel tubing of four inch diameter, and were argon arc welded. The kovar seals were argon arc welded around the circumferences of the end flanges as shown in Figure 25 A, by using a 0.040 inch tungsten welding electrode. A l l demountable -joints use aluminum gaskets f i t t e d as shown in Figure 25 B, with the exception of the main joint at the diffusion pump Which uses soft solder held by an aluminum retaining ring. The' magnet was constructed from a mild steel ("boiler plate") yoke with armature and pole pieces from Armco iron. The area of each pole piece ! i ' • 2 is 800 cm and the distance between them is 0.7500 ±0.0005 inches. The two magnet pole faces are mounted together in a rigi d unit, adjustable in two directions by fixtures on the support. This provides the possibility of adjusting for optimum focal conditions without moving either the magnet or the tube. 89 [— 14" H -• Fig. 24 Outlines of analyser tube and i t s relative position to magnet vacuum side Fig. 25 Details 1 of tube construction A Setting of' kovar leads for argon arc welding B Gasket arrangement between stainless steel flanges 90 Fig. 26 Approximate impedance characteristics of the magnet measured at i t s terminals. 91 The reluctance of the magnetic circuit is substantially that of the a i r gap and therefore the magnetizing curve is linear except at very high f i e l d s . -'The magnet coils contain a total of 21,000 turns of No. 18l/2•-A.-W.Gi copper wire having a total resistance, when connected in series, of 720 ohms. The-inductance of the magnet calculated from the above figures is 2100 Henry. Howeverj the effect of eddy currents becomes significant at frequencies higher than about 0.2. cycles per second. The impedance of the magnet is shown as a function of »fireo;uen«y"-i-a->Flgure--'26-.•••••The-data for this figure were obtained at M'gher frequencies by determining the amplitude and phase of the current produced by a small sinusoidal driving voltage, and at lower frequencies by analysing the response due to a voltage step. The characteristic applies at zero D.C. current and can be expected to be different in detail at high f i e l d s . Magnet current »power'supply -(Figure 27) -The circuit incorporates two regulating devices; the saturable reactor which is a slow device,1 and the transistor circuit which has a fast response. One important function of the saturable reactor c i r -cuit i s to keep t"hei,--voltage',a,cross-"-t,he-,-cont'r61:-"transis,tor at a suitable level. However; i t also contributes -significantly to the regulating action. Thus the regulation is seriously impaired i f the saturable reactor is removed and the power transformer supplied a suitable fixed" input voltage by means of a variable voltage transformer (e.g. a Variac). The supply operates as follows. The mains voltage is fed through an auto transformer to the primary of the power transformer connected in series with the saturable reactor. The auto transformer 92 F i g . 27 Magnet current supply Condenser values are given i n microfarads, r e s i s t o r s i n ohms, Zener diodes are i d e n t i f i e d by t h e i r breakdown voltages. Transformer and r e a c t o r numbers are Hammond type numbers„ 93 adjusts the supply voltage close to the"-optimum value for the different current ranges employed. Since the saturable reactor can i t s e l f pro-duce changes in the transformer voltage by approximately a factor of ten, the auto transformer could probably be omitted, but the regulation would be somewhat impaired, particularly at very high and very low currents. A f u l l wave r e c t i f i e r is employed with a condenser input f i l t e r . The f i l t e r i n g circuit used is in the same servo loop as the saturable reactor which has a long time' constant resulting from the large inductance of the D.C. winding. A conventional LC f i l t e r introduces additional phase lags and causes a deterioration in the transient characteristics of the overall system. Thus the tuned 120 cycle rejection circuit shown is used to reduce ripple. The remaining ripple, the order of one volt, is easily handled by the transistor regulator- c i r c u i t . The output from the fi l t e r e d power supply is connected to the magnet windings in series with the 2M57 control transistor 'and-' a-•series ^ resistor, the voltage drop across •• which resistor serves as a measure of the magnet current. The sum of the voltages across the control resistor and transistor regulates the bias of the triode operated -6AQ5 vacuum tubes controlling the •saturable reactor current. This serves to hold the voltage drop across the control transistor to a reasonable value (about 30 volts) at a l l magnet- current ranges, and-well -below the 60 volt maximum rating of the transistor. The control transistor is rated at a current of 5 amps maximum and a power dissipation the order of 50 watts. Therefore, the maximum magnet current that can be delivered by the supply is here" limited to 500 mililamperes only by the- trans-former, choke and saturable reactor ratings. 94 Two 2N369 transistors are used as a difference amplifier, comparing a fraction of the voltage developed across the series control resistor with the 5.36 volts produced by four mercury cells which serve as the basic reference. This reference voltage is high enough so that'pickup and d r i f t in the difference-amplifier inputs can easily be made"negligible. -The-writer feels that the use in this circuit of a very low reference voltage (5Gmv)^  as "Suggested by Garwin, would not only exaggerate problems of noise but sacrifice loop gain needlessly and therefore require additional gain in the electronic circuits increasing the problems of s t a b i l i t y . In this circuit the-mercury cells, are-quite stable and have a long l i f e , and therefore, seem to-be a convenient reference; i f the use of batteries were con-sidered undesirable, a Zener diode would provide an excellent reference. A third 2N569 isolates the power transistor from the difference amplifier'. This transistor is essentially an emitter follower, -matching the A.C. and D.C. impedances of the difference amplifier to 'the power transistor. -'The 10 volt Zener diode stabilizes the collector supply voltage for onetransistor i n the difference amplifier otherwise the gain would be greatly reducedby negative feedback. The difference amplifier is designed so that the collector , voltages are almost identical at a l l operatingcurrents. The 20 volt Zener diode adjusts the operating point of - the 6AQ5, and thereby adjusts the range of voltages across the control transistor. The 22k resistor provides some positive feedback, thereby increasing the loop gain. * Gerwin, R.L., 1958, Efficient precision current regulator for low-voltage magnets, Rev. Sci. Instrum. 29, 225. Garwin, R.L., Hutchinson, D., Penman, S., and Shapiro, G., 1959. Efficient precision current regulator for high-power magnets. 95 The 54 volt Zener diode protects the circuit against switch-ing or other transients that might result in- exceeding the voltage rating of the 2N457. The M-500 Sarkes Tarzian diodes protect the magnet windings from a voltage transient when the supply is shut off.. Many diodes- would be equally- suitable, but- -in-any-ease-best-results-can be obtained by measuring the reverse characteristics and selecting diodes of low reverse current. The characteristics of these s i l i c o n diodes were quite-variable. In the construction of the unit, care must be taken to eliminate A.-C. -pick-up in the inputs to the difference amplifier. In particular i t was found worthwhile to use b i f i l a r would control resistors. The1 transistor amplifier was built on a heat sink made from 3" x 2" x 1/8" aluminum channel, four inches long which plugs into an octal socket. The performance of the remainder of the circuit can be easily checked by unplugging this unit and replacing i t by a variable power resistor of 50-30G ohms resistance. Thus the saturable reactor circuit can be checked independently of the transistor c i r c u i t . ^Similarly replacing-;:'ttie s6AQ5' tribes by a suitable resistors-permits- the transistor unit to ^ e tested' independently of the saturable reactor control loop. The A.C. power for the supply is obtained from a Sola regulating transformer and the D.C. from a 250 volt Sola D.C. supply. These supplies were available for other parts of- the mass spectrometer; otherwise the D.C. could be obtained from a simple circuit with s i l i c o n r e c t i f i e r s . The -performance- of-the- supply does not deteriorate very much i f the Sola regulating transformer i s not used. 96 A to- volt step in -the input line voltage (without the A.C. Sola regulator) results in a 0.012 per cent change in magnet current'. The s t a b i l i t y of the current- is the order of one part in 25,000 for variations of periods in the order of seconds or minutes, and- the ripple current -is less -than this - if-the f i l t e r is reasonably well tuned so that the transistor amplifier is not overdriven. This is not only because the ripple voltage is greatly reduced by the A.G.-degeneratlve"-feedback y but also because of the high-coll impedance at 120 cps (see Figure 26). The difference amplifier transistors are individually quite temperature sensitive. The slow d r i f t resulting from this can be minimized by arranging that the temperatures of these transistors are always equal. A further improvement in long term s t a b i l i t y might be achieved with s i l i c o n transistors in this part of the c i r c u i t . Fewer losses are quite small-. At maximum current there are about 15 watts dissipated on the power transistor, five watts on the control resis*"orf five watts on the 6A05 filaments 1 and- about 25 watts in the 6AQ5 plate c i r c u i t . The power delivered to the magnet is about 180 watts, giving an overall efficiency of more than 75 per cent. The cost of the electronic components was the order of $150 not. including the meter, which is a mirror scale instrument of'one per cent accuracy imported from Denmark at a cost of $40 and the magnet coils which were manufactured commercially at a cost of $650. The meter has, in addition to the current scale, a scale proportional to the square of the current which is useful as a mass indicator. 97 A- number of modifications could be considered in the design. The use of 400 cycles or higher frequency power is possible and this would be worthwhile. The substitution of transistors for the 6AQ5 is an obvious change that would require only a change of resistance of the D.C. winding of the saturable reactor. Thyratron regulators could substitute for the saturable reactor, as grid-controlled r e c t i f i e r s . Since the supply was built s i l i c o n controlled rec t i f i e r s have appeared on the market and these also seem to offer an attractive substitute for the saturable reactor; In either case the speed of response and gain of the "slow" loop could be greatly increased, and this i n turn would make the f i l t e r design less c r i t i c a l . The filament emission control (Figure 28) This circuit is a straightforward regulated power supply using a series transistor as the regulating element. This transistor can easily regulate the five ampere current used at a voltage drop of less than 2.5 volts. Obviously no vacuum tube w i l l operate at these currents and voltages and therefore this type of circuit is only i possible with transistors. A three phase 120/208 volt service was available at this laboratory, and therefore a three phase input-six phase output trans-former was used for the filament supply to give a low ripple without additional f i l t e r i n g (approximately 4.5 per cent R.M.S. in the D.C. voltage). A similar result could be obtained by using a f u l l wave single phase r e c t i f i e r with a large' condenser (several tens of thousands of microfarads) or an LC f i l t e r , or else the large ripple of an unfiltered supply might not be considered objectionable. 98 10LF (6) Fig. 28 Filament emission control. Component identification is the same as i n Fig. 27. Three phase transformer secondaries are 9 volt center-tapped windings insulated to 5 kilovolta. The resistance of the filament and 5A meter, including interconnecting cables, must not exceed 1 ohm. 99 The 2N456 control transistor is connected with a second similar transistor in the Darlington compound arrangement to provide a large value of current gain and therefore a sufficiently low base current (typically 0.7 ma). This pair is driven by a 2N369 voltage amplifier-'whrch 'in- turn-is -driven 'by -a- -2N366 emitter• follower. The forward voltage drops across the emitter-base diodes of the latter two transistors ^approximately'cancel- and therefore at balance the error signal, which is applied between the base of the 2N366 and the emitter of the 2N369, is very nearly zero. The error voltage is the difference between the voltage developed across the 100 k helipot and i t s series resistor, and the reference provided by the 2 7 volt Zener diode. In addition there are two supplies of very low power. One is a glow tube-stabilized supply to provide the electron accel-erating voltage and the trap-bias i n the mass spectrometer source,-a s well a s providing current for the 2 7 volt Zener. This supply uses the smallest •available filament transformer connected backwards so that 7.2 volts R.M.S. from the three phase transformer is applied across i t s 6.3 volt winding, the voltage developed in the primary being rect i f i e d with- s i l i c o n diodes in a simple doublercircuit. This arrangement'''^ transformer insulated for the 5^000 volt accelerating voltage. The second supply is a low voltage three phase s i l i c o n r e c t i f i e r circuit which supplies the emitter follower. The three phase power was available from the main transformer secondary, and using i t simplified the f i l t e r i n g . 100 This supply i s physically the smallest used in the mass spectrometer;, ••••The1 electronic units, including the two subsidiary power supplies and "the required heat sinks for the power transistors and s i l i c o n - r e c t i f i e r s f i t entirely into a chassis 2" x 8" x 10". This chassis floats to 5,000 volts when the mass spectrometer is operating,1-and-is suspended inside-a somewhat larger grounded chassis, thus 6l-iml-irat±ttg""1yh!e"«po&si-bili-ty• -ot--corona--di-ac-harges • around small components"and'providing increased-safety. The transformer is mounted on the subsidiary chassis and the meters are mounted separately on the front panel- with the various controls. The supply input power is less than 30 watts The supply cost approximately $100 plus the cost of the meters. The regulation of the supply is adequate; the measured variations i n electron current being the order of one or two parts in 10,000. The supply requires no warm-up time and exhibits remarkably l i t t l e d r i f t i n view of the fact that the base current of the 2N366 shunts the 100k control helipot. The electrolytic condenser in parallel with the error signal is necessary for a well damped system and has not caused any of the anticipated d i f f i c u l t i e s , perhaps because the maximum voltage across i t is tens of m i l l i v o l t s . The measuring system The measured voltage coefficients of the High-Meg resistors are plotted i n Figure 29. Corrections in the measurements for the honlihearity of these resistors are not necessary. The ion current preamplifier is shown in Figure 30. There are two such units i n the mass spectrometer differing only in the values of the grid leak resistors. The circuit diagram shows the • 101 No. 1 preamplifier only and the collector electrodes connected as the measurements were made without the grid ratio system. The supply voltages marked on the diagram are obtained from batteries. A Zener diode stabilized supply was used originally with sl i g h t l y different voltage taps. At the same time the negative side of the filament of the electrometer tube was directly connected to the base of transistor Q,j . The output signal was obtained between the emitter of Qg and a +2.4 volt reference voltage from the Zener stabilized supply. That arrangement was modified to the illustrated version giving a much smaller 60 cycle ripple on the signal. The whole unit is built on a small chassis which is mounted on the steel angle-iron frame of the mass spectrometer. Shielded cables give interconnections to the leads on the collector flange and to the main consol containing the other parts of the servo voltmeter. The diagram shows an interesting cathode follower operation of the electrometer tube. Transistor Q 3 drives the interconnected plate and screen grid with the same voltage variations that appear on the cathode of the tube. This results i n a unity gain cathode follower in which a l l the tube electrodes follow the grid. Thus the tube does not add i t s interelectrode capacities to the stray capacity of the collector. This circuit was built and operated, and the expected reduction in capacity was obtained. However, a certain amount of noise was generated in the transistor circuit and there was not sufficient time to eliminate this. At present the plate is rather connected to a steady +7.5 volt tap on the battery, marked with an arrow head on the diagram. . Although the gain is only about 0.5 102 4— 1 — — • — i — — r - — r — i — 1 — — — — 0 0 10 15 20 28 80 85 ••• VOLTS •• Fig. 29 Voltage coefficient of Victoreen resistors. The voltage'coefficient given here is defined as l/R (dR/dE). The plotted values were obtained by measuring the unbalanced voltage of the Wheatstone bridge incorporating the High-Meg resistors. Changes i n the unbalancedvoltage were determined when the voltage on the bridge was varied. * * Whittles, A.B.L., Voltage coefficient of Victoreen High-Meg resistors. Rev. Sci. Instr. 31_, 208-209, 1960. 103 * + 22.5 V o Relays _ t _ M ^ Fig. 30 Ion current preamplifier Push-button switches Fig. 31 Lead equalizer and attenuator 104 60 cps power transformer 1 Meg |~~] Control windings of motor Fig. 32 Original circuit of output stage in Brown, No. 358816 chopper amplifier. Feedback 7.3k Fig. 33 Modified Maxwell bridge for velocity damping. 105 -13v 0-3 mA SHUNT SELECTOR 5 position push-button switch Gen. Contr. MPB-S B14 HC i r V W W V - o 6 4 o—AAAAA/V—< 5k , W 1 ^ 6 T ° 344 K A / Y Q ^ V - O Motor 1 (333 rpm) 20k 6 2o—\AA/VW-< 324k 1 ©-WAA/Vv— 1 909k 10 turn Helipot 7603 20k Coarse DP 4 Pos. 10k ^  rotary-switch 10k, 10k 10 turn double Helipot Ser.. 7600 (± 0.025$ linear) 10k lOv Dial on shaft 3 in.dia. Beckman BP-352 10k 10k Motor 2 (162 rpm) RANGE •Fine 10k 3 turn Feedback No.2 Chart recorder (adjusted to lOv f u l l scale sensitivity) Feedback No.l Fig. 34 Servo motor driven potentiometers and attenuators in measuring ci r c u i t . 106 the improvement in noise i s worth the sacrifice until more development work can be carried out. The effects of the higher time constant of the grid circuit are now compensated by suitable equalizer networks which permit the use of higher gain in the chopper amplifier. A lead network of 33 kiloohms and 4 microfarads is between the preamplifier and attenuator (see Figure 31). This attenuator ensures a constant loop gain of the servo system when the measuring range is changed. (Fpr Ishe collector No.2 system no such elaborated attenuator was necessary; only a 2 kiloohm potentiometer was used). Figures 32 and 33 show the original and the improved version of the chopper amplifier. The D.C. decoupling reduces the j i t t e r of the motor. The carefully balanced bridge circuit permits a higher feedback ratio than was possible with the original components which were not balanced well. The back e.m.f. of the motor is proportional to i t s speed and a fraction of this signal is fed back to cause a damping proportional to the velocity. Figure 34 shows the complete range determining network of the measuring system. A l l the resistors are IRC wire wound 0.1 per cent pieces. The multipliers giving the relative sensitivities in different ranges were determined by a Leeds Northrup, Type 4230 precision Wheatstone bridge. Repeated checks were made other ways too, with similar results. The multipliers (often called shunt coefficients) are given i n Table 11. At present, when the second servo system is not working, the input of the No.l resistor assembly is directly supplied from the -13 volt battery. This gives a f u l l scale sensitivity of 1.3 x 10~^° amperes in the range 5. Note that the input Impedance 107 of the feedback divider of the No.l system i s not quite constant and that the values of the resistors should be trimmed for ratio mea-surements. The circuit as shown would produce errors of 0.03 per cent. TABLE 11 Switch Shunt coefficients 1 1 2 3.3211 1 3 10.014 3.0152 1 4 33.547 10.101 3.3501 5 99.970 30.101 9.9832 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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


Related Items