BULLETIN OF THE GEOLOGICAL SOCIETY OF AMERICA VOL. ee PP. 1131-1148 SEPTEMBER 1955

ISOTOPIC COMPOSITION AND DISTRIBUTION OF LEAD, URANIUM, AND THORIUM IN A PRECAMBRIAN GRANITE

By GEORGE R. TILTON, CLAIRE PATTERSON, HARRISON BROWN, MARK INGHRAM, RICHARD HAYDEN, DAVID HESS, AND ESPER LARSEN, JR.

ABSTRACT The isotopic compositions and concentrations of lead and uranium have been deter- mined in some separated minerals and the composite of a granite from Monmouth town- ship, Haliburton County, Ontario. The chemical and mass spectrometric methods that were used are described. The age of the zircon from the granite is 1050 million years. Much of the lead, uranium, and thorium exists in chemically unstable and presumably interstitial phases of the granite. A comparison of the observed amounts of uranium, thorium, and lead in the various minerals with those amounts that should have been pres- ent, had these three elements existed within the minerals as closed systems, shows a non- balance of these elements in every case. It appears that the granite as a whole has closely approximated a closed system since it was formed with respect to uranium and its decay products, but has been an open system with respect to thorium and its decay products. Interpretations concerning the relationship of these data to lead ores are discussed.

CONTENTS TEXT Page Conclusions 1147 Page References cited 1147 Introduction 1131 Acknowledgments 1132 TABLES Rock description 1132 Table Page Chemical procedures 1134 1 Chemical analysis, norm, and mode of Contamination 1134 granite from near Tory Hill, Ontario.. . 1133 Reagents and apparatus 1135 2. Semiquantitative spectrographic analysis Isolation of lead 1135 of granite from Tory Hill, Ontario, and Isolation of uranium 1136 of three minerals separated from the Isolation of thorium 1137 granite 1134 Mass spectrometry of uranium and lead 1137 3. Uranium, thorium, and lead concentration General 1137 and isotopic compositions of lead in vari- Lead procedure 1138 ous constituents of the granite 1140 Uranium procedure 1139 4. Comparison of lead and thorium values ob- Accuracy 1139 tained by different methods 1141 Results 1139 5. Ages for the granite and the Wilberforce Discussion 1141 pegmatite 1142 Uranium-lead age of granite 1141 6. Contribution by each mineral to the ura- Potassium-argon age of the granite 1142 nium, thorium, and lead content of the Loss or gain of uranium, thorium, and lead total granite 1145 in the granite 1143 7. Comparisons between the observed and Distribution of uranium, thorium, and lead calculated amounts of radiogenic leads, in in the granite 1144 various constituents of the granite 1145 Material balance 1145 8. Calculated isotopic compositions of lead in Internal transfer of uranium, thorium, and the granite at various times 1146 lead 1145 9. Average composition of lead ores at various Relation to lead ores 1146 times 1147

INTRODUCTION in amounts of but a few parts per million, the It has long been recognized that if techniques range of materials amenable to precise radio- could be developed to determine precisely the active dating would be enormously increased. isotopic compositions and concentrations of lead In addition, such techniques would permit more and uranium, when these elements are present detailed studies of the time and space relations 1131

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 1132 TILTON ET AL—LEAD, URANIUM, AND THORIUM IN A GRANITE

of radioactive elements and their decay ACKNOWLEDGMENTS products. The isotope dilution technique of quantita- We wish to express our gratitude that the tive analysis, first described by Hayden, Rey- facilities of the Argonne National Laborato- nolds, and Inghram (1949, p. 1500) and later ries, the Department of Terrestrial Magnetism applied (Inghram et al., 1950, p. 916) to the of the Carnegie Institution, the U. S. Geologi- study of calcium and argon in Stassfurt sylvite, cal Survey, the Institute for Nuclear Studies has been applied in this work to lead, thorium, of the University of Chicago, and the California and uranium. The techniques described in this Institute of Technology could be combined to paper make possible both the precise determi- make these studies possible. We are indebted nation of as little as 0.01 part per million of to The Geological Society of America for a uranium and thorium and 0.1 part per million grant which has supported some of this work. of lead in silicates, and the determination of The portion of the work undertaken at the the isotopic composition of lead totaling only a California Institute of Technology was sup- few micrograms in a sample. ported by the U. S. Atomic Energy Commis- Briefly, the isotope dilution method of anal- sion (Contract #AT-11-1-208). We sincerely thank Kenneth Jenson of the Argonne National ysis operates as follows: If the uranium content Laboratories for the careful and precise potas- constituting about one part per million of a sium analyses, Claude Waring of the U. S. mineral is to be determined, approximately 5 Geological Survey for making available the re- grams of the mineral are dissolved and a few 235 sults of the spectrographic analyses for lead micrograms of U containing but a small frac- to which he has devoted considerable time and 238 236 tion of U are added to the solution. The U effort, and H. S. Armstrong of McMasters Uni- becomes mixed with the uranium from the un- versity for providing the sample of pegmatite known which is primarily U238. The mixture of perthite. We are grateful for the advice and uranium from the unknown and the uranium criticism of our colleagues, in particular Gunnar tracer is extracted from the solution, and the Kullerud of the Mineralogisk-Geologisk Mu- new ratio of U235 to U238 is determined. Knowing seum, Oslo, Norway, and Leon Silver, Cali- the initial sample weight, the weight of U236 fornia Institute of Technology. tracer added, and the original and final ratios of U235 to U238, one can readily compute the ROCK DESCRIPTION concentration of uranium in the unknown. The It was agreed that the granite to be used for application of the technique to lead, although this investigation should be of Precambrian age more complicated, is essentially the same. and should fulfill the following requirements: The main obstacles to be overcome in this it should not have been altered or recrystal- work were contamination during the isolation lized, but it should be moderately radioactive; of such small quantities of material and diffi- it should contain a moderate amount of lead culty in measuring accurately the isotopic com- and at least average amounts of the accessory position of extremely small samples. Therefore minerals zircon, apatite, sphene, and magne- extreme care was taken to control contamina- tite; its potash feldspar should be rich in lead tion and a combination of surface ionization so that the isotopes of the lead of the original and electron multiplier detection techniques magma could be determined. The late Dr. H. were used in the mass spectrometer. V. Ellsworth supplied granite that satisfied The techniques developed for uranium, tho- most of these requirements from a source he rium, and lead have been applied to a study described as follows: of their distribution in a Precambrian granite. "On Map No. S2a, Haliburton area, Province of The studies were not exhaustive, but were un- Ontario, (Satterly, 1943) the area between Tory Hill and Essonville marked with the symbol 'Ra' dertaken primarily to test the method and represents roughly the area from which the granite demonstrate its applicability to geochemical samples were taken and the symbols probably the actual openings. This granite is from the Brower- and geochronological problems. Simpson property."

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 ROCK DESCRIPTION 1133

The geology of the area between Tory Hill Semiquantitative spectrographic analyses of and Essonville is shown on a map by Adams the granite and of apatite, sphene, and zircon and Barlow (1910) and on a modified map by separated from the granite are given in Table 2.

TABLE 1.—CHEMICAL ANALYSIS, NORM, AND MODE or GRANITE FROM NEAR TORY HILL, ONTARIO Chemical analysis by A. M. Sherwood, U.S. Geological Survey, Washington, D.C. Chemical Analysis Norm Mode*

SiO2 72.55 Q 24.24 Quartz 24 ± 1 TiO2 0.25 or 30.02 Perthite 52 ± 2 A1203 14.29 ab 41.39 Plagioclase 20 ± 1 Fe203 1.41 an 0.83 Hornblende 2 ± 1 FeO 0.54 C 0.41 Magnetite 0.4 ± 0.2 MnO 0.02 mt 1.16 Zircon 0.04 ± 0.01 MgO 0.06 il 0.46 Apatite 0.02 ± 0.01 CaO 0.17 hm 0.64 Sphene 0.4 ± 0.1 Na20 4.92 hy 0.20 Pyrite 0.02 K2O 5.06 Calcite trace H2O- 0.09 + H2O 0.14 P206 0.01 U 0.001 Total 99.50 * The mode was estimated from the chemical analysis, thin sections, and the amounts of separated minerals.

Satterly (1943); both, on a scale of 2 miles to About 50 pounds of the granite were ground the inch, show metasedimentary rocks of the by hand to pass 15 mesh and sized. The miner- Grenville series at Tory Hill and Essonville als were separated by using bromoform, meth- and granite on the road between the two vil- ylene iodide, and the Frantz Isodynamic sepa- lages, with the contact between the Grenville rator with special precautions to avoid lead and the granite about a mile east of the road. contamination (Larsen, Keevil and Harrison, In October 1952, Larsen investigated the excel- 1952, p. 1047). The final zircon sample was lent new road cuts and found them to be chiefly hand-picked, and all mineral samples were more in rocks of the Grenville series, with some small than 99 per cent pure. bodies of granite. There may be more granite The sphene, which is more abundant in this to the west. The granite specimens used in this granite than in most granites, occurs in crystals, investigation undoubtedly came from some of several mm long. It is reddish brown in trans- the small bodies of granite. It is rather uni- mitted light. Its optical properties are: a = formly red and coarse-grained, with some 1.89, |3 = 1.90, 7 =2.01, and 2V small, which gneissic structure. The microscopic texture is indicate an ordinary sphene. granoblastic with many grains of strained and The zircon occurs in stout prisms that aver- broken plagioclase and quartz. Whether this age about 1 mm in length. Several hundred is merely a protoclastic texture in early crystal- mg of these grains were ground to approxi- lized grains, or whether it reflects a post-crys- mately 200 or 300 mesh to obtain the powder tallization deformation cannot be demon- used in analytical work. Optical examination strated. The larger grains of feldspar may be of the powder revealed wide variations in re- as long as 10 mm, with an average length of fractive index (&>), which ranged from 1.92 to about 3 mm. Table 1 gives the chemical anal- 1.82 and averaged about 1.85. Thin sections ysis, the norm, and the mode of the granite. of several of the larger grains indicated exten-

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 1134 TILTON ET AL—LEAD, URANIUM, AND THORIUM IN A GRANITE

sive zoning of the mineral. Some of the outer- (1) The laboratory was mopped and vacu- zone zircon is very near ordinary fresh zircon umed at frequent intervals. A positive pressure with co of about 1.915 and with strong bire- of filtered air was maintained in the laboratory fringence. Some parts of a zircon crystal may and drafts were minimized.

TABLE 2.—SEMIQUANTITATIVE SPECTROGRAPHIC ANALYSIS OF GRANITE FROM TORY HILL, ONTARIO, AND OF THREE MINERALS SEPARATED FROM THE GRANITE Analyses by C. S. Annell, U.S. Geological Survey, Washington, D.C. Over 10% 1-10% 0.1-1.0% 0.01-0.1% 0.001-0.01% 0.0001-0.001% Granite AISi FeKNa CaMn Sr Ga Pb Cu MgVBa Ti Apatite CaP Si Zr Fe Pb VTi Yb Tm Ga Ba Cu Ag Mn Na Sr Mg Ho AlYLa Sphene Si Ca Ti FeAl Mn Th Na Pb Mg Sr Ga Lu Tm Ba ZrYLa SnVYb Ba Ho Cu Zircon ZrSi FeYHf Mn Yb Tm Sr Ga Ba AgCu AlPb Ho Lu Ti Mg

be nearly or completely isotropic with n about (2) The bench tops and hoods were con- 1.84. All intermediate indices of refraction seem structed of stainless steel and kept clean by to be present, as if mixtures in all proportions frequent scrubbing and rinsing. of the two end materials were present. (3) All cleaned glassware and reagent bottles Hurley and Fairbairn (1953, p. 659) have were sealed with a semiadhesive sheet plastic, found that the value of 2d for diffraction from "Parafilm".1 The seals were broken only when the lattice plane (112) of zircon is 35.635° for materials were being transferred, and were im- the undamaged mineral and approaches 35.1° mediately resealed after an operation. for complete metamictization, accompanied by diminishing peak intensities and sharpness. An (4) Hydrofluoric acid evaporations were car- X-ray study of the present zircon gave a low, ried out in a stainless steel box. Tank nitrogen, broad diffraction from the (112) plane with the filtered through glass, was kept flowing through maximum at about 35.3° indicating extensive the box. metamictization in agreement with the optical (5) All other evaporations were carried out data. The alpha activity of the zircon is in glass tanks. Fumes were swept out of the 1100/mg/hr. The powdered mineral has been tanks by a continuous flow of filtered nitrogen. leached for 30 minutes in hot 1:1 aqua regia, Under these conditions, when the processing and 35 per cent of the activity was removed. of a sample required 10-15 hours of evapora- tions and several days of handling, the lead CHEMICAL PROCEDURES contamination varied from a few tenths to slightly more than 1 microgram per sample. If Contamination air contamination during evaporation was not Contamination was introduced from three controlled, the amount of lead introduced per sources: reagents, apparatus, and the labora- sample, varied erratically from 1 to more than tory air. The removal of lead, thorium, and 10 micrograms. The uranium and thorium con- uranium from the reagents and apparatus is taminations were generally about .02 micro- discussed under Reagents and Apparatus. Con- gram or less per sample. tamination from the air was a critical factor in the case of lead. Air contamination was con- 1 Manufactured by Marathon Corp., Menasha, trolled as follows: Wisconsin.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 CHEMICAL PROCEDURES 1135

Reagents and Apparatus Pb208 tracer processed by the reagents, was iso- lated at the same time. Water was triple distilled. The final distilla- The feldspars and quartz were dissolved by tion was carried out through a Pyrex, triple heating first with hydrofluoric acid, and then cold wall reflux condenser. Nitric and perchloric with a mixture of borax and perchloric acid. acids were vacuum distilled in a transparent Zircon and sphene were dissolved by fusing silica still containing a 60-cm vigreux column. with borax, leaching with hydrochloric acid, Solid reagents were recrystallized and extracted and dissolving the residue in the same manner repeatedly with dithizone in choloroform to as for the feldspars and quartz. Magnetite and remove lead. Organic liquids were extracted apatite were dissolved in hydrochloric acid and repeatedly with dilute nitric or hydrochloric perchloric acid respectively. The granite com- acids and water. Hydrofluoric acid was pre- posites were treated in the same manner as the pared by passing the gas through a filter of 2 feldspars and quartz, the insoluble residues be- fine "Teflon" shavings and a "Teflon" tube ing collected and treated in the same manner bubbler into water in an ice-cooled platinum as the zircon and sphene. vessel. The isolation of lead from these solutions Hydrochloric acid and ammonium hydroxide consisted of complexing interfering ions with were prepared in the same manner as hydro- citrate and extracting lead dithizonate into fluoric acid except Pyrex and "Tygon" tubing, chloroform at a pH of 9.5. When the iron con- glass filters, and Pyrex vessels were used. Ura- tent was high, as in the case of magnetite, the nium, thorium, and lead tracer solutions were solutions were first oxidized with nitric acid standardized both gravimetrically and by iso- and then ether-extracted. When some interfer- tope dilution. ing anions were in large excess, as in the case Pyrex glassware was used throughout. This of phosphate from apatite, large excesses of was cleaned by scrubbing with scouring powder, citrate and high dilutions were employed. The rinsing with distilled water, complete immer- procedures are modifications of those devel- sion for 20 minutes in a 10 per cent solution of oped by G. Lundell and H. B. Knowles (1920, potassium hydroxide held near the boiling p. 1440), P. A. Clifford and N. J. Wickmann point, rinsing with distilled water, complete (1936), K. Bamback and R. E. Burkey (1942), immersion for 20 minutes in concentrated nitric and E. B. Sandell (1950, p. 388-412) and were acid held near the boiling point, and rinsing developed with the aid of radioactive tracers. with double distilled water. The glassware was The yields ranged from 30 to 80 per cent. A handled only with stainless steel tongs that were detailed discussion of some of the lead analyti- kept clean and frequently rinsed. Stopcocks and cal work has appeared in a declassified Atomic ground stoppers were used without grease. Energy Commission document (Patterson, Platinum-ware was cleaned by scrubbing with 1951). The specific analysis of the acid-washed scouring powder, etching a fresh surface with granite was as follows: aqua regia and rinsing with double distilled (1) Ten grams of powdered rock were placed water. The lead gaskets encountered in tanks in a large centrifuge bottle, 50 ml of ether was of compressed gases were removed, the fittings added and the mixture swirled for 3 or 4 min- cleaned, and gaskets of "Teflon" substituted. utes. The ether was centrifuged off and dis- carded. This was repeated twice with ether, Isolation of Lead three times with 6 molar HC1, and four times with water. The sample was dissolved and divided into 208 (2) The cleaned material was dried at 110°C two aliquots, a known amount of Pb tracer and weighed into platinum dishes. One hun- was added to one aliquot. The two aliquots dred ml of concentrated hydrofluoric acid was were processed chemically to isolate the lead. added and evaporated to near dryness at low A blank, consisting of a known amount of heat over a long period of time. One hundred 2 ml of concentrated perchloric acid was added Manufactured by E. I. duPont de Nemours and Co. and the mixture warmed until HF fumes ceased

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 1136 TILTON ET AL.—LEAD, URANIUM, AND THORIUM IN A GRANITE

to evolve. Fifty ml of a saturated solution of dissolved by fuming with perchloric acid in the borax was added and the mixture heated presence of borate ion (feldspars and quartz). strongly until a cake was almost ready to After the sample was dissolved, the resulting form. It was then cooled, stirred with 200 ml solution was equilibrated with 30 micrograms of warm water, and centrifuged. of essentially pure UM6. The principal purifica- (3) The centrate was reserved and the resi- tion of uranium was then accomplished by due transferred to the top of a 1-gram bead of hexone extraction of a nearly saturated solu- fused borax in a platinum crucible. The bead tion of the nitrate salts (Blaedel et al, 1945). was fused again, cooled, and dissolved in 50 The few remaining impurities were then pre- ml of warm 3 molar HC1. This solution was cipitated on ferric hydroxide by ammonium made ammoniacal, centrifuged, the centrate carbonate, the carbonate ion holding the ura- reserved, and the residue transferred to a plat- nium in solution by complex ion formation. A inum dish. final hexone extraction from 10 M ammonium (4) The residue was treated with hydrofluo- nitrate solution gave uranium which apparently ric and perchloric acids and borax as described contained no more than a few micrograms of in (2), except that one-tenth of the quantities impurity and which was pure enough in all of reagents was used. There was no visible res- cases for good spectrometer runs. idue at this point. All solutions were combined In this study the contamination of a mineral and diluted to 500 ml in a volumetric flask. analysis was obtained by comparing the results (5) A 100-ml aliquot was withdrawn and for two different weights of sample which were mixed with an aliquot containing 25 micro- processed in an identical manner using identi- grams of Pb208 tracer. This solution was treated cal quantities of reagents. In newer work as described below, the remaining 400 ml being straight blanks using ferric hydroxide as a treated in a similar manner except that propor- carrier have been used since this has been tionately larger amounts of reagents were used. found to be a superior method of measuring (6) Twenty-five ml of a 30 per cent citrate the small contaminations of 0.01-0.001 micro- solution was added, then 10 ml of ammonium grams of uranium usually encountered in an hydroxide. This solution was extracted with 50 analysis. ml of chloroform containing 2.5 mg of dithizone. A detailed discussion of the uranium analyti- (7) The chloroform layer was separated and cal work has appeared in a declassified Atomic extracted with 50 ml of 2 per cent HN03. Energy Commission document (Tilton, 1951). (8) The water layer was separated and 10 The complete analysis of the acid-washed ml of a solution containing 15 per cent concen- composite rock sample is as follows: Approxi- trated ammonium hydroxide and 2 per cent mately 17 g of rock powder ground to pass 40 KCN were added to it. This solution was then mesh was washed three times with ether, extracted with 10 ml of chloroform containing washed for three 5-minute periods in 6 M hy- 0.2 mg of dithizone. The chloroform layer was drochloric acid and finally rinsed three tunes separated and evaporated to dryness. One ml with distilled water. The powder was dried at each of water, concentrated nitric acid, and 115°C for 3 hours. The sample was weighed to concentrated perchloric acid was added and the nearest mg into platinum and treated with the solution evaporated to dryness. The yield 120 g of hydrofluoric acid at about 70°C for was about 70 per cent. 3 hours until a moist paste remained. Eighty ml of perchloric acid and 40 ml of water were Isolation of Uranium added and the solution was heated until it fumed on a hot plate at medium temperature Three methods of sample decomposition were and fumed for 15 minutes. Borax (7.5 g in 20 employed: digestion in mineral acids (magne- ml of water) and 100 ml of perchloric acid were tite and apatite), fusion in borax followed by then added. The solution was concentrated to digestion in dilute hydrochloric acid (zircon acid fumes and fumed strongly for about 90 and sphene), and treatment with hydrofluoric minutes at which point crystals began to sepa- acid after which the insoluble fluorides were rate from the thick syrup. Two hundred ml of

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 CHEMICAL PROCEDURES 1137

water was added and the mixture was steamed beaker. The residue was heated a few minutes for 10 minutes. Five to ten mg of black and with the hot plate at high temperature to de- brown granules representing unattacked zircon stroy the ammonium nitrate present. A thin and possibly other accessory minerals remained black film of uranium oxide then remained. at this point. The solution was decanted into a 250 ml volumetric flask; the granules were Isolation of Thorium washed into a beaker with water and the water evaporated off. The undissolved granules were The thorium chemistry has been described fused 15 minutes in 3 g of borax at the maxi- elsewhere (Tilton et al., 1954). Briefly, the work mum heat obtained from a Meeker burner (ca. was carried out as follows: Approximately 1 232 1200°C). The bead was digested 1 hour in 25 mg of thorium with a Th^/Th ratio of four ml of 6 M hydrochloric acid. This solution was was isolated from a Colorado Plateau uranium combined with that in the volumetric flask and ore. This material served as a carrier to deter- the contents were diluted to the mark with mine the thorium content of the minerals 232 water. Aliquots of 50.0 and 25.0 ml of this (which consisted solely of the isotope Th ), 235 solution were pipetted into two 250 ml centri- as the U served for the uranium determina- fuge bottles where they were equilibrated with tions. Chemical concentrations of thorium were 29.60 micrograms of uranium with a U^VU238 accomplished by various combinations of three ratio of 1752 by weight. Each sample was procedures: precipitation of thorium on lan- diluted to 1 liter and the hydroxides were pre- thanum oxalate as carrier; extraction at a pH cipitated by bubbling ammonia gas into the of two into a benzene solution of trifluoro- solution. The precipitate was dissolved in nitric thenoylacetone (TTA) (Hagemann, 1950, p. acid and a second hydroxide precipitation was 769); and extraction into hexone from a nearly performed. This step was necessary to prevent saturated solution of aluminum nitrate con- the separation of perchlorates when the solu- taining 10 per cent nitric acid (Levine and tion was later concentrated for extraction. Grimaldi, 1951). The second hydroxide precipitate was dis- The mass spectrometric determinations were solved in 50 ml of concentrated nitric acid and checked colorimetrically using the sodium salt each sample was concentrated to a volume of of 2-(2-hydroxy-3, 6-disulfo-l-napthylazo) ben- about 30 ml. Ammonia gas was passed into the zene-arsonic acid (Thorn) using the proce- solution until the ferric ion present turned from dures described by Thomason, Perry and light yellow to deep orange in color (pH 2 to 3). Byerly (1949, p. 1240) and the 24.1 day beta 234 234 The salted nitrate solutions were each ex- activity of Th -Pa as a tracer to correct tracted three times with 35 ml of hexone (4- yield losses in processing. The two methods methyl-2-pentanone). The hexone was then ex- checked within the limits of error. tracted twice with 25 ml of water. The aqueous extract was concentrated to 10 ml and 0.5 mg MASS SPECTROMETRY or URANIUM of Fe+++ as ferric nitrate was added. The iron AND LEAD was precipitated by dropwise addition of a saturated solution of ammonium carbonate, General adding one to two drops in excess. The ferric The isotopic analysis of lead and uranium hydroxide with occluded impurities was centri- from separated minerals of granites has not fuged off and discarded. been previously attempted, largely owing to Each supernate from the carbonate treat- the fact that the minimum sample size which ment was acidified and concentrated to a can be readily handled using the classical tech- volume of a few drops. Fifteen ml of 10 M ammo- niques is of the order of several mg. Thus, in nium nitrate solution was added and the result- order to obtain sufficient lead from a mineral ing solution was extracted twice with 15 ml of containing one part-per-million of the element, hexone. The hexone was washed with 2-10 ml several kg of sample would have to be processed. portions of water. This aqueous extract was In view of the difficulties of preparing satisfac- concentrated, taking to final dryness in a 5 ml tory mineral separates and of undertaking the

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 1138 TILTON ET AL.~LEAD, URANIUM, AND THORIUM IN A GRANITE

special chemical porcedures here outlined, the currents are small, thus necessitating the use amount of work involved in processing such of an ion detector of high sensitivity. For this large quantities would be prohibitive. The reason the usual ion detector was replaced alternative was chosen of developing tech- with an internally placed electron multiplier niques of micro mass spectrometry such that similar to that used by Cohen (Cohen, 1943; only gram quantities of mineral separates are Allen, 1939). required. The standard 60° sector deflection instru- The sample size necessary for an isotopic ment, in which transmissions of several per analysis by use of the lead iodide evaporation cent are obtainable, was selected, but the size method (Nier, 1939b, p. 155) is prescribed by was increased to improve both resolution and several factors. Many mass spectrometers can transmission. The mass spectrometer finally satisfactorily resolve the isotopes of lead in the evolved is a 12-inch radius of curvature single Pb+ ion position, but few operate satisfacto- focussing instrument using the surface ioniza- rily in the Pbl+ position. But in the former tion source of Inghram and Chupka (1953, p. position the spectrum is complicated by the 519). For large samples of 10~6 g, only the presence of the mercury isotope of mass 204, if central filament is used. For smaller uranium mercury pumps are used, and the sample size samples, it is operated with the sample loaded must be such as to swamp this peak. Mass spec- on a side filament. The advantages of the mul- trometers in which oil pumps are used do not tiple filament assembly have not to date been suffer from this limitation since the background realizable with lead. peak at this mass is much smaller, but the ad- The machine is equipped with differential vantage is partially cancelled by the fact that pumping to shorten the pump-down time be- the hydrocarbon background is variable with tween samples. The time required is about 1 time, thus making corrections difficult to apply. hour instead of 6-8 hours as is required with a In principle, however, the lead iodide evapora- single pump. Both permanent magnets and tion method could be utilized for smaller sam- electromagnets have been used to produce the ples than are now required if a machine of 7000 gauss field. Either is satisfactory. With sufficient resolving power were used to permit the machine finally evolved it is possible to operation in the Pbl+ region where the back- detect isotopes of lead if present to 10~10 g, grounds are much smaller. and of uranium to 10~13 g. The limitation in sample size with the Pb (CH3)4 method (Collins et al., 1951, p. 966) Lead Procedure is largely one of preparing the volatile com- pound in sufficient quantity and purity for sat- Three compounds were used in the surface isfactory analysis. The recent work using this ionization source: the oxide, the sulfide, and method indicates that several mg are necessary. the sulfate. Either of the first two compounds This is little improvement over the sensitivity give reliable results, although all samples in the obtained by Aston (1933, p. 535) in his early present work used the oxide. In particular: studies of radiogenic leads using Pb(CHs)4 as (1) The sample of Pb(ClO4)2 is placed in a the source gas. The method does not appear to 5 ml beaker, evaporated to dryness, and heated be extendable to micro-samples. for one-half hour at 350°C to remove all traces After several possible ion sources were con- of acid. sidered, the surface ionization source was (2) A sample of borax weighing one-half as selected because it is very selective in its ion- much as the lead sample is dissolved in 25 ization. It will not ionize mercury, and hydro- microliters of water and added to the above. carbons are ionized only very weakly. The (3) Concentrated HNO3, sufficient to neu- resultant spectrum is quite clean, thus remov- tralize the borax, is added and the 25 micro- ing partially one of the major limitations of liters is swished about the beaker to dissolve the lead iodide evaporation source. the lead. Although the surface ionization source gives (4) The sample is taken up in a pipette and a cleaner spectrum than the older sources, ion transferred to the 0.030 inch x 0.001 inch tung-

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 MASS SPECTROMETRY OF URANIUM AND LEAD 1139

sten filament of the surface ionization source, to a narrow range of isotopic compositions; evaporated to dryness, inserted in the mass spec- (2) The sample sizes required are a million or trometer, and pumped down. more times less than with gas sources. (5) The temperature of the filament is raised One of the most important factors limiting slowly to about 600°C where emission of Pb+ accuracy is the background correction. The ions starts. At first the spectrum is complicated presence of thallium, bismuth, and W0+ ions by the peaks of W0+ resulting from tertiary complicates the correction of the lead data. In processes, (Hess et al., 1951, p. 838), but this general the background was determined in the clears up with time if the acid concentration region above Bi209 and interpolated under the has been kept sufficiently low. Data must ei- lead peaks. In most cases the corrections were ther be taken on a ratio recorder or plotted on less than 2 or 3 per cent of the Pb204 peak and a linear time scale so that the slow drift in ion were accurate to about 10 per cent. intensity can be calculated from the data. The ion intensity from a surface ionization (6) After the data is calculated, corrections source drifts with time. To correct this the must be applied for the nonlinearity of the spectra are recorded on a linear time base so source and electron multiplier ion detector. that intensities can be interpolated. A ratio- These corrections must be determined by run- recording system (Stevens and Inghram, 1953, ning a standard isotope mixture. p. 987) has also been used on samples other The sulfide procedure is less reliable from the than those reported here. transfer standpoint since the sulfide must be The problem of isotopic fractionation during transferred as a slurry from the bottom of a the evaporation process is less serious, for with centrifuge tube. The sulfate can be formed by refractory salts evaporation takes place from a adding weak sulfuric acid to the filament after solid rather than from a liquid surface. Thus, the nitrate is loaded, but care must be taken to little mixing takes place and the fractionation avoid excessive attack of the filament. is very small compared with the theoretical maxima. Uranium Procedure The uranium procedure is similar to that RESULTS given for lead except that no borax is used and the filament material consists of outgassed tan- The uranium, thorium, and lead data are talum instead of tungsten. The most intense listed in Table 3. Assigned errors in the values peaks are UC>2+, when the single filament source for the composition of the lead and the amounts is used, and U+, when the multiple assembly of uranium and thorium in the composite gran- is used. Uranium ions are emitted at much ite samples are due to possible losses in dissolv- higher temperatures than are lead ions, and ing sphene and zircon, which comprise less than the currents are much more stable. half of one per cent of the sample. Measured isotope ratios have been assigned absolute er- Accuracy rors of one per cent which are the result of uncertainties in absolute corrections. An uncer- With great care, the accuracy of isotopic tainty of ±50 per cent was assigned to the analysis of uranium and lead by use of a mass amount of lead contamination introduced in spectrometer equipped with a surface ionization processing a sample. When the ratio of sample source is about 0.5 per cent of the abundance lead to contamination lead was small, probable of an isotope. This gives less accuracy than errors in the measured lead concentrations were does a gas source, since rapid intercomparison due mainly to uncertainties in the amounts of with identical ion geometries is not possible. lead contamination. This disadvantage, however, is counteracted by Independent analyses of the lead in three several factors: (1) There is no "memory" samples were made by C. L. Waring, of the whereby measurement of one sample is affected U. S. Geological Survey, using emission spec- by previous samples, a serious problem in the trography. Independent analyses of the tho- case of gas sources, usually limiting a machine rium in three samples were made by use of

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 TABLE 3.— URANIUM, THORIUM AND LEAD CONCENTRATIONS AND ISOTOPIC COMPOSITION OF LEAD IN VARIOUS CONSTITUENTS OF THE GRANITE Isotopic composition of leadb Isotopic composition of lead Mineral (atom ratios) (atomic per cent) 0 abundance Total U Total Th Total PB»>« in Sample Pbin ! (weight in PPM" in PPM" PPM» Pb»7 2 4 per cent) PPM Pb °» Pb*» Pb » Pba>« Pb»> Pb*°» Pb*« Pb»< Pb»« 0.04 ±0.01 2650 ± 40 2180 ± 20 461 392 ± 10 <0.01 85.35 ± 1.5 6.39 ±0.1 8. 26 ±0.15 Sphene Id 0.4±0.1 303 ±5 240 68.4 ± 3 39.13 17.11 79.38 0.732 28.65 12.52 58.10 Sphene IP 0.4 ±0.1 2205* 113 39.9 ±1 108.6 22.18 146.2 0.360 39.07 7.98 52.59 n 3170h 222 44.5 ±1 21.78 15.73 69.30 0.928 20.20 14.59 64.28 Magnetite 0.4 ±0.2 2. 57 ±0.04 1.4i 0.325 ±0.1 36.7 ±3 16. 8 ± 1 97.4 ±8 0.658 24.2 11.1 64.1 Apatite 0.02 ±0.01 90.5 ± 1.3 136k 34.5 ±4 31.9±5 16.3 ±3 76.3 ± 10 0.797 25.4 13.0 60.8 Pyrite 0.02 370" 20.34 15.92 54.34 1.09 22.21 17.38 59.32 Quartz 24 ±1 0.13 ±0.01 5.5m 1.36 ±0.05 20 ±1 0.20 ±0.02 3.8 0.94 ±0.1 18.16 15.48 40.02 1.34 24.32 20.74 53.60 Perthite 52 ±2 0.22 ±0.03 0.410 ±0.008 9.5 2. 35 ±0.1 18.58 15.74 39.52 1.34 24.82 21.03 52.81 2. 74 ±0.05 41. 88 ±0.60 9.3 2. 18 ±0.05 20.25 15.65 48.73 1.17 23.65 18.27 56.91 1.80 ±0.20 24.33' 5.6 1.35 ±0.1 21.20 15.56 49.35 1.15 24.33 17.89 56.63 8.4 2.05 ±0.05 16.81 15.28 36.02 1.45 24.32 22.11 52.12 (a) The errors assigned to the concentrations result from errors in the mass spectrometric and chemical procedures, uncertainties in the amounts of contamination, and nonreproducibility of leaching procedures. (b) Except where indicated, the absolute errors of the lead ratios are about one per cent of the values. (c) All the samples, except the unwashed composite, have been stirred with 6 M HC1, centrifuged, and rinsed with water. (d) A coarse, heterogenous sample; uranium and lead determinations were made on aliquots of single dissolved sample. (e) A finely ground, homogenous sample; thorium and lead determinations were made on separate solid samples. (f) Concentrations are given in terms of micrograms leached per gram of sphene. (g) Obtained by difference of thorium in acid leachings and thorium in untreated sample (5375 ± 75 ppm). (h) Colorimetric determination; yields determined with radioactive tracer. (i) Obtained by difference of thorium in acid leachings (17.55 ± 0.25 ppm) and thorium in untreated sample. (j) Errors are the result of low sample lead to contamination lead ratios. (k) Errors result from uncertainties in estimated corrections for unshielded electron multiplier and indirect measurements of composition. All of the sample was sacrificed during these initial and less accurate runs. At the time, it was thought that lead tracer was needed as a carrier as well, and two samples containing different proportions of tracer were used. (1) Colorimetric determination; yields determined with radioactive tracer. (m) The lead composition was assumed to be identical to that in the perthite. All of a large sample of highly purified material was sacrificed for a single concentration run before it was discovered that the lead concentrations were so high. In an initial run, in which a highly discriminating electron multiplier was used on a sample suspected of being contaminated, the quartz lead composition was different from the compositions of perthite and plagioclase lead samples run under the same conditions. (n) Fluorometric uranium determinations of unleashed and acid-leached sphene (finely ground) showed approximately 20 per cent of the uranium to be removed by the acid treatment. Analysts: F. Cuttitta, P. Scott and M. Delevaux, U. S. Geological Survey, Washington, D. C.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 DISCUSSION 1141

colorimetric procedures. The spectrographic, The studies that have been made on the spectrophotometric, and isotope dilution values sphene may, perhaps, be related to those of are compared in Table 4. the zircon. The loss of Pb208 from the dilute- A determination of the U238/U235 ratio in acid-washed sphene is extreme. The analysis uranium isolated from sphene was made for of the acid washings, given in Table 4, con- the purpose of ascertaining whether uranium in granite possesses the same isotopic composi- TABLE 4.—COMPARISON or LEAD AND THORIUM tion as uranium in ore deposits. In sphene, the VALUES OBTAINED BY DIFFERENT METHODS following ratio was found: Total lead Total thorium TJ238/U235 = 138.4 ± 0.6 (ppm) (ppm) Sample Emis- a value which agrees within experimental error Isotope sion Isotope Colon- dilution spectrog- dilution metric with the best values for ore uranium (Nier, raphy analysis 1939a, p. 151; Inghram, 1946, p. 35). Zircon 461 449 2170 2180 DISCUSSION Sphene I 240 245 Sphene II 5375 5550 Uranium-Lead Age of the Granite Composite granite 41.9 44.0 The ages of pegmatitic uraninites from Wil- Perthite 9.5 5 berforce, Haliburton Co. have been determined by Nier (1939b, p. 159) and by Collins, Lang, clusively demonstrates that small amounts of Robinson, and Farquhar (1952, p. 21). These uranium lead and major amounts of thorium are listed in Table 5 together with the ages for and thorium lead are very loosely bound, chem- zircon and sphene which were calculated by ically, and that there is a net deficiency of assuming the isotopic composition of the lead Pb208 in the mineral. These mineral grains are in sphene at the tune it was formed was that doubtlessly heterogeneous, and the mass of the found in the macro pegmatitic perthite crystal. phase containing loosely bound thorium and lead The following constants were used: is small compared to the mass of the entire grain. TjTJ236 = 7.13 X 108 yrs. (Fleming et al. 1952, The low thorium-lead age for the zircon could p. 642). have been the result of addition of thorium to TiTh232 = 13.9 X 10" yrs. (Kovarik and Adams, the mineral, which would not affect the ura- 1938, p. 413). nium-lead age seriously, since the Th/U ratio TlU238 = 4-51 x io» yrs. (Fleming et al., 1952, of the interstitial material of the rock is about p. 642). 24. The sphene data indicate, however, that U238/TJTO J37.8 (Inghram, 1946, p. 35). = the cause may be the loss of Pb208 instead. The errors associated with the lead data for Selective loss of Pb208 is possible if there has apatite and magnetite are so large that ages been a segregation of thorium and uranium in calculated from these values are not considered some regions of the mineral grains. Probably significant. part of the thorium is concentrated in the The close agreement of the U238/Pb206 age for skin or surface inclusions of the mineral where zircon with the U236/Pb207 and Pb207/Pb206 ages Pb208 has been lost. Preliminary studies have is quite significant in view of the highly meta- shown that the powdered mineral loses 35 per mict character of the zircon. These differences cent of its radioactivity and 30 per cent of its are within the effect of combined experimental lead when it is leached for 30 minutes in hot errors, which include an uncertainty of 2 per 1:1 aqua regia3. It may be noted in this con- cent in the half-life of U235. The obvious dis- nection that Buttlar and Houtermans (1951, crepancy of the Th^/Pb208 age for zircon re- p. 43) have demonstrated a fractionation of quires that there has been either addition of thorium without any significant addition of 3 Recent studies at the California Institute of 208 Technology have demonstrated that substantial uranium or loss of Pb without any loss of fractions of the radioactive elements can be leached Pb206 or Pb207. from granites with acid strengths as low as 0.1 M.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 1142 TILTON ET AL.—LEAD, URANIUM, AND THORIUM IN A GRANITE

thorium and uranium within a zircon crystal several phases of widely different chemical and 5 cm in diameter from Renfrew County, On- physical properties which are associated with tario, in which the Th/U ratio ranged from each mineral, and as a result, different amounts zero to two in different regions. of these elements have been lost during the isolation of each mineral in the laboratory. The TABLE 5.—AGES FOR THE GRANITE AND THE Tj235//pt,207 age of the zircon, 1060 m.y., is the WlLBERFORCE PEGMATITE most probable age of the granite and this (In millions of years) agrees with the ages determined for associated Pb!07 U235 USB Th»> pegmatites. Formation of the red granite which Pb20. Pb207 Pb20' Pb» marks the end of major igneous activity involv- ing the Grenville sediments (see Engel and Uraninite (a) 1015 1055 1060 1000 Engel, 1953) occurred 1060 m.y. ago and the Uraninite (b) 1025 associated pegmatites were formed at about the 1032 same time. 1090 Zircon 1090 1060 1030 390 Potassium-Argon Age of the Granite Sphene I (c) (c) 910 Sphene II 1090 450 Measurement of the amount of K40 and the 40 (a) Cardiff Township, Haliburton County, amount of radiogenic A in a rock together with the known values of the half life and Ontario (Wilberforce); Nier (1939) 40 (b) Same location as (a); Collins, Lang, Robin- branching ratio of K allows the potassium- son, and Farquhar (1952) argon age of the rock to be calculated. Assum- (c) The fraction of radiogenic Pb207 was too ing there has been no subsequent leakage or small to warrant age calculations involving that leaching of argon or of potassium from the rock, isotope. this age gives the length of time since the crys- tallization of the potassium-containing phases Previous thorium-lead age determinations of the rock. If potassium crystallizes in com- have been lower for the most part than ura- parable amounts and at different times in sev- nium-lead age determinations where the two eral phases of the rock, the potassium-argon methods have been applied to the same min- age of the gross rock is a weighted average of eral. In Nier's earlier work (1939b) and (1941), the ages of the components. the two methods agree for only three samples To measure the amount of radiogenic A40 in the thorium-lead ages are lower for seven and the Haliburton Co. granite, the gases in the samples. However, Holmes, Leland, and Nier granite were first released by adding, in a vac- (1950, p. 87) have summarized four additional uum, 9.24 g of granite into 30 g molten NaOH analyses in which the thorium-lead ages are which was contained in a nickel tube. The re- higher in three cases. It is perhaps significant sultant gases were mixed with 6.10 X 10~6 cc that all monazite minerals studied by Nier et at stp tracer argon which was 96.5 per cent A38 al. (1941) have given, as has the present work, and 3.5 per cent A40. The resultant mixture was radically lower thorium-lead ages where com- then purified by passing successively in a parison with uranium-lead ages is possible. greaseless system through traps of anhydrous Because the substantial differences which oc- Mg(ClO,t)2, heated CuO, vacuum moistened cur among some of the age values are well NaOH, anhydrous Mg(ClO4)2, and finally outside experimental error, the existence of through a calcium vapor furnace. The remain- loosely bound Pb208 demonstrated, and the ing gases were collected on 1.2 g of activated Pb207/Pb206 ages for all of the minerals agree coconut charcoal at liquid nitrogen tempera- well, we have concluded that: either (1) ura- ture. nium, thorium, and their decay products The argon so collected and so purified was have existed only with varying degrees of ap- introduced from a greaseless sample system proximation to closed systems within each into a 60° deflection, 6-inch radius of curva- mineral since its formation; or (2) these ele- ture gas analysis mass spectrometer, and its ments have a heterogeneous distribution in isotopic composition was measured. Results of

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 DISCUSSION 1143

the analysis were, after correction for spec- to some other ion the calculated value is once trometer discriminations but not for possible more too low. In addition to these possible impurities, that the relative amounts of the errors, all tending to lower the observed age, various argon isotopes collected were: small errors in either direction could be intro- duced in the measuring (using mass spectro- A36 = 0.0153; —40 = 0.0436 metric isotope dilution techniques) of the tracer A argon sample, in the chemical analysis, and in Two gravimetric analyses of different samples the mass spectrometric analyses. of the granite by Kenneth Jensen of the Ar- If the age of this sample as determined by gonne National Laboratory gave values of 4.82 the lead-uranium method is assumed correct, per cent and 4.86 per cent for the amount by and if no argon or potassium has been lost, this data may be used to compute an apparent K40 weight of K2O extracted from the granite. The average value of 4.84 per cent was used. branching ratio of 0.0587. This is well below In addition it was assumed that the relative the most reliable measured values and indicates abundance of K40 in potassium is 0.0119 per that leakage of argon is probably the source of cent (Nier, 1950, p. 793), that the half life of the discrepancy between this age and the lead- K40 is 1.27 X 10" yr (W. F. Libby, personal uranium ages. communication, 1952), that the branching ratio (number of K-captures) Loss or Gain of Uranium, Thorium, - — 3 for K40 is 0.085 (Was- (number of beta decays) and Lead in the Granite serburg and Hayden, 1955, p. 51), and that the If there has been no transfer or exchange of normal abundances of the argon isotopes are uranium, thorium, and their decay products A36 = 0.337%, A38 = 0.063%, A40 = 99.600% between the granite and its surroundings, the (Nier, 1950, p. 792). These normal abundances isotopic composition of lead in the granite at were checked to 1 per cent on the mass spectrom- the time it was formed can be calculated from eter with which the analyses were made. These the total concentrations of uranium, thorium, data indicated that the age of the sample is 800 and lead and the isotopic composition of lead m.y. Approximately 80 per cent of the A40 meas- now present in the rock. The initial composi- ured by the mass spectrometer was radiogenic. tion of the lead, after correcting for decay over Possible sources of error are numerous. First, a period of 1050 m.y., is calculated to be: A40 and/or K40 may have leaked or been leached out of the rock over periods of several hundreds Pb206/pb204 = 16 4 of millions of years. Argon leakage would make Pb»7/PbiM = is.4 the observed age too low; loss of potassium Pb208/pb204 = 30.2 would make the observed age too high. Sec- Of all the constituents of the rock that were ondly, although almost no rock residue was completely studied, only the perthite and the found in the NaOH after the gas extraction, plagioclase have ratios of uranium and thorium still all the rock argon might not have been to lead such that no appreciable addition of able to mix with the tracer argon owing to oc- radiogenic lead would occur in a billion years. clusion in the nickel, in the molten NaOH, or The average isotopic composition of the lead on the glass of the vacuum system. This would in these minerals is: lower the apparent age. Thirdly, the small 36 Pb206/pb204 = lg 3 peak observed in the mass spectrometer might 204 36+ Pb^VPb = 15.6 in part be due to ions other than A . For ex- p a>8/py»4 _ 39 g ample, HC135+ frequently gives peaks at this b mass. This particular contamination seems un- which values are considerably more radiogenic likely in this instance since no Cl+ peaks were ob- than the initial ones calculated. This marked served. The 36 peak was observed to decay with discrepancy cannot be resolved by analytical time proportionately to the 38 and 40 peaks, errors and can only result from either the loss but this is not a positive proof that it was com- of uranium and thorium and the gain of lead posed solely of A36. If part of the 36 peak is due by the granite, or the internal transfer of ura-

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 1144 TILTON ET AL.—LVAV, URANIUM, AND THORIUM IN A GRANITE

nium, thorium, and their decay products among may be quite complicated; indeed the differ- the mineral constituents of the granite. To dis- ences between the isotopic compositions of cover the cause, some lead isolated from an these leads may be highly significant. However, associated pegmatite mineral was studied. If there is a general agreement between the ob- both the granite and the associated pegmatites served isotopic compositions of uranium leads in this area originated from a common source formed in this region 1000 m.y. ago and those and were formed at nearly the same time, the calculated for the lead originally present in the initial isotopic composition of lead in both granite when it was formed, and a corresponding should have been the same. Furthermore, the disagreement for the thorium leads. It would amount of internal transfer of trace constitu- appear that the sample of granite studied has ents among very large pegmatitic minerals closely approximated a closed system since it should be much less than among the relatively was formed with respect to uranium and its fine-grained granitic minerals. Lead, isolated decay products, but has been an open system from a single large perthite crystal which was with respect to thorium and its decay products. taken from an associated pegmatite,4 was found to have an isotopic composition of: Distribution of Uranium, Thorium and Lead Pnm/PbM = 16.81 in the Granite 15.28 Marked differences exist in the distribution 36.02. of uranium, thorium, and lead in the granite. Determinations of the amounts of uranium and Most of the stably bound uranium and tho- thorium in this mineral are in progress; how- rium was concentrated in the accessory min- ever by analogy with previous data they are erals zircon, sphene, and apatite. This result is presumed to be insignificant. As an indication consistent with the observations of previous that the isotopic composition of this pegma- workers, notably Larsen, Keevil, and Harrison titic lead is not anomalous, a comparison may (1952) who measured the distribution of radio- be made with the isotopic compositions of some activity in mineral separates from a number of lead ores formed at the same time in this gen- granitic rocks. In contrast, most of the stably eral region. The mean composition of the Tet- bound nonradiogenic lead, indicated by the 204 reault, Ontario lead ore (Nier et al., 1941, p. concentration of Pb , was found to reside in 115) and the Frontenac County, Quebec, lead the feldspars and quartz, resisting strongly any ore (Collins et al., 1953, p. 410), both formed tendency to concentrate in zircon or magnetite. 1000 m.y. ago, is: Thus, there are enormous variations in the 204 ratios of uranium and thorium to nonradiogenic Pb^/Pb = 16.53 lead in the various minerals of the granite. The 15.28 35.55 ratios range from values so low that the iso- topic composition of the lead would not have The relationship among the leads in the peg- been altered appreciably by radioactive decay matitic perthite, in the ores, and in the granite in a billion years, to values so high that the 4 The perthite specimen was collected in June of lead found is almost entirely of radiogenic 1952 by H. S. Armstrong while in charge of a field origin. party for the Ontario Department of Mines. The A substantial fraction of the uranium and crystal was about 2 inches in diameter and 6 inches long and was taken from a pegmatite cutting a even greater fractions of thorium and non- crystalline limestone-biotite paragneiss complex radiogenic lead were found to exist in chemi- north of Tory Hill, Monmouth Twp., Ontario. This locality is a new (summer, 1952) road clear- cally unstable environments in the rock. Ura- ing and can be located on Satterly's map, No. 52a, nium, thorium, and lead thus present are so Haliburton Area, published by the Ontario Depart- loosely bound that they can be removed from ment of Mines. It is immediately adjacent to the road shown on this map. The specimen was taken the pulverized rock by a 5 minute stirring with a few hundred feet north of the jog in the road be- 6 M hydrochloric acid in the cold. The fractions tween the double "1" in "Hill" of Tory Hill and the of the uranium, thorium, and lead in the com- "R" of the "Ra" occurrence, the red dot of which is superimposed on two black squares representing posite rock observed to be extracted by this houses. weak acid treatment were approximately 34 per

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 DISCUSSION 1145

cent, 42 per cent, and 40 per cent, respectively. thorium, and uranium concentrations in the The different uranium and thorium concentra- acid-washed rock. This lack of material balance tions found in the untreated and in the acid- may have resulted from one or more effects. washed granite agree with the observations of Errors in estimating mineral abundances could Hurley (1950, p. 3), and Picciotto (1950, TABLE 7.—COMPARISONS BETWEEN THE OBSERVED TABLE 6.—CONTRIBUTION BY EACH MINERAL TO AND CALCULATED AMOUNTS OF RADIOGENIC THE URANIUM, THORIUM AND LEAD CONTENT LEADS IN VARIOUS CONSTITUENTS OF THE or THE TOTAL GRANITE GRANITE (Parts per million of total rock) Ratio of Ratio of radiogenic Pb2°« radiogenic Pb208 2 2 found to found to Acid washed Pb204 Pb!06 PbM! U " Th* Sample radiogenic Pb206 radiogenic Pb308 minerals calculated from calculated from uranium present thorium present Zircon 0.0000 0.157 0.02 1.06 0.87 Sphene 0.0069 0.273 0.24 1.21 8.8 Leachings from 0.45 Apatite 0.0002 0.007 0.16 0.02 sphene (II) Magnetite 0.0000 0.001 0.00 0.01 Sphene (II) 0 43 Perthite 0.0652 1.222 2.61 0.11 0.21 Sphene (I) 0.84 Plagioclase 0.0103 0.189 0.41 0.04 Zircon 0.97 0.36 Quartz 0.0178 0.326 0.70 0.03 Leachings from com- 0.60 0.62 posite rock Total. . . 0.1004 2.18 4.14 2.48 (9.9) Plagioclase 2.3 Perthite 6.4 17.1 Acid-washed 0.0629 1.35 3.16 1.80 24.33 Pyrite (Probably rock »!)• " The amount of radiogenic Pb208 present in the p. 175). Picciotto, using refined radio-auto- acid washed mineral would require about 0.2 per graphic techniques, has found that a large frac- cent thorium, which is improbably high. tion of the radioactive elements present in some granitic rocks reside in the mineral interstices cause the discrepancy for uranium or thorium, and intra-crystalline fractures, most of the re- but it is not possible to reduce the calculated maining fraction residing in accessory minerals, concentrations for lead sufficiently by changing and on this basis it may be assumed that the the mineral abundances by any reasonable loosely bound uranium, thorium, and lead orig- amount. The discrepancies may be related to inate from the inter-crystalline interstices. The the fact that the mineral separates consisted presence in the interstices of such large fractions of particles that were considerably larger on the of trace elements with contrasting lattice dis- average than the same unseparated minerals in tributions suggests that interstitial material, the composite. There may be a variation either like pegmatitic veins, may contain high concen- in trace element concentration or in leaching trations of other trace elements. effects with particle size. The small composite samples may not have been representative of Material Balance the very large sample which was the source of the separated minerals. Table 6 shows the contribution made by each mineral to the uranium, thorium, and lead Internal Transfer of Uranium, Thorium, content of the total granite, as determined by and Lead multiplying the concentrations of uranium, tho- rium, and lead in each mineral by the weight- One may calculate the amount of lead gen- fraction of the mineral in the original rock. In erated in 1050 m.y. by the decay of the ura- all cases the sums of the contributions of lead, nium and thorium in each mineral and compare uranium, and thorium in the individual acid- these values with the observed radiogenic lead washed minerals are different from the lead, in each mineral which is found by subtracting

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 1146 TILTON ET AL.—LEAD, URANIUM, AND THORIUM IN A GRANITE

the nonradiogenic lead originally present. The to the time it was formed and at earlier times composition of this nonradiogenic lead was as- in the material from which the rock was de- sumed equal to that in the pegmatitic perthite. rived, it is found that lead ores formed from The comparisons listed in Table 7 are made by representative samples of this lead would have dividing the amount of radiogenic lead found the compositions at various times given in in each mineral by the corresponding amount Table 8. of radiogenic lead that should have been gen- erated by the uranium or thorium present in TABLE 8.—CALCULATED ISOTOPIC COMPOSITIONS or each mineral. LEAD IN THE GRANITE AT VARIOUS TIMES All these ratios deviate from unity, and it is Time (m.y. ago) Pb!0«/Pb204 Pb»VPb*M Pb»/Pb*M clear that in no single constituent of the rock that has been isolated in the laboratory and 0-50 20.3 15.7 48.7 examined, has radioactive decay operated 1060 16.4 15.4 30.2 within a completely closed system. Both defi- 3500 4.7 11.1 «0) ciencies and excesses of radiogenic lead are demonstrated for systems of uranium, thorium, The average composition of lead ores formed and lead existing in stably bound, or supposedly within the last 50 m.y., formed 1050 m.y. ago, intra-crystalline environments. Similarly, defi- and formed 3500 m.y. ago (determined by least ciencies of radiogenic lead are demonstrated for squares analyses of the variations of the iso- systems of uranium, thorium, and lead existing topic compositions of lead ores with time; in chemically unstable, loosely bound, or sup- Collins et al., 1953, p. 413), and the isotopic posedly interstitial environments that have very composition of meteoritic lead (Patterson et al., small masses relative to the masses of the 1953, p. 1234) are given in Table 9. minerals or the rock as a whole. A satisfactory Since it is improbable that a sample of ter- explanation of these effects is not known at restrial lead can be less radiogenic than pri- present; however, it seems apparent that either M 204 mordial lead, and since the Pb VPb and complete chemical systems were not isolated in 207 204 208 204 Pb /Pb ratios, as well as the Pb /Pb the laboratory from various parts of the granite ratio, are improbably low for lead isolated from or transfer of uranium, thorium, and lead has the granite source material 3500 m.y. ago, it is occurred on a micro scale within the granite. apparent that much lower ratios of uranium It is probable that extensive transfer on a macro and thorium to lead were present for a consid- scale has also occurred for thorium or its decay erable period of time in the granite source products. material as a whole than now exists in the granite itself. In effect, a pronounced enrich- Relation to Lead Ores ment of uranium and thorium relative to lead In connection with discussions of the world- occurred when the granite was formed. It is wide variations of the isotopic composition of possible that ore leads may be representative lead ores with time, a number of writers, of the isotopic composition of the lead in the notably Holmes (1947), Houtermans (1947), material which was a source of both the ores Bullard and Stanley (1949), and Collins, and their associated rocks at the time they Russell, and Farquhar (1953), have assumed were formed, but it is obvious that the isotopic that lead ores are products of representative compositions of the ore leads and leads from isolation of lead at different times from rocks this granite are not directly related after the where, in any given area, the relative ratios of time of formation of the granite. Such differen- uranium and thorium to lead have not changed tiation processes may invalidate the assump- except by decay processes since the earth's tion that lead ores, which are formed at dif- crust was formed. If the relative ratios of ura- ferent times at different places in the earth's nium and thorium to lead and the isotopic com- crust, may be directly correlated by a single position of the lead now present in the granite genetic history. are used to determine the variation of the iso- The physical and chemical processes by which topic composition of the lead, both in the rock an ore lead is formed may be extremely impor-

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 DISCUSSION 1147

tant in relation to its isotopic composition, in REFERENCES CITED view of the marked variations of the isotopic Adams, F. D., and Barlow, A. E., 1910, Geology of composition of lead found to exist in the dif- the Haliburton and Bancroft areas, Province ferent components of this granite and the large of Ontario: Ottawa Govt. Print. Bur., Canada Dept, of Mines Mem. No. 6, 419 p. variations in the chemical stabilities associated Allen, J. S., 1939, The detection of single positive ions, electrons and photons by a secondary TABLE 9.—AVERAGE COMPOSITION OP LEAD ORES electron multiplier: Phys. Rev., v. 55, p. 966- 971. AT VARIOUS TIMES* Aston, F. W., 1933, The isotopic constitution and 1 !07 ! 2 atomic weight of lead: Proc. Roy. Soc. Lond., Time (m.y. ago) Pb'M/Pb " Pb /Pb °< PbM/Pb " v. 140A, p. 535. Bamback, K., and Burkey, R. E., 1942, Micro 0-50 18.5 15.6 38.4 determination of lead by dithizone: Anal. 1060 16.7 15.5 36.3 Chem., v. 14, p. 904-907. Blaedel, W. J., Post, Walling, and Sedlet, 1945, 3500 11.4 13.5 31.1 Atomic Energy Comm., Rept. CN 2766. Meteoritic 9.4 10.3 29.2 Bullard, E. C., and Stanley, J. P., 1949, The age of troilite the earth: Veroffentl. d. Finn. Geodatischen Institutes No. 36, p. 33-40. * Ore values calculated from the data of Collins, Buttlar, H. V., and Houtermans, F. G., 1951, Photographische Messung des U and Th Russell, and Farquhar (1953) gehaltes nach der Auflage methode: Geochim. et Cosmochim. Acta, v. 2, p. 43-62. Clifford, P. A., and Wickmann, N. J., 1936, Dithi- with these components. The large variations zone methods for the determination of lead: J. of the isotopic composition of the Sudbury ore Assoc. Official Agr. Chem., v. 19, p. 140-145. leads, reported by Russell, et al. (1954, p. 307), Cohen, A. A., 1943, The isotopes of cerium and rhodium: Phys. Rev., v. 63, p. 219. may illustrate the result of the discriminatory Collins, C. B., Freeman, J. R., and Wilson, J. T., action of such processes upon a system similar 1951, A modification of the isotopic lead method to the one studied here. Considerable caution for determination of geological ages: Phys. Rev., v. 82, p. 966-967. should be exercised in assuming mechanisms of Collins, C. B., Lang, A. H., Robinson, S. C., and ore formation that ignore these possibilities. Farquhar, R. M., 1952, Age determinations for some uranium deposits in the Canadian Shield: Proc. Geol. Assoc. Canada, v. 63, p. CONCLUSIONS 15-41. Collins, C. B., Russell, R. D., and Farquhar, R. M., The present work has demonstrated that a 1953, The maximum age of the elements and the age of the earth's crust: Can. Jour. Physics, study of the amount of uranium, thorium, and v. 31, p. 402^18. the isotopes of lead in granites and their mineral Engel, A. E. J., and Engel, C. G., 1953, Grenville Series in the N. W. Adirondack Mountains, constituents may help to trace rock histories N. Y.: Geol. Soc. America Bull., v. 64, p. 1013. and to explain related geochemical phenomena. Fleming, E. H., Jr., Ghiorso, A., and Cunningham, Since the U/Pb and Th/Pb ratios may vary B. B., 1952, The specific activities and half- lives of U234, U236, U238: Phys. Rev., v. 88, p. widely from one mineral to another, the var- 642-652. ious leads will develop different isotopic com- Hagemann, French, 1950, The isolation of actinium: positions with time. This "labels" the lead in J. Am. Chem. Soc., v. 72, p. 768-771. Hayden, R. J., Reynolds, J. H., and Inghram, M. one phase of the rock with respect to that in G., 1949, Reactions induced by slow neutron another, and thus provides internal lead irradiation of europium: Phys. Rev., v. 75, p. tracers. This makes it feasible to apply criteria 1500-1507. Hess, D. C., Weatherill, G., and Inghram, M., 1951, and study aspects of problems which are im- Suppression of spurious ions in the mass possible with elements that are not involved in spectrometer: Rev. Sci. Inst., v. 22, p. 838-839. radioactive decay processes. Holmes, Arthur, 1947, A revised estimate of the A co-ordinated study of uranium, thorium, age of the earth: Nature, v. 159, p. 127-130. Holmes, A., Leland, W. T., and Nier, A. O., 1950, and lead in a variety of igneous and sedimen- The age of uraninite from Gordonia, S. Africa: tary rocks by the same methods will permit Am. Jour. Sci., v. 248, p. 81-94. conclusions concerning the past history of these Houtermanns, F. G., 1947, Das Alter des Urans: elements in geochemical cycles and possible Zs. f. Naturforschung, v. 2a, p. 322-328. Hurley, P. M., 1950, Distribution of radium in magmatic fractionations which have taken granites and possible relation to the age meas- place in the earth's crust. urement: Geol. Soc. America Bull., v. 61, p. 1-7.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021 1148 TILTON ET ^i.—LEAD, URANIUM, AND THORIUM IN A GRANITE

Hurley, P. M., and Fairbairn, H. W., 1953, Radia- Patterson, C. C., Brown, H. S., Tilton, G. R., and tion damage in zircon: A possible age method: Inghram, M. G., 1953, Concentration of ura- Geol. Soc. America Bull., v. 64, p. 659-674. nium and lead and the isotopic composition of Inghram, M. G., 1946, Vol. 14, Manhatten Project lead in meteoritic material: Phys. Rev., v. 92, Technical Series, Div. 2, Gaseous Diffusion p. 1234-1236. Project, Chap. V., p. 35. Picciotto, E. E., 1950, Distribution de la Radio- Inghram, M. G., Brown, H., Patterson, C., and activit^ dans les Roches firuptives: Bull, de la Hess, D. C., 1950, The branching ratio of K« Soc. beige de Geol. de Paleontol. et. d'Hydrol., radioactive decay: Phys. Rev., v. 80, p. 916. v. 59, p. 170-199. Inghram, M. G., and Chupka, W. A., 1953, Surface Russell, R. D., Farquhar, R. M., Cumming, and ionization source using multiple filaments: Rev. Wilson, J. T., 1954, Dating galenas by means Sci. Inst., v. 24, p. 518-520. of their isotopic compositions: Trans. Am. Kovarik, A. F., and Adams, H. L., Jr., 1938, The Geophys. Union, v. 35, no. 2, p. 301-309. disintegration constant of thorium and the Sandell, E. B., 1950, Colorimetric determination of branching ratio of ThC: Phys. Rev., v. 54, p. traces of metals: N. Y., Interscience Pub- 413-421. lishers, Inc., 673 p. Larsen, E. S., Keevil, N. B., and Harrison, H. C., Satterly, J., 1943, Mineral occurrences in the 1952, Method for determining the age of Haliburton area: Ontario Dept, of Mines, 52nd igneous rocks using the accessory minerals: Ann. Rept., v. 52, pt. 2, TV. Geol. Soc. America Bull., v. 63, p. 1045-1053. Stevens, C. M., and Inghram, M., 1953, Ratio re- Levine, H., and Grimaldi, F. S., 1951, Application cording in isotopic analysis: Rev. Sci. Inst., v. of mesityl oxide to the determination of 24, p. 987-989. thorium: Atomic Energy Comm. Rept. AECD Thomason, P. F., Perry, M. A., and Byerly, W. M., 3186. 1949, Determination of microgram amounts of thorium: Anal. Chem., v. 21, p. 1239-41. Lundell, G. E. F., and Knowles, H. B., 1920, The Tilton, G. R., 1951, The distribution of trace quan- determination of zirconium and titanium in tities of uranium in nature: Atomic Energy zirconium ores: J. Amer. Chem. Soc., v. 42, p. Comm. Rept. AECD 3182. 1439-1448. Tilton, G. R., Aldrich, L. T., and Inghram, M., Nier, A. 0., 1939a, The isotopic composition of 1954, Mass spectrometric determination of uranium and the half-lives of the uranium thorium: Anal. Chem., v. 26, p. 894-898. isotopes. I: Phys. Rev., v. 55, p. 150-153. Wasserburg, G. J. and Hayden, R. J., 1955, A4»-K«) •— 1939b, The isotopic composition of radiogenic Dating, Geochim. et Cosmochim. Acta, v. 7, leads and the measurement of geological time. no. 1, p. 51-60. II: Phys. Rev., v. 55, p. 153-163. DEPARTMENT OF TERRESTRIAL MAGNETISM, CAR- — 1950, A redetermination of the relative abun- NEGIE INSTITUTION or WASHINGTON, WASHING- dances of the isotopes of carbon, nitrogen, TON, D. C.; CALIFORNIA INSTITUTE or TECH- oxygen, argon and potassium: Phys. Rev., v. NOLOGY, PASADENA, CALIF.; CALIFORNIA 77, p. 789-793. INSTITUTE OF TECHNOLOGY, PASADENA, CALIF.; Nier, A. O., Thompson, R. W., and Murphy, B. F., DEPARTMENT OF PHYSICS, UNIVERSITY OF 1941, The isotopic composition of lead and the CHICAGO, CHICAGO, ILL.; ARGONNE NATIONAL measurement of geological time: Phys. Rev., LABORATORIES, CHICAGO, ILL.; ARGONNE NA- v. 60, p. 112-116. TIONAL LABORATORIES, CHICAGO, ILL.; U.S. Patterson, C. C., 1951, The isotopic composition of GEOLOGICAL SURVEY, WASHINGTON, D. C. trace quantities of lead and calcium: Atomic MANUSCRIPT RECEIVED BY THE SECRETARY OF THE Energy Comm., Rept. AECD 3180. SOCIETY, APRIL 24, 1954.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/66/9/1131/3431742/i0016-7606-66-9-1131.pdf by guest on 27 September 2021