Lawrence Berkeley National Laboratory Recent Work

Title RADIOISOTOPE DATING WITH ACCELERATORS

Permalink https://escholarship.org/uc/item/0kh119x5

Author Muller, R.A.

Publication Date 1979-02-01

eScholarship.org Powered by the California Digital Library University of California Published in Physics Today LBL-8331(I, . ~ Preprint

RADIOISOTOPE DATING WITH ACCELERATORS

ECEIVED Richard A. Muller ."' lAWRENCE t!1:~K61:t'l I.ABORATORY MAY 30 \919 February 1979 LIBRARY AND OGC:UME.NTS SE.CTION

Prepared for the U. S. Department of Energy under Contract W-7405-ENG-48

TWO-WEEK LOAN COpy

This is a Library Circulating Copy which may be borrowed for two weeks. (l For a personal retention copy, call Tech. Info. Dioision, Ext. 6782 DISCLAIMER

This document was prepared as an account of work sponsored by the United States Government. While this document is believed to contain correct information, neither the United States Government nor any agency thereof, nor the Regents of the University of California, nor any of their employees, makes any warranty, express or implied, or assumes any legal responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by its trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof, or the Regents of the University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof or the Regents of the University of California. Radioisotope dating with accelerators

Counting accelerated ions rather than decay events increases Jhe sensitivity by several orders of magnitude so we can find the ages of much older and smaller samples.

Richard A. Muller

A new method of detecting radioactive materials because therat'io of C14 to C12 technique is to take advantage of this very isotopes promises to have a revolutionary is approximately constant with time for large number, and to estimate the ratio impact on the field ofradioisotope dating. carbon in the biosphere. This ratio is 04/C12 in the sample by counting atoms The technique, which was developed over determined by the equilibrium between rather than decays. Even with losses due the past two and a half years, consists of the production of C14 by cosmic rays, and to inefficiency in accelerating gas atoms counting individual atoms of radioactive its decay with a half-life of 5730 years. into a beam (0.25 to 10-5) the accelerator isotopes that have been ionized, acceler­ When a living object is taken out of equi­ method allows one to use much smaller ated to high energies, and then selected librium with the biosphere (when it dies) samples and to measure much older dates. and identified. By detecting all-or a the C14 continues to decay without being Dates as old as 40 000 years have now substantial fraction-of the atoms in the replaced by fresh atmospheric C14. The been measured with only 0.015 gram of beam, this method has much greater initial rate is 14 decays per minute per carbon (as compared with the 1 to 10 sensitivity than the standard method, gram of carbon, and decreases with time, grams that are necessary for decay de­ which detects only the tiny fraction of the t, as 2-t /5730 yrs = e-t/8270 yrs. Measure- tection), and soon one may be able to date atoms that decay during the counting ment of the decay rate gives an estimate back 100000 years. period. The new "direct detection" of the "age" of the sample; the use of 0 4 Other cosmic-ray-produced radioiso­ method therefore will allow one to use for this pUrpose was developed by Willard topes3 accessible to the new method in­ much smaller samples and to measure F. Libby and coworkers1 in the late 1940's. clude H3, Be10, A126, Si32, Cp6, Ar39 and much greater ages than the older "decay Alternately, with objects of known age Kr81. The longer the half-life, the greater detection" method. (such as tree rings) one can study varia­ is the advantage of direct detection over One remarkable fact about direct ra­ tions in the intensity of the cosmic rays or decay detection. For example, in the case dioisotope detection with accelerators is in the amount of diluting 0 2; such studies of Be10 (half-life of 1.5 X 106 yr) one decay that six: groups have already made very are yielding a fascinating record of the per minute-implies 1012 atoms of Be10 in sensitive measurements with virtually climate history of the Earth over the last the sample. Beryllium-10 may prove to unmodified existing accelerators. Dedi­ 8000 years.2 be a particularly useful isotope in the cated accelerators are now being planned study of sedimentary geology. It is pro­ at Oxford University, the University of Decay detection duced by cosmic rays and some of it set­ Rochester and the University of Arizona. Modern carbon-14 laboratories detect tles in the oceans and is trapped in the Several additional groups either plan to the disintegration of the remaining C14 accumulating sediments. Although it has begin measurements with existing accel­ atoms, typically by oxidizing the carbon been detected in both sedimentary de­ erators, or have already begun. Samples to CO2 and counting the decays in a pro­ posits and manganese nodules, its general a thousand times smaller than previously portional counter. Because of the low usefulness has been severely limited by required have already been used to obtain decay rate, the undistinctive nature of the the very low decay rate. Potential ap­ \1" dates. The radioisotopes carbon-14, tri­ decay (the emitted electron has a broad plications of the other radioisotopes span ~ tium, beryllium-10, and chlorine-36 have beta-decay spectrum with an average many areas of research, from astrophysics served for measurements inconceivable energy of only 45 ke V) and problems with to theology. (It with the older methods. background radioactivities, typically 1 to Carbon-14 is the most famous of the 10 grams of carbon must be counted for 1 History of direct detection dating radioisotopes. It can be used for to 10 hours or more. Except in unusual Perceptive scientists have recognized determining the ages of formerly living cases when large amounts of material are for decades that direct detection is po­ available and long counting periods taken, tentially more sensitive than decay de­ one cannot obtain dates older than 30 to Richard A. Muller is an associate professor of tection, and that the problem was to in­ physics at the University of California at 40 thousand years. vent a practical method for accomplishing Berkeley. He holds a joint appointment at the For every decay per minute, there are it. The key to the solution was the use of Lawrence Berkeley Laboratory and the Space 4 X 109 atoms of carbon-14 in the sample. high-energy beams. In this section I will Science Laboratory. The fundamental idea of the accelerator try to reconstruct the history of the cross

0031·9228/79/020023-08/$00.50 © 1979 American Instftute of Physics Reprinted from Physics Today fertilization that brought nuclear physi­ cists into the radioisotope-dating busi­ ness. I have discussed this history with, and corresponded with, most of the sci­ entists who played a major role in the so­ lution of the problem, and although memories differ on some points and opinions differ on the relative importance of some events, I will do my best to relate the history leading directly to the use of high-energy beams for dating and to the successes of the past two and a half years. One of the early sensitive attempts at direct detection was made by Michael Anbar and coworkers4 at SRI Interna­ tional, using a mass spectrometer spe­ cially designed for that purpose. Unfor­ tunately, a mass spectrometer has diffi­ culty distinguishing the radioactive ion from background ions with the same charge-to-mass ratio. For example, if (CI4)6+ ions are used, then there are in­ evitably (N14)6+ ions in the beam in suf­ ficient numbers to swamp the carbon. Anbar used (C14NI5)- molecular ions, and although his group solved many problems, they were ultimately limited by back­ ground (Si29)- and lack of funds; they never achieved sufficient sensitivity to work at this time, I discovered later that at this meeting and afterwards they and detect C14 at natural levels. my familiarity with the need for direct several others discussed informally the The clue to direct detection lay in the detection was due to a colleague who had possible use of a tandem Van de Graaff revival of the method of high-energy mass read Anbar's proposal.) In a colloquium accelerator for direct detection of car­ spectrometry by Luis W. Alvarez in a at the Lawrence Berkeley Laboratory in bon-14. They had already been thinking search for quarks of integral charge, which July 1976, r reported the first age ob­ about direct detection for some time. he proposed to our group in 1974. In our tained using the new idea: we had dated Earle Nelson of Simon Fraser University search we used the 88-inch cyclotron at a sample of deuterium known to be about heard of Middleton's thoughts from Berkeley as a mass spectrometer. (The 25 years old (about one tritium mean-life) Gordon Brown. He immediately became only prior use of a cyclotron for mass by measuring the ratio of tritium to deu­ interested in the application of accelera­ spectrometry, as far as I know, was re­ terium. I also described the techniques tors to dating, and he says that he decided ported by Alvarez and R. Cornog in that would be required for the direct de­ to pursue the technique after learning 1939.5) Although the cyclotron is an ex­ tection of beryllium-IO, carbon-14 and from my article8 how particle-identifica­ tremely good mass spectrometer (resolu­ other natural radioisotopes. Attending tion techniques could be used to eliminate tion of 10-3 to 10-4 with negligible tails at the colloquium was Grant M. Raisbeck, the expected backgrounds. Nelson, beam currents of 10 microamp), the key who initiated a program using a cyclotron Ralph Korteling and William Stott made feature was the high energy of the for direct Be lO detection upon his return extremely sensitive measurements of C14 emerging beam, whi~h allowed the atomic to France. Edward Stephenson and at natural levels, in their first experiment number Z of the individual atoms to be Terry Mast joined the effort at Berkeley, at the McMaster tandem.9 determined through ionization measure­ and together we worked on the develop­ A month after the archeometry con­ ments. Thus potential "quarks" could be ment of the beam line and particle de­ ference, Litherland met with Purser and distinguished from background ions of tectors necessary for the efficient detec­ Harry Gove, of Rochester, at the Wash­ higher charge on a particle-by-particle tion of BelO and C14. ington meeting of The American Physical basis. About a year after the quark A key event in the next phase of ex­ Society, and they planned an experiment search was begun, Kenneth Purser6 of the periments was an archeometry conference at the Rochester tandem. (They were not General Ionex Corporation proposed and in Pennsylvania in March 1977. Among aware of our paperlO at the same confer­ patented a system using a tandem Van de the speakers there familiar with our work ence in which we described our first re­ Graaff accelerator combined with a mass was R.E.M. Hedges, who spoke about the sults with carbon-14, nor had they seen spectrometer for the detection of atmo­ plans at Oxford for direct detection of my article, which had appeared in Science spheric chlorine and fluorine com­ carbon-14 using a system that combines several days earlier.) Others soon joined pounds. laser enrichment with a mass spectrom­ this Toronto- General Ionex-Rochester - The quark search concluded in 1976 eter. Hedges had spent several months ("TIR") collaboration, and in their first with the result? that the ratio of integrally in Berkeley collaborating with C. Bradley experimentll they detected 0 4 with great charged quarks to other singly charged Moore on isotope enrichment; although sensitivity. Within a short time they - nuclei (hydrogen) was less than 3 X 10-19 he had visited our laboratory and read my were dating objects known to be as old as in the mass range of 3 to 8 amu. I was preprint8 it was not yet obvious that the 40000 years, using samples of 15 mg and struck by this remarkable sensitivity, mass-spectrometer section of his plans less. 12 much better than that of virtually any should be replaced by an accelerator, The quick successes of the three groups other method of trace-element analysis. especially since we had not yet obtained inspired others to try, and in April 1978 a In looking for other potential applica­ a date using carbon-14. topical conference was held at Rochester13 tions, I realized that the new technique According to Roy Middleton of the at which six groups reported results. The could be used for radioisotope dating. University of Pennsylvania and Ted TIR group described new measure­ (Although I was "unaware" of Anbar's Litherland of the University of Toronto, mentsll with C1 4 and C136; our group in

24 PHYSICS TODAY / FEBRUARY 1979 Steering magnet

Particle filters and identifiers of the Lawrence Focussing magnet Berkeley cyclotron dating system (opposite page). Richard Muller (left) and Terry Mast are checking out the system. The large box in the center contains the xenon range filter, whose pressure is adjusted so that nitrogen-14 ions are absorbed while carbon-14 ions pass through to be counted. The filter reduces the nitrogen background by better than 10-14. Figure 1

The dating system at the Lawrence Berkeley Laboratory's 88-inch cyclotron (this page). Positive ions are accelerated by the cyclotron, which also acts as a q/ m filter. The range filter removes much of the background ioncurrent, and the remaining particles are identified by their Cyclotron energy-loss rate and total energy. The diagram L-______--' is only schematic. Figure 2

Berkeley presented resultsl4 with C14, through ionization and kinetic-energy tion is required. In the experiments done BelO , and C136. Nelson and his col­ measurements, and so far with H3, Be IO, Cl4 and Cl36 this leaguesl5 described their ongoing work in ~ normalization with respect to a stable separation has been accomplished by the Canada; G. Raisbeck and his collaborators isotope. range-filter method described in reference at Orsay and Grenoble reportedl6 mea­ No separate stage of mass spectrometry 7, although other approaches are possible. surements of BelO with the Grenoble cy­ is needed; the acceleration itself serves (See figures 1 and 2). clotron. Hedges and collaboratorsl7 at that purpose: only those particles with The range filter is based on the obser­ Oxford had incorporated an existing the proper charge-to-mass ratio (qlm), vation that the range of an ion in matter tandem in their system, and without iso­ given by the cyclotron resonance equa­ is approximately proportional to 11Z 2 tope enrichment had detected Cl4 at tion, are accelerated. The ionization where Z is the ion's atomic number. natural levels; and Robert Andrews and source can be external to the cyclotron, or Thus the range of H3 is approximately his coworkersl8 had detected natural Cl4 located at t.he center. The sample can be four times that of the background BlO; with the Chalk River tandem. Optimism introduced as a gas (C02 or CH4 for Cl4 and the range of Cl4 is approximately 1.36 was rampant at the conference, and most dating) or as a solid (Be metal or BeO for that of the background N14. For H3 and participants felt that the accelerator BelO dating) and sputtered. Be3, a simple solid foil is sufficient to stop technique would soon be making impor­ The sharp resolution and absence of the background ions, and only the radio­ tant new measurements in areas inac­ "tails" in the cyclotron resonance plot is isotopes emerge to be counted by silicon cessible to the decay-dating method. dramatic; one can tune the frequency of detectors or gas ionization chambers (see Indeed, a few months later the TIR the cyclotron 1% away from the resonant figure 2). For C14 the background current groupl9 had increased their sensitivity for frequency for a 10 microamp Cl2 beam, of Nl4 was so large (1-10 nanoamp) and Cl36 to the point where it was virtually and observe no accelerated particles in an the range difference so small that the assured to be an important new tool for hour of operation. This sharp resonance necessary separation of 106 was not ob­ dating old ground water, with an imme­ response is a consequence of the many tained with a simple foil. Irregularities diate application to nuclear-waste storage (100 to 200) turns made in the cyclotron in the solid were responsible for sufficient problems. And in July 1978 Raisbeck during acceleration. Thus when we ac­ straggling that we could not obtain a date, and his collaborators20 reported mea­ celerate C14 the only other isotope accel­ although we had sufficient sensitivity to surements of BelO in 1000 and 5000 year­ erated is N14; no high energy C12, C13, Nl5 measure Cl4 at approximately natural old Antarctic ice, in samples that were or other isotopes appear in the beam. levels.lo unmeasurable by any other technique. If the current is low enough, one can We finally used gas for the stopping They conclude their article with the ob­ send the beam directly into a set of "par­ material to reduce the straggling from servation that accelerator dating had ticle-identification" detectors, which irregularities.14,21 Separating the gas moved "out of the phase of exploration measure the ionization rate and kinetic from the cyclotron vacuum required and into that of exploitation." energy of each ion; these two quantities windows strong enough to support the gas serve to determine the atomic number Z pressure yet thin enough to contribute Cyclotron techniques and the atomic weight A of the ion. (The little to straggling. The solution was to The five basic stages involved in the ionization rate is determined by the nu­ use 1/3 -micron gold foil for the window cyclotron technique are clear charge Z rather than by the initial material. It holds an atmosphere of ~ ionization of the sample, charge q of the ion, because the ion is pressure when supported with a set of ~ acceleration of the isotope of interest, stripped of its remaining electrons SODn grids (see figures 3 and 4) fine enough that ~ separation in the high-energy beam of after entering the particle detector.) In the foil never spanned more than Va mm. the radioisotope from intense stable most cases the beam current from the The nuclear charge-exchange reaction backgrounds, stable background ion is too intense to NI4(Z, Z + 1)C14, could have produced a ~ identification of the individual ions allow this, and some preliminary separa- contaminating signal, but all the materials

PHYSICS TODAY / FEBRUARY 1979 25 inserted into the beam (xenon, gold, number of hours devoted to experiments; tungsten) have a high nuclear charge and the result is about $250 per hour, or $250 the beam energy was kept low enough per date. There are numerous cyclotrons (:560 MeV) that the Coulomb barrier 10-' around the world that could be used for suppressed the reaction. The filter to­ radioisotope dating; however, one can buy tally eliminated the nitrogen; only car­ a new cyclotron of the minimum required bon-14 emerged from the xenon gas. We 0. size for about $500 000. The cost is E estimate the suppression to be better than .!':! comparable for tandem Van de Graaff 14 f- 10 . In several hours of coincident de­ Z 10-15 accelerators. Other types of heavy-ion w tection, not a single nitrogen ion pene­ a:: accelerators could also be adapted for a:: trated the xenon cell to be counted as a => dating. coincidence in our particle-identification u detectors. Tandem techniques To determine the age of the sample, the 10-21 Many of the techniques used at the ' detection rate of the radioisotope must be tandem accelerators are similar to those normalized with respect to a stable iso­ used at cyclotrons, and the general dis- j tope. There are several ways to accom­ 30 40 50 cuss ion given for the cyclotron need not plish this: THICKNESS (mg/cm') be repeated. Negative ions are acceler­ ~ by switching the cyclotron frequency ated from ground to positive high voltage, so that the stable isotope is accelerated, Transmission through ttie xenon gas range fil­ stripped of orbital electrons, and then and then measuring that isotope in the ter, as a function of the mass of xenon in the accelerated as they move from the posi­ same beam line (although the sensitive beam. We plot the typical background nitro­ tive potential back to ground. Because of gen-14 current (black) along with the carbon-14 the high charge of the stripped ion, the silicon detectors must be replaced with current (color) for a modern sample. Figure 3 Faraday cups capable of handling the energy of the particle in MeV can be sev­ microampere beam currents); eral times larger than the value of the high ~ by switching the sample with unknown The carbon-14 "background" comes voltage measured in millions of volts. age to one of known age, and measuring from the cyclotron components them­ Since the tandem itself provides no mass the ratio of radioisotopes in the two sam­ selves. The cyclotron dee is protected selection, a separate mass spectrometer ples (the ratio of radioisotope to stable from the intense internal beam by must be employed. Figure 5 shows the isotope is already known in the reference graphite "beam scrapers." Graphite was Toronto-General Ionex-Rochester sys­ sample); chosen in the original cyclotron design tem, which uses one stage of charge-to­ ~ by monitoring the beam currents exit­ because it remains relatively inactive mass (qlm) selection before acceleration, ing from the ion source. For example, one when exposed to intense beams; virtually and three afterwards. Each "mass spec­ could magnetically separate the C12 from the only long-lived radioactivity produced trometer" is simply a bending magnet C14 prior to acceleration, and use the on graphite is 0 4 , and of course nobody with collimation slits. In the TIR system current of this isotope for normalization had anticipated use of the cyclotron for a final stripping foil changes the charge (provided that the acceleration efficiency radiocarbon dating. The carbon-14 from state from 4+ to 6+; molecular ions that is known). Possible isotope fractionation the graphite has sufficient vapor pressure survive to this point are broken up in the effects in the ion source may affect these to contribute a small but annoying cur­ foil and eliminated in the final stage of experiments and one needs to account for rent of background Cl4 to our runs; no q 1m selection. The final beam enters a them to obtain the fullest accuracy. corresponding backgrounds are observed set of detectors that identify the nuclear For the heavier isotopes the cyclotron for our H3, BeIO, or Cps measurements. charge Z and the atomic mass A of the technique is more diffi~ult to implement, This background is presently limiting the ions through ionization and total energy both because one cannot obtain high sensitivity of our 0 4 measurements, and measurements. beam currents with present cyclotrons, we have deferred the measurement of Several elements do not form stable and because the fractional range differ­ valuable and irreplaceable samples until negative ions; s'o the use of a negative ion ence between neighboring elements de­ it has been eliminated. We plan to do source has important consequences for creases as liZ. In the case of Cps there this by pre-accelerating the sample ions dating. In particular, because the nega­ is an additional problem, because there is with an external ion source which is kept tive nitrogen-14 ion is not stable, the pri­ a stable isotope with lower Z (S3S), which free of contaminants. Such an ion source mary stable background ion for carbon-14 is not removed by the range filter. is now under construction, and will be dating is eliminated right at the source. However, we found that in our measure­ installed on the 88-inch cyclotron in a few This suppression of the nitrogen-14 ments the background S3S in the sample months. background was a driving consideration was low enough that the beam emerging The 88-inch cyclotron at Berkeley is in the plans of the tandem groups; the from the range cell could be sent directly much larger than required for dating same advantage can be gained by using into the particle identifiers, which clearly purposes; in fact we sometimes operate at negative ions in a cyclotron. Molecular distinguished the two elements. such low energy that it is difficult to ad­ backgrounds such as (02H2)- are broken By switching rapidly from an "un­ just the magnetic field to the required low up when they pass through the stripping known" to a reference sample, we have level, due to hysteresis in the magnet iron. foils, and then suppressed by the qlm made carbon-14 measurements on several But the flexibility of this cyclotron has mass-spectrometers. Particle identifi­ samples with ages up to 10 000 years. been very useful in exploring the range of cation is still necesary, because some The real proof, however, is to measure a techniques and isotopes accessible to ac­ stable background ions such as C12, C13, :; "blind" date, that is, one for which one celerator dating. In spite of its large size, and 02H2 do leak through the mass does not know the answer beforehand. the cost of obtaining a date at this cyclo­ spectrometers. It is the high energy of We have been the only group to try this so tron is comparable to that of the decay the beam that allows the use of the par­ far; perhaps we were incautious; although method; although the machine is expen­ ticle-identification detectors and the we obtained the correct answer14 in one sive to operate, the date can be obtained stripping foils. sample preViously dated by Rainer Berger in less than an hour of running. A rough To normalize thE count rate to that of of UCLA, we also had a failure,22 which we estimate of the cost of this hour can be a stable isotope, the magnets can be believe was due to a systematic shift in 0 4 calculated by taking the total operating switched to select that isotope. However, background level as we switched from the budget of the cyclotron (scientists and it is also possible to include a wide-enough unknown to a reference sample. engineers included) and dividing by the selection in the mass spectrometer so that

26 PHYSICS TODAY / FEBRUARY 1979 one can measure the stable and unstable isotopes simultaneously in different de­ tectors, as Nelson has proposed: Dedi­ cated tandems for radioisotope dating will certainly do this. The table below shows a comparison of the most sensitive carbon-14 measure­ ments reported to date by the TIR group12, compared with measurements made using decay dating by the US Geo­ logical Survey. The most impressive feature of the measurements is that the sample size was only 15 mg, except in the Mt Hood sample for which it was 3.5 mg. We show a typi­ cal output plot from their particle-iden­ tification detectors in figure 6; note the very clean separation between the C14 and other isotopes. It is premature to be concerned about the relatively large errors in the measurements. As systems are developed, stabilized and improved, the Microphotograph of the tungsten grid used to support the %-micron-thick gold foil. The larger grid supports the finer grid, which supports the foil. The hexagons of the 'fine grid are Y3 mm in diameter. errors are certainly going to be reduced. The combination can support a pressure difference of greater than one atmosphere between the We can estimate the present "maximum filter chamber and the cyclotron vacuum. Figure 4 detectable age" for the Rochester tandem to be about 50 000 years: this is the measured age for the graphite blank known to contain no C14. The back­ ground C14 counts are believed to come 8 MVTandem from contaminant C14 in the tandem, just as the background 0 4 in our cyclotron comes from radioactive graphite in the cyclotron vacuum; it should be absent in a new machine dedicated to radioisotope dating. It may also be possible to elimi­ nate the contamination in existing tan­ dems with a specially designed ion source, the same approach we are pursuing at Berkeley. Ion source

Applications Focussing magnet Many applications have been suggested for accelerator dating, especially for those radioisotopes that have already proven valuable with decay dating. An extensive discussion appears in the proceedings of the Rochester conference; I will draw heavily from this without individual at­ tribution of the ideas.23 For carbon-14 dating, it appears likely Stripping foil that the most important immediate ap­ plication is to samples that are too small to have been dated before. In this cate­ gory I include invaluable artifacts for which only a small bit can be sacrificed. Examples include manuscripts, bone The TIR dating system at the Rochester tandem Van de Graaff accelerator. The sketch i~ not to fragments, cave paintings and the Shroud scale and shows only the basic components. (Adapted from reference 11 .) Figure 5 of Turin. Just as important is the ability to select, carefully, the carbon in the sample that is used, so that those car­ bon-bearing chemicals which exchange Accelerator dates carbon with the surroundings can be avoided. Thus, for example, one can use Accelerator Decay date only the collagen in bone, or perhaps only Carbon Sample Running time (min) age (yrs) age (yrs) specific amino acids. C. V. Haynes gives 220 ± 300' 220 ± 150 the example that with the selection of Mt Hood 90 Mt Shasta 180 5700 ± 400 4590 ± 250 certain organic compounds an entire Lake Agassiz 90 8800 ± 600 9150 ± 300 mammoth skeleton might be required for Hillsdale 400 41000 ± 1100 39500 ± 1000 decay dating; for the direct method a toe Graphite 65 48000 ± 1300 bone would suffice. One can date arti­ Based on reference 11 facts by selecting entrapped plant frag­ • assumed rate (used for calibration) ments, seeds, or insect remains. The

PHYSICS TODAY / FEBRUARY 1979 27 long compared to the estimated trapping time of cosmic rays in the Milky Way galaxy. Other radioisotopes from deeply buried samples (such as Kr81 or Pb205) may someday be measurable with suffi­ cient sensitivity to study neutrino inter­ actions. One of the first "practical" applications of accelerator dating may be for hydro­ logical tracing. A 1977 workshop on the dating of old ground water25 concluded that the development of a chlorine-26 (3X105 yr) method of dating and the continued development of carbon-14 . dating present the most promise for very old ground water. This conclusion was based largely on the expected geochem­ istry of these elements, and particularly on the fact that dissolved chlorine is not expected to exchange rapidly with sur­ rounding soil. The most immediate ap­ plication of chlorine-36 dating is for the determination of the hydrological isola­ tion of potential storage sites for nuclear wastes. Carbon-14 dating has already proved to be very important in studies of the circulation times of deep ocean water. Other radioisotopes however may be Measurements made in the particle identifier of the TIR system, for their 4590 ± 250 year-old , better matched to the overall circulation sample from Mt Shasta. The logarithm of the heavy-ion counts is shown as a function of the total 3 32 heavy-ion energy and the energy deposited in the final ionization dete.ctor. The clean separation time of 10 years; these include Si (650 of the carbon-14 signal from the various backgrounds is evident. The nitrogen-14 peak is from yr) and Ar39 (269 yr). the N14H- molecule, (From reference 12 ,) Figure 6 Limitations After such an optimistic discussion of small samples used for racemization then the C14 content will be low, whereas, the glories of future accelerator dating, it dating can be cross-checked by compari­ if recent plants are the source, the C14 is appropriate to recall the limitations of son with carbon-14 dates. Glaciallayers content will follow (with a delay) the the technique. For the immediate future, have been dated by trapped CO2; in the thermonuclear in crease. the following general rule will undoubt­ future only a few liters of ice, rather than Many other radioisotopes are produced edlyapply: several tons, will yield a useful date. The by cosmic rays, but their detailed geo­ Any object that can be dated using the ability to date ice layers can have impor­ chemistry is not yet worked out. Of these standard decay technique should be tant impact on our kriowledge of the cli­ isotopes, J. R. Arnold suggests that Be lO dated in that way. Any iS0tope that mate over the past 100 000 years, because holds the most interest, partially because can be detected using an ordinary the oxygen-18 concentration in these its half-life of 1.5 million years matches mass spectrometer should be detected layers is related to the global temperature the time scale of the evolution of Man and with that method. at the time they were deposited. The of the ice ages. Be lO has been detected in "Standard" radioisotope dating and potential 'of the new 0 4 techniques to sea-floor sediment and manganese nod­ mass spectrometry are highly developed date beyond the present 40000 year limit ules. Measurements of BelO in individual arts; accelerator dating is a new technique, may allow the study of the climate in the layers of manganese nodules should elu­ and we are only about to learn the prob­ period immediately preceding the most cidate their history and help determine lems associated with it. The main ap­ recent glaciation, which began about whether they have always sat on the sea plications of accelerator dating in the next 80 000 years ago. floor or have been periodically buried. few years will be to obtain approximate The C14 fluctuations in tree rings of Since the Be10 cOl)centration depends on answers in areas where approximate an­ known age (the" de Vries effect") allow us not only the age but also the sedimenta­ swers are adequate and where no answer to study climates over the past 8000 years, tion rate, it is necessary to have an addi­ is obtainable otherwise. Presently it takes 10-20 adjacent tree tional measurement to untangle these. Because mass spectrometry can detect rings to obtain enough ~aterial to mea­ One possibility is "double dating"3.7 with trace elements at concentrations of 10-6 sure the C14 content; with accelerators a another radioisotope such as Aj26. Ac­ to 10-9, it is unlikely that accelerators will single ring should be sufficient, allowing cording to J, C. Higdon and R. E. replace them for stable isotopes. Virtu­ us to study the climate year by year. Lingenfelter24 an anomaly in the ratio of ally all elements appear at this level as Measurement of small samples will also Be lO to Aj26 suggests the detonation of a traces in vitually all other substances. simplify studies that make use of the nearby supernova in the recent past; it Except for exotic molecules such as C13D4 sudden doubling of the atmospheric C14 will be interesting to study the effect in (see references 8 and 26) , the only isotopes ' content in the 1950's due to atmospheric more detail. at sufficiently low level to require an ac­ thermonuclear bomb testing. (Such Longer-lived radioisotopes such as Kr81 celerator for direct detection are the ra­ studies have been a primary concern to an (2X 105 yr) and p 29 (1.7X107 yr) can be dioactive ones. ongoing program of small-sample count­ used to study the constancy of the cosmic The ultimate usefulness of several of ing at the National Bureau of Standards.) rays, provided they are detected in ma­ the radioisotopes, including BelO and C136, Particularly interesting is the origin of terials of known age, such as materials will depend on their geochemistry, and atmospheric gases and aerosols, including that have been potassium-argon dated. presently this is poorly known. Questions methane and CO2. If these originate Iodine-129 is particularly interesting for to be answered include: What path does from the burning of ancient hydrocarbons this measurement, because its half-life is the radioisotope take from the upper at-

28 PHYSICS TODAY / FEBRUARY 1979 mosphere to the Earth's surface? In 5. L. W. Alvarez, R. Cornog, Phys. Rev. 56, what chemical and physical forms does it 379 (1939). arrive? Is it ingested by living· organ­ 6. K. H. Purser, US Patent 4037100, filed isms? What is the appropriate stable March 1976, issued July 1977.· element to use for normalization? How 7. R A. Muller, L. W. Alvarez, W. R Holley, uniformly is the radioisotope distributed E. J. Stepherison, Lawrence Berkeley over the surface of the Earth? Fortu­ Laboratory Report LBL-5399, 1976; nately, the same technique that makes 'Science 196, 521 (1977). these questions interesting also makes 8. R A. Muller, LBL-5510 (1976); Science them experimentally tractable. Much of i96, 489 (1977). Most of the ideas ap­ peared in an internal memo, Physics Note the work to be done with accelerators in 319 (July 1976). the next few years will be devoted to an­ 9. D. E. Nelson, R y. Korteling, W. R Stott, swering these questions: Science 198,507 (l977). One limitation to dating samples that 10. E. J. Stephenson, L. W. Alvarez, D. J. are many half-lives old is the background Clark, R A. Gough, W. R. Holley, A. Jain, of spurious isotopes introduced by recent R A. Muller, Bull. Am. Phys. Soc. 22, 579 contamination of the old material. For a (1977) .. sample t half-lives old, contamination by 11. K. H. Purser, R B. Liebert, A. E. Lither­ modern material in an amount 2-1 will land, R P. Beukens, H. E. Gove, C. L. cause a systematic error in the age deter­ Bennett, M. R Clover, W. E. Sondheim, mination by one half-life. For example, 2nd Int. Conf. Electrostatic Acce!. Tech., a contamination of a carbon sample 100 Strasbourg, France, 1487 (1977); C. L. thousand years old by 2-(100/5.73) = 6 parts Bennett, R P. Beukens, M. R Clover, H. per million of modern carbon will cause E. Gove, R B. Liebert, A.E. Litherland, K. H. Purser, W. E. Sondheim, Science 198, the calculated age to be underestimated 508 (1977). by 5730 years. It is too early to know 12. C. L. Bennett, R P. Beukens, M. R Clover, whether one can in practice maintain the D. Elmore, H. E. Gove, L. Kilius, A. E. extreme level of purity required to mea­ Litherland, K. H. Purser, Science 201, 346 sure such great ages. It is equally im­ (1978); see also ref. 13, page 70. portant to determine whether at the low 13. Proceedings of the First Conference on levels encountered for the older dates, Radiocarbon Dating with Accelerators, (H. there might be a process other than cos­ E. Gove, ed.), University of Rochester, 20, mic radiation (such as radioactivity in the 21 April 1978. Earth) which can produce the radioiso­ 14. R A. Muller, E. J. Stephenson, T. S. Mast, tope even after the sample "died." Science 201, 347 (1978); see also ref. 13, Accelerator dating is still a young field, pages 33-37, 239-244, 197-195, and and it is, as yet, impossible to predict what LBL-7585. the most important applications will be. 15. D. E. Nelson, R G. Korteling, D. G. Burke, I suspect we also don't yet know which J. W. McKay, W. R Stott, ref. 13, page radioisotopes and which experimental 47. techniques will prove the most fruitful. 16. G. M. Raisbeck, F. Yiou, M. Fruneau, M. As a physicist, what I have enjoyed most Lieuvin, J. M. Loiseaux, ref. 13, page 38 .. about accelerator dating has been talking 17. E. T. Hall, R E. M. Hedges, N. R White, to ,scientists in many other fields and H. R McK. Hyder, D. Sinclair, ref. 13, page learning about their unsolved problems. 257. I hope that my enjoyment in the future 18. H. R Andrews, G. C. Ball, RM. Br~wn, N. will come from helping them solve some Burn, W. G. Davies, Y. lmahori, J. C. D. Milton, ref. 13, page 114. of these problems. 19. D. Elmore, B. R Fulton, M. R Clover, J. R Marsden, H. E. Gove, H. Naylor, K. H. * * * Purser, L. R Kilius, R P. Beukens, A. E. Litherland, to be published in Nature. This article is an adaptation of an address to 20. G. M. Raisbeck, R Yiou, M. Fruneau, J. M. the American Institute of Physics 1978 Loiseau, Science 202, 215 (1978); G. M. Meeting of Corporate Associates, Columbus, Raisbeck, F. Yiou, M. Fruneau, M. Lieu­ Ohio, 28 September 1978. vin, J. M. Loiseau, Laboratoire Rene Ber­ References nas preprint LRB-78-10 (Orsay, to be published) 1978. 1. W. F. Libby, Phys. Rev. 69,671 (1946); E. 21. E. J. Stephenson, T. S. Mast, R A. Muller, C. Anderson, W. F. Libby. S. Weinhouse, LBL-7579 (June 1978), to be published in A. F. Reid, A. D. Kurshenbaum, A. V. Nucl. lnst. Methods. Groose, Phys. Rev. 72, 931 (1947). See also W. F. Libby, Radiocarbon Dating, U. of 22. R A. Muller, ref. 13, page 34. Chicago Press, Chicago (1955). 23. The contributers to the applications dis­ '. 2. For a recent review see P. E. Damon, J. C. cussion in ref. 13 include C. V. Haynes, L. Lerman, A. Long, Ann. Rev. Earth and A. Currie, J. R Arnold, W. S., Broecker and Planet. Sci. 6,457 (1978). T. Pengo 3. Reviews of cosmogenic nuclides can be 24. J. C. Higdon, R E. Lingenfelter, Nature found in Handbuch der Physik 46/2, 246,403 (1973). Springer, Berlin (1967): D. Lal, B. Peters, 25. Workshop on Dating Old Ground Water page 551; M. Honda, J. R Arnold, page (S. N. Davis, ed.) U. of Arizona, Tucson 613. . (1978). 4. M. Anbar, Proceedings of the 22nd Con­ 26. G. Cowan, D. Ott, A. Turkevich, L. Machta, ference on Mass Spectrometry, (Phila­ G. Ferber, N. Daly, Science 191, 1048 delphia Pa., 1974). (1974). 0 '0

This report was done with support from the Department of Energy. Any conclusions or opinions expressed in this report represent solely those of the author(s) and not necessarily those of The Regents of the University of California, the Lawrence Berkeley Laboratory or the Department of Energy. Reference to a company or product name does not imply approval or recommendation of the product by the University of California or the U.S. Department of Energy to the exclusion of others that may be suitable.

," ,. · ~ TECHNICAL INFORMATION DEPARTMENT LAWRENCE BERKELEY LABORATORY UNIVERSITY OF CALIFORNIA BERKELEY, CALIFORNIA 94720