Radioisotopes trace out mechanisms important both to … · 2016-10-21 · Isotope and Nuclear...

Migration ofRadioisotopes

in theEarth’s Crust

written by Roger Eckhardt for theIsotope and Nuclear Chemistry Division

Radioisotopes trace out mechanismsimportant both to underground

nuclear waste storage and to ourplanet’s geology.

w ater moves steadily in the earth beneath our feet,flowing in aquifers through the pores, fractures, andgrain boundaries of rocks and minerals. This ubi-quitous “universal solvent” interacts with the rock,

helping gradually to reshape the geology of our planet’s crust; it isthe fluid most commonly associated with an important process calledgeochemical migration, that is, the selective movement of chemicalelements in rock. A better understanding of the nature of this processhas implications for a variety of human activities.

For example, burial of hazardous materials has often been chosenin the past as the safest method of disposal. But this choicesometimes ignored the pervasive nature of geochemical migration,leading to a number of serious cases in which hazardous materialresurfaced in drinking water. A particularly important disposalproblem is what to do with nuclear reactor waste. Storage in deepgeologic sites is being considered, but it is imperative to evaluate thepotential for migration of long-lived, highly radioactive waste awayfrom the site and into the accessible environment.

There are also implications for locating ore deposits. A widevariety of ore bodies are thought to be formed as the result ofselective dissociation of elements from a large body of rock followedby transport in water and eventual concentration and precipitation atanother site. What conditions favor the dissolved forms of theelements. what favor the precipitated forms? A full understanding ofsuch processes would serve as a rational basis for predicting thelocation of ore bodies in different geologic settings.

25

The complex phenomena that constitutegeochemical migration have long intriguedand challenged geochemists. Much relevantinformation has come from studies of vol-canism, of the processes that alter a rock’smineral composition, and of age determina-tion techniques. Much remains to be learnedif we are to understand the complexities ofgeochemical migration and its relationship toall the other geologic processes that shapeour planet.

Radioisotopes

At Los Alamos we draw upon the Labora-tory’s background in nuclear physics andradiochemistry to study geochemical migra-tion using radioisotopes. These methodshave much to recommend them. startingwith the fact that even trace amounts ofradioisotopes are easily detectable. This isimportant because elements are frequentlypresent in water only to the extent of partsper billion. The ease of detection is due, ofcourse, to the radiation emitted by theseisotopes. This radiation, characterized bytype and energy, also helps establish aunique signature for any source of radio-isotopes, especially when coupled with massspectrometric measurements that give therelative isotopic abundances of both theradioactive and nonradioactive isotopes in asample. The signature can then be used notonly to identify the source of an element butalso to follow its migration through theenvironment.

For example, the atmospheric nuclearweapons tests that took place over the oceanin the fifties injected large quantities ofchlorine-36, a long-lived beta emitter, intothe environment. This isotope is normallyproduced by cosmic rays in the atmosphereand had already been of use as a tracer inhydrology. The much larger quantities, how-ever, could be uniquely associated with testsde tonated in proximi ty to seawater :chlorine-36 is not a direct fission or fusionproduct but is generated by thermal neutron

26

activation of the chlorine-35 isotope abun-dant in seawater. A recent study by re-searchers at the University of Arizona, NewMexico Institute of Mining and Technology,and the University of Rochester has shownthat these tests caused a global “pulse” ofchlorine-36 to be injected into the environ-ment. This pulse lasted from 1953 to 1976and, at its peak, rose to 1000 times thenatural background level of chlorine-36 inthe mid latitudes. In contrast to carbon-14and tritium, which are still being generatedby China’s atmospheric testing, fallout ratesfor chlorine-36 have returned to naturalbackground levels, making this isotope idealfor use as an environmental tracer. Becauseits long half-life eliminates ambiguities as-sociated with loss of the isotope throughradioactive decay, chlorine-36 should beespecially valuable for studies of hydraulicflow and dispersive mixing as it incorporatesitself into the water and geology of theearth’s crust. Although these processes con-tinuously reduce the concentrations of thetracer, recent advances in analyticalchemistry, namely accelerator-based massspectrometry. make possible its detection inamounts of 107 atoms per liter or less.

There are further advantages in the use ofradioisotopes. As is well known from radio-active dating, individual radioisotopes decayat specific rates characterized by their half-lives; ages are determined using the ratio ofthe amount of radioisotope, or parent, to thatof the decay product, or daughter. Butbecause the daughter is usually a differentelement that is chemically distinct from theparent, fractionation processes can occurthat physically separate the two. For exam-ple, conditions may have been such that thetwo elements had considerably different sol-ubilities in the groundwater moving throughthe rock and one was transported awaywhile the other remained in place. Deficien-cies of parent or daughter elements in geo-logic samples compared to amountspredicted from nuclear physics may thusindicate geochemical migration. Samples

from other geologic locations may reveal thedestination of the transported element. Evenmore important, the chemical nature of theelements sheds light on the chemical natureof the fractionation process that took placein the sample. And comparison of parent todaughter ratios helps determine the geologictime at which the fractionation occurred(Fig. 1).

A particular advantage of analytical tech-niques based on mass spectrometry stemsfrom the fact that amounts of both unstableand stable isotopes are measured. Evenbillions of years after a particular geologicalor nuclear event, stable isotopes and long-lived radioisotopes remain. Depending on thehistory of a sample, the relative abundancesof these isotopes differ, providing the re-searcher a unique signature of the source.For example, the products of a uraniumfission chain reaction can be distinguishedeasily from chemically identical productsfound naturally.

Initial motivations for much of the radio-isotope migration research at the Laboratoryhave been related to nuclear weapons testingand nuclear energy. For example, how much,if any, of the radioactive material generatedin the underground testing at the NevadaTest Site will be carried away by thegroundwater? Or if radioactive nuclearwastes are stored underground, whatmechanisms may cause migration of theradioisotopes and how fast would this hap-pen? Yet, because geochemical migration isactually a complex group of interactingphenomena, answers to these particularquestions also shed light on more generalquestions important to geology. (See“Mechanisms of Migration” for highlights ofthis broad picture.)

Migration from Underground Tests

From 1957 to the present there have beenover five hundred underground nuclear testsat the Nevada Test Site, of which at leastninety-five were detonated at or below the

Summer 1983 LOS ALAMOS SCIENCE

Migration of Radioisotopes in the Earth’s Crust

Fig. 1. Simple examples of fractionation of radioisotopes.Because the daughter isotope (solid circles) of a radioactivedecay process is a different element than the original radio-isotope (open circles), chemical and physical processes mayeventually separate the two. The figure depicts a radioisotope-containing mineral in the same state at the top of each column.In (a) considerable decay occurs before fractionation, whereasin (b) only a small amount of decay occurs first. The frac-tionation process for these two columns is identical, say

t e r

dissolution and transport of the daughter isotope bygroundwater. The result of the difference in the time whenfractionation occurs is seen at the bottom of each column.Column (a) has a high depletion of daughter in the mineral,whereas in (b) the depletion is much less. In (c) there is againconsiderable decay, but the fractionation process transportsthe original radioisotope, leaving an excess of daughter in themineral.

LOS ALAMOS SCIENCE Summer 1983 27

M ost people are aware that rainfall,as it seeps into the earth, becomesladen with dissolved material or

solutes. What they may not realize as fully isthat the composition of the solutes canchange drastically as the water passes fromone geochemical environment to another,causing significant relocation of particularelements in relatively short geologic times.

Consider granite, a rock with a very lowpermeability to water. If a cube of such rock,1 kilometer on a side and without fractures,has a typical hydraulic pressure gradientalong one axis of 10 meters per kilometer,then 30 billion liters of water will flowthrough the rock in only 10 million years. Ifthe water picks up uranium from the rock tothe extent of 5 parts per billion, which is only5 milligrams per 1000 liters, then a total of150 kilograms of uranium will be trans-ported out of the block. Although the degreeof permeability and the strength of thedriving forces determine how fast waterflows, we see that large volumes can movethrough even very impermeable rocks, trans-porting significant quantities of trace ele-ments.

The forces that act to produce migration

are gravitational, chemical, and thermal.Gravity is the force responsible for hydraulicgradients. The rate of flow of groundwater isdetermined by this gradient, which, in turn, isgenerally a function of the topographic relief.A totally flat plain at sea level has a zerohydraulic gradient and zero flow.

For a given hydraulic gradient, flow is afunction of permeability, which reflects tex-ture and structure. Fractures may providerelatively low-resistance pathways throughrock while a fine-grained or plastic rock maybe relatively impermeable. In sedimentaryand volcanic ash sequences, the permeabilityof individual strata may differ widely.

Another factor affecting flow is porosity, ameasure of the space available in the rock forwater. Porosity is generally related to per-meability so that highly permeable rockstend to be quite porous and vice versa. Butfor geochemical migration it is connectedporosity that is important. This parameterdetermines the degree to which the entirebulk of the rock, as opposed to the surfacesof fractures, is exposed to flowing ground-water. Highly porous, permeable rock. suchas some tuffs, are totally exposed, allowingelements to be leached from all minerals in

the rock.Above the water table, rock is un-

saturated, and flow is limited although notstopped. As air replaces water, flow is gradu-ally restricted to surface films that must beconnected for liquid flow to occur. When thefilm becomes discontinuous, water move-ment is restricted to vapor transport in theair-tilled pores, and transport in water be-comes negligible.

A thermodynamic measure of the chemi-cal force is the chemical potential, which, inaqueous systems, is directly related tovolubility. For a given element, the chemicalpotential varies with the element’s concentra-tion and the environmental conditions suchas temperature and pressure, acidity, thetendency for reduction or oxidation reac-tions, and the presence or absence of com-plexing ions or molecules in the water. Ingeneral, elements move from higher to lowerpotentials. For example, if the amount ofdissolved mineral is less than the equilibriumconcentration, then dissolution will takeplace, albeit often very slowly; if the amountexceeds the equilibrium volubility, thenprecipitation occurs.

The chemical processes contributing to

water table (Fig. 2). Considerable thoughtwent into the engineering of the tests toinsure that no radioactivity!; escaped into theenvironment, not even by way of under-ground water. However. it was realized thataspects of this latter path, such as thecomplex chemical partitioning of a radio-isotope between the liquid, gaseous. and solidphases. were poorly understood. Hard datawere needed.

Starting in 1972, an AEC (now DOE)program evolved that included the Los Ala-mos and Lawrence Livermore National Lab-oratories. the Desert Research Institute ofthe University of Nevada. the U.S. Geologi-

28

cal Survcy, and the Reynolds Electrical andEngineering Company. The program’s mis-sion was to gather and interpret datanecessary for an understanding of radio-isotope migration at the Test Site. Studieswere eventually conducted at eight areas,including seven nuclear test sites and onearea in which tritium was detected severalhundred meters from the nearest under-

ground test cavity. One of the most in forma-tive of these projects. and the one we willdiscuss here, measured migration by impos-ing an artificial hydraulic gradient on thenormal underground water flow from a nu-clear test cavity.

The cavity, the rcsult of a test shot calledCambric. was located 73 meters below thewater table within a Test Site water-supply

aquifer. Since the detonation had occurred in1965, nine years before the start of the study.it was predicted that the cavity, and thechimney resulting from later collapse abovethe cavity, had long since filled with water tothe preshot lcvel of the water table. Anyradioisotopes present would constitute a po-tential source for transport, and it was feltthat the postshot debris would contain suffi-cient amounts of radioisotopes to compareconcentrations between rubble and ground-watcr. “Table I gives the Cambric source



geochemical migration can be very complex.A specific element can, in fact, move up-stream against its potential treated as anisolated system if there is a net zero ornegative potential change for the total sys-tem of interacting elements. It is quitepossible in complex systems for one or moreminerals in a rock to be dissolving whileothers are forming.

An illustration of this complexity is thevariety of possible associations betweenmetal cations in the minerals and the anionsthat may be available in groundwater, suchas chloride, fluoride, carbonate, sulfate, orphosphate. For some conditions anions formstable complexes with cations, such

increase the metal’s volubility. For otherconditions the anions do the opposite andform insoluble compounds, such as CaCO3,MgS04, HgCl, or Th3(PO4)4.

Temperature has a major effect on chemi-cal potential. A mineral stable at 200 degreesCelsius may be unstable at 300 degrees.Solubilities typically increase with tempera-ture but may be retrograde and decreaseinstead, Geothermal systems are particularlyactive zones for geochemical migration be-

cause thermal gradients combine with gravi- Because sorption is a surface effect, thetational forces to produce convective crystal structure of a mineral is very impor-groundwater circulation that exposes the tant. With densely packed structures, onlyminerals to a wide range of chemical en- the external surface of the crystals can sorbvironments. species from solution. However, other min-

Important chemical processes take place erals, such as zeolites, have a very openeven in solid or crystalline materials+ Trace lattice structure that permits relatively easyelements such as the rare earths are fre- access to interior passages. These showquently found as dissolved components in preferential sorption for species whose sizesolid solution. These impurities can either be matches the openings most closely. incorporated into the solid when it forms or Because the surfaces of the mineral in thediffuse into the lattice later. Conversely, pores are predominantly negatively charged,diffusion out of the solid can release trace anions are typically repelled and cannotelement impurities into the groundwater. enter a small pore as readily as a neutral

One of the most important aspects of molecule or a cation. (Only a few mineralsgeochemistry is the role of surfaces. For sorb anions, and then not very strongly.)example, silicates tend to have a negative This repulsion leads to the surprising resultssurface charge, due to exposed oxygen that anions can migrate through the rockatoms, that attracts cations. Depending on more rapidly even than the water. The waterthe mineral, the cation, and the chemical molecules exchange with the stationary fluidenvironment, this attraction can cause a in small water-filled pores, while the anionsparticular cation to spend most of its time continue moving in the flowing stream.sorbed on a surface, retarding its transport The task, then, of the geochemist trying toby moving water. Van der Waals forces, due understand migration in a given geologicto dipole interaction, and chemisorption, in setting is complicated. He must sort from awhich one or more chemical bonds are variety of factors those particular ones thatformed between the solute species and the apparently caused the equilibria to shift insurface atoms, also play an important role.

term, that is, the expected activity of selectedradioisotopes if radioactive decay but nomigration had occurred during the ten yearsbefore re-entry into the cavity. One impor-tant feature of this source term was theamount of tritium — enough to act as aneasily measurable tracer for water from thecavity.

The geologic medium surrounding theCambric cavity is tuffaceous alluvium, adeposit of gravel-like or sand-like debris thatwas originally volcanic material. This mate-rial was believed to constitute a good me-dium for hydrologic studies because it isquite permeable and does not have large

LOS ALAMOS SCIENCE Summer 1983

fissures or cracks through which the watermight flow selectively. Because the yield ofCambric was relatively low (about 0.75kilotons). the test was expected to have hadlittle effect on the local hydrology.

NATURAL MIGRATION. Two wells weredrilled: the first. the satellite well, 91 metersfrom the cavity in the direction of normalwater flow for the aquifer; the second, the re-entry well, at an angle through the lowerchimney and cavity of the Cambric test (Fig.3). The first question addressed was theextent of natural radioisotope migrationsince the detonation. To answer it. core

samples were taken from the walls of the re-entry well starting just below the groundsurface and continuing to 50 meters belowthe detonation point. Then, by using in-flatable packers to isolate a zone in the welland perforating the casing in that zone,water samples were taken from five levelsthat ranged from just above the chimney tobelow the cavity.

Several interesting facts were determinedabout the distribution of the radioactivematerial. Ten years after the test, most of theradioactivity was still in the cavity region.No activity above background was found 50meters below the point of detonation. Al-

29

Crater

Fig. 2. Sequence of events in an underground nuclear ex-plosion. (a) The explosion vaporizes the rock immediatelysurrounding the weapon, melts and spans rock somewhatfurther away, and generates a compressional shock wave thatcreates a cavity and may fracture rock for several hundredmeters. (b) After the explosion the gases cool and condense,

forming puddle glass, or resolidified rock, at the bottom of thecavity that incorporates a significant fraction of the refractoryradioisotopes, such as the rare earths, plutonium, zirconium,and the alkaline earths. At this point the cavity begins tocollapse, and volatile radioactive species, such as the inert

gases, move upward. Partly volatile species, such as the alkalimetals, uranium, antimony, and iodine, may condense onrubble surfaces in the lower chimney. A [so note that the shockwave has perturbed the water table, generating a mound in thehydraulic head. (c) A chimney of rubble forms above thecollapsed cavity and, if it reaches ground level, forms the cratertypical of underground tests. Eventually, the cavity and part ofthe chimney refill with water, dissipating the hydraulic headmound. The time for this last process ranges from days toyears.

though some krypton-85 and tritium were (HTO), and the krypton-85 was dissolved in solid material. As a result, only strontium-90found in the collapsed zone above the cavity, the water. In the lower cavity region, where and tritium were present in the water fromthese two isotopes were concentrated in the the puddle glass was concentrated, stron- this region of highest radioactivity at concen-cavity region. More than 99.9 per cent of the tium-90, cesium-137, and plutonium-239 trations above the federal guidelines fortritium was in the form of tritiated water were almost entirely incorporated into the drinking water in uncontrolled areas.

30 Summer 1983 LOS ALAMOS SCIENCE


TABLE I

Calculated Source Term

Perforations

Accelerated /Flow

Fig. 3. The Cambric radioisotope migration study. The Cambric underground test wasdetonated about 9 years earlier at a depth of 73 meters below the water table.Accelerated flow of radioisotopes from the cavity (large arrow) was induced bypumping water from a satellite well 91 meters away in the direction of the normal flowof groundwater.

LOS ‘ALAMOS SCIENCE Summer 1983

Retention factors were calculated to meas-ure how effectively strontium-90. for exam-ple, had been retained in solid material. Theratio of strontium-90 activity in a solidsample to tritium activity in a water samplefrom the same location was compared to thesame ratio for the Cambric source term. Thehigh retention factors for strontium-90,

ruthenium- 106, antimony- 125, cesium-137,promethium- 147, and plutonium-239 (TableI) indicate that these isotopes were eitherfused into or highly sorbed on the solidmaterial, or both.

ACCELERATED MIGRATION. The nextstep was to generate an artificial hydraulicgradient by pumping water from the satellitewell. It was planned that pumping woulddraw water from the cavity and permit thestudy of radioisotope migration under fieldconditions. Pumping began in October 1975at a rate of 1 cubic meter per minute; twoyears later the rate was increased to 2.3cubic meters per minute. Early in 1978, aftera total of 1.4 million cubic meters had beenpumped, significant amounts of tritium weredetected, signaling the arrival of water fromthe cavity region.

The tritium concentration grew, peaked atjust below the maximum permissible concen-tration for drinking water in controlled areas(3 nanocuries per milliliter), and is now de-creasing. By October 1982, 42 per cent ofthe tritium available from Cambric had beenremoved by this pumping. The concentrationdata are shown in Fig. 4 along with atheoretical tit to the data of a typical modelfor fluid flow through a dispersive medium.

Neutron-activated chlorine-36 produced

31

by the Cambric explosion was also detectedby mass spectrometry in water pumped fromthe satellite well. The same flow model alsofit these data. The major difference betweenthe migrations of tritiated water and chlorinewas that the chlorine arrived and peaked 20per cent sooner. This difference was felt to bedue to the anion exclusion effect in whichanions are repelled by the negative surfacecharge in the small pores of thealuminosilicates. As a result, anions cannotenter these pores as readily as cations orneutral water molecules. The net result isthat the anions migrate on a more direct paththan the water.

Another radioisotope of great interest wasiodine- 129. Because of its biological activityand lack of sorption on most minerals, thisradioisotope is considered a particular haz-ard. A method was devised, using neutron

activation followed by counting of theiodine- 130 product, that measured the aque-ous concentration of iodine- 129. Signifi-cantly, water from the cavity region had aniodine- 129 concentration that was at least afactor of five lower than that recommendedfor drinking water in uncontrolled areas.

Iodine- 129 was also detected in waterdrawn to the satellite well-at concentra-tions 5000 times lower than in the cavitywater. These data are also plotted in Fig. 4.Fluctuations in the low concentrations makeit difficult to determine a maximum: how-ever. iodine- 129 appeared to arrive at thesatellite well at about the same time as thetritium. If iodine was present as an anionicspecies. its arrival and peak, like those ofchlorine-36. should have preceded tritium’s.Examination of the complex behavior ofiodine in dilute solutions offers a possibleexplanation: for the pH and oxygen contentof Cambric water, iodine may exist pre-dominantly as the neutral molecule HIO andthus experience little or no anion exclusioneffect.

The only other radioisotopes detected atthe satellite well have been krypton-85 and,possibly. minute amounts of ruthenium- 106,

32

1.0

0.5

0.0

●

Chlorine-36— T r i t i u m ●

\

Iodine-129

4 6

Volume Pumped (million m3 )

Fig. 4. Migration of radioisotopes from the Cambric cavity. The detection of tritium(black) signaled the arrival of water from the cavity. Concentrations of chlorine-36(red) arrived earlier and peaked sooner than the tritium, most likely because ofrepulsion between the negative chloride ions and the negatively charged pore surfacesof the rock. On the other hand, iodine-129 (blue) appeared not to be accelerated bythis effect, probably because its dominant species in the groundwater was the neutralmolecule HIO. The solid curves are calculated from a model of dispersive fluid flow.

The krypton, dissolved in water, has been

arriving with concentrations about one-thirdas great as expected based on tritium. Be-cause ruthenium- 106 has the lowest reten-tion factor (Table I) and a very complexchemistry, it might be expected to migrate inminute amounts,

The main conclusion from the accelerated

migration study is that, in general, the sorp-tion of radioisotopes is sufficiently high topreclude the accelerated migration of mostradioactive material. particularly cationicspecies. in the near future, However, pump-ing and analysis of the water will be con-

tinued to investigate the possible arrival ofother poorly retarded species.



Migration from aFission Reactor

Natural

A unique opportunity for the study ofradioisotope migration presented itself in themid ’70s with the discovery of the remnantsof natural fission reactors in the Oklouranium mines in equatorial Africa. Abouttwo billion years ago—nearly half the age ofthe earth—thick lenses of uraninite wereformed that were compact enough and hadthe right mix of neutron-moderating elementsso that fission chain reactions started. Thereactors operated for a few hundred thou-sand years, leaving behind a set of nuclearreaction products different from those nor-mally present in uranium ore.

Several features made the study of thesefossil reactors important. The migration orcontainment occurred over geologic timeperiods. Radioisotopes present at Oklo wereidentical to those of modern reactor wasteproducts. The average thermal loadingwithin the reactor zones during operationwas estimated to be about 50 watts persquare meter, several times greater than thelocal thermal loadings proposed for reactorwaste repositories. Most important, this wasan opportunity to quantitatively characterizethe migration or retention of more thanthirty elements, the products of nuclear fis-sion, neutron capture. and radioactive decay.

Starting in 1974, samples from Oklo weresent to the Laboratory for geochemical andmigration studies. It was hoped that by

examining various decay-product signaturesa fair amount could be learned about whichelements migrated, when the migration oc-curred, and what mechanisms operated tocause the migration.

REACTOR CHARACTERISTICS. Naturaluranium consists essentially of two isotopes:the abundant uranium-238 and the fis-sionable but less abundant uranium-235. Theratio of uranium-235 to uranium-238 hasbeen found to be 0.0072 with only slightvariations throughout the world. However,


within the thirteendiscovered at Oklo,

reactor zones so farthere were significant

deficiencies of uranium-235. Based on thenormal ratio of uranium isotopes, it wasestimated that about 12,000 kilograms ofuranium-235 had fissioned. In fact, the nu-clear product inventory at Oklo correspondsto 1 to 2 per cent of the plutonium andfission products and 10 to 20 per cent of theuranium in the United States spent fuelinventory as of 1980.

The rock surrounding the reactor zones isgenerally pelitic sandstone, a fine-grainedsedimentary rock. The uranium is uraninite,a crystalline uranium dioxide. Examinationof this mineral revealed crystallite charac-teristic of those produced by radiation. Acorrelation between the abundance of thesecrystallite and the degree of uranium-235depletion suggested that the grains ofu r a n i n i t e h a d b e e n t h e r m a l l y a n dmechanically stable since the reactorsstarted. Also, the discovery of uranium sam-ples with quite different degrees of depletionlocated within a few tens of centimeters ofeach other clearly showed that the uraniumhad not been chemically homogenized in the2 billion years since the reactor operated.

The two uranium isotopes decay ul-timately to different stable lead isotopes atdifferent rates. This fact is frequently used todate rocks, and in this case analysis ofuranium and lead isotopes showed thaturanium mineralization occurred 2 billionyears ago. Likewise, analysis of the uraniumfuel and the fission product neodymiumindicated that the reactors became criticalvery soon after mineralization.

What were these zones like during thetime of reactor criticality? Neutron fluenceswere as high as 102] neutrons per squarecentimeter, a remarkably high fluence. Waterin the rock was the most likely neutronmoderator for the reactors, and this water,heated by the fission energy, generated hotaqueous solutions that circulated convec-tively through the rock. In fact, fluid in-clusions in mineral grains formed at

temperatures between 450 and 600 degreesCelsius have been found up to 30 metersfrom the reactor zones. Much of the uniquephysical character of the rock surroundingthe reactors can be attributed to the dissolu-tion and modification of minerals from heatand radiation and the chemical redistributionassociated with circulating hydrothermalfluids.

FRACTIONATION. The isotopic compo-sition in the reactor zones can be visualizedas two components: a natural componentfound in materials throughout the earth anda nuclear component generated by the vari-ous nuclear processes. Total isotopic abun-dances were measured in the Oklo samplesby mass spectrometric techniques and thenatural component was then subtracted. Thenuclear component remaining was comparedto the predicted abundance resulting fromnuclear reactions. Any difference, that is,any loss or gain for a given isotope, wasindicative of migration.

Of particular interest was the evidence forfractionation between technetium andruthenium. Technetium does not occur nat-urally on the earth in significant quantities,but the Oklo reactors generated approx-imately 730 kilograms of technetium-99.This isotope is of special concern in thestorage of radioactive wastes because its 0.2million year half-life implies storage times onthe order of millions of years.

Natural ruthenium is composed of sevenisotopes. These can be split into three groupsdepending on the relationship of each isotopeto the fission process (Figs. 5 and 6). Threeisotopes, the “natural-only” isotopes (96, 98,and 100), are not produced by fission.Knowing the amounts of these allows thenatural component of all seven to be easilyaccounted for. Another three, the “instan-taneous” isotopes of ruthenium (101, 102,and 104), are produced by the decay ofshort-lived technetium fission products.Amounts of these isotopes were generatedwhile the reactors were operating. The last is

33

‘1o nxl


Fig. 5. Ruthenium isotopes and their relationship to the Oklo reactor. The three“natural-only” isotopes on the left were not produced by any of the fission reactionsand thus remained at their original concentrations except as modified by neutron-capture reactions. Fission reactions increased the concentrations of the three“instantaneous” isotopes during the 100,000 years of reactor operation. Because theseisotopes are stable, their concentrations remained constant after reactor shutdown.Ruthenium-99, the “delayed” isotope, was produced by the gradual decay of reactor-generated technetium-99. The half-life of technetium-99 is 0.2 million years so thatabout 96 per cent changed to ruthenium-99 during the first million years aftershutdown. the “delayed” isotope ruthenium-99, which,

because it is the decay product of tech-netium-99, serves as a marker for fission-generated technetium. Any fractionation thatseparates ruthenium and technetium beforetechnetium-99 decays would be observedtoday as a depletion or enrichment of de-layed ruthenium relative to instantaneous

235U Thermal Fission ruthenium.A detailed analysis based on the ideas

Fig. 6. Relative abundances of ruthenium isotopes in normal rock (white) and,likewise, as generated by the thermal-neutron-induced fission of uranium-235 (black)and of plutonium-239 (grey). Samples from the Oklo reactor zones usually containedruthenium with the isotopic composition of the uranium-235 fission process; however,a few per cent of the fissions took place in plutonium-239 that had been produced

from uranium-238 by neutron capture and subsequent beta decay. Samples from theperiphery contained various proportions of natural and fission-generated ruthenium.Natural ruthenium was easily corrected for in a given sample using ratios based onabundances of the natural-only isotopes (96, 98, or 100).

only briefly outlined here showed that tech-netium-ruthenium fractionation had indeedoccurred at the Oklo reactors. In fact, thehalf-life of technetium-99 places an upperlimit on the time between reactor shutdownand fractionation. The basic conclusion wasthat chemical or physical processes causedtechnetium to be lost in greater abundancethan ruthenium from the reactor zones atsome time less than a million years afterreactor shutdown. Analysis of samples fromoutside the reactor zones showed that thelost technetium (as well as other lost fissionproducts) were contained within a few tensof meters of their source.

TRANSPORT MECHANISM. How weretechnetium and other elements carried out ofthe reactor zones? One clue was the physicaldistribution of technetium (now ruthe-nium-99). Generally, the direction of move-ment of this isotope was upward, consistentwith convecting hydrothermal fluids as thecarrier.

To examine this feature in more detail, itwas postulated that the insoluble oxides ofruthenium and technetium were further ox-idized to form soluble oxyion complexes(Table II). Because virtually nothing isknown about the chemical conditions atOklo during the thermal period, only thissimple case of oxyions was considered. A pHand oxygen content typical of the reactorgeology were assumed, and thermodynamicdata were used to calculate the solubilities ofthe species as a function of temperature. Ascan be seen in Fig. 7, the solubilities vary


widely (encompassing 50 orders of magni-tude). The band in Fig. 7 around 10-’0 molesper liter represents estimates of the concen-trations of technetium, ruthenium. and neo-dymium in the aqueous fluid at Oklo. Thewidth of the band is the uncertainty in theestimates. Also included is the volubilitycurve for the neodymium cation, which isretrograde.

Examination of Fig. 7 shows that at sometemperature the thermodynamically con-trolled solubilities of neodymium, tech-netium, and ruthenium fall within the con-centration band. Above the band, thevolubility would cause the transport away oftoo much material, and below the band, oftoo little. However, this conjunction occursat different temperatures for each species.

around 150 degrees. Because of its retro-grade volubility, neodymium achieves thenecessary volubility below 25 degreesCelsius. The volubility temperature rangesfor ruthenium and technetium agree withtemperatures estimated for the reactor zoneduring its thermal period, Thus, the transportof ions by convecting hydrothermal fluidscan account for the deficiency of these twoelements but not for that of neodymium.

But there is a further complication. If thefluid temperature was 150 degrees Celsius,just hot enough to remove the appropriatequantity of technetium, then neither

Temperature (oC)

Fig. 7. The temperature dependence of the equilibrium solubilities for various ionicspecies of ruthenium, technetium, and neodymium. The range of estimated concentra-tions of these three elements in the groundwater at Oklo is superimposed on thediagram as the horizontal band centered at 10-10 moles per liter. Solubilities withinthis band indicate chemical processes and conditions that could have removedelements from the reactor zones at the appropriate rate. Solubilities above the bandmay have also occurred if other processes limited the rate of dissolution. The fact that

solution of tetravalent RU0 2 was not a factor in the transport of ruthenium.(’Uncertainty ranges for the volubility curves have been eliminated for clarity.)



ruthenium nor neodymium would have beentransported away. On the other hand, if thefluid temperature was 350 degrees, hotenough to remove the necessary quantities ofruthenium, the volubility for TcO-4 wouldhave been so high that all the technetiumwould be gone. Since roughly equal quan-tities of each were removed from the reactorzone. it is likely that the rate of migrationwas controlled by diffusion of each elementfrom the host mineral rather than by subse-

quent transport away from it as determinedby the solubilities.

The retrograde volubility of neodymiumsuggests that it was redistributed after thethermal period. The fact that it appears tohave been redistributed on a smaller scalethan ruthenium or technetium agrees withthis suggestion.

Although ruthenium and technetium mi-grated during the thermal period of thereactor, the overall transport was relativelyinsignificant. These elements were trans-ported only a few tens of meters within amillion years after reactor operation. Rates

of transport were thus on the order of 1O-5

meters per year whereas fluid movementcould have been up to 5 meters per year. Inother words, the pelitic sandstones surround-ing the reactors had a remarkable ability tocontain these nuclear products even afterthey were released from the uraninite. Theextreme temperature dependence of thesolubilities suggests a suitable mechanism forthis retention capacity. As the fluids left thehot reactor zones and began to cool, theybecame supersaturated with respect to theruthenium and technetium oxyions, and theirreduced form could have been deposited inthe rocks adjacent to the hot regions. If thismechanism is correct, aureoles reflectingtemperature profiles around the hot zonesshould be found. That is, supersaturation forR u O-2

4 occurs around 200 degrees higherthan for TcO4, and ruthenium should befound closer to the reactors than technetium.This temperature dependence may also helpexplain the apparent fractionation of

ruthenium and technetium during thethermal period.

In general, our studies have revealed thebroad outline of a possible mechanism forthe geochemistry of the Oklo reactors. Muchwork remains before the details are under-

stood. However, the possibility of long-termstorage of radioactive waste in undergroundsites has certainly been demonstrated by theexistence of these natural reactors.

Nuclear Waste Storage

The National Waste Terminal program,initiated in 1976 and presently directed byDOE, is investigating several sites as poten-tial repositories for high-level nuclear waste.One of these, Yucca Mountain, is locatednear the southwest corner of the NevadaTest Site and is being studied jointly by theLos Alamos, Sandia, and Lawrence Liver-more National Laboratories and the US.Geological Survey. Our role in the researchat this site is to develop part of the scientificbase to be used in deciding whether YuccaMountain will be appropriate as a repository.

Such a determination necessarily involvesa wide range of studies. Rock samples atvarious depths to 6000 feet need to be

analyzed for thermal and mechanical prop-erties, porosity, permeability, extent of frac-tures, mineral composition, sorptionproperties for different elements, and soforth. The water there is very mild, contain-ing mostly low concentrations of sodium andbicarbonate ions. Nevertheless, its physicaland chemical properties, such as theamounts of dissolved oxygen, organic mate-rial, microorganisms, and other constituentsthat might effect migration, must beevaluated. One of the most important tasksis the realistic modeling of the geochemistryso that predictions can be made with con-fidence. Here we will discuss only a couple ofexamples to illustrate the link between labo-ratory or field experiments and model devel-opment.

RETARDATION. The rock underlyingYucca Mountain is primarily welded andnonwelded tuff, which are volcanic materials.Welded tuff is hard, dense, and has excellentthermal properties. Nonwelded tuff belowthe water table there contains large amountsof clay and zeolitic minerals, and the rock ishighly porous but impermeable. In fact, thislatter rock can hold an average of 25 percent water by volume. Radioisotopes in theflowing water in fractures can diffuse into therock matrix and be trapped in the staticwater or be sorbed on pore walls. Retarda-tion describes the group of processes that,through interactions with rock, remove ele-ments from solution in moving water and isan important parameter to measure andmodel.

A variety of experimental techniques havebeen used on Yucca Mountain tuff tomeasure sorption, an important process inretardation. They include the batch processon either crushed rock or tablets of wholerock, flow through columns of crushed mate-rial, continuous circulation of groundwaterthrough crushed material, and radioisotopetransport through cylinders or blocks of rockthat contain real or artificial features. Quitesimilar sorption parameters are obtained forall these techniques if care is taken in bothexperimental procedure and in proper mathe-matical interpretation of the measurements.

For example, Fig. 8 shows batch sorptiondata for cesium as a function of depth in a

series of layered volcanic materials. Al-though the sorption ratio can easily beexperimentally reproduced within a range of20 per cent, it varies with depth by morethan two orders of magnitude, depending onthe mineral composition of each sample.

The principal phases that aid in sorptionand cause the dramatic changes in Fig. 8include hydrated volcanic glasses; smectiteclays, and zeolites. The glasses are the least

important for sorption, but they are veryreactive and can alter to other minerals withchanges in hydration if heated in the pres-ence of water. Smectite clays are reversibly


expandable with highly sorptive properties,although the properties can be modified in adetrimental way by prolonged exposure toheat. Zeolites, with their open lattice struc-ture, are the most highly sorptive minerals,

Because tuff samples may be composed ofseveral sorbing minerals, we are currentlydeveloping ways to predict sorption ratios bycombining the effects of several minerals.One approach uses the sorptive mineralcontent, or SMC, a weighted sum defined by

0

where W i is the weighting factor for eachmineral phase and Xi is the per cent abun-dance of each phase.

Figure 9 demonstrates the results of thisapproach for the cesium sorption data.Weighting factors were determined relativeto a zeolite, clinoptilolite, by directly measur-ing the sorption of cesium on pure mineralsor by inferring it from measurements withmixtures. The resulting values are plotted inFig. 9 along with the theoretical relationship(solid line) based on thermodynamic data.Note that almost all the data lie within theenvelope (dashed lines) for an uncertaintyfactor of 3. Procedures similar to this shouldeventually allow us to predict sorptiveproperties for the tuffs along pathways to theenvironment. However, sorption of cesium isprobably the simplest test of this concept.Any such procedure is critically dependentupon accurate determination of the mineralphases available to the groundwater.

KINETICS OF SORPTION. Another aspectthat must be understood before an adequatetransport model can be developed is thekinetics of sorption. This is especially true oftuff because its high porosity makes diffusionof waste elements into the rock matrix a verysignificant retardation process for radio-

38

Fig. 8. The sorption ratio for cesium as a function of depth into earth. Because therock at the Yucca Mountain site is made up of series of layers of volcanic material,each with its own mix of minerals, the sorption of cesium varies dramatically.Experimental reproducibility of the data shown here is about 20 per cent.

isotopes.Figure 10 shows the sorption of strontium

onto a sample of tuff as a function of time(circles). Three types of fits were made to thedata. The first and least successful. labeledsimple diffusion, is based on diffusion into aplane sheet. In this case the equilibriumconcentration of strontium between the grainsurfaces and the pore water, defined by thesorption parameters. is reached instan-taneously and sorption is treated as beinglinear with concentration. However, thenumber of available sorption sites may belimited so that at higher concentrations sorp-tion eventually drops below the linear curve.This nonlinearity complicates the diffusionequations by giving the diffusion coefficient aconcentration dependence. A finite-dif-ference solution of the nonlinear diffusionproblem was made and is also plotted in Fig.10. The agreement with experimental data isbetter in this case but still does not reproducethe early time points. Attempts to adjust the

Sorptive MineralContent (%)

Fig. 9. Sorption ratios for cesium arehere plotted as a function of sorptivemineral content (SMC), a parameterthat weights the sorptive qualities ofeach mineral in a sample. The solid lineis calculated from thermodynamicparameters; the dashed lines indicate anuncertainty factor of +3.



Equilibrium Sorption

Simple Diffusion

Relative Time

MODEL EQUATIONS

Simple Diffusion Slow Sorption

Fig. IO. Modeling the kinetics of sorption of strontium on tuff. The fraction ofstrontium sorbed on the rock approaches equilibrium amounts as relative timeincreases (relative time is Dt/L2 where D is the apparent diffusion coefficient for thematerial and L equals half the sample thickness). Simple diffusion involves a constantdiffusion coefficient, whereas nonlinear diffusion involves a concentration-dependentcoefficient. The best fit is achieved with a slow sorption model that includes areversible immobilization of part of the material in the rock matrix.


fit led to values for the adjustable parameters

that were not consistent with other data.The third type of fit. labeled slow sorption

in Fig. 10. is like the first model except thatthe equilibrium concentration between grainsurfaces and pore water is not reachedinstantaneous>. The process was consideredto be a reversible reaction with the solid tuffin which part of the dissolved species is freeto diffuse within the material and part isimmobilized in the tuff. Mathematically, aterm representing the rate of removal of theimmobilized component was subtractedfrom the simple diffusion equation. Theagreement between the data and the shape ofthe curve based on this approach was quitegood. Moreover, the parameters determinedfrom the best fit agreed with other data.

The consequences of these kinetic studiesmay be profound for migration of waste inwhich the radioisotopes are carried byrapidly moving water in fractures. The sorp-tion rate constants determined by us will beused in calculations with the nonequilibriumversion of our three-dimensional transport

code (TRACR3D) to model the results offlow in fractured rock. Our models arebecoming detailed enough so that we canextrapolate from the short time scales oflaboratory experiments to the long timescales required for storage of nuclear waste.

In general, our studies of the geochemistryof the Yucca Mountain site, although farfrom complete, so far indicate that the

geochemical properties and setting of the sitewill strongly inhibit the movement of radio-isotopes by flowing groundwater to the ac-cessible environment.

We have discussed examples from threeareas in the study of radioisotope migration.There are many more published examples ofsuch Los Alamos studies, and there areconsiderable possibilities for future investiga-tions. The field remains vital and excitingbecause these radioisotopes act as tracers,illuminating processes generally hidden fromus by complexity and by time in the earth’scrust. ■

39

Further Reading

H. W. Bentley, F. M. Phillips, S. N. Davis, S. Gifford, D. Elmore, L. E. Tubbs, and H. E. Gove, “Ther-monuclear 36CI Pulse in Natural Water,” Nature 300, 737 (1982).

Darleane C. Hoffman, Randolph Stone, and William W. Dudley, Jr., “Radioactivity in the UndergroundEnvironment of the Cambric Nuclear Explosion at the Nevada Test Site,” Los Alamos ScientificLaboratory report LA-6877-MS (July 1977).

IDarleane C. Hoffman, “A Field Study in Radionuclide Migration," in Radioactive Waste in GeologicStorage, Sherman Fried, editor (American Chemical Society, Washington, D. C., 1979), p. 149.

William R. Daniels, editor, “Laboratory and Field Studies Related to the Radionuclide Migration Project,October 1, 1981 - September 30, 1982,” Los Alamos National Laboratory report LA-9691-PR (May1983). Previous reports in this series are LA-8670-PR (February 1981) and LA-9 192-PR (February1982).

E. A. Bryant, G. A. Cowan, W. R. Daniels, and W. J. Maeck, “Oklo, an Experiment in Long-TermGeologic Storage,“ in Actinides in the Environment, Arnold M. Friedman, editor (American ChemicalSociety, Washington, D. C., 1976), p. 89.

George A. Cowan, “A Natural Fission Reactor,” Scientific American 235, 36-47 (July 1976).

David B. Curtis, Timothy M. Benjamin, and Alexander J. Gancarz, “The Oklo Reactors: NaturalAnalogs to Nuclear Waste Repositories,“ in The Technology of High-Level Nuclear Waste Disposal, Vol.1, prepared and edited by Peter L. Hofmann, Battelle Memorial Institute, Project Management Division(Technical Information Center, Oak Ridge, Tennessee, 1983), p. 255.

B. R. Erdal, W. R. Daniels, D. C. Hoffman, F. O. Lawrence, and K. Wolfsberg, “Sorption and Migrationof Radionuclides in Geologic Media, “ in Scientific Basis for Nuclear Waste Management, Vol. 1, G. J.McCarthy, editor (Plenum Press, New York, 1979), p. 423.

B. R. Erdal, B. P. Bayhurst, W. R. Daniels, S. J. DeVilliers, F. O. Lawrence, E. N. Vine, and K.Wolfsberg, “Parameters Affecting Radionuclide Migration in Geologic Media,” in Scientific Basis forNuclear Waste Management, Vol. 2, C. J. Northrup, Jr., editor (Plenum Press, New York, 1980), p. 609.

B. R. Erdal, B. P. Bayhurst, W. R. Daniels, S. J. DeVilliers, F. O. Lawrence, J. L. Thompson, and K.Wolfsberg, “Parameters Affecting Radionuclide Migration in Argillaceous Media,” in The Use ofArgillaceous Materials for the Isolation of Radioactive Waste (Nuclear Energy Agency of the

IOrganization for Economic Cooperation and Development, Paris, 1980), p. 55.

B. R. Erdal, B. P. Bayhurst, B. M. Crowe, W. R. Daniels, D. C. Hoffman, F. O. Lawrence, J. R. Smyth,J. L. Thompson, and K. Wolfsberg, “Laboratory Studies of Radionuclide Transport in Geologic Media,”in Underground Disposal of Radioactive Wastes, Vol. 11 (International Atomic Energy Agency, Vienna,1980), p. 367.


Many people from the Isotope and Nuclear Chemistry’ Division have contributed to thework described in this article. Those pictured above are: Tim Benjamin, Ernie Bryant,John Cappis, Dave Curtis, Clarence Duffy, Alex Gancarz, Darleane Hoffman. EdNorris, Al Ogard, Jim Sattizahn, and Kurt Wolfsberg. Others, who weren’t availablefor the photograph, are Bill Daniels, Bruce Erdal, Bob Rundberg, and RosemaryVidale.


Radioisotopes trace out mechanisms important both to … · 2016-10-21 · Isotope and Nuclear...

Documents

Transcript of Radioisotopes trace out mechanisms important both to … · 2016-10-21 · Isotope and Nuclear...