In the OSTI Collections: Isotope Separation

Dr. Watson computer sleuthing scientist.

Article Acknowledgement:

Dr. William N. Watson, Physicist

DOE Office of Scientific and Technical Information


Alternate Text Placeholder

The atoms that constitute the matter we observe with our unaided senses differ in their numbers of electrons, protons, and neutrons. 


While electrically neutral atoms have as many negatively charged electrons in their periphery as they have positively charged protons in their nuclei, atoms can also have more or fewer electrons than that.  In fact, atoms’ transferring or sharing electrons among themselves is one way that they chemically bond with each other.  The number of protons in each atom’s nucleus determines the positive charge of that nucleus, which determines the strength of its electrical attraction of negative electrons, which almost completely determines the electrons’ possible arrangements and thus each atom’s possible chemical reactions.   


Except for atoms of the least massive elements, the number of neutrons in their nuclei affects their chemical behavior much less.  The positive and negative electric charges of a neutron’s components cancel each other out, so their electric fields scarcely affect the electrons.  Each neutron also has a magnetic field (as do the protons and electrons), which has a small effect on the atom’s electron distribution.  Both protons and neutrons also have significant nonelectromagnetic interactions with any electrons that happen to be within a femtometer[Wikipedia] or so of them—just millionths of the atom’s width.  The most significant chemical effect that adding neutrons to a nucleus would have is to make the nucleus more massive, making it oscillate less around the atom’s center of mass, thus keeping it closer on average to each electron, and thereby increasing the strength of the electromagnetic interaction—though not as much as the addition of another proton to the nucleus would strengthen it.  In short, electrons in atoms that have equal numbers of protons but different numbers of neutrons generally have such similar behavior that their chemistries are nearly identical; such atoms are therefore considered to be atoms of the same element.  Since they’re represented at the same place on the periodic table, they’re known as isotopes[Wikipedia] of the same element (Greek isotopos, “in the same place”).  A particular isotope of a given element is designated by the element name or chemical symbol together with the total number of protons and neutrons in that isotope’s atomic nucleus—for instance, “oxygen-16” or “16O” represents the isotope of oxygen in which the atom’s nuclei have eight protons and eight neutrons. 


On the other hand, the number of neutrons in a nucleus has a significant effect not only on its mass, but on whether that nucleus can undergo particular nuclear or electromagnetic interactions.  Thus when certain nuclear masses or interactions are called for, it’s useful to extract the appropriate isotopes from among others of the same element.  However, many kinds of chemical reactions that separate different elements from each other are much less effective at separating different isotopes of the same element unless the element’s atoms are among the least massive ones like hydrogen, helium, &c. 


To separate any element’s isotopes by their differing nuclear interactions would present problems.  As noted above, for a nuclear interaction to have any significant chance of occurring, the particles involved have to be within about a femtometer of each other, so these interactions can’t be used to manipulate isotopes remotely.  Besides, such interactions often change nuclei into nuclei of a different isotope of the same element or even a different element.  The mass differences are much more useful for isotope separation; as mentioned, these mass differences have only a slight chemical effect for most elements, but extremely sensitive chemical processes can treat different isotopes differently enough to separate them.  Many other separation processes involve subjecting mixtures of atoms to the same forces, which accelerate the less massive atoms more, thus making the atoms of different isotopes move apart from each other. 


Alternate Text Placeholder

A possible miniaturization of one process


Several isotope separation methods are briefly mentioned in a 2012 Los Alamos National Laboratory slide presentation[SciTech Connect] as context for a newer method named in the presentation’s title, “On Isotope Separations by Micro-Scale Thermal Diffusion”.  One potential application of the new method is to separate helium-3 from the much more common helium-4 for several purposes—e.g., as a neutron detector, in cryogenic refrigerators that cool to temperatures very close to absolute zero, and for showing human lung airways by magnetic resonance imaging.  Another application of the separation method is to extract a radioactive power source for spacecraft generators and heaters, plutonium-238, from among other isotopes of plutonium. 


The separation method, micro-scale thermal diffusion, is a variant of the industrial-scale thermal diffusion process that was used during the Manhattan Project[Wikipedia; Wikipedia] as a first-stage process for increasing the fissile-isotope fraction of uranium.  In that process, uranium hexafluoride molecules circulating between a steam-heated surface and a water-cooled surface tended to diffuse in different directions according to their mass:  the less massive molecules containing uranium-235 diffused toward the hot surface while the more massive molecules containing uranium-238 diffused toward the cold surface.  In the Manhattan Project plant, the surfaces were pipes:  an inner pipe contained the steam, the uranium hexafluoride flowed through another pipe surrounding the steam pipe, and the water circulated through an outer pipe surrounding the uranium-hexafluoride pipe.  The micro-scale thermal diffusion process works on the same principle, but its heat sources are lasers trained on small spots roughly a tenth of a millimeter across instead of steam heating an entire pipe.  The slide presentation describes details of Los Alamos experiments with the micro-scale process on helium-3 and helium-4, and mentions experiments with other gases.  Results were similar in all cases, and look typical of many early-stage experiments with a new technology:  redistribution of the isotopes between the product and waste streams apparently occurred, but not consistently; in one of the three helium experiments, the product stream actually had more helium-4 rather than less. 


Figure 1

Figure 1.  Left[Wikimedia Commons]:  The dark building next to the powerhouse with the smokestacks is the S-50 Thermal Diffusion Process Building in Oak Ridge, Tennessee.  This facility was used during the Manhattan Project as a first-stage isotope separator whose product, a gas whose percentage of uranium-235 was higher than that of the gas fed into it, was in turn fed into the project’s gaseous-diffusion separation plant for further enrichment.  (The S-50 plant’s waste product had a lower percentage of uranium-235 than its input did.)  Right[Wikimedia Commons]:  Sets of coaxial pipes through which uranium hexafluoride gas was circulated to separate the molecules comprising uranium-235 atoms and uranium-238 atoms.  The innermost pipe contained steam, the next pipe contained the uranium hexafluoride gas, and the outer pipe contained water.  The less massive molecules diffused toward the hot inner pipe while the more massive ones diffused toward the cooler outer pipe.  The lighter gas tended to rise while the heavier one fell, effecting their separation.  (From Wikimedia Commons; public domain works produced for the United States Government by official photographer Ed Wescott[Wikipedia].) 


The presentation concluded that the micro-scale process was “very immature” at the time, that a different method using zero-viscosity superfluids[Wikipedia] and distillation to separate helium-3 from the helium-4 in natural gas was “more ready”, and that quite different means of producing plutonium-238 made micro-scale thermal diffusion “even more immature” for that purpose.  However, the possibility that micro-scale thermal diffusion could be developed into an effective technique is of interest for an additional reason:  micro-scale plants would add to the difficulty of nuclear-weapon antiproliferation efforts because any such small plants used to separate isotopes for nuclear weapons would be more difficult to observe.  On the other hand, the presentation notes that additional processes, which can be more easily observed, are still required to produce weapon materials, thus limiting the effectiveness of obscuring separation plants.  The last slide before the presentation’s reference section recommends a small micro-scale thermal diffusion (MSTD) research program as “prudent to characterize the observable signatures of MSTD activity and to lay ground work for possible MSTD application to other materials of interest.” 



Alternate Text Placeholder

Separation to measure isotope content


One way to determine precisely how small a trace of one isotope is in a material requires using that material to dilute a different one that’s enriched with a large, known amount of the isotope, and then making the easier measurement of the mixture’s isotope ratio.[Wikipedia]  For that measurement to be useful, the isotope amounts in the enriched material have to be precisely known.  How to produce precision-enriched materials is the subject of two articles in the Journal of Radioanalytical and Nuclear Chemistry by Idaho National Laboratory researchers.  The earlier article, “Advancement of isotope separation for the production of reference standards”[SciTech Connect], describes the upgrading of an isotope separator that had been used in the 1970s to produce suitably enriched materials.   The separator works by projecting atoms that have more protons than electrons, which thus have a net positive charge, into a magnetic field.  If the atoms have the same positive charge and enter the field with the same velocity, they will experience equal sideways forces; but if they also have different masses, they will not be equally accelerated by the same forces.  In particular, if the atoms are all of one element but are different isotopes of that element, the atoms with more neutrons will be deflected less, and will thus wind up at different locations from the less massive atoms.  The later article[DoE PAGES] described “Production of highly-enriched 134Ba for a reference material for isotope dilution mass spectrometry measurements”—in this case, measurements of a sample’s barium-137 content that was produced by radioactive decay of the sample’s cesium-137 content.  Cesium-137 has a half-life of about 30 years, so accurately measuring a sample’s barium-137 decay product amounts to measuring its age.  This report describes experiments in collecting already-separated barium-134 atoms so they can be used for isotope-dilution measurements of barium-137. 


Figure 2

Figure 2.  Top:  photo (left) and schematic diagram (right) of a device to separate different isotopes by mass.  Plasma emits electrically charged atoms of multiple isotopes that enter the magnetic field of a separation magnet.  If the different-mass isotopes all entered the magnet with the same velocity, they would experience equal forces and different deflections and be cleanly sorted by mass; since the initial velocities vary, some isotopes of greater mass will be deflected toward the same spot as isotopes of lesser mass, but the atoms collected at each endpoint will be more nearly all of one isotope than they were at the source.  (From “Advancement of isotope separation for the production of reference standards”[SciTech Connect].)  Bottom:  Blocks for collecting atoms behind slits at the endpoints of the atoms’ paths (left) and removable collection foils made for the slits (right).  (From “Production of highly-enriched 134Ba for a reference material for isotope dilution mass spectrometry measurements” [DoE PAGES], p. 269.) 



Alternate Text Placeholder

Isotope separation at a new facility


The operating principle of Idaho National Laboratory’s isotope separator is also to be used in the new Facility for Rare Isotope Beams (FRIB) being built at Michigan State University.  This facility is designed to advance understanding of nuclear physics, nuclear astrophysics, and fundamental interactions of nuclei by enabling experiments with rare isotopes.  The rare isotopes are to be produced by a technique that has been used elsewhere:  more common atoms are accelerated to high speed and smashed into a target, which will fragment their nuclei into many different smaller ones, including those of the desired rare isotopes; these will be separated from the others by passage through a magnetic field that deflects atoms with different charge/mass ratios differently.  Since the fragmentation products will fly off in many directions, only some of them will simply pass through the magnetic field; others will irradiate the electromagnet that produces the field by colliding with it and heating it, so the magnet will have to withstand this irradiation.  Some of the considerations explored by the magnet designers are documented in “High Radiation Environment Nuclear Fragment Separator Magnet ”[SciTech Connect]


The designers analyzed the option of using “high-temperature”[Wikipedia; OSTI] superconducting wire to build the magnet, which would require no power consumption to overcome electrical resistance and less refrigeration than “ordinary”[Wikipedia] superconducting wire would need to maintain its resistance at zero.  Making the magnets curved to follow the path of the selected isotopes presented a manufacturing problem since wire coils with a curved inner segment tend to unwind while they are being wound.  Two different winding procedures were tested, one of which involved making wire coils with a straight inner segment; both procedures overcame the problem.  A different problem involved the design of the coils’ mechanical support.  Any electromagnet’s coils have to be held firmly in place while the magnet is on to keep the coils from distorting themselves (and their magnetic fields) by the strength of their own magnetism, but the support for these coils not only has to withstand their mechanical forces but also the heat from nuclear-fragment collisions.  At the same time, the magnet coils can’t get too hot or they’ll suddenly lose their superconductivity and severely distort themselves as their magnetic field collapses.  Coil supports made of organic heat-insulating materials are used in other superconducting magnets to minimize the required refrigeration, but their rapid degrading under nuclear-fragment irradiation would make them unsuitable for use in the FRIB.  Thus two metallic-support designs were explored that compared the amount of tolerable coil deformation and heat leakage.  While tests showed that such supports could work, the FRIB project’s time and funding constraints made the use of high-temperature superconducting coils risky, so the facility’s electromagnets will use resistive copper wire instead.  However, the solutions explored may be practical for other projects that can incorporate them from the beginning. 


Individual experiments at the Facility for Rare Isotope Beams will generally involve only a few of the hundreds of isotopes produced by smashing atoms into a target.  Many of the other isotopes, though, have long half-lives and could have practical uses in medicine, national security, or the environment.  The report “Preparing for Harvesting Radioisotopes from FRIB”[SciTech Connect] describes how researchers at Hope College in Holland, Michigan and at Washington University in St. Louis, Missouri designed and built a system for capturing these byproducts in small volumes of water and chemically extracting them from the water.  At the time of their report, they had successfully extracted sodium-24 and copper-67, and were to try extracting vanadium-48 and krypton-85 next.  The researchers also held workshops on harvesting isotopes at FRIB at which participants from multiple institutions discussed candidate radioisotopes to harvest, the need for harvesting them, and how to harvest them at FRIB.  Their consensus was that such harvesting would supply novel isotopes for the nation’s research needs, and they accordingly recommended that the Facility for Rare Isotope Beams be enhanced with radiochemistry equipment to realize the opportunities. 



Alternate Text Placeholder

Chemistry and isotope separation


If the motion of atoms’ negative electrons and positive nuclei distributes them asymmetrically at some instant, their attractions and repulsions of neighboring atoms’ electric charges can induce enough asymmetry in those charges’ distribution to result in an overall attraction between the atoms, if opposite charges in the different atoms are induced to be closer to each other than like charges.  This attraction, known as a van der Waals force[Wikipedia], can hold the atoms together, even if the atoms are of noble gases like helium, which have such low chemical reactivity that relatively few compounds of them have been discovered.  Being among the least massive of atoms, helium atoms have so little mass that a one-neutron difference between two isotopes has a significant effect on how close their nuclei stay to their electrons—and thus on the strength of their van der Waals interactions.   The report “Novel Concepts for Isotopic Separation of 3He/4He”[SciTech Connect] describes how the size of this difference was calculated for several helium compounds to see if it might provide a practical means of separating large amounts of the very rare isotope helium-3 from the less rare helium-4 component of natural gas.  The results suggested that helium-xenon compounds could be used in a separation process.  Determining whether such a process would best operate on the compound as a liquid or gas, or at a high enough temperature and pressure to remove the liquid-gas distinction altogether, required actual experiments yet to be designed at the time of the report.  An ideal system would also have a means of separating and recycling the xenon. 


The reports described above discuss how people have figured out various ways to separate isotopes artificially to accomplish particular ends.  A different application of human ingenuity has been to see how certain isotope separations that already occur in nature can help us understand the associated natural phenomena. 


Different isotopes of the same elements have been found to diffuse through volcanic liquids at different speeds.  How fast the different isotopes diffuse is affected by which elements they’re isotopes of and by the composition of the volcanic liquids.  To better understand the exact mathematical relationship between isotope diffusion speeds, element identity, and liquid composition was the object of experiments reported in “Influence of liquid structure on diffusive isotope separation in molten silicates and aqueous solutions”.[SciTech Connect]  The relationship is a significant one, since variations in the isotopic content of rocks and minerals provide clues about geochemical transport processes, temperatures, and even the classification of planets and meteorites. 


The experiments, like many initial explorations in the history of research, involved artificially simplifying a complex natural phenomenon to clarify its essential features.  In this case cylinders composed of two different minerals in well-defined proportions stood in for arbitrarily shaped volcanic liquids of complex multimineral composition.  Cylinders made of 99% or 85% albite[Wikipedia] and 1% or 15% of a second mineral (either anorthite[Wikipedia] or diopside[Wikipedia]) were pressed together at 800 ± 10 megapascals[Wikipedia] (roughly 8,000 times atmospheric pressure) and heated by 2.5 kelvins[Wikipedia] per second to 1723 ± 3 kelvins (1450 ± 3 °C).  Keeping the cylinder pairs at the high temperature for two hours allowed different components of the minerals to diffuse from whichever cylinder had the higher concentration of it into the other cylinder.  Afterward, the cooled cylinders were sliced apart and the slices’ compositions examined to see how much each component had diffused and how far.  Differences found in the amounts of each single element’s isotopes were noted and related to the cylinders’ compositions. 


A mathematical expression that fits the data suggests that, if silicon atoms from the liquid matrix are more mobile than atoms of another element diffusing through it, less massive isotopes of that element won’t diffuse much faster than more massive ones, likely because what is actually diffusing are complexes of those atoms plus the several other atoms of the liquid matrix that each of them is strongly bound to—complexes that are more similar in mass than unattached atoms.  Conversely, diffusive separation of isotopes is more efficient when the isotopes’ atoms are more mobile than matrix atoms, being less strongly bound to them and able to move more as separate entities.  The researchers’ mathematical model of the data and their inference of the underlying physical causes provide a tool for predicting just how differently different isotopes will diffuse in many geochemical environments, and a basis for a more comprehensive theory of isotope separation in liquid solutions.  Such a theory could be applied to both natural and artificial isotope separations. 

Figure 3

Figure 3.  Left:  Chemical concentration profiles of major components of molten albite-diopside cylinders after diffusion at high temperature and pressure.  The cylinder on the left was initially a mixture of 85% albite and 15% diopside, while the cylinder on the right was a 99% albite—1% diopside mixture.  Dashed lines represent the starting concentrations, circles represent the concentrations measured from slices of the cylinders at the end of the experiment, solid lines are the best fit curves according to the experimenters’ mathematical model of the diffusion process, and gray shading represents estimates of uncertainty in the model’s diffusion coefficients.  Right:  Differences in the diffusion profiles of different isotopes of calcium (40Ca and 44Ca) and of magnesium (24Mg and 26Mg).  Circles represent measured differences, gray horizontal bar represents initial constant isotope ratio along the two cylinders, and solid lines are best-fit curves according to the experimenters’ mathematical model.  (From “Influence of liquid structure on diffusive isotope separation in molten silicates and aqueous solutions”[SciTech Connect], pp. 3 and 5.) 




Alternate Text Placeholder






Wikimedia Commons



Additional references



Reports available through OSTI’s SciTech Connect



Reports available through DoE PAGES



Alternate Text Placeholder