A University of Cambridge-led team publishing in Nature Geoscience has combined 9,000 rock samples with seismic images of Earth's interior to show that CO2-rich carbonatite magmas, the primary host of rare earth element deposits, form preferentially along the steep margins of ancient, thick continental roots. The finding offers exploration geologists a globally applicable predictive framework for the first time, with direct implications for supply diversification efforts as China's export controls continue to disrupt non-Chinese rare earth supply chains.
Introduction
A study published May 22, 2026, in Nature Geoscience has produced what its authors describe as a global predictive framework for rare earth element discovery, one rooted not in surface geology but in the deep architecture of Earth's oldest continental structures. Led by researchers at the University of Cambridge's Department of Earth Sciences and published under the DOI 10.1038/s41561-026-01990-7, the paper demonstrates that the CO2-rich carbonatite magmas responsible for hosting more than half of the world's REE resources form systematically along the steep margins of cratons, the ancient, thick lithospheric cores that underpin major continents.
The findings arrive at a moment of acute supply chain pressure. China mines approximately 60% of global rare earths, processes roughly 90% of them, and manufactures around 94% of REE-containing magnets used in clean energy and electric vehicles, according to the International Energy Agency's 2025 outlook. Export controls introduced in April and October 2025 on seven heavy rare earth elements, including dysprosium, terbium, and yttrium, triggered licensing approval rates below 25% for European firms and drove European spot prices to as much as six times their Chinese equivalents. Against that backdrop, a geological tool capable of systematically directing exploration capital toward craton-margin zones in under-explored jurisdictions carries real strategic weight.
The Cambridge paper, authored by Dr. Emilie Bowman, Professor Sally Gibson, Dr. Siyuan Sui, and Professor Sergei Lebedev, is the product of a multi-year program called REE-LITH, which was inspired by Gibson and Lebedev's hypothesis that differences in lithospheric properties might govern not only where carbonatites form but also their degree of rare earth enrichment. That hypothesis has now been confirmed at a global scale, and the team is already preparing to extend its model backward in time to rocks older than 200 million years, which host the majority of the world's economically significant REE mines.
The Science: Seismic Imaging Meets Geochemistry
The methodological core of the study is a deliberate fusion of two disciplines that had rarely been applied together at global scale. Dr. Bowman assembled chemical data on 9,000 igneous rock samples from around the world, all enriched in dissolved CO2, a key ingredient that enhances the potential for rare earth element concentration. Simultaneously, Dr. Siyuan Sui developed new computer-generated images of lithospheric thickness and structure by processing seismic data extracted from earthquakes. The two data sets were then cross-referenced to test whether magma chemistry varied systematically with the underlying lithospheric architecture.
The answer was unambiguous. As Professor Lebedev explained: "Using seismic waves from earthquakes, we can create a slice-through image of the lithosphere, much like a sonar can pick out features on the seabed. From this mapping, we can see that lithospheric thickness plays a guiding role in where we find these deposits." The geophysical imaging drew on the extensive bank of seismic data held at Cambridge's Bullard Laboratories, and the statistical analysis of magma geochemistry was designed to test the co-variation between CO2 content and lithospheric thickness across the full sample population.
The gradient that emerged is striking in its regularity. Basanites, which carry less than 5 wt% CO2, erupt through seismically slow, thin, non-cratonic lithosphere typically shallower than 100 kilometers. Moving along the lithospheric thickness spectrum, nephelinites, melilitites, and ultramafic lamprophyres occur within progressively faster and thicker lithosphere. At the far end, lamproites and kimberlites, which can carry up to 20 wt% CO2 and are the primary hosts of diamond deposits, are emplaced on cratonic lithosphere exceeding 160 kilometers in thickness. Carbonatites, which exceed 25 wt% CO2 and are associated with economic deposits of REEs as well as phosphate, fluorite, niobium, and tantalum, occupy the steep gradient zones at craton margins, rather than craton interiors or thin off-craton settings.
Professor Gibson summarized the interpretive significance: "We needed to put together these two pieces of the puzzle, the rock chemistry and seismic data, to make the connection. Rocks with the right chemistry for enrichment occur only in very specific places, mainly along the steep edges of Earth's thickest and oldest lithosphere."
The Mechanism: Why Thick Lithosphere Concentrates Rare Earths
The physical explanation for the observed pattern lies in the pressure and thermal regime imposed by thick lithosphere on the underlying mantle. Where continental lithosphere is exceptionally deep-rooted, it holds the mantle rocks beneath at high pressures and relatively cool temperatures, suppressing broad-scale melting. Only tiny, volatile-saturated pockets of mantle material can melt under these conditions. These small-volume melts become progressively enriched in CO2 and incompatible trace elements, including the lanthanide series that constitutes the rare earth elements, as they evolve.
Critically, thick lithosphere acts as a physical trap. It prevents the small magma pockets from rising quickly to the surface, forcing them to stagnate at depth and undergo prolonged fractional crystallization and fluid exsolution. REE-rich hydrothermal fluids separate from the cooling carbonatite magma and infiltrate surrounding crustal rocks, producing the concentrated mineral assemblages that define economic deposits. As the research team describes it, the thick lithosphere essentially creates the chemical pressure cooker in which these metals accumulate.
This mechanism also explains why the analogy to diamond-bearing kimberlites is more than metaphorical. Gibson noted: "We know that lithospheric thickness matters for other special igneous rocks that host diamonds. Typically, diamond-hosting kimberlite rocks only occur in areas where the lithosphere is particularly thick. I thought it was time we tested if there was a similar relationship for carbonatites." The REE-LITH study confirms the structural parallel while also clarifying why carbonatites cluster at craton margins rather than craton interiors: the steep lithospheric gradient at the margin provides both the necessary thickness for CO2-rich melt generation and the structural conduits, often ancient fault systems, that allow those melts to ascend into the crust.
The finding aligns with the known distribution of major REE deposits. Bayan Obo, the world's largest light REE-iron-niobium deposit, sits on the northern margin of the North China Craton. Maoniuping occupies the rift belt along the western margin of the Yangtze Craton. Mountain Pass in California is located in a continental marginal depression along a cratonic boundary. More than 527 carbonatite occurrences have been recorded globally, but only approximately 30 host economic REE resources, and those 30 cluster, without exception, in cratonic margin settings.
Exploration Implications: A Predictive Tool for Under-Explored Jurisdictions
For the exploration community, the study's most consequential contribution is not the confirmation of what is already known about existing deposits but the framework it provides for identifying what is not yet known. Dr. Bowman stated: "Our research is beginning to provide a kind of predictive power for where we can expect these rocks and, by extension, their associated rare earth element deposits, to form." Professor Gibson was direct about the goal: "Having some kind of model that could predict the most likely locations for rare earth deposits is really the ultimate goal for many geologists."
The current study is deliberately limited to carbonatites and CO2-rich magmas younger than 200 million years, focusing on deposits formed after the main phases of continental breakup. The reason is methodological: older rocks have been disturbed by mountain building, continental rifting, and metamorphism, making geochemical and seismic interpretation more complex. But those older rocks are commercially more important. The team now plans to extend the model to pre-200-million-year sequences, which host most of the world's economically significant REE deposits and mines. As Gibson explained: "Now we have established this systematic behavior exists, we can go back further in time. It's going to be more challenging, but I'm hopeful that this will be a key step in predicting mineral occurrences."
Even in its current form, the framework opens systematic targeting of craton-margin zones in jurisdictions that have received comparatively little exploration attention. Cratonic margins in central and western Africa, northern Canada, Scandinavia, and parts of South America share the lithospheric architecture that the study identifies as favorable. A complementary data point: a 2025 study in ScienceDirect documented a newly discovered carbonatite occurrence at the Terghat structure in the peri-cratonic terrain of Morocco's Reguibat Shield, a 9.5-kilometer ring-shaped intrusion with REE-bearing phases, providing exactly the kind of real-world validation that the Cambridge framework would predict.
A separate but related finding published in Science Advances in April 2026 by Professor Carl Spandler and colleagues at Adelaide University identified ancient subduction zones as an additional dominant control on REE deposit formation. The two studies are complementary rather than competing: the Cambridge paper addresses carbonatite-hosted REE systems governed by lithospheric thickness, while the Adelaide work illuminates a different ore-forming pathway. Together, they signal a rapid maturation of the predictive geoscience underpinning global REE exploration.
Market Context: Geology Meets Geopolitics
The timing of the Cambridge study's publication is inseparable from the supply chain crisis reshaping global REE markets. China's April 4, 2025, export controls on seven heavy rare earth elements, including various derivatives of terbium, dysprosium, gadolinium, and lutetium, were followed by further restrictions announced in October 2025 and additional dual-use controls targeting Japan in early 2026. The effect on non-Chinese supply chains was immediate: export volumes fell sharply in April and May 2025, forcing some carmakers to cut utilization rates or temporarily shut down factories. European prices for affected elements reached up to six times Chinese domestic levels.
David Merriman, research director at Project Blue metals consultancy, assessed the durability of that disruption: "The ex-China market will continue to face bottlenecks in the supply of HREE products over 2026 and 2027 as alternative suppliers of HREEs are constructed and commissioned. Key elements facing disruption are yttrium, lutetium, terbium and dysprosium, flowing into the supply chains of magnet manufacturers, aerospace components, and electronics." Chris Berry, independent battery metals analyst, reinforced the price outlook: "As long as China continues its saber-rattling regarding dual use and export restrictions, this will serve to impede trade flows and elevate prices. As European nations rebuild their defense capabilities, the AI and data center thematic continues its enormous capex binge, and traditional demand drivers such as magnets and robotics continue their ascent, I can't see lower rare-earth pricing in 2026 or 2027."
The demand backdrop compounds the urgency. Global rare earth market volume is projected to grow from approximately 197 kilotons in 2025 to 273 kilotons by 2031, a compound annual growth rate of 5.6%. Magnetic rare earth demand specifically is projected to triple from 59 kilotons in 2022 to 176 kilotons by 2035, driven by EV traction motors and offshore wind turbines. The IEA's baseline scenario projects a 45% increase in rare earth demand between 2024 and 2030. Market value estimates place the global REE sector at roughly $14 billion in 2025 and project it reaching $41 billion by 2034. Against those numbers, the strategic logic of accelerating geological discovery is self-evident.
Building on my analysis of the REalloys-Tanbreez offtake deal in June 2026, a recurring theme in non-Chinese supply chain development is the acute shortage of heavy rare earth supply outside Chinese jurisdiction. The Cambridge framework's identification of craton-margin zones as the most geologically favorable targets for carbonatite REE discovery is directly relevant to that gap: the majority of undiscovered carbonatite potential in under-explored regions likely coincides with heavy-REE-enriched deposit types in precisely the jurisdictions, notably central Africa, parts of Greenland, and northern Canada, where exploration activity has been accelerating most rapidly.
What the Map Cannot Do: Processing Remains the Binding Constraint
The scientific achievement of the Cambridge study is substantial, and industry observers have largely received it as credible and potentially valuable. Seismic tomography combined with geochemical databases represents a legitimate and novel application of existing methodologies at a scale and rigor that prior work had not attempted. The systematic quantification of the CO2-to-lithospheric-thickness relationship across 9,000 samples, spanning the full compositional range from basanites to carbonatites, gives exploration geologists a quantitative basis for prioritizing targets that previously rested on qualitative geological intuition.
However, a candid reading of the study and its commercial implications requires acknowledging what the map cannot resolve. Gibson herself identified the central open question: while the team can now narrow down the regions where carbonatites should occur, they still need to establish why only certain carbonatites generate economically important rare earths. More than 527 carbonatite occurrences are on record; only about 30 are economically significant. The geological conditions that produce REE enrichment within a carbonatite system, including the degree of mantle metasomatism, the fluid chemistry during crystallization, and the secondary enrichment processes at the surface, remain imperfectly understood even in well-studied deposits.
Beyond the geological uncertainty, the structural reality of the rare earth supply chain imposes its own constraints. Hayley Channer, a rare earths expert at the United States Studies Centre at the University of Sydney, has noted that China's dominance was built through decades of integrated investment in mines, mineral processing, refining, alloying, and magnet manufacturing simultaneously: end-to-end supply chains that took generations to construct. Discovery of a favorable carbonatite outcrop in a cratonic margin zone in, say, the Democratic Republic of Congo or northern Quebec initiates a process that runs through feasibility studies, metallurgical testing, permitting, project finance, construction, and ramp-up before a single kilogram of separated rare earth oxide reaches a magnet manufacturer. As of early 2026, no Western facility has achieved commercial-scale processing of heavy rare earths.
The Cambridge study is best understood as a tool for improving the quality and efficiency of the exploration pipeline, not as a near-term fix for supply chain vulnerability. It makes the earliest stage of that pipeline, geological targeting, more rigorous and potentially more capital-efficient. That matters enormously over a 10-to-20-year horizon. But it does not compress the timeline for the processing, separation, and downstream manufacturing investments that represent the real binding constraints on Western supply chain independence.
Conclusion: A Framework That Reshapes the Starting Line
The Cambridge REE-LITH study represents a genuine inflection point in the geoscience of rare earth exploration. By establishing, for the first time at a global quantitative scale, that carbonatite formation and REE enrichment are systematically governed by lithospheric thickness, the team has converted what was previously a qualitative geological heuristic into a mappable, testable, and extendable predictive framework. The analogy to kimberlite-diamond geology is apt: once that lithospheric relationship was understood for diamonds, exploration targeting improved dramatically. A comparable improvement in carbonatite-REE targeting is now plausible.
The next phase of the work, extending the model to rocks older than 200 million years, is where the commercial stakes are highest. The world's major REE mines, including Bayan Obo, Mountain Pass, and the African carbonatite systems, are hosted in pre-200-million-year geology. Demonstrating that the lithospheric thickness relationship holds in those older, more structurally complex sequences would transform the tool from a proof of concept into a globally deployable exploration platform.
For critical minerals policymakers and supply chain strategists, the study offers a reminder that the geological endowment of non-Chinese jurisdictions is likely larger than current discovery records suggest, and that the limiting factor in exploration has been predictive capability as much as physical access. As the DOE's recent investments in processing technology, covered in my earlier analyses of the $45.7 million midstream processing awards and the ORNL technology licensing pipeline, begin to address the separation and refining gap, the upstream discovery pipeline needs to be filling simultaneously. The Cambridge treasure map does not locate the treasure precisely; it narrows the search to the right kind of terrain. In exploration economics, that is worth a great deal.