Earth’s mantle is a roughly 2,900-kilometer-thick shell of mostly solid rock that sits between the thin crust we live on and the liquid iron core at the planet’s center. It makes up about 84 percent of Earth’s volume, and almost everything dramatic that happens at the surface, from earthquakes and volcanic eruptions to the slow drift of continents, traces back to what is going on inside it. The mantle is not a uniform slab of rock, though. It is layered, chemically varied, and full of slow-motion currents that have been reshaping the planet for billions of years.
Where the Crust Ends and the Mantle Begins
The boundary between Earth’s crust and the mantle is called the Mohorovičić discontinuity, or “Moho” for short. It was identified over a century ago because seismic waves from earthquakes speed up sharply when they cross it. Beneath continents, the Moho sits at roughly 30 to 50 kilometers deep, while under the oceans it can be as shallow as 5 to 10 kilometers. The speed of compressional waves just below the Moho under continents averages about 8.2 kilometers per second where the boundary is 30 kilometers deep, dropping to around 8.0 kilometers per second where it deepens to 50 kilometers. That velocity decrease with depth is steep enough to indicate the start of a low-velocity channel, a zone where the rock is slightly softer and closer to its melting point, right at the top of the mantle beneath continental regions.1Geophysical Journal International. Wave Velocities below the Mohorovičić Discontinuity
That softer zone is part of the asthenosphere, a region in the upper mantle where rock behaves in a slightly more fluid way over geological time. It is not liquid; it just deforms more easily than the rigid lithosphere above it. The lithosphere, which includes the crust and the very top of the mantle, rides on the asthenosphere somewhat like a stiff raft on a slowly flowing river. This distinction between rigid lithosphere and weaker asthenosphere is what makes plate tectonics possible in the first place.
What the Mantle Is Made Of
The upper mantle is dominated by olivine, a magnesium-iron silicate mineral. But as you go deeper and pressure climbs, olivine can no longer hold its crystal structure together. At about 410 kilometers down, it transforms into a denser mineral called wadsleyite. Deeper still, around 520 kilometers, wadsleyite converts to ringwoodite. These phase transitions do not change the chemical recipe of the rock very much; they rearrange the atoms into tighter configurations. The transitions matter because they alter the rock’s physical properties in ways that seismologists can detect. Laboratory measurements show that wadsleyite and ringwoodite conduct electricity almost a hundred times better than olivine, which supports the idea that the upper mantle has a distinct two-layer structure separated by the 410-kilometer boundary.2PubMed. Electrical conductivity of olivine, wadsleyite, and ringwoodite under upper-mantle conditions
Below 660 kilometers, the mineral landscape changes more dramatically. The dominant mineral in the lower mantle is bridgmanite, a dense magnesium silicate. At the very bottom, near the core-mantle boundary around 2,700 kilometers deep, bridgmanite can transform into an even denser phase called post-perovskite. The chemistry of that transition is sensitive to iron and aluminum content: adding certain forms of iron or aluminum shifts the pressure at which the change happens and broadens the depth range over which both phases coexist.3Journal of Geophysical Research: Solid Earth. The Post‐Perovskite Transition in Fe‐ and Al‐Bearing Bridgmanite: Effects on Seismic Observables Recent diamond-anvil-cell experiments have also found that the balance of iron oxidation states in bridgmanite shifts with depth: above about 2,300 kilometers, iron tends to be in its oxidized form, but at greater pressures the reduced form dominates, which changes how the mineral interacts with water and other volatiles.4PubMed Central. Pressure stabilizes ferrous iron in bridgmanite under hydrous deep lower mantle conditions
What Actually Drives the Plates
A common picture of plate tectonics imagines convection currents in the mantle dragging plates around like items on a conveyor belt. The reality is more nuanced. The single largest force moving tectonic plates is slab pull: the weight of a cold, dense oceanic plate sinking into the mantle at a subduction zone. Earthquake data and tectonic models consistently show that slab pull is much larger than ridge push, the gentle shoving that happens when new rock forms at mid-ocean ridges and slides downhill.5Reviews of Geophysics. Slab pull and the seismotectonics of subducting lithosphere
Quantitative modeling puts numbers on this. The net slab pull force that actually accelerates a plate is only about 8 to 12 percent of the total gravitational pull on the sinking slab, but that is still roughly twice the ridge push force. Most of the slab’s gravitational energy, around 70 percent, goes into driving mantle flow around the retreating slab, with another 15 to 30 percent spent bending the plate as it dives into the trench.6Geophysical Research Letters. Quantifying the net slab pull force as a driving mechanism for plate tectonics So the plates are less like passengers being carried by convection and more like active participants whose own sinking pulls the system along.
How Slabs Sink Through the Mantle
Not all subducting slabs behave the same way once they reach the transition zone between 410 and 660 kilometers. Some punch straight through into the lower mantle. Others flatten out and stagnate horizontally for hundreds of kilometers before eventually sinking deeper. Seismic imaging beneath the western Pacific, for instance, shows slabs beneath the Kurile and Bonin arcs bending to nearly horizontal near the 660-kilometer boundary and stretching over 1,000 kilometers before eventually descending. The Java slab, by contrast, penetrates the boundary more directly but then spreads out at greater depth, reaching about 1,200 kilometers.7Journal of Geophysical Research: Solid Earth. Subducting slabs stagnant in the mantle transition zone
What determines the outcome? Slab age and temperature play a role. Numerical models show that younger, thinner slabs tend to fold inside the transition zone and then continue sinking, while older, thicker slabs may initially punch into the lower mantle but then bend back and lie flat above 660 kilometers for a time.8Earth and Planetary Science Letters. Subducting-slab transition-zone interaction: Stagnation, penetration and mode switches The resistance at 660 kilometers comes from a mineral phase change (the post-spinel transition) and a jump in viscosity. Together, these create a partial barrier, not an impenetrable wall, but enough to influence the geometry of mantle flow for millions of years.
One Big Circulation or Two Separate Layers
Whether the mantle convects as a single system from top to bottom or as two separate layers divided near 660 kilometers has been debated for decades. Both models can explain certain observations, but neither fits everything perfectly. Models of whole-mantle convection do a good job reproducing the long-wavelength geoid, the subtle gravitational bumps and dips on Earth’s surface caused by density variations deep below. Layered convection, however, better explains the relatively small amplitudes of dynamic topography at the surface, and some models place the convective barrier about 250 kilometers deeper than the 660-kilometer phase boundary to match both geoid and topography simultaneously.9Earth and Planetary Science Letters. Layered mantle convection: A model for geoid and topography
A significant problem with strict layering, though, is that it traps heat. A thermal boundary layer between two convecting layers acts as insulation, making it harder for the planet to cool. Parameterized convection studies find that whole-mantle convection satisfies thermal constraints across a wide range of assumptions, whereas layered convection severely narrows or eliminates the parameter space that produces a reasonable thermal history for Earth.10Geochemistry, Geophysics, Geosystems. Cooling of the Earth: A parameterized convection study of whole versus layered models The current thinking leans toward a hybrid picture: material generally circulates through the full mantle, but the transition zone acts as a filter that slows and sometimes temporarily traps sinking slabs, creating behavior that looks partially layered in seismic snapshots.
Giant Blobs in the Deep Mantle
Two continent-sized regions in the lowermost mantle, one beneath Africa and one beneath the Pacific, show up in seismic data as zones where waves travel unusually slowly. These large low-velocity provinces, or LLVPs, extend hundreds of kilometers above the core-mantle boundary and are among the most prominent features inside the planet. They are thought to contain a significant amount of recycled oceanic crust that was carried down by subducting slabs over hundreds of millions of years and then stirred into dense piles by mantle convection.
Simulations of mantle circulation over the past billion years show that two antipodal LLVPs develop naturally as a result of Earth’s subduction history combined with the gravitational settling of dense former ocean floor. Although both provinces look similar in seismic wave speed, their histories differ. The Pacific LLVP has been continuously fed by circum-Pacific subduction since about 300 million years ago, while the African LLVP consists of older, more thoroughly mixed material. The Pacific domain may store over 50 percent more recycled oceanic crust from the last 1.2 billion years than its African counterpart, potentially making it denser and less buoyant.11Scientific Reports. Unique composition and evolutionary histories of large low velocity provinces
Right at the core-mantle boundary, even more extreme patches appear where seismic velocities drop dramatically. These ultralow-velocity zones are thin, sometimes only tens of kilometers thick, and their properties can be explained by iron-rich post-perovskite silicate that forms through chemical reactions between the silicate mantle and the liquid iron core.12PubMed. Iron-rich post-perovskite and the origin of ultralow-velocity zones The bridgmanite-to-post-perovskite transition itself happens fast enough on geological timescales to be seismically detectable: thick layers of coexisting bridgmanite and post-perovskite produce measurable reflections of seismic waves.13PubMed Central. Kinetics and detectability of the bridgmanite to post-perovskite transformation in the Earth’s D″ layer
Water and Carbon Hidden in the Deep
The mantle is not as dry as its rocky reputation suggests. Water rides into the deep Earth inside hydrous minerals in subducting slabs. A relay of mineral phases, including serpentine, chlorite, and a succession of high-pressure hydrous minerals, can carry water to transition-zone depths. Once there, the minerals wadsleyite and ringwoodite can dissolve substantial amounts of water into their crystal structures, giving the transition zone a large theoretical water-storage capacity.14Chemical Geology. Hydrous minerals and the storage of water in the deep mantle Hydrous ringwoodite has even been found as an inclusion in ultra-deep diamonds from Brazil, offering direct evidence that at least some parts of the transition zone do contain water.
That said, “wet” at these depths does not mean soaking. Under the cold conditions of a subducting slab’s core, the olivine-family minerals themselves hold very little water. Experiments show that wadsleyite coexisting with other hydrous phases at slab-core temperatures contains only about 200 parts per million of water, and ringwoodite as little as 8 to 46 parts per million.15Earth and Planetary Science Letters. Limited water contents of wadsleyite and ringwoodite coexisting with hydrous minerals in cold subducting slabs The bulk of the water being transported is carried by dedicated hydrous minerals rather than dissolved in the main silicates.16Physics of the Earth and Planetary Interiors. Water transport into the deep mantle and formation of a hydrous transition zone
Carbon follows a different path. Oxygen isotope measurements of mineral inclusions in diamonds from Guinea, spanning depths from the shallow mantle into the lower mantle, show that carbonated oceanic crust is the main carbon carrier down to transition-zone depths. Carbonate-rich slab material melts in this zone, forming diamonds with distinctive isotopic signatures. Deeper than 660 kilometers, however, the story changes: the lower mantle is too chemically reducing for carbonate to survive, and carbon is instead hosted in metallic phases. When water from the subducting slab reaches these metals, it destabilizes them and frees carbon to crystallize as diamond, but without imprinting any crustal isotopic fingerprint.17PubMed. The lithospheric-to-lower-mantle carbon cycle recorded in superdeep diamonds Diamond inclusions with carbon isotopes consistent with surface-derived carbon confirm that this deep carbon cycle extends all the way into the lower mantle.18PubMed. Deep mantle cycling of oceanic crust: evidence from diamonds and their mineral inclusions
The Mantle’s Grip on Earth’s Magnetic Field
Earth’s magnetic field is generated in the liquid outer core, but the mantle controls how fast the core loses heat, and that heat flow governs the vigor of the dynamo. The thermal evolution of the core is effectively set by mantle convection, which determines the total rate of heat escaping across the core-mantle boundary.19Physics of the Earth and Planetary Interiors. Thermal and magnetic evolution of the Earth’s core Even a modest heat flow above what would simply conduct down the temperature gradient is enough to power the geodynamo through thermal convection alone. For a plausible present-day heat flux, compositional convection, driven by light elements released as the inner core solidifies, accounts for roughly two thirds of the energy dissipated in the core, with thermal convection supplying the remaining third.20Journal of Geophysical Research: Solid Earth. On the thermal evolution of the Earth’s core
Because the deep mantle is not uniform, with structures like LLVPs and slabs altering heat flow in different regions, the pattern of heat extraction from the core is patchy. Coupled core-mantle evolution models can reproduce features like the frequency of geomagnetic polarity reversals and the westward drift of the magnetic field by accounting for these time-varying heat-flow patterns across the core-mantle boundary.21Progress in Earth and Planetary Science. A coupled core-mantle evolution: review and future prospects In short, the lumpy, evolving mantle is one of the reasons the magnetic field flips its polarity irregularly rather than on a neat schedule.
When Did Plate Tectonics Begin
Plate tectonics as we know it, with a global network of rigid plates separated by narrow boundaries, has operated since at least the early Paleoproterozoic, around 2.2 billion years ago. Before that, things get murky. Higher mantle temperatures in the Archean probably made stable subduction difficult: hotter rock is weaker and more buoyant, which works against the kind of sustained sinking that modern subduction requires. The period from about 3.2 to 2.3 billion years ago appears to have been a long transition from an earlier tectonic mode, possibly a stagnant or sluggish lid where the surface barely moved, to the mobile-plate system of today. One hypothesis is that mantle plumes punching through the lid could have initiated the first subduction zones, kick-starting the whole process.22Annual Review of Earth and Planetary Sciences. Plate Tectonics and the Archean Earth
Earth’s Long Cooldown
The mantle has been losing heat since the planet formed. Geochemical evidence from the compositions of ancient lavas suggests that mantle temperatures have dropped by around 250 degrees over the past three billion years.23Geochemistry, Geophysics, Geosystems. Thermal History of the Earth: On the Importance of Surface Processes and the Size of Tectonic Plates Models that try to reproduce this cooling using simple convection scaling laws run into a problem: they predict a thermal runaway when extrapolated back in time, with the early Earth implausibly hot. The fix turns out to involve the plates themselves. When plate boundaries are allowed to evolve, form, and disappear over time rather than staying fixed, the models produce a cooling rate that matches the geochemical record without blowing up in the past.24Earth and Planetary Science Letters. Thermal evolution of the Earth: Secular changes and fluctuations of plate characteristics Plate tectonics, in other words, is not just a consequence of mantle convection; it regulates the pace of planetary cooling.
Why Venus Went a Different Way
Venus is nearly the same size and bulk composition as Earth, yet it has no plate tectonics. Its surface is a single stagnant shell. Several explanations have been proposed: Venus may lack the water that weakens Earth’s rocks, or its grain sizes may differ. But one compelling mechanical argument focuses on fault friction. On Earth, where the surface is relatively cool, faults can slip rapidly and unstably, which weakens them dramatically and allows plates to break and subduct. On Venus, where surface temperatures hover around 460 degrees Celsius and the lithosphere is hotter throughout, fault motion is slow and stable, producing little dynamic weakening. The result is that Venus’s shell stays strong and resists breaking into plates.25Scientific Reports. Dynamics of fault motion and the origin of contrasting tectonic style between Earth and Venus Venus may be convecting underneath, but its lid is too tough to participate.
Supercontinents and the Mantle’s Pulse
The continents have repeatedly assembled into supercontinents and then broken apart over cycles lasting several hundred million years. This is not random wandering; it is linked to the structure of mantle convection. When the mantle settles into a pattern dominated by a single large-scale convective cell, the continents get swept together.26Earth-Science Reviews. Supercontinents, mantle dynamics and plate tectonics: A perspective based on conceptual vs. numerical models Numerical simulations show that assembly and breakup happen cyclically as long as a sufficient amount of dense, chemically distinct material exists at the base of the mantle. Hot plumes rising from the edges of this deep dense material drive continental motion, and the continents in turn push the dense piles around, which moves the plumes, eventually breaking up the supercontinent they just assembled.27Geosciences. Supercontinent Cycle and Thermochemical Structure in the Mantle: Inference from Two-Dimensional Numerical Simulations of Mantle Convection It is a feedback loop between what happens at the surface and what is simmering thousands of kilometers below.
Mineral Wealth From the Deep
Mantle plumes do more than break apart continents. When hot upwellings from the deep mantle reach the base of the lithosphere, they can trigger rifting, flood volcanism, and the formation of economically important mineral deposits. Plume-related magmatism has been linked to nickel-copper-platinum-group-element deposits, iron-titanium-vanadium ores, and iron-oxide-copper-gold deposits, among others. The Bushveld Igneous Complex in South Africa, one of the world’s richest ore bodies, is a classic example of plume-driven mineralization.28Geological Society, London, Special Publications. Mineral systems and their putative link with mantle plumes
When plume activity coincides with subduction, the combination can be especially productive for gold. In Patagonia’s Deseado Massif, a mantle plume enriched the deep lithosphere in incompatible elements including gold, while the nearby subduction zone oxidized portions of that enriched mantle, making it easier for rising magmas to carry dissolved gold upward. An extensional tectonic setting then allowed those magmas to reach crustal levels, forming a major gold-silver province.29Nature Communications. Plume-subduction interaction forms large auriferous provinces The mantle, in this sense, is not just the engine of earthquakes and volcanic eruptions; it is the ultimate source of many of the metals and resources we depend on.
How Scientists See Inside the Earth
Nobody has drilled more than about 12 kilometers into the Earth. Everything we know about the mantle’s structure comes from indirect methods, and seismic tomography is the most powerful. The technique works by comparing the actual travel times, waveforms, and dispersion patterns of seismic waves against what models predict, then iteratively adjusting the three-dimensional picture of the interior until the mismatch shrinks. This process has revealed the shapes of subduction zones, plume conduits, and the LLVPs described earlier.30ScienceDirect (Earthquake Research Advances). The evolution of seismic tomography in earth sciences—advancements, limitations, and its AI-enabled future (A critical review)
The other major window into the deep Earth is the laboratory. Laser-heated diamond anvil cells can squeeze tiny samples to pressures exceeding those at the core-mantle boundary while heating them to thousands of degrees, recreating deep-mantle conditions on a tabletop. Recent work has produced the first direct measurements of how thermal pressure distributes inside these cells, improving the accuracy of the density measurements that link lab minerals to seismic observations.31Journal of Geophysical Research: Solid Earth. Thermal Pressure in the Laser‐Heated Diamond Anvil Cell: A Quantitative Study and Implications for the Density Versus Mineralogy Correlation of the Mantle Between the seismologists mapping wave speed anomalies and the mineral physicists squeezing tiny grains in diamond vises, the picture of the mantle keeps getting sharper, and every new detail reminds us how much the ground beneath our feet depends on processes unfolding far below it.