Stone melts the same way ice does: add enough heat, and its solid crystal structure breaks apart into liquid. Every volcanic eruption on Earth is proof that rock routinely crosses this threshold. The temperatures required vary enormously depending on what kind of rock you are dealing with, but the range runs from roughly 600°C for some water-rich granitic compositions up past 1,900°C for dry mantle rock. What makes the science interesting is that heat is only one of several knobs nature and humans can turn to push rock across the line from solid to liquid.
What It Takes to Melt Different Rocks
Rock is not a single substance. It is a mix of minerals, each with its own melting behavior, so “the melting point of rock” is always a range rather than a sharp number. At the low end, wet sedimentary rocks rich in silica and alkali minerals can begin to soften and partially melt around 650–700°C. Basalt, the dark volcanic rock that makes up most of the ocean floor, melts more completely around 1,000–1,200°C. Peridotite, the dense olivine-rich rock of Earth’s upper mantle, does not fully liquefy until roughly 1,700–1,900°C under surface pressure. Thermodynamic models now exist that can predict melting behavior across this entire spectrum, from granite to peridotite, at pressures ranging from near-surface conditions up to about 70 kilobars deep in the Earth.
1Oxford Academic (Journal of Petrology). Melting of Peridotites through to Granites: A Simple Thermodynamic Model in the System KNCFMASHTOCrPressure complicates the picture. Under normal conditions, raising pressure makes it harder for rock to melt because the atoms are squeezed together more tightly. Deep inside the Earth, temperatures that would easily melt surface rock instead keep minerals locked in a solid state. But there are surprising exceptions, and those exceptions drive much of the volcanism we see at the surface.
How Nature Melts Rock Without a Furnace
You might assume that melting rock underground requires something to actively heat it up, like a giant burner beneath a volcano. In reality, much of the melting inside Earth happens because pressure drops or water arrives, not because the temperature suddenly spikes. The mantle is already extremely hot. What usually prevents it from melting is the immense weight of the rock above it.
At mid-ocean ridges, where tectonic plates pull apart, the mantle rises to fill the gap. As it ascends, the pressure on it decreases. That drop in pressure lowers the melting point enough for the rock to start liquefying even though its temperature barely changes. This process, called decompression melting, is how oceanic crust is continuously created along the roughly 65,000-kilometer chain of ridges winding across the ocean floor.2Lithos. The timescales of magma evolution at mid-ocean ridges It is one of the most prolific rock-melting mechanisms on the planet, and it runs nonstop.
Water plays an equally dramatic role. At subduction zones, where one tectonic plate dives beneath another, minerals in the sinking slab carry water down into the mantle. As those minerals break down under heat and pressure, they release their bound water into the overlying mantle wedge. Water radically lowers the melting temperature of silicate rock, so the mantle in that zone begins to melt even though it is not as hot as the mantle beneath a mid-ocean ridge. The resulting magma is buoyant and rises, eventually fueling the chains of volcanoes that line subduction zones around the Pacific Rim and elsewhere.3Annual Review of Earth and Planetary Sciences. The Role of H2O in Subduction Zone Magmatism Water also changes the chemistry of the melt, making subduction-zone lavas distinctly different from mid-ocean ridge basalts. This water-assisted melting is ultimately how most of Earth’s continental crust was built over billions of years.
Even the chemical environment matters. Recent experiments show that changing the oxygen content of deep mantle rock by a few orders of magnitude can shift its melting temperature by hundreds of degrees. Under more oxidizing conditions at around 16 gigapascals of pressure, the melting onset dropped by over 450°C compared to more reduced conditions.4Nature Geoscience. Melting at the base of a terrestrial magma ocean controlled by oxygen fugacity This finding has reshaped how scientists think about where and when melting occurred early in Earth’s history.
When Rock Melts in Seconds
The geological processes described above unfold over thousands to millions of years. But rock can also melt almost instantaneously when enough energy is dumped into it all at once. Nature and human technology both offer examples.
Lightning strikes deliver tens of thousands of amperes in microseconds. When a bolt hits sand, soil, or exposed rock, the energy is concentrated enough to melt minerals along the strike path, creating hollow tubes of glass called fulgurites. These are literal pieces of melted earth, and they turn up everywhere from beaches to mountain summits.5Geophysical Research Letters. Generation of shock lamellae and melting in rocks by lightning‐induced shock waves and electrical heating
Meteorite impacts deliver energy on a much larger scale. When a space rock slams into Earth’s surface at tens of kilometers per second, the shock wave can compress and heat minerals to the point of melting or even vaporization. Quartz and feldspar in the target rock get selectively transformed into glass, while some tougher minerals nearby survive intact. The resulting “impactites” contain telltale features like tiny metal spherules mixed into the glass.6PubMed. Shock effects in certain rock-forming minerals
Earthquakes can also melt rock, though this one surprises most people. During a large earthquake, rock surfaces on either side of a fault can slide past each other at speeds of a meter per second or more. The friction generates enough heat to melt a thin layer of rock right along the fault plane. This frozen melt, called pseudotachylyte, is essentially earthquake glass.7Journal of Structural Geology. Microstructural evolution of pseudotachylyte-bearing rocks during increasing temperatures: Evidence from rock-heating experiments Laboratory experiments replicating these conditions have found something unexpected: minerals melt at temperatures well below their known melting points during rapid frictional sliding. Quartz, for instance, melted at around 1,000–1,200°C rather than its expected 1,720°C, and olivine melted near 1,300°C instead of about 1,700°C.8Scientific Reports. Frictional melting mechanisms of rocks during earthquake fault slip The rapid, chaotic nature of frictional sliding apparently disrupts crystal lattices in ways that make them give up and liquefy at lower temperatures than calm, steady heating would require.
Then there is the most extreme human-made example. On July 16, 1945, the first nuclear bomb test at Trinity Site in New Mexico released enough thermal energy to melt the surrounding desert sand into a green glassy material now called trinitite.9Geology Today. Trinitite—the atomic rock The sand, an arkosic mix rich in feldspar and quartz, was swept into the fireball and fused into glass that rained back down across the test site.10American Mineralogist. Trinitite redux: Mineralogy and petrology Similar glasses form during any ground-level nuclear detonation. Trinitite remains one of the most vivid demonstrations that with enough energy, turning rock to liquid is trivially easy.
What Happens When Molten Rock Cools Back Down
Melting rock is only half the story. What you get when it solidifies again depends on how fast it cools. Cool it slowly, deep underground over thousands of years, and you get coarse-grained rocks like granite, where individual mineral crystals have time to grow large enough to see with your eyes. Cool it quickly at the surface, and you get fine-grained volcanic rock like basalt. Cool it very quickly, and you can skip crystallization altogether, producing natural glass.
Obsidian is the best-known natural glass, and it comes with a persistent myth. The common story is that obsidian forms because lava cools extremely rapidly, freezing before crystals can form. A recent study overturned this idea. Dense obsidian actually forms through slow cooling, on timescales of a month to a decade. The key insight is that water-poor rhyolite melt takes an astonishingly long time to nucleate crystals: at the temperature where nucleation is fastest, the minimum delay before crystals begin forming is roughly 45 years. So the lava has a generous window. It cools slowly enough for trapped gas bubbles to dissolve back into the melt, producing dense glass, yet still fast enough that no crystals have time to appear.11Nature Communications. Obsidian forms by slow cooling
The viscosity of the molten rock also matters for what happens at the surface. Rhyolite lavas are extremely viscous, and there is a widespread assumption that they cannot flow far from their source. While it is true that high-viscosity lava flows slowly, it can travel a great distance if it retains heat efficiently and enough magma keeps feeding it from below.12Journal of Volcanology and Geothermal Research. Extended cooling and viscous flow of large, hot rhyolite lavas: implications of numerical modeling results Some ancient rhyolite flows extend tens of kilometers. By contrast, basalt lavas are relatively runny and can spread rapidly over vast areas. The enormous flood basalt eruptions of the geologic past covered millions of square kilometers, releasing staggering amounts of sulfur and chlorine gas in the process: one estimate puts it at roughly 3.5 teragrams of sulfur dioxide and 1 teragram of hydrogen chloride for every cubic kilometer of lava erupted.13PubMed. Sulfur and chlorine in late Cretaceous Deccan magmas and eruptive gas release When individual flows were hundreds or thousands of cubic kilometers in volume, the environmental consequences were catastrophic.
Melting at the Bottom of the World
The deepest known rock melting on Earth happens nearly 3,000 kilometers below the surface, at the boundary between the mantle and the outer core. Pressures there reach about 135 gigapascals, and temperatures hover around 3,800–4,000 kelvin. Experiments simulating those conditions found that subducted basalt, the ocean-floor rock that gets pushed down at subduction zones over hundreds of millions of years, begins to melt at about 3,800 kelvin at core-mantle boundary pressure. That is roughly 350 degrees below the temperature at which surrounding mantle rock would melt.14PubMed. Melting of subducted basalt at the core-mantle boundary The implication is that pockets of ancient ocean floor piled up at the base of the mantle could be partially molten right now, forming thin layers of magma that seismologists have detected as “ultra-low velocity zones” in their imaging of Earth’s deep interior.
Entire Planets Made of Liquid Rock
Earth’s current magma production is modest compared to what happened during planetary formation. In the early solar system, the energy released by accretion, giant impacts, and radioactive decay was so intense that nearly all large rocky bodies went through at least one “magma ocean” phase, in which most or all of the planet’s rocky shell was molten.15PubMed Central. Magma oceans as a critical stage in the tectonic development of rocky planets Earth, the Moon, Mars, and Mercury all likely experienced global magma oceans. On early Earth, the magma ocean would have been hundreds of kilometers deep, and its crystallization set the chemical template for the mantle and crust we have today.
Recent work suggests that the depth and duration of these magma oceans depended on the oxygen content of the early mantle. As noted earlier, more oxidizing conditions lower the melting temperature of mantle rock by hundreds of degrees. That means a more oxidized early Earth could have sustained a deeper, longer-lasting magma ocean than previously calculated.4Nature Geoscience. Melting at the base of a terrestrial magma ocean controlled by oxygen fugacity The chemistry of the atmosphere, the timing of core formation, and even the potential for early plate tectonics all hinge on how long that molten stage lasted.
Melting Rock on Purpose
Humans have been deliberately melting rock for thousands of years, starting with smelting metal ores. Modern industry has found uses for the molten rock itself, not just the metals extracted from it.
Basalt fiber manufacturing is one of the more direct examples. Basalt is melted and drawn into fine filaments that are used for thermal insulation, acoustic insulation, and as a reinforcement material in composites. There are two main industrial approaches: a duplex method where raw basalt is the sole ingredient, and a Cupola furnace method where basalt is mixed with limestone, dolomite, or slag to adjust the chemistry.16Procedia Materials Science. Indian Program for Vitrification of High Level Radioactive Liquid Waste – Section: Abstract The resulting fibers are chemically resistant and perform well in high-temperature environments, making them an alternative to traditional glass fiber or asbestos.
A more sobering application is nuclear waste management. High-level radioactive waste, the most dangerous byproduct of nuclear power and weapons production, is immobilized by mixing it into molten glass. The preferred medium is borosilicate glass, which can incorporate a meaningful amount of radioactive material while remaining chemically stable and resistant to leaching over geological timescales.17Procedia Materials Science. Indian Program for Vitrification of High Level Radioactive Liquid Waste The resulting glass logs are stored in stainless steel canisters and eventually destined for deep geological repositories. India, France, the UK, and the United States have all operated vitrification plants. The logic is straightforward: if you need to lock something away for tens of thousands of years, a glass matrix made from melted rock is one of the most durable containers humans can make.
Looking further ahead, melting rock may become essential for building on the Moon. Lunar regolith, the powdery soil covering the Moon’s surface, is essentially pulverized basaltic rock. Researchers have demonstrated that a focused laser can melt lunar regolith simulant and manufacture large interlocking elements suitable for paving, without needing furnaces or molds.18Scientific Reports. Laser melting manufacturing of large elements of lunar regolith simulant for paving on the Moon The appeal is obvious: instead of hauling construction materials from Earth, you melt what is already there. The same principle could apply to roads, landing pads, and radiation shielding.
Drilling Into Magma by Accident
One of the stranger intersections of rock melting and human activity involves geothermal drilling. Engineers drilling into volcanic regions to tap geothermal energy have, on multiple occasions, accidentally punched into shallow magma bodies they did not expect to find. These encounters are rare, but they keep happening as geothermal exploration pushes deeper and into more volcanically active areas.19European Review. Accessing Magma: A Necessary Revolution in Earth Sciences and Renewable Energy
The most famous case was Iceland’s IDDP-1 well in 2009, which hit rhyolite magma at about 2,100 meters depth. Rather than being a catastrophe, the well ended up producing superheated steam at record-breaking temperatures and pressures. The experience sparked serious discussion about whether drilling into or near magma bodies could be done intentionally, creating a new class of ultra-high-temperature geothermal energy. The engineering challenges are immense: conventional drill bits and casing materials are not designed for temperatures above roughly 300–350°C, let alone the 900°C+ range you find near molten rock. But the energy payoff of a single well tapping supercritical fluids near magma could be five to ten times that of a conventional geothermal well, making it one of the most tantalizing prospects in renewable energy research.
These accidental encounters have also given scientists an unexpected gift: direct samples of magma from known depths, something that is nearly impossible to obtain otherwise. Most of what we know about magma chemistry comes from erupted lava that has already lost gas and changed composition on its way to the surface. Drill-intercepted magma provides a much less altered snapshot of what is happening underground. As geothermal projects expand in volcanic regions from Iceland to Japan to East Africa, the odds of more magma encounters keep rising, turning an engineering headache into a scientific opportunity.