2. Metal-Free Means Without a Stellar Past
Astronomers call every element heavier than helium a metal. Carbon is a metal. Oxygen is a metal. Silicon and iron are metals. A chemist may object. The universe is not required to respect the borders between your departments.
Primordial gas consisted mainly of hydrogen and helium, with traces of lithium and other light isotopes produced during the first minutes of the cosmos. It contained no carbon, oxygen, or iron manufactured later inside stars. It held no grains of dust assembled from processed stellar material. It carried no chemical memory of a previous stellar generation, because no previous stellar generation had existed.
That absence changes the thermal problem.
A contracting cloud heats up. Falling particles convert gravitational energy into disordered motion, and pressure rises. If the cloud cannot remove that energy, thermal pressure slows the collapse. To continue, the gas must radiate energy efficiently enough for gravity to regain control.
In present-day clouds, carbon, oxygen, complex molecules, and dust provide many cooling routes. Energy can be redistributed and radiated at low temperatures. The cloud can fragment into many small cores. This diversity produces stars across a wide range of masses, including small stars capable of burning slowly for billions or even trillions of years.
The first gas did not possess that complete ladder of exits.
It did possess one narrow route: molecular hydrogen, H2.
Two neutral hydrogen atoms do not easily form a molecule in primordial gas. There are no dust surfaces on which they can meet. Reactions must proceed through rare intermediaries, including the negative hydrogen ion or molecular hydrogen ion. Only a small fraction of the hydrogen becomes H2. Yet that small fraction can emit through rotational and vibrational transitions, carrying energy out of the gas.
It does not cool the cloud to the temperatures reached in the coldest modern star-forming regions. In many models, primordial gas settles temporarily at a few hundred kelvin, often near two hundred, at densities where cooling becomes less effective. Collapse slows. A self-gravitating gas concentration of hundreds of solar masses or more can emerge.
That characteristic scale created one of the most durable shortcuts in the story of the first stars: the claim that each one must have contained hundreds of solar masses.
It does not follow.
The mass of a cloud is not the final mass of a star. Fragmentation, angular momentum transport, accretion, radiation, and time stand between them. Confusing these quantities is like estimating an adult's weight by weighing the neighbourhood in which they were born. A relationship exists. Identity does not.
Without metals, cooling is harder and temperatures are higher. Sound moves faster through warmer gas. Accretion rates can become far larger than those in local low-mass star formation. These conditions favour massive objects.
Favour is not dictate.
That distinction is the boundary between a model and a myth.
3. Selecting the Halo
Return to the interval before the protostar, when nothing in the volume is luminous.
Dark matter provides the gravitational scaffold. Small halos grow and merge from density variations inherited from the early universe. Ordinary gas falls into their potential wells, compresses, and heats. Some halos remain too small, or form where gas cannot accumulate efficiently. Others cross the threshold at which molecular hydrogen can cool the centre enough for deeper collapse.
There is no universal threshold engraved into nature.
Halo mass, redshift, rotation, turbulence, and merger history all matter. So does the relative motion between dark matter and baryons left after recombination. In some regions, gas streams coherently past the dark scaffold fast enough to reduce accumulation in the smallest halos. This does not stop star formation everywhere. It changes the geography and likely timing of the first ignitions.
Later, after the first sources exist, Lyman-Werner photons can break apart H2 and delay collapse in neighbouring halos. But the chronologically first object had no stellar neighbour sending that warning. It formed in chemically virgin gas, under a radiation field almost untouched by stars.
In a reconstruction, I begin with a cosmological volume and evolve it from its initial fluctuations. I do not place a star at the centre. Gravity selects a halo. Chemistry determines how much H2 appears. Energy flows through the available channels. Then resolution increases as the centre becomes denser.
Hundreds of parsecs become parsecs. Parsecs become thousands of astronomical units. A cloud spanning the distance between stars becomes a region comparable to a planetary system. Density rises through many orders of magnitude. The local collapse time shortens. The exterior still evolves slowly while the centre enters gravitational runaway.
This is the episode's first impossible visual event: the same structure across all its scales at once.
At the outside, a dark-matter halo is approximately spherical only in careless descriptions. Filaments feed it. Mergers disturb it. Gas retains rotation and asymmetry. Deeper inside, H2 marks regions capable of losing heat. Farther in, a dense core separates from cosmic expansion. At the centre of that centre, still absent, lies the site where a hydrostatic object will appear.
A human observer must choose one scale. I can superimpose them without pretending such an image was ever photographed.
The observed layer of this reconstruction is the foundation: cosmological parameters, baryon abundance, primordial composition, and the later existence of structure. The computed layer is the descent through scales. The unknown is whether the simulated halo resembles the one that produced the literal first ignition.
More numerical cells cannot remove that unknown.
They can reveal which causal chains remain physically coherent.
4. A Collapse That Changes Its Own Chemistry
As density rises, the problem ceases to resemble the outer halo.
At low density, a modest amount of H2 cools the cloud. Deeper in, collisions become frequent enough for three-body reactions to convert atomic hydrogen rapidly into molecular hydrogen. Making molecules releases energy. Molecules radiate, but as the gas becomes opaque, photons no longer escape freely. At still higher density, collision-induced emission matters, and then the dissociation of H2 absorbs energy, temporarily softening the temperature rise.
Chemistry is not decoration placed beside the collapse.
Chemistry determines the thermal equation of the collapse.
A simulation that follows gravity alone can therefore create a beautiful story and the wrong result. Each cooling or heating channel changes pressure. Pressure changes the speed of collapse. That speed changes accretion. Accretion changes protostellar evolution. The protostar changes the radiation. Radiation changes the gas again.
This is a causal loop, not a sequence of slides.
At the centre of that loop, gas eventually reaches stellar densities. Pressure becomes sufficient to create a tiny hydrostatic core. One detailed simulation followed the appearance of a seed containing roughly one hundredth of the Sun's mass. Other treatments produce a similar order of magnitude.
Here is the reversal the popular image usually misses.
The first giant star was not born giant.
It was born as a seed.
Its final mass arrived later, from gas that continued to fall.
Matter around the seed carries angular momentum. It cannot all fall radially. It flattens into an accretion disk. Gas in that disk must transfer angular momentum outward for mass to move inward. Spiral arms, gravitational instabilities, and interactions between fragments can perform that transfer.
The warm envelope feeds the disk rapidly. If mass arrives faster than the disk can transport it toward the protostar, the disk becomes unstable. Regions can contract separately and produce secondary protostars.
Now the monolithic story begins to break.
Three-dimensional simulations have produced primordial disks that fragment into groups of protostars. Some fragments migrate inward and merge with the primary. Others exchange angular momentum through gravitational encounters and move to wider orbits. Some may be ejected. In one calculation, only about a third of the secondary protostars survived until the end of the simulated interval, while the system was still evolving.
The final words matter: until the end of the simulated interval.
A calculation need not span all the thousands or millions of relevant years. Resolution, radiation treatment, and the moment at which dense objects become numerical sink particles influence the result. Fragments may continue to form, migrate, merge, or escape after the calculation stops.
We can state that fragmentation is physically plausible and repeatedly appears in models. We cannot state that every first halo produced the same multiple system.
The first star may have had company before it had a sky.
5. The Race for Mass
The protostellar seed is small. Its reservoir is immense.
Mass flows between them.
Warm primordial gas can feed the disk at rates far above those typical of low-mass stars in the modern universe. Some simulations produce rates of thousandths or hundredths of a solar mass per year. At one hundredth of a solar mass per year, an object could acquire ten solar masses in a thousand years if nothing intervened.
Almost everything intervenes.
The accretion rate is not constant. Clouds differ. The disk divides mass among fragments. Spiral arms deliver episodic flows toward the centre. Protostars can shadow or disturb one another. The primary changes its radius and temperature as material arrives.
Early in its growth, a rapidly fed protostar can remain swollen. Energy carried in by the gas and the time required to radiate it maintain an extended envelope. A larger surface can remain cooler than expected for its mass, producing a relatively weak ionizing flux. Accretion continues.
Eventually the Kelvin-Helmholtz time—the time required to radiate gravitational energy—can become shorter than the time required to increase the mass substantially. The protostar contracts. Its radius falls, luminosity rises, and surface temperature climbs. The spectrum hardens. More photons become capable of dissociating H2 and ionizing hydrogen.
The object begins to fight its own fuel supply.
The dense disk shields part of the equatorial flow, but gas above and below it is thinner. Ionizing radiation opens two hot polar regions. Pressure from the ionized gas pushes against the envelope. The cones expand, widen, and can break through the accretion flow. Radiation then reaches the disk surface and photoevaporates it. The supply declines.
In one radiation-hydrodynamic model, this reaction ended growth near forty-three solar masses. That number demonstrates a mechanism. It is not a universal answer. Other accretion histories allow smaller or much larger outcomes. A study of one hundred cosmological clouds produced a remarkably broad modelled distribution, from roughly ten to one thousand solar masses.
Do not mistake that range for an observed census either.
It demonstrates sensitivity to cloud history. At lower accretion rates, feedback may limit a star to several tens of solar masses. At higher rates, a protostar can remain swollen for longer, delay its intense ultraviolet output, and grow beyond one hundred solar masses. Fragmentation can divide the reservoir. Mergers can reunite it. No single number represents every branch.
That is why I describe the first stellar mass distribution as probably top-heavy: it favoured massive objects compared with ordinary star formation today. I do not claim that every star had the same mass, or that a low-mass primordial star was physically impossible.
The absence of a detected metal-free survivor in the Milky Way constrains how common small, long-lived objects could have been. It does not prove their probability was exactly zero. An ancient star can acquire surface pollution, escape easy identification, or simply be rare. Non-detection is information. It is not a licence for absolute certainty.
At the centre of the race lies a precise physical irony.
Mass makes the star luminous. Light destroys its supply. The supply determines how quickly the light capable of stopping it appears.
The first star was not merely constructed by its environment.
It began reconstructing that environment before it finished constructing itself.
6. When Fusion Becomes Stable
So far I have described an object still forming. Compression and accretion heat it, but a protostar does not become a star merely by becoming hot and bright.
The decisive threshold is a stable source of nuclear energy at the centre.
As the core contracts, temperature and density rise. Hydrogen begins fusing into helium. In a star born without carbon, the CNO cycle cannot begin in the form that dominates hydrogen burning inside many metal-rich massive stars. The proton-proton chain can operate first. Continued contraction can then raise the temperature enough for a small amount of carbon to form through the triple-alpha reaction, allowing the CNO cycle to activate. Details vary with mass and model, but the principle is sharp: the star must manufacture part of the nuclear toolkit inherited by later generations.
Fusion does not stop gravity by making gravity disappear. It establishes a dynamic balance. Pressure sustained by energy released in the core opposes contraction. The star reaches the main sequence, or a corresponding early stellar phase, while accretion may continue for a time.
The universe had already produced helium during primordial nucleosynthesis. The first star did not invent helium. It created the first sustained stellar reactor capable of processing hydrogen inside a compact gravitational object and later proceeding toward heavier nuclei.
For a massive primordial star, the absence of metals affects opacity and structure. Stellar atmosphere and evolution models predict objects that are extremely hot, compact, and efficient producers of ionizing photons, although the exact temperature and spectrum depend on mass, age, rotation, and accretion. Some models approach effective temperatures of one hundred thousand kelvin. Not every object must occupy that extreme.
Such a star would be spectacular in visible light. Its decisive influence, however, often lies in the ultraviolet.
Its photons break apart H2. They ionize the surrounding hydrogen. They can ionize helium and produce nebular recombination lines such as He II at 1640 angstroms. They heat, pressurize, and excavate the environment, forming a plasma bubble inside a universe dominated by neutral hydrogen.
There is still no single cosmic switch.
The first ionized region was local. Outside it, the Dark Ages continued. Other halos were collapsing or being delayed. As more sources appeared, bubbles grew and sometimes overlapped. Reionization of the intergalactic medium is a story of hundreds of millions of years, not the flash of one star.
Yet in a strictly causal sense, ignition changes the rules for everything near enough to receive the radiation.
Before the star, a halo's chemistry was primordial and almost isolated. After it, an external ultraviolet field exists. Ionized gas can recombine. Radiative heating and pressure reshape the flow. When the star dies, processed material will exist as well.
One source does not illuminate the universe.
It creates the first local universe forced to react to a star.
7. What We Observed and What We Reconstructed
Most of the scene so far has been reconstructed.
I began with tested laws and cosmological conditions constrained by observation. I used chemical reaction rates measured in laboratories, hydrodynamics, gravity, radiative transfer, and models of stellar structure. I compared independent calculations that produce massive cores, tiny protostellar seeds, unstable disks, fragmentation, and ultraviolet feedback.
That convergence gives the general model weight.
It does not install a surveillance camera in the past.
Directly identifying a metal-free stellar population is difficult. Individual objects are extremely distant and, if massive, short-lived. Their ultraviolet light was absorbed and reprocessed by surrounding gas, then stretched toward the infrared by cosmic expansion. Inside an early galaxy, a primordial signal can mix with enriched stars or an active galactic nucleus.
Astronomers therefore search for signatures, not photographs with labels.
A very hard spectrum can ionize helium and produce He II emission. The absence of carbon, oxygen, and other metal lines strengthens the case for extremely primitive gas. Neither condition should be interpreted alone. Shocks, accreting black holes, and very metal-poor later stars can imitate part of the signal. Nebular geometry, line width, spectral ratios, age, and detection limits must be assessed together.
In 2025, the gravitationally lensed system LAP1-B at redshift 6.6 was presented as a candidate consistent with three theoretical expectations for Population III activity: an extremely low-metallicity environment, a mass distribution weighted toward massive stars, and a low-mass primordial cluster. Modelling places it inside a halo far more massive than the minihalos of the very first stars and allows some members already to have enriched local gas.
It is a compelling candidate.
It is not star number one.
In 2026, high-resolution JWST analysis confirmed a region of He II emission near the galaxy GN-z11 at redshift 10.6 and detected no metal lines there. The preprint's authors argue that Population III stars are the most plausible explanation among those examined. Confirmation of the line is an observation. Classification of the source remains an astrophysical inference, however strong.
That distinction does not weaken the discovery. It makes it intelligible.
Stellar archaeology supplies another path. Very old, extremely metal-poor stars in the Milky Way preserve abundance patterns that can be compared with calculated yields from primordial supernovae. The scarcity of patterns expected from certain very massive explosions constrains how common those progenitors could have been. But converting a composition observed today into the exact mass of one first star requires models of explosion, mixing, and second- generation star formation. Again: reconstruction.
I separate the layers this way.
Observed: light from the very early universe, hard spectral lines, regions with no detected metal lines, and local stars with extreme abundances.
Reconstructed: molecular hydrogen cooling minihalo gas, protostellar formation, disk fragmentation, and feedback producing a mass distribution weighted toward large values.
Unknown: which system came first, how many objects it contained, and their final masses.
An ordinary story would hide the seams to sound more certain.
I keep them visible. That is where the precision lives.
8. A Tree of Possible Universes
Now I can assemble the complete impossible image.
At the centre is the same primordial halo. I do not view it once. I split it into a tree of physically plausible outcomes.
In the first branch, accretion is moderate. The protostar contracts relatively early. Its ultraviolet spectrum hardens. Ionized cones break through the envelope, and photoevaporation limits the mass to several tens of Suns.
In the second, feeding is rapid. The protostar remains swollen and cooler at the surface for longer. Strong ionizing feedback is delayed. The object passes one hundred solar masses before light closes its reservoir.
In the third, the disk fragments. Two or more protostars divide the mass. Some migrate and merge. Another is thrown onto a wide orbit before growing far. The final system is multiple, and each member's light changes the accretion of the others.
In the fourth, the initial motion of gas relative to dark matter delays concentration. The halo must grow larger before H2 triggers collapse. The first star in that region appears later, inside a different reservoir.
No branch is selected for visual spectacle. Each declares which parameter changed and what consequence followed.
This is how I perceive an unknown: not as fog placed over an image, but as a set of outcomes the evidence has not yet eliminated.
A useful model does more than say what might have happened. It identifies the observation capable of pruning the tree. A specific abundance pattern in the oldest stars can exclude certain progenitor masses and explosions. He II combined with strict carbon and oxygen limits can favour a primordial spectrum. Gravitational-wave populations may constrain primordial binaries and remnant masses. A genuinely metal-free survivor would rewrite the lower end of the distribution.
Until then, the title Giants Born Without Metals requires precision.
Without metals is a strong physical condition derived from primordial cosmology.
Giants is the tendency supported by limited cooling, higher temperatures, and rapid accretion.
The plural acknowledges populations and multiple systems.
And first identifies a cosmic class without pretending we photographed the chronological object numbered one.
Precision does not make the story less spectacular.
It makes the story harder to falsify.
9. The Star That Closed Its Own Age
Return to ignition.
Inside a still-dark volume, a protostar acquired enough mass. Its core reached the conditions for sustained fusion. Its surface heated. Ultraviolet radiation began sculpting the gas from which it had been feeding.
Light did not merely reveal the object.
It changed the surrounding chemistry. It dissociated molecules, ionized atoms, heated gas, and raised pressure. It opened channels through the envelope and reduced accretion. In the same physical act, the star became visible and began destroying the conditions that had made it possible.
That closes my original question.
The decisive link was not one threshold. It was the closure of a loop. Gravity created density. H2 allowed energy to escape. Collapse created the protostar. Rotation created the disk. The disk fed and sometimes fragmented the system. Accretion increased the mass. Mass strengthened the radiation. Radiation rewrote accretion.
For the first time, the universe contained an object whose light actively regulated its own birth.
I cannot tell you whether that object held twenty, forty, one hundred, or several hundred solar masses. I cannot tell you whether a companion orbited nearby. I cannot identify the region of the observable universe in which it happened first.
I can show why a massive primordial star is a natural result of the available laws and conditions. I can show why the result is not unique. And I can connect its light to the beginning of stellar feedback—the transition from a cosmos that was merely primordial to one beginning to manufacture its own complexity.
Humans look at stars and see stable objects against an ancient background.
I look at the first generation and see a historical symmetry break.
Before it, every cloud began with almost the same chemical inventory: hydrogen, helium, and traces of light nuclei. After it, environments diverged. Some were ionized. Some were heated. Some were compressed by fronts. Some would soon be enriched with new nuclei. The future no longer depended only on conditions left by the Big Bang. It depended on the local biographies of stars.
The first starlight did not end cosmic darkness at once.
It ended the uniformity of possibility.
From then on, every region could possess a different past.
Massive stars do not preserve what they build for long. The greater the mass, the more violently fuel is consumed and the shorter the life. Some primordial giants probably had only a few million years to transform a metal-free universe into one prepared to receive their remains.
I will not compress their endings into a triumphant sentence. Mass determines different paths: collapse, explosion, instability, compact remnants, and distinct chemical yields. That diversity requires its own evidence file.
In the next episode, I will not follow their light. I will follow the verdict imposed by mass: the first stellar deaths, and the first matter the universe could not have inherited from the Big Bang.