Time’s seemingly inexorable march has always provoked interest in, and speculation about, the far future of the cosmos. The usual picture is grim. Five billion years from now the sun will puff itself into a red giant star and swallow the inner solar system before slowly fading to black. But this temporal frame captures only a tiny portion—in fact, an infinitesimal one—of the entire future. As astronomers look ahead, say, “five hundred and seventy-six thousand million years,” as humorist Douglas Adams did in The Restaurant at the End of the Universe, they meet a cosmos replete with myriad slow fades to oblivion. By then the accelerating expansion of space will have already carried everything outside our galaxy beyond our view, leaving the night sky ever emptier. Lord Byron captured the prospect of such a celestial wasteland in his 1816 poem “Darkness”: “The bright sun was extinguish’d, and the stars/Did wander darkling in the eternal space.”

But here’s the good news: oncoming darkness captures only half the story. Star formation has indeed long since passed through its most glorious epoch, but the universe has life in it yet. Strange new beasts will enter the astronomers’ zoo. Outlandish phenomena that now occur rarely, if at all, will become routine. Cosmic conditions favorable to life may, if anything, become even more abundant.

Scientific eschatology—the study of the far future—has a distinguished history in cosmology and physics. Fascinating in its own right, this endeavor also offers a conceptual testing ground for new theories, plus an opportunity to make abstract ideas more concrete. One of the most abstract ideas of all, the shape of space, may prove easier to grasp when cosmologists describe what it implies for the fate of the universe. Physicists who seek to reconcile their disparate theories about fundamental particles and forces predict processes that will occur only after trillions of years, or even longer, such as the decay of protons and the evaporation of black holes. Increasingly, astrophysicists bring the far future into their models of stellar and galactic evolution as well. During the past decade they have attempted to reconstruct the ways that the formation and composition of stars and galaxies have changed since the big bang. Their growing knowledge of the past allows them to extrapolate trends into the far future.

Forgot to Turn It Off
Among this subject’s pioneers is Greg Laughlin, an expert on star formation at the University of California, Santa Cruz. As a graduate student, he created a computer code to calculate the evolution of extremely low mass stars and forgot to flag it to turn off after reaching the present age of the universe. Left to its own devices, the program ran and ran, producing trillions of years of future predictions—quite wrong, as things turned out, but sufficient to get him hooked on the subject.

To know the future of stars requires understanding how they form. Stars are born within interstellar clouds of gas and dust, which contain hundreds of thousands to several million times the mass of our sun. Such stellar nurseries, sprinkled throughout the Milky Way, gave birth to its few hundred billion stars and will eventually produce tens of billions more. Yet this success consumes the future: the raw material for new stars is being used up. Even though massive stars die in supernova explosions that return some material into interstellar space and even though galaxies can also accrete fresh gas from intergalactic space, the new material cannot replenish all the gas that stars have locked up. The interstellar gas within our galaxy now totals only a tenth or so of the mass in stars.

Today stars form in the Milky Way at a rate close to one solar mass per year, but at its peak, eight to 10 billion years ago, the rate was at least 10 times higher. Laughlin estimates that star formation will decrease by approximately a factor of 10 for every factor of 10 in time, so that in 100 billion years it will slow to a tenth of its present rate, and a trillion years from now stars will form at only about one one-hundredth of the current rate.

That said, impressive changes could disrupt the steady march toward stellar obscurity. We in the Milky Way, for example, must soon—“soon,” as in a few billion years—confront the arrival of the onrushing Andromeda system, the closest giant spiral galaxy to our own. The dense central regions of these two galaxies will either collide or begin to orbit their common center of mass. Their interaction will produce “Milkomeda.” By churning and stirring the interstellar gas and dust, the creation of Milkomeda will reinvigorate star formation temporarily, producing what astronomers call a starburst. Once this growth spurt dies away, the merged system will closely resemble an elliptical galaxy, a mature system with a low density of star-forming material and a consequently low rate of star formation.

In addition to forming in smaller numbers, stars of the future will show the effects of the changes that will occur in their raw material. The fiery furnace of the big bang forged hydrogen, helium and lithium. All the heavier elements have been created by stars themselves, typically late in their lives—either within red giant stars, which shed their outer layers as they age, or during supernova explosions. Red giants provide most of the lighter and more abundant heavy elements, such as carbon, nitrogen and oxygen, whereas supernovae produce a wider range, all the way up to uranium. All of these mix into the existing elemental mulch of interstellar gas, allowing subsequent generations of stars to begin life with more of these materials. The sun, a comparative youngster at five billion years old, has 100 times the heavy-element abundance of stars that formed over 10 billion years ago; indeed, some of the oldest stars contain almost no heavy elements at all. Coming generations of stars will be even more enriched, which will alter their inner workings and outward appearance.

New Abodes for Life
The steadily increasing abundance of heavy elements within newborn stars produces two noticeable effects. First, it augments the opacity of a star’s outer layers. Hydrogen and helium are nearly transparent, but even a modest pinch of heavy elements traps radiation, lowering the star’s luminosity. The balance of forces within the star shifts because the lower luminosity means that the star consumes its nuclear fuel at a lower rate. If only this effect were in operation, a star rich in heavy elements would live longer than a star of the same mass that lacks those elements. But a second effect counteracts the first: the heavy elements are nuclear deadweight. Because they do not participate in nuclear fusion, they reduce the amount of nuclear fuel available within a star of a given mass and tend to shorten its life.

Laughlin and his colleague Fred Adams of the University of Michigan made the initial study of these two effects in 1997. They found that the first will dominate for the next trillion years or so, as the increase in heavy elements within new stars raises their opacities and thus lengthens their lifetimes. Eventually, however, the heavy elements will constitute a significant fraction of stars’ masses and will begin to shorten their lifetimes. The crossover point will occur when the heavy-element fraction within a newborn star reaches about four times the current value.

The extraheavy elements should also favor the birth of planets, along with stars, and thus the prospects for life in the universe. Astronomers have measured the elemental abundances in the stars around which more than 700 (and counting) Jupiter-like planets have now been discovered. Their results show that stars with greater heavy-element abundances are more likely to have one or more giant planets in orbit around them. “Jovian-type planets show a definite correlation with [heavy-element abundances],” says John Johnson, a planet-hunting expert at the California Institute of Technology. “Because the interstellar medium is getting steadily enriched [in heavy elements], planet occurrences will probably increase.”

What about Earth-like planets? Although space-based telescopes are only now beginning to provide similar data for smaller worlds, their formation, too, should be correlated with their stars’ heavy-element abundances, even more so as Earth-like planets consist almost entirely of the heavier elements. In short, the universe of the far future should be filled with planets. Despite the diminishing rate of star formation, perhaps half or two thirds of all the planets that will ever exist have yet to be born.

At first, the proliferation of planets does not seem promising for life. Most of the stars of the far future will be much less massive and less luminous than the sun. Fortunately, even a low-mass, dim star can allow life to flourish. A star with as little as one one-thousandth of the sun’s luminosity can maintain temperatures that allow liquids to exist on close-in planets, satisfying what seems to be a requirement for living things to exist.

Planets should not only grow generally more common but also be enriched in the stuff of life. In addition to requiring a liquid bath, life on Earth, as well as almost all other forms of life that scientists speculate about, depends on the existence of carbon, nitrogen and oxygen. As time goes on, the increasing relative abundance of these elements should yield planets more hospitable to life. Therefore, as star formation steadily diminishes, every newborn star should appear with a progressively greater probability of lighting one or more potential life-bearing planets. Some of these new stars will have the low masses and tiny luminosities that allow them to last for hundreds or thousands of billions of years (not that such immense lifetimes seem necessary for the origin and evolution of life). However full or empty of life the universe may be today, it should teem with more abundant and more varied forms of life in the future.

When Worlds Collide
Planetary systems will endure so far into the future that new considerations will come into play. We take the stability of our solar system for granted; no one worries that Earth’s orbit will soon grow chaotic and cause us to collide with Venus. That confidence evaporates when we look to multibillion-year time­­scales. In 2009 Jacques Laskar and Mickael Gastineau of the Paris Observatory conducted sev­eral thousand computer simulations of the future orbits of the sun’s four inner planets, varying the planets’ initial positions by a tiny amount—just a few meters—between each simulation. They found a probability of about 1 percent that Mercury would smack into Venus during the next five billion years, setting the stage for even more horrific collisions that would probably involve Earth. Over a trillion years such collisions would become highly probable.

The pot will be stirred when the Andromeda galaxy merges with the Milky Way, an event that will reconfigure both galaxies’ gravitational fields and could well trigger a wholesale restructuring of the solar system. As Laughlin commented in reviewing Laskar and Gastineau’s simulations, “What now remains is to understand the extent to which the hand of dynamical chaos that so lightly touches our solar system has molded the galactic planetary census.”

The orbital chaos within a star’s planetary family will also occur on much larger scales. The stars in closely bound double-, triple- and higher-multiple star systems orbit the center of mass of each system under their mutual gravitational influences. Much the same is true for star clusters and even entire galaxies. Stars in all these structures almost never make contact; huge expanses of space separate them despite their astronomical neighborliness.

Over long expanses of time, however, “almost never” ratchets up to “sometimes” and ultimately to “almost always.” Every double-star system will eventually experience either disruption, as the result of external gravitational forces, or merger, if the two stars orbit so closely that gravitational radiation saps the system of energy. Naturally enough, widely separated double-star systems face the former fate, whereas close-in binaries confront the latter.

When two stars merge, they may temporarily produce a more massive, more luminous star [see “When Stars Collide,” by Michael Shara; Scientific American, November 2002]. Even a planet such as Jupiter can cause a similar effect, though on a smaller scale. Consider a modest star, with just one tenth of the sun’s mass and a lifetime close to a trillion years, and suppose it has a Jupiter-like planet. If the planet has an orbital period greater than few days, it will probably eventually be lost from the system. But if it moves on a tighter orbit, the planet could eventually merge with the star, contributing a fresh supply of hydrogen that would temporarily boost the star’s energy output dramatically, producing a novalike outburst. In the future such stellar eruptions will punctuate the slow decline in star numbers and brightnesses. Astronomers even a trillion years from now will observe some strange events among the ever declining numbers of stars in their host galaxies.

Live Slow, Die Old
Even after tens and hundreds of billions of years have elapsed, even when star formation has slowed to a trickle, enormous numbers of stars will continue to shine. Most stars in the universe have low masses and extremely long life expectancies. Stars’ lifetimes depend on their masses in a strikingly inverse manner. High-mass stars are so luminous that they burn themselves out quickly and explode after a few million years. Medium-mass stars such as our sun shine modestly and last for billions of years. Stars with significantly less than the sun’s mass can endure for hundreds of billions of years or even longer. These stars consume their fuel so slowly that even their meager supplies can feed their nuclear fires through these immense spans of time.

Stars of different masses die in different ways. The sun will become a red giant and, as its outer layers dissipate completely into interstellar space, reveal its core as a white dwarf—a dense, Earth-size stellar corpse made almost entirely of carbon nuclei and electrons. But in stars with less than about 50 percent of the sun’s mass, the core temperature never rises high enough to trigger the nuclear reactions that lead to the red-giant phase. Instead astronomers think these stars eventually become helium white dwarfs. Such beasts, as the name suggests, consist almost entirely of helium, with little if any hydrogen and just a smattering of other elements. In today’s universe they are occasionally born when two close binary stars strip each other’s outer layers before they can ignite their helium cores, but astronomers have yet to discover any that arose in the normal course of stellar evolution, because not enough time has passed since the big bang so far. Isolated helium dwarfs are a prime example of a novel phenomenon that our distant descendants (may they live in peace) will one day see for the first time.

Stars with larger masses undergo far more dramatic deaths. The collapse of a massive star core that forms either a neutron star or a black hole triggers a shock wave that blasts the star’s overlying layers into space in a supernova explosion. As massive stars disappear from the skies, so, too, will most of these explosions that now punctuate the cosmos. But a second kind of supernova will still occasionally light the skies. This class, called type Ia supernovae, arises in binary-star systems in which one star has become a white dwarf. According to astronomers’ most favored models, in some of these stellar pairs hydrogen-rich material from the companion star collects on the white dwarf’s surface until its sudden nuclear fusion produces a supernova. Such events will take place as long as there are sufficiently massive companions, perhaps for another 100 billion years or so.

In another supernova model, which has been gaining in popularity, two white dwarfs orbit their common center of mass in close proximity. As they do so, their orbital motions cause the binary system to emit gravitational radiation. This emission robs the system of energy and shrinks the size of the white dwarfs’ orbits. The approach of the dwarfs proceeds ever more rapidly, until their death spiral melds them in a brief, final paroxysm. Such events might continue to occur for trillions of years.

Even brighter than supernova explosions are gamma-ray bursts (GRBs). These megaexplosions come in two distinct varieties, which apparently originate in two entirely different scenarios. Long GRBs, those whose eruptions of energetic radiation last for two seconds or more, are believed to occur when a massive star’s core collapses to form a neutron star. Short GRBs, whose outbursts last for less than two seconds, are thought to result from the merger of a neutron star with either another neutron star or a black hole. Over the coming eons the long variety will become exceedingly rare, as massive stars cease to form, but short bursts might punctuate the heavens for trillions of years.

Trillions and Trillions
When we measure cosmic time not by billions but by trillions of years, we enter an epoch when star formation will have ended. All but the lowest-mass stars will have burned themselves out, ending their lives either by blowing up or by withering into white dwarfs. Not counting dark matter, whose composition remains a mystery, our galaxy—and all others in the universe—will then consist primarily of black holes, neutron stars, white dwarfs and extremely faint red stars, so dim that none of them would be visible without a telescope, even at distances less than the current distances from the sun to the nearest stars. How sad, how degenerate, how uninteresting.

And yet, among these dead or fading objects, nature will on occasion produce an enormous outburst, a brief reminder of the nuclear fury that once spangled the heavens with the light from billions of stellar furnaces. If the surviving stars have planets in close propinquity—and we may expect that many or most of them will—then liquid water, along with various forms of life, could appear and endure on their surfaces. Any life that might arise on those planets will have the possibility (already present around the faintest stars) of lasting for epochs well beyond easy imagination, provided they can avoid being blasted into eternity by nearby supernovae or GRBs.

This survey of the far future leaves a great and indeterminate issue. Could highly advanced civilizations, if they exist and persist, change the course of the cosmic history? More than 30 years ago Freeman Dyson of the Institute for Advanced Study in Princeton, N.J., reviewed the situation. The grand leader in this sort of cosmic speculation, he stated, “I think I have shown that there are good scientific reasons for taking seriously the possibility that life and intelligence can succeed in molding this universe of ours to their own purposes.” In our present ­epoch, not even 14 billion years after the big bang, little evidence exists that living things have affected the cosmos on a grand scale. But time’s train has barely left the station. In the future the survi­­val of life will require that it commandeer an ever greater fraction of the cosmos’s resources [see “The Fate of Life in the Universe,” by Lawrence M. Krauss and Glenn D. Starkman; Scientific American, November 1999]. All of the universe will become our garden.

Bound on this journey for a brief moment, we have little chance of attaining absolute certainty about what will actually happen. Our unfettered minds remain free to roam as far into the future as we choose. As W. H. Auden wrote in his 1957 poem, in an entirely different context: “Were all stars to disappear or die/I should learn to look at an empty sky/And feel its total darkness sublime/Though this might take me a little time.”

This article was published in print as "The Far, Far Future of Stars."