Sven Erik Matzen

Software Architect | Cloud & Security Expert | AI-enabled Solutions

The Accelerating Cosmos: Type Ia Supernovae, Dark Energy, and Whether Einstein's Constant Is Wavering

🎧 Listen to this article

Astrophysics · 2026-08-25

EU label: fully AI-generated content Fully AI-generated article (no prior review).

The Hook: The Light That Was Too Faint

In the late 1990s, two rival teams of astronomers were hunting for the end of the world — or, more precisely, for the pace at which it approaches. Both started from the same assumption, one that seemed utterly self-evident at the time: the universe has been expanding since the Big Bang, but the gravity of all the matter within it must be slowing that expansion down, just as Earth's pull decelerates a ball thrown upward. The only open question seemed to be whether the braking is strong enough to one day reverse the expansion and collapse everything in a "Big Crunch," or whether the cosmos will keep growing forever, if ever more slowly. The teams set out to measure this deceleration. For that, they needed distant, extremely bright "beacons" of known luminosity whose light had been traveling toward us for billions of years.

When they analyzed their measurements, something was wrong. The most distant supernovae they had observed were not brighter than expected, as a decelerating expansion would have implied — they were fainter. Noticeably fainter. That meant these stellar explosions were farther away than they should have been in a slowing universe. And that permitted only one unsettling conclusion: the expansion of the universe is not slowing down. It is accelerating.

It was as if one had thrown a ball into the air and watched it, instead of falling back, race upward ever faster. Both teams initially distrusted their own data. They searched for errors, for cosmic dust that might dim the light, for peculiarities of early supernovae. But the result held. In 1998 the journal Science named the discovery its "Breakthrough of the Year," and in 2011 the Nobel Prize in Physics went to the three central figures. The cosmos, they had shown, is being driven apart by something unknown — a force that acts against gravity and that we have called dark energy ever since.

This article tells how a too-faint light signal overturned our picture of the universe. It explains why a particular kind of stellar explosion is the most reliable measuring rod in the cosmos, how a number Einstein discarded returned unexpectedly, and why an instrument with 5,000 robotic eyes in the Arizona desert has, since 2024, ignited perhaps the greatest physics debate of the present day: Is dark energy truly an unchanging constant — or does it evolve with time?

Part 1: Measuring the Unreachable

Astronomy's Fundamental Problem

All of cosmology rests on a deceptively simple, in truth treacherous question: How far away is an object in the sky? On Earth we measure distances with a tape measure or a laser. In the starry sky we have only the light that reaches us — and light alone does not reveal whether it comes from a faint candle nearby or an immense searchlight far away.

The solution is a concept of striking elegance: the standard candle. Suppose there existed a class of objects that always shine with the same, precisely known luminosity. Then, from the apparent brightness with which such an object reaches us, one could directly infer its distance. For light thins out with the square of distance: an object twice as far away appears four times fainter. Whoever knows the true luminosity and measures the apparent one can compute the distance. A familiar 100-watt bulb glimmering faintly on the horizon must be far away — the physics does the rest.

The only problem is: where does one find cosmic objects with guaranteed identical luminosity? Stars will not do; they vary by many orders of magnitude. Entire galaxies even less so. For decades this was cosmology's central Achilles' heel — the notorious "cosmic distance ladder," in which each rung is built upon the one below it and measurement errors propagate upward.

The Perfect Explosion

The answer came from one of the most violent events in the universe: the Type Ia supernova. To understand why, of all things, a stellar explosion makes a reliable standard candle, one must know its ignition mechanism.

At the beginning stands a white dwarf — the burned-out, Earth-sized but Sun-heavy stellar remnant of a star like our Sun. It consists mostly of carbon and oxygen and is held stable no longer by nuclear fusion but by the quantum-mechanical degeneracy pressure of its electrons. That pressure, however, has a limit. When the mass of the white dwarf reaches the famous Chandrasekhar limit of about 1.44 solar masses, it can no longer support itself.

Here lies the key to the uniformity. In a binary system, a white dwarf can, over a long time, siphon matter from a companion star (the "single-degenerate" scenario) or merge with a second white dwarf ("double-degenerate"). In either case it approaches the same critical mass threshold. And because ignition always sets in at nearly the same mass, the ensuing catastrophe always follows the same script: the carbon in the interior begins to fuse abruptly, a thermonuclear burning front races through the star, and within seconds it is torn completely apart. The explosion releases so much energy that a single Type Ia supernova can, for weeks, outshine an entire galaxy of hundreds of billions of stars.

The real reason for the uniform luminosity is chemical: the explosion produces large amounts of radioactive nickel-56. This decays via cobalt-56 to iron-56, heating the outrushing cloud of debris from within — this radioactive decay, not the explosion itself, produces the weeks-long afterglow that we see as a supernova. Since a similar initial mass yields a similar amount of nickel-56 (typically about half to three-quarters of a solar mass), Type Ia supernovae reach a remarkably uniform peak brightness.

From "Standard" to "Standardizable" Candle

Nature, however, is not quite that perfect. Look closely at many Type Ia supernovae and their peak brightnesses do scatter noticeably. Were they exactly equally bright, they would be worthless as a precision cosmic tool, because the scatter would drown out the subtle distance differences.

The rescue came in 1993 from the astronomer Mark Phillips. He discovered an astonishingly tight relationship, the Phillips relation named after him: how brightly a Type Ia supernova shines at maximum betrays itself in how quickly it fades afterward. The intrinsically brighter explosions cool more slowly; their light curves are broader. The fainter ones fade more rapidly. The technical measure is Δm₁₅(B) — the number of magnitudes by which the supernova declines in blue light 15 days after maximum. Physically, this too traces back to nickel-56: more nickel means not only more light but also a higher opacity of the shell, which holds the light in longer.

"Brighter is broader" — this simple rule of thumb transformed the Type Ia supernova from a merely approximate into a standardizable candle. One measures the shape of the light curve, corrects the observed brightness accordingly, and the scatter shrinks so far that distances in the universe can be determined to within a few percent. I am of the opinion that this step — turning a messy reality into a precision instrument by means of an empirical relation — ranks among the most elegant tricks in observational astronomy.

Part 2: The Discovery No One Expected

Two Teams, One Race

In the mid-1990s two groups took up the chase. The Supernova Cosmology Project under Saul Perlmutter at Lawrence Berkeley National Laboratory was the older one and had co-developed the technique of systematically catching supernovae: one photographs a patch of sky shortly after new moon, then again three weeks later, and subtracts the images. What remains are new points of light — candidates for supernovae that one then follows up specifically with the world's largest telescopes.

The younger but fast-rising High-z Supernova Search Team was led by Brian Schmidt at Australia's Mount Stromlo Observatory; one of its driving forces was the young Adam Riess. Both teams pursued the same goal: to find the most distant Type Ia supernovae possible, whose light dates from a time when the universe was younger and smaller, in order to read off from it the deceleration of the expansion.

The Wrong Sign

When the results ripened in 1998, both teams independently ran into the same riddle. The distant supernovae appeared about 25 percent too faint. In the language of cosmology this meant: they stood farther away than even a universe with no braking at all should have permitted. The expansion must have, in the time since that light was emitted, not slowed down but accelerated.

The High-z team published its analysis of an initial ten distant supernovae in 1998 in the Astronomical Journal; the Supernova Cosmology Project followed in 1999 with 42 supernovae in the Astrophysical Journal. That two competing groups, with different methods, different telescopes, and different objects, arrived at exactly the same, unwelcome result was decisive. No single error, no quirk of software, no overlooked veil of dust could have fooled two teams at once in the same direction. The acceleration was real.

What Is Accelerating

An accelerating expansion demands a cause. Matter — whether ordinary or dark — pulls together through its gravity and acts as a brake. To drive the cosmos apart, one needs something with negative pressure, a kind of antigravitational tension that stretches space itself. In Einstein's General Relativity this is by no means absurd: there, not only mass and energy density contribute to gravity, but also pressure. A substance with strongly negative pressure can therefore act repulsively.

This mysterious substance was given the name dark energy. It is not the same as dark matter. Dark matter is invisible stuff that, through its ordinary attraction, holds galaxies together and shapes the cosmic web. Dark energy is its antagonist: a smooth, omnipresent energy of space itself that drives it apart. And it dominates the universe: according to the precision data of the Planck satellite from 2018, the cosmos consists of about 68 percent dark energy, roughly 27 percent dark matter, and only barely 5 percent ordinary matter — everything we see, touch, and are ourselves made of. We inhabit the thin, bright rim of a predominantly dark universe.

Part 3: Einstein's Worst Decision — and Its Resurrection

The Constant No One Wanted

The most astonishing thing about dark energy is that its simplest mathematical description had been on hand for over eighty years — unloved and discarded. In 1917, shortly after completing General Relativity, Albert Einstein applied his field equations to the entire universe. To his dismay, they admitted no static, unchanging cosmos such as the worldview of the time demanded: under the influence of gravity, the cosmos had to either shrink or expand.

To artificially bring his universe to rest, Einstein added an extra term to his equations, a Greek lambda: the cosmological constant Λ. It acted like a subtle repulsive force meant to balance gravity exactly and keep the cosmos at rest. It was a purely computational device without physical motivation.

A few years later the basis for that device collapsed. The Russian mathematician Alexander Friedmann showed that Einstein's equations do very much admit dynamic, expanding solutions. And in the late 1920s Edwin Hubble demonstrated observationally that distant galaxies are receding from us — and the farther away, the faster. The universe was in fact expanding. Einstein's static model was superfluous, and with it the cosmological constant. According to legend, Einstein later called the introduction of Λ his "biggest blunder." He struck it from his equations and wanted nothing more to do with it.

The Return from Retirement

All the greater the irony, then, that this very discarded constant returned triumphantly in 1998. For mathematically, a cosmological constant behaves exactly like a dark energy of constant density: an energy inherent to space itself, whose density stays the same even as space expands. While matter "dilutes" as it expands, dark energy remains constant per cubic meter of space. Early on, when the cosmos was small and dense, the braking matter therefore prevailed; only when space grew large enough did the ever-constant dark energy gain the upper hand and drive the expansion into acceleration. This effortlessly explains why the switch from braking to acceleration set in only a few billion years ago.

The modern standard model of cosmology even carries this constant in its name: ΛCDM ("Lambda Cold Dark Matter"). Λ stands for the cosmological constant as dark energy, CDM for cold dark matter. This model describes, with astonishingly few parameters, an overwhelming wealth of observations — from the cosmic background radiation, through the distribution of galaxies, to the abundance of the light elements. One of the discoverers of the acceleration remarked dryly that perhaps there was more insight in Einstein's blunder than in the best efforts of ordinary mortals.

Physics' Most Embarrassing Problem

But the resurrection of the constant brought with it a problem that remains unsolved to this day and that some physicists call the "worst prediction in the history of physics." If dark energy is an energy of empty space, then it is natural to identify it with the vacuum energy of quantum field theory. According to quantum mechanics, the vacuum is by no means empty but seethes with virtual particles constantly coming into and out of existence. These fluctuations carry an energy — exactly the kind of energy inherent to space that could stretch it.

Only: compute how large this vacuum energy ought to be, and out comes a value that exceeds the observed dark energy by about 120 orders of magnitude — a 1 with 120 zeros. Were the vacuum energy really that large, it would long ago have torn space apart so violently that galaxies, stars, or humans could never have formed. Why the actual dark energy is so unfathomably much smaller, without being exactly zero, no one knows. This "cosmological constant problem" is one of the deepest open wounds in theoretical physics — a rift between our two grandest theories, quantum mechanics and General Relativity.

Part 4: What Is Dark Energy Really? The Equation of State

A Single Number That Decides Everything

To distinguish the various proposals for the nature of dark energy, cosmologists use a single, powerful figure: the equation of state, abbreviated w. It is defined as the ratio of pressure to energy density of a substance. For ordinary matter w is practically zero (matter exerts virtually no pressure, at least cosmologically); for radiation w = 1/3.

For dark energy, everything turns on the sign and exact value of w, and the decisive threshold lies at w = –1. If dark energy is exactly a cosmological constant — an unchanging energy of space — then w = –1, precisely, at all times. That is the simplest possibility. The Planck data of 2018 yielded, for a constant value, w₀ = –1.03 ± 0.03, thus compatible with exactly –1. Until recently, everything looked as though dark energy really were Einstein's pure constant.

Quintessence and Phantom

But there are alternatives, and they bear resonant names. If w is not rigidly –1 but changes with cosmic time, then dark energy cannot be a mere vacuum but must be a dynamic field. Models of this kind are grouped under the term quintessence — after the "fifth element" of ancient natural philosophy. In quintessence, dark energy is a slowly changing scalar field permeating space; its value of w then lies above –1 (that is, between –1 and –1/3).

Even more exotic is the region below –1, so-called phantom energy (w < –1). It would mean that dark energy actually increases in density over time — a scenario that could, in the distant future, end in a "Big Rip," in which even atoms are finally torn apart. Pure, physically simple scalar-field models normally do not reach this phantom regime; it is regarded by many theorists as problematic because it leads to instabilities.

The exciting question on which the entire present debate turns, then, is this: Is w constant at –1 (Einstein's constant) — or does it change? And if so, might it even have crossed the magic boundary of –1 over the course of cosmic history? To test this, one often parametrizes w as w = w₀ + wₐ·(1 – a), where w₀ is today's value and wₐ describes how strongly w changes with the expansion factor a of the universe. If wₐ equals zero, we have a constant. If it differs from zero, dark energy is alive.

Part 5: DESI — 5,000 Eyes on the Darkness

A New Tool, a New Measuring Rod

To measure w₀ and wₐ, supernovae alone no longer suffice. One needs a second, independent method of surveying the universe — and that is supplied by the Dark Energy Spectroscopic Instrument, DESI for short. It sits on the 4-meter Mayall Telescope at Kitt Peak Observatory in Arizona and is a technical marvel: in its focal plane are 5,000 optical fibers, each guided by a tiny robot that can realign itself in seconds. Thus DESI records, in a single exposure, the spectra of 5,000 galaxies or quasars at once and measures their redshift — that is, their distance. Over the years the instrument is building the largest three-dimensional map of the universe to date, reaching back about eleven billion light-years into the past.

DESI's measuring rod is not the supernova but a subtler phenomenon: the baryon acoustic oscillations (BAO). In the hot, dense early universe, sound waves propagated through the plasma of matter and radiation. When the cosmos cooled about 380,000 years after the Big Bang and became transparent, these waves "froze" and left behind a characteristic preferred distance in the distribution of matter: galaxies cluster with slightly elevated probability at a separation of roughly 500 million light-years. This frozen sound wavelength — the "sound horizon" — is a cosmic standard length, a ruler of known physical size.

By measuring how large this ruler appears in the sky at different distances (that is, at different cosmic epochs), DESI can reconstruct the expansion history of the universe. Where supernovae as standard candles determine distances via brightness, BAO works as a standard ruler via geometry — two completely independent routes that check one another.

The Results That Made Everyone Sit Up

In April 2024 the DESI collaboration presented its first cosmological results; in March 2025 the second data release (DR2) followed, based on three years of observation and more than 14 million galaxies and quasars. The precision of the BAO measurement reached about 0.24 percent — roughly twice as accurate as before.

Taken on their own, the DESI data are still compatible with the plain ΛCDM model, but show a mild tension of about 2.3 σ between the parameters derived from BAO and those from the background radiation. It becomes truly exciting when one combines DESI with other data sets — the cosmic background radiation and large supernova catalogs. Then a pattern crystallizes: the data fit better a model with w₀ > –1 today and wₐ < 0, that is, dark energy would have behaved in the past like a phantom energy (w < –1) and would have crossed the critical line w = –1 a few billion years ago, around a redshift of z ≈ 0.5, from below to above. In short: the data whisper that dark energy is weakening.

How strong is this whisper? That depends sensitively on which supernova catalog one adds. Depending on the data set used — the three big ones are Pantheon+, Union3, and the five-year sample of the Dark Energy Survey (DESY5) — the statistical significance for a preference toward a time-varying dark energy comes out between about 2.8 and 4.2 σ. The highest value (with DESY5) would already lie close to the threshold at which physicists speak of a "discovery" (5 σ); the lowest (with Pantheon+) is more of a hint than a proof.

Caution Is the Mother of Cosmology

Precisely this dependence on the chosen supernova data set counsels restraint. Several analyses have already examined whether the apparent dynamics of dark energy might be an artifact — for instance through unrecognized systematic errors in the supernova calibrations, or through the particular choice of the parametrization w = w₀ + wₐ(1 – a), which is only one of many possible descriptions. The question of whether DESI DR2 really shows "dynamical dark energy" is the subject of an intense, still-open scholarly debate. A result between 3 and 4 σ has, in the history of physics, often looked tempting and later turned out to be a statistical fluctuation or an underestimated systematic effect.

And yet: should the finding harden — and to that end the coming data releases of DESI, along with new instruments such as the Vera C. Rubin Observatory and ESA's Euclid mission, will contribute — then cosmology would face its greatest upheaval since 1998. A dark energy that changes over time would no longer be a cosmological constant. Einstein's Λ would have to give way to a dynamic field, and the door to a new physics beyond the standard model would stand wide open. I am of the opinion that this is one of the very few fields in which we may, within a few years, witness a decades-old dogma being confirmed or toppled.

Part 6: The Bigger Picture — A Cosmos Full of Open Questions

Two Tensions, One Suspicion

The possible dynamics of dark energy are not the only crack in the seemingly flawless ΛCDM model. For years cosmology has been plagued by a second, stubborn problem: the Hubble tension. Measure today's expansion rate of the universe via the nearby distance ladder (among other things, with those very Type Ia supernovae), and one obtains a markedly higher value than when deriving it from the early background radiation. This discrepancy is by now so significant that it can hardly be dismissed as measurement error. Whether the changing dark energy and the Hubble tension might be two symptoms of the same unknown cause is one of the most exciting questions of the present day — and the reason I expressly recommend the dedicated chapter The Cosmic Tension: Why the Universe Seems to Have Two Rates of Expansion.

Why Standard Candles and Rulers Belong Together

The story of dark energy is at heart a story of distance measurement — and thus related to many other surveying problems of the cosmos. When two neutron stars merged in 2017, they delivered, via their gravitational waves, an entirely novel distance measurement independent of any luminosity: the "standard siren," which likewise contributes to determining the Hubble constant and plays a leading role in Cosmic Gold: Neutron Star Collisions, the r-Process, and the Origin of the Heavy Elements. The fast radio bursts, too, whose dispersion weighs the scattered matter between the galaxies, have become a cosmic measuring tool — as recounted in Milliseconds from the Cosmos: Fast Radio Bursts and the Universe's Missing Matter. And the most extreme objects that make this surveying possible in the first place, from neutron stars to the gravity traps themselves, connect this article with The Clocks Made of Neutrons: Pulsars, Pulsar Timing Arrays, and the Hum of Spacetime and The Shadow of the Invisible: Black Holes, the Event Horizon, and the Event Horizon Telescope.

The Deep Irony

In the end a philosophical punchline remains. Five percent of the universe consists of the matter we know and understand. The remaining 95 percent — dark matter and dark energy — bear names that above all veil our ignorance. "Dark" here does not mean black but simply: unknown. We have measured their effects exquisitely, determined their proportions to within a percent, and reconstructed their history over billions of years — and still do not know what they are. Modern cosmology is thus at once a triumph of precision and a monument to humility. Rarely has physics been so exact and so baffled at the same time.

The Central Takeaway

Perhaps the most important lesson of this story is methodological in nature and valid far beyond astrophysics: A measurement is only as good as the model with which one interprets it — and a surprising result deserves both serious attention and merciless skepticism. The discoverers of the acceleration at first did not believe their own data and hunted for errors for months before bringing themselves to publish. That their finding held was owed to the fact that two independent teams arrived at the same conclusion by different methods. We must apply exactly the same standard to the DESI results today: a 3-to-4-σ hint that depends sensitively on the chosen auxiliary data set is a strong motive to keep researching — but not yet a reason to rewrite the textbooks.

For one's own work — whether in data analysis, in engineering, or in any decision under uncertainty — this means three things. First, independent confirmation beats any single measurement, however elegant. Second, the choice of parametrization or reference data set can shape a result more strongly than the data themselves — always check how robust an effect is against these choices. Third, the most powerful model is worthless if one forgets that 95 percent of the system lies in the dark. Whoever knows the limits of their knowledge measures better than the one who ignores them.

A Question to Reflect On

Cosmology has learned to describe a universe whose dominant constituents it does not understand — and nonetheless to make exact, testable predictions. Where in your own work do you rely on a "dark" element that demonstrably works, but whose inner nature you do not know at all — and how would you notice that this silent foundation had just begun to change?

Cross-References in the Vault

Sources

← All articles