Did Dark Energy Just Disappear? | Space Time | PBS Digital Studios

A fresh analysis of Type Ia supernova data has revived a familiar headline—“dark energy may have disappeared”—but the underlying conclusion hasn’t flipped. The updated results still favor an accelerating universe driven by a positive cosmological constant; they just weaken the certainty enough that a non-accelerating expansion history is no longer ruled out as decisively by supernovae alone. In other words, dark energy hasn’t vanished, yet the statistical case for it from supernova measurements by themselves is less airtight than before.

The story traces back to 1998, when two independent supernova teams reported that the universe’s expansion was accelerating rather than slowing under the pull of matter. They used Type Ia supernovae—exploding white dwarfs with predictable brightness—to infer distances and reconstruct the expansion history over billions of years. Those early datasets were small, but the pattern was striking: instead of the expected deceleration, the inferred expansion rate increased for roughly half the universe’s age. That discovery earned a Nobel Prize for Adam Riess, Brian Schmidt, and Saul Perlmutter.

The new work, published in October 2016 by Nielsen, Guffanti, and Sarkar in Nature, reanalyzes updated supernova compilations. The sample size is dramatically larger—740 Type Ia supernovae compared with about ten in the Riess/Schmidt analysis and 49 in the Perlmutter dataset. With more events, the expectation is higher confidence. Yet the analysis finds the data can be consistent with no acceleration, meaning the “no dark energy” scenario sits within the broader range of acceptable expansion histories.

The key nuance is statistical, not physical. The study reports about 3-sigma confidence for a positive cosmological constant. In practical terms, that level of significance corresponds to a false positive occurring roughly 0.27% of the time—about 1 in 300 hypothetical repetitions—when the universe truly has no dark energy. For a claim as consequential as dark energy’s existence, that’s not enough. Scientists generally look for around 5-sigma, where false positives are rarer (about once per 3.5 million trials). The supernova-only result therefore reads as a strong hint, not a proof.

Crucially, the earlier Nobel-winning supernova papers also hovered around similar low significance when using supernova data alone. What pushed the case over the threshold was combining supernova evidence with other cosmological constraints. Dark energy can’t be observed directly; it’s inferred from how it alters cosmic expansion. That means the analysis must account for the competing effects of matter (which slows expansion through gravity) and dark energy (which accelerates it). The relevant parameters are the fractions of the universe’s total energy in matter and dark energy, commonly denoted Omega m and Omega Lambda.

Supernova contours alone allow a sliver of parameter space where dark energy is near zero, but that region implies an almost empty universe with almost no matter. Independent measurements of matter density—by counting galaxies and weighing dark matter—place Omega m around 0.3 and at least about 0.2, ruling out the low-matter corner. Geometry provides another decisive cross-check: the cosmic microwave background indicates the universe is very close to flat, which constrains the allowed combinations of Omega m and Omega Lambda. When supernova constraints are combined with cosmic microwave background geometry (and also other probes such as baryon acoustic oscillations), the “no dark energy” region becomes far too remote to remain viable.

So the headline is misleading. The updated supernova analysis reduces the certainty from supernovae alone, but the broader cosmological picture still requires something that counteracts gravity and yields the observed near-flat geometry. The episode closes by emphasizing the scientific process: even widely accepted results get re-tested, and the community now needs to understand why the confidence level shifted in the new analysis.