Skip to content
Starts With A Bang

Ask Ethan #83: What if dark energy isn’t real?

If our “standard candles” aren’t so standard, is dark energy still real?

Some say the world will end in fire,
Some say in ice.
From what I’ve tasted of desire
I hold with those who favor fire.
But if it had to perish twice,
I think I know enough of hate
To say that for destruction ice
Is also great
And would suffice.” –
Robert Frost

Every once in a while, some Earth-shattering discoveries come along that forever change our view of the Universe. Back in the late 1990s, observations of distant supernovae made it clear that the Universe wasn’t only expanding, but that distant galaxies were actually speeding up as they moved away from us, a Nobel Prize-worthy discovery that told us the fate of our Universe. But among your questions and suggestions this week was one from João Carlos, who pointed out a new study, and asked:

I read this on Eurekalert! and thought you should, too. I can’t wait to see [y]our comments about this.

The “this” in question was from a University of Arizona press release — a place where I was a postdoc just a few years ago — that said the following:

Image credit: screenshot from

This is potentially a very, very big deal for our understanding of all there is, and how our Universe will end. Let’s go back nearly 100 years to a lesson we should have learned, and then come forward to today to see why.

Image credit: European Southern Observatory (ESO), via

Back in 1923, Edwin Hubble was looking at these obscure, faint “spiral nebulae” in the sky, studying novae occurring in them and trying to add to our knowledge of just what these objects were. Some people contended that they were proto-stars within the Milky Way, while others believed them to be island Universes, millions of light years beyond our own galaxy, consisting of billions of stars apiece.

While observing the great nebula in Andromeda on October 6th of that year, he saw a nova go off, then a second, and then a third. And then something unprecedented happened: a fourth nova went off in the same location as the first.

Image credit: Edwin Hubble / Carnegie Observatories, via

Novae do sometimes repeat, but it usually takes hundreds or thousands of years for them to do so, as they occur only when enough fuel builds up on the surface of a collapsed star to ignite. Of all the novae we’ve ever discovered, even the most rapidly replenishing takes many years to go off again. The idea that one would repeat in only a few hours? Absurd.

But there was something we knew about that could go from very bright to dim to bright again in just a few hours: a variable star! (Hence, his crossing out of “N” for nova and excitedly writing “VAR!”)

Images credit: ESA / Hubble, of the star RS Puppis, via

The incredible work of Henrietta Leavitt taught us that some stars in the Universe — Cepheid variable stars — get brighter-and-dimmer with a certain period, and that period is related to their intrinsic brightness. This is important, because it means that if you measure the period (something easy to do), then you know the intrinsic brightness of the thing you’re measuring. And since you can easily measure the apparent brightness, then you can immediately know how far away that object is, because the brightness/distance relationship is something we’ve known for hundreds of years!

Image credit: E. Siegel.

Now, Hubble used this knowledge of variable stars and the fact that we could find them in these spiral nebulae (now known to be galaxies) to measure their distances from us. He then combined their known redshift with these distances to derive Hubble’s Law and figure out the rate of expansion of the Universe.

Remarkable, right? But unfortunately, we often gloss over something about this discovery: Hubble’s conclusions for what that expansion rate actually was were totally wrong!

Image credit: E. Hubble, 1929.

The problem, you see, was that the Cepheid variable stars that Hubble measured in these galaxies were intrinsically different than the Cepheids that Henrietta Leavitt measured. As it turned out, Cepheids come in two different classes, something Hubble didn’t know at the time. While Hubble’s Law still held, his initial estimates for distances were far too low, and so his estimates for the expansion rate of the Universe were far too high. In time, we got it right, and while the overall conclusions — that the Universe was expanding and that these spiral nebulae were galaxies far beyond our own — didn’t change, the details of the expansion definitely did!

And now, we come to the present day.

Image credit: NASA/ESA, The Hubble Key Project Team and The High-Z Supernova Search Team.

Far brighter than Cepheids, supernovae can often outshine — albeit briefly — the entire galaxy that hosts it! Instead of millions of light years away, they can be seen, under the right circumstances, more than ten billion light years distant, allowing us to probe farther and farther into the Universe. In addition, a special type of supernova, type Ia supernovae, arises from a runaway fusion reaction taking place inside a white dwarf.

When these reactions occur, the entire star is destroyed, but more importantly, the light curve of the supernova, or how it brightens and then dims over time, is well-known, and has some universal properties.

Image credit: S. Blondin and Max Stritzinger.

By the late 1990s, enough supernova data had been collected at large enough distances that two independent teams — the High-z Supernova Search Team and the Supernova Cosmology Project — both announced that based on this data, the Universe’s expansion was accelerating, and that there was some form of dark energy dominating the Universe.

Like many people, I was skeptical of this, as if supernovae weren’t as well-understood as we’d thought, these entire conclusions would go away.

Images credit: NASA / CXC / M. Weiss.

For one, there were two different methods by which supernovae could occur: from accretion of matter from a companion star (L), and from a merger with another white dwarf (R). Would both of these result in the same type of supernova?

For another, these supernovae at great distances may have been occurring in very different environments from the ones we see close by today. Are we positive that the light curves we see today reflect the light curves at great distances?

For still another, it’s possible that something happened to this light during their incredible travels from great distances to our eyes. Are we sure there isn’t some new type of dust or some other light-dimming property (like photon-axion oscillations) at work here?

Image credit: NASA/Swift/P. Brown, TAMU.

As it turns out, these issues were all able to be resolved and ruled out; these things aren’t issues. But recently — and what João Carlos’ question is all about — we’ve discovered that these so-called “standard candles” may not be so standard after all. Just like the Cepheids come in different varieties, these type Ia supernovae come in different varieties too.

Imagine you had a box of candles that you thought were all identical to one another: you could light them up, put them all at different distances, and immediately, just from measuring the brightness you saw, know how far away they are. That’s the idea behind a standard candle in astronomy, and why type Ia supernovae are so powerful.

Image credit: NASA/JPL-Caltech.

But now, imagine that these candle flames aren’t all the same brightness! Suddenly, some are a little brighter and some are a little dimmer; you have two classes of candles, and while you might have more of the brighter ones close by, you might have more of the dimmer ones far away.

That’s what we think we’ve just discovered with supernovae: there are actually two separate classes of them, where one’s a little brighter in the blue/UV, and one’s a little brighter in the red/IR, and the light curves they follow are slightly different. This might mean that, at high redshifts (large distances), the supernovae themselves are actually intrinsically fainter, and not that they’re farther away.

In other words, the inference we drew — that the Universe is accelerating — might be based on a misinterpretation of the data!

Image credit: Ned Wright, based on the latest data from Betoule et al. (2014), via

If we’ve got the distances wrong for these supernovae, maybe we’ve got dark energy wrong, too! At least, that would be the big worry. The smaller worry would be that dark energy is still real, but there might be less of it than we previously thought.

So which of these worries are valid? As it turns out, only the small one, and not the big one! You see, in 1998, we only had supernova data pointing towards dark energy. But as time went on, we gained two other pieces of evidence that provided evidence that was just as strong.

Image credit: ESA and the Planck Collaboration.
Image credit: Planck Collaboration: P. A. R. Ade et al., 2013, A&A Preprint; annotations by me.

1.) The Cosmic Microwave Background. The fluctuations in the leftover glow from the Big Bang — as measured by WMAP and later, to higher precision, Planck — strongly indicated that the Universe was about 5% normal matter, 27% dark matter, and about 68% dark energy. While the microwave background doesn’t do a great job by itself of telling you what the properties of this dark energy are, it does tell you that you have about 2/3 of the Universe’s energy in a form that isn’t clumpy and massive.

For a while, this was actually a problem, as supernovae alone indicated that about 3/4 of the Universe’s energy was dark energy, so it’s possible that these new revelations about supernovae could help the data line up better.

Image credit: Zosia Rostomian, Lawrence Berkeley National Laboratory.

2.) The way galaxies cluster. In the early Universe, dark matter and normal matter — and how they do-and-do-not interact with radiation — govern how galaxies wind up clustered together in the Universe today. If you see a galaxy anywhere in the Universe, there’s this odd property that you’re more likely to have another galaxy about 500 million light years away from it than you are to have one either 400 or 600 million light years away. This is due to a phenomenon known as Baryon Acoustic Oscillations (BAO), and it’s because normal matter gets pushed out by radiation, while dark matter doesn’t.

Travel the Universe with astrophysicist Ethan Siegel. Subscribers will get the newsletter every Saturday. All aboard!

The thing is, the Universe is expanding due to everything in it at all times, including dark energy. So as the Universe expands, that preferred scale of 500 million light years changes. Instead of a “standard candle,” BAO allows us to have a “standard ruler,” which we can also use to measure dark energy.

Image credit: NASA / JPL-Caltech.

As it turns out, the measurements from BAO are just as good at present as the measurements from supernovae, and seem to give the same results: a Universe that’s about 70% dark energy, and consistent with a cosmological constant and not domain walls, cosmic strings, or many other exotic types.

In fact, if we combine all three data sets, we find that they all point roughly towards the same picture.

Image credit: Supernova Cosmology Project, Amanullah, et al., Ap.J. (2010).

What we’ve learned from this is that the amount of dark energy and the type of dark energy we infer from supernovae may change slightly and in a subtle manner, and this may actually be good for bringing the three methods — supernovae, the CMB and BAO — into better alignment. This is one of those great moments in science where one incorrect assumption doesn’t cause us to throw all our results and conclusions out, but rather where it helps us more accurately understand a phenomenon that’s puzzled us since we first discovered it.

Dark energy is real, and thanks to this new discovery, we just might come to understand it — and its effects on the Universe — better than ever before. Thanks, João Carlos, for the opportunity to take on such an interesting discovery, and if you’ve got a question or suggestion for the next Ask Ethan, send it in here!

Leave your comments at the Starts With A Bang forum on Scienceblogs!


Up Next