Making Sense of the Unseen

Dark Energy

What is the shape of the universe? What is its ultimate fate, our ultimate fate? Do we face an ultimate crunch as the universe folds back in upon itself, a “big whimper” as the universe continues to expand into nothingness, or some kind of cosmic perpetual bounce of bang, crunch, bang, crunch?

These are some of the big questions addressed in the still relatively young field of cosmology.

Tam Hunt

It was only a few decades ago that we realized that we inhabit just one enormous galaxy among billions of other galaxies in the observable universe. Our sense of scale has consistently been enlarged over the centuries, adding to the wonder of being alive at this exciting time.

The 2011 Nobel Prize for Physics was awarded to Saul Perlmutter, at Lawrence Berkeley National Laboratory; Brian Schmidt, at Australian National University; and Adam Riess, at Johns Hopkins University, for their work suggesting that the universe is not only expanding but is actually accelerating in this expansion. “Dark energy” is the common explanation for this accelerating expansion, though exactly what this strange form of energy consists of is still very much up for debate.

Robert Kirshner
Nikcholas Suntzeff

Robert Kirshner, Clowes Professor of Science at Harvard, was the PhD advisor to both Schmidt and Riess and has been involved intimately with this work since its inception. Kirshner was part of the High-z Supernova Search Team led by Schmidt, and Perlmutter was chief of the Supernova Cosmology Project (SCP) at Lawrence Berkeley. These two rival teams eventually came to very similar conclusions with different data, providing fairly compelling (but not unassailable) evidence that they’re onto something real with this accelerating-expansion thing.

I’ve been a long-time observer of cosmology and physics more generally, reading avidly in this field since my teen years. I was privileged to interview Kirshner recently about his book, The Extravagant Universe, which describes in accessible terms the interesting history of his and the SCP team’s work on supernovae detection, and the race to be first to announce rigorous results.

What inspired you to go into physics as a career? Was there any particular person or problem involved?

When I was a kid, my mother and father let my big brother and me do things related to science, even when they were risky. My mother was kind of busy with five kids plus her own interests in community theater, so she didn’t supervise us very closely. My father had worked for the Signal Corps and had liberated a significant store of electronics parts at the end of WW II. He traveled a lot in his job at the MITRE Corporation, so he didn’t supervise us either. In October, 1957, we watched Sputnik orbiting overhead and picked up its radio transmissions on an old military receiver. I remember wondering exactly how it stayed up.

We flew model airplanes, ignited home-built rockets, and built transmitters as amateur operators. For an 8th grade science project I built a 40,000 volt Tesla coil from the plans in Popular Science. You could hold a fluorescent bulb in your hand near this thing and it would light up. Or you could catch a wasp in a set of alligator clips and zap it with the high voltage. I might have injured myself with all this freedom, but fortunately that didn’t happen.

When I got to Harvard, I took a freshman seminar in astronomy. That was great – Alan Maxwell was the instructor, and Joe Taylor (who won the Nobel Prize for his work testing General Relativity) was one of the teaching assistants. I always felt I was below average in my college classes, but now I know the average was very, very high. Richard Gott, now a professor at Princeton; Bill Press, who was the youngest professor tenured at Harvard until Larry Summers; Ned Wright, a distinguished professor at UCLA, were all a year ahead of me and we took many of the same classes.

How gratifying was it to have been part of the team that won the Nobel Prize in 2011?

It was pretty clear to me that there was going to be a Nobel Prize for the discovery of cosmic acceleration. Both Brian Schmidt (’93) and Adam Riess (’96) did their PhD work with me, using supernovae to measure cosmic distances, which is exactly what was needed. When the High-z Team got going, I had Peter Garnavich as my postdoc, Saurabh Jha as a beginning student, and Pete Challis as a staff member supported on my grants. We went all-in on the High-z project. Others on the team included Chris Smith (at Cerro Tololo), who had also been my PhD student, and Bruno Leibundgut (at the European Southern Observatory). And, hedging all bets, I also had Pilar Ruiz-Lapuente (Barcelona) as a postdoc and later Tom Matheson, who were on the competing team.

Based on the Shaw Prize to Saul Perlmutter, Brian, and Adam, and then the Gruber Prize where Brian and Adam convinced the Grubers to give their prize to all of us, any rational person could see how a Nobel Prize would be awarded. Either they would find a way to include all of us or it would be split with Saul getting half and Brian and Adam dividing the other half. I certainly knew how the rules worked, so whether this was really the fair way to award the prize or not, I had plenty of time to get used to it.

Your book, The Extravagant Universe, describes the history of the major discoveries that led to the Nobel Prize in 2011, in an interesting and engaging way. I really enjoyed the read, and appreciated the fact that it’s not just a dry account of science in the alleged ivory tower; rather, you discuss real people and real emotions. Could you briefly describe a few highlights of your book and your main point(s)?

I am delighted you’ve read The Extravagant Universe. As I mentioned, many of the members of the High-z Team had been my students or postdocs and I wanted to make sure our story got told. There had been a lot of press accounts of this research and not all of it accurate. Plus I wanted to show how much fun we have doing research.

The key moment came at the beginning of 1998. Adam Riess had been calling me up through November and December, saying it looked like the data were pointing toward cosmic acceleration. This could be the result of a “cosmological constant,” or something like it. Einstein had invented the cosmological constant to make a static universe, then abandoned it when Slipher, Humason, and Hubble showed there were big redshifts that got bigger with distance (as Lemaitre had shown you would get in a solution to Einstein’s equations in an expanding universe.) So the cosmological constant was a kind of theoretical poison ivy that nobody wanted to touch. Or, as my mother said, “Bobby, do you think you’re smarter than Einstein?”

But Adam wasn’t intimidated. After Brian had a chance to check the results, we had a lively discussion inside the team on what to do. Some of the e-mails are in the book. We could wait for more data, but we knew Saul’s team [at Lawrence Berkeley Lab] had more data in hand. I thought it was worse to be wrong than second. Saul’s team had already published a result in July of 1997 that said the supernovae showed the universe was slowing down. Were we confident enough to contradict them? Anyway, we went ahead. You can read about it in the book.

Compare this to the guys with the superluminal neutrinos – they decided to go ahead and publish, but they ended up with egg on their faces. Maybe they should have wiggled a few more cables. Or, on the other side, we have the decision to go ahead and show the recent data from Large Hadron Collider, which may or may not be the Higgs, but surely is something real.

Anyway, the main idea is that we can now make measurements to test our ideas about the universe. This is mostly due to advances in technology: telescopes like Hubble, CCD detectors like the one built for Cerro Tololo, computers to help sift haystacks of data for the sharp needles of supernovae. We’re not smarter than Einstein or Hubble, but we have much better data!

Can you briefly explain the science of “standard candles” and the Hubble shift data that led to the now widely accepted conclusion that our universe is not only expanding, but is accelerating in that expansion?

The great breakthrough in understanding the Universe came in the 1920s when Edwin Hubble figured out how to measure the distances to galaxies from the apparent brightness of the stars they contain. If you know the energy output of a certain type of star, you can gauge its distance from the energy you receive from it each second. If you look out over a desert highway, you see distant cars with faint lights and nearby ones with brighter-looking lights. If they’re all really the same, you can do a good job of judging distances. For our work, we used a particular kind of exploding star, which erupts in an immense nuclear explosion. These are more-or-less the same intrinsic brightness, but a million times brighter than the stars that Hubble used. The apparent brightness goes down like the square of the distance, so this means we can see these Type Ia supernovae a thousand times farther away than the Cepheid variable stars Hubble employed.

Instead of measuring distances to the nearby galaxies of a few million light years as Hubble did, we can use supernovae to measure the distances to distant galaxies a few billion light years away – roughly halfway back to the Big Bang.

The other half of the expanding universe [methodology] is to measure the velocities of galaxies. This was done in the 1920s by Vesto Slipher at the Lowell Observatories, and then later, with Hubble’s encouragement, by Milton Humason at Mount Wilson. The amazing thing is that on average, the velocity of a galaxy is proportional to its distance, as measured from Cepheids or other methods. We interpret this as evidence for an expanding universe.

But the details of the relation between velocity and distance depend on the history of cosmic expansion. The light from distant objects has been traveling to us for a long time, so it reflects the conditions when it left. By carefully tracing the relation between velocity and distance over a large span of cosmic time, we can find out whether the universe has been slowing down (consensus favorite in 1990), staying the same, or speeding up. By 1998, our High-z Team had enough measurements, each of which was precise enough, to show that the expansion of the universe has actually been speeding up over time – there’s cosmic acceleration.

I get the impression from your book, and from other accounts, that you have a very hard nose when it comes to aiming for the most rigorous science. In other words, you’re always pushing for more data and making sure everyone is extremely comfortable with conclusions before they’re published. Would you describe your role on the High-z Team, and more generally, in this manner?

Well, it certainly is true that I hate to be wrong. It’s OK for a theorist to be wrong, as long as they are interesting, but an observer (which I am) should be a reliable source of information. I told Adam Riess that the penalty for being wrong should be as big as the reward for being first. But when you are sitting on top of a big discovery, like cosmic acceleration, you don’t want to be too slow, either.

As you know, there were two teams at work and the other guys had already published a paper in the Astrophysical Journal, in July 1997, that claimed their measurements of distant supernovae showed the universe was slowing down and there was no need for a cosmological constant. When we started to see the opposite, in the fall of 1997, we had to overcome our own reluctance to contradict this work, plus my personal revulsion for the cosmological constant (after all – after all, Einstein had said it was unsatisfactory and banned it from serious discussion in 1931!), to be certain we had a valid result. In my book, there are some verbatim e-mails from the discussion inside our group. In the end, we decided that we should go for it, but not make any public statement until we had the scientific paper describing our work ready to submit to a professional journal.

So we made poor old Bruno Leibundgut say “We’re working on it” in January at the Moriond meeting where he reported on our work, and waited until the Marina Del Rey meeting in February to announce that we saw acceleration. The results were quite astonishing – the press really picked up this announcement and pretty soon Adam Riess was on The News Hour. We submitted our paper in March and it appeared in The Astronomical Journal in September 1998. The other guys submitted a paper in September 1998 that came out in the middle of 1999 with similar results.

Along the same lines, you mention the problem of cosmic dust many times in your book and you leave this issue unresolved in your book because the science was still unsettled by the time you went to press. Has this issue been settled now or is it still possible that other explanations may be offered, such as interstellar dust, for what is currently interpreted as evidence for accelerating expansion?

Yes, dust is really important. It can only make objects dimmer, which you might mistakenly interpret as a greater distance, from the standard-candle idea. So you must have a way to measure dust if you are going to learn about cosmic acceleration (which also makes the more distant objects appear fainter).

When Adam was a graduate student, this is what he worked on with Bill Press and me – how to measure distances accurately even when there is some dust in the way. It turns out dust absorbs blue light more than red light, so you can tell how much dust there is by measuring how it changes the color of a distant supernova. You have to measure the supernova through different color filters and figure out how much dust. Adam figured out a neat way to do this. So did our colleagues working in Chile.

Our competitors had not absorbed this lesson, and made their earliest measurements through just one filter. With just one measurement you cannot tell the difference between a supernova that is dimmed by dust and one that is dimmed by the effect of an accelerating universe.

More recently, we’ve been measuring the light from supernovae at wavelengths beyond visible light – out in the infrared. It turns out that the dust is less important there, and, as an unexpected bonus, the supernovae are also better standard candles at infrared wavelengths. This means the infrared is the place to make the most accurate measurements of cosmic acceleration, and use those to figure out the properties of the dark energy. Is it the cosmological constant or something else? You need very reliable measurements to find out. I am the Principal Investigator of a big program on the Hubble Space Telescope that will use infrared measurements of distant supernovae for this purpose. Maybe I will have something interesting to tell you about this in a year or two.

I look forward to learning more about your work on infrared imaging of distant supernovae as this develops. You mentioned in your book that the issue of pink dust (what you also call playfully “pixie dust”) was still being worked on at that time. Has that issue been resolved or is this what your work on infrared imaging hopes to achieve?

I think this issue is pretty well settled by the wide-wavelength span of our observations. But there still is the puzzling matter of different behavior of dust in cases where there’s a lot of it. In those cases, it looks as if light is scattered more than once, leading to some significant differences in the right way to correct for dust. Overall, these have very little effect on the cosmological evidence for acceleration, but we will need to do better if we want to get really reliable constraints on the nature of the dark energy.

More philosophically, how do you personally conceive of dark energy? Do you view it as a repulsive property of space itself, akin to some type of “new ether” (the phrase Einstein used later in his career to explain the properties of space)? How convinced are you that dark energy is the right theoretical construct to explain the supernova data?

Well, calling it the dark energy is meant to convey the possibility that it could be something different from Einstein’s cosmological constant. But I would say the cosmological constant is the leading contender. The modern view of the cosmological constant would be a pressure (negative pressure) associated with the vacuum, just as Lemaitre discussed in the 1930s. Another possibility is a light scalar field that has similar properties. (The Higgs field is a scalar field – so this idea has some traction.) But there are so many possibilities, it is hard to know how to weed out the ones that are wrong. So the best we can do for the present is to see if the properties of the dark energy do or do not match the predictions of the cosmological constant. If the observations show something different, that would be really interesting.

There’s also the possibility that gravity does not behave exactly as Einstein imagined. While it is hard to come up with a theory that is better than General Relativity and that passes all the known experimental tests, some people are exploring this path.

I must admit I’m always interested in maverick physics theories, as it certainly seems to me that there is plenty of room for debate about what is still a pretty nascent state of the art in cosmology, and what is a maverick view one day often becomes mainstream a few years later. When it comes to the accelerating expansion of the universe, have you followed developments in plasma cosmology at all? In plasma cosmology, developed by Nobelist Hannes Alfven and others like Eric Lerner, the electromagnetic influence of plasmas (which is of course what all stars are comprised of) at the largest spatial scales is considered to be far more important than gravity – an idea which is hard to accept for most physicists who are used to gravity being the only relevant force at the largest scales. Do you think there is any merit to plasma cosmology?


The correct solution to understanding why the cosmological constant is so small eludes us at the moment. When the correct explanation is developed, it will be a new idea, and may seem strange at first. But this does not mean that all strange ideas are correct. The most important thing is to work out their consequences and confront the ideas with the data.

Also, the convergence of many independent lines of evidence – from galaxy clustering, from the properties of the cosmic microwave background, and from the history of cosmic expansion that the supernovae trace – all point to a mixed dark matter/dark energy universe with a pinch of ordinary matter that allows us to see what is going on. There’s a lot less room for speculation in cosmology than there was 30 years ago. This is a good thing – it shows we’re beginning to understand how the universe works.

I hope you don’t mind if I shift toward the philosophy of science in my next few questions. As a non-physicist but someone trained in science (I have a background in biology as well as law) and someone who has read widely in physics for many years, I can’t avoid the impression that today’s cosmology lacks a strong foundation. The word “epicycles” comes up a lot in Richard Panek’s book The Four Percent Universe: The Race to Discover the Rest of Reality. Many of the physicists quoted used this term in reference to dark matter and dark energy because these are, of course, very substantial modifications to general relativity. In other words, they’re late additions to make Einstein’s equations of general relativity, our prevailing theory of gravity, match observations of universal expansion and galactic rotation and formation. Do you feel this charge is unfair (most of Panek’s discussants end up accepting these additions to general relativity at some point)? Are you optimistic that we are well on our way to firming up the foundations of modern cosmology?

The observations are firm enough – there’s strong evidence that the masses of galaxies are much bigger than the mass of the luminous matter they contain. But it is true that the nature of the dark matter remains obscure.

It certainly would help if we detected the dark matter particles. People take these ideas very seriously and have built subtle experiments to find the mist of dark matter particles through which the Sun and Earth are moving as we orbit in the Milky Way. This could happen in the next few years.

Similarly, while the data are now quite solid for cosmic acceleration, it would be very helpful if there were a strong theoretical notion for the mechanism that makes the cosmological constant small (but not zero.) There are many theoretical ideas, but none of them seems compelling. Still, I think the way forward is to make more reliable measurements that could tell the difference between a truly constant cosmological constant and something that is subtly different. Looking ahead, it could be that 10 years from now, when we’ve pinned down the properties of dark energy with 10 times the precision, it still looks just like the cosmological constant. At that point, I suppose people will begin to look elsewhere for exciting observational programs.

Do you have a preferred approach to the philosophy of science? Thomas Kuhn’s approach, fleshed out in The Structure of Scientific Revolutions and other works, suggests that scientific paradigm changes occur (abruptly rather than incrementally) not due to rational debate and accumulation of evidence, but more due to an accumulation of evidence contrary to the prevailing view that is ignored for a long time but at some point becomes too strong to ignore and, so to speak, the dam bursts and a new paradigm is born. Do you see any merit to Kuhn’s thinking with respect to physics? Do you view the now widely accepted notion of accelerating expansion as a paradigm shift?

I suppose you could talk about the dark energy in those terms. In 1997, I certainly was unhappy that our data were pointing toward a cosmological constant – this was both contrary to the published work of the other team and leading us into the theoretical poison ivy that Einstein had banished from the discussion. But another way to think about it is that the theorists (Lawrence Krauss & Michael Turner for example, or Jerry Ostriker and Paul Steinhardt in a Nature paper in 1995) could see how the pieces of the cosmological puzzle would fit together neatly if only there were a non-zero cosmological constant. Then you could get the Hubble constant to agree with the age of the Universe and have a flat, Euclidean geometry for the universe even though all the evidence was that there was too little matter (including the dark matter in galaxies and galaxy clusters) to do this. The balance would be made up by dark energy. So the problem was not conceptual – it was that Perlmutter’s group kept saying at conferences, and finally in an Astrophysical Journal paper in 1997, that they had done the measurement with supernovae and found the universe was decelerating. So I think the important step forward was not an irrational flash of insight, it was a correct measurement of the properties of supernovae.

Everybody was ready for this to be the correct answer.

The cosmological constant and dark energy debate is an interesting case of how physics evolves over time. The history is well-known: Einstein proposed the cosmological constant in his initial work on cosmology in order to preserve a static universe, which was the prevailing view at the time and a view that matched his intuitions about how nature should be. However, he later rejected the cosmological constant as the biggest blunder of his career due to new evidence suggesting that the universe was in fact expanding. Now, many decades later, there is a broad consensus that Einstein was actually right the first time due to the evidence from supernovae that your team and others have collected in recent years. In light of this history, and many other similar evolutions in physics, how confident are you that we generally “have it right” in our views today? Do you see physics changing very dramatically in the next few decades?

It would be pretty silly to say “everything has been turned upside down since 1990, but now we understand everything perfectly.” Surely, a more modest view is appropriate. The evidence for an evolving universe where the visible world traces the effects of dark matter and dark energy is very strong. But, as you suggest, there’s a lot we don’t know. This is a great situation – for a scientist, ignorance is opportunity. Fortunately, we have terrific ideas for more powerful telescopes and experiments that can tell us which picture is right, and help guide our thinking. Unfortunately, some of these instruments are very expensive, but I think that deep down, people want to know how what the world is made of and how it works. Aside from the economic benefits that investment in science and technology brings, it would be a good thing to continue this quest.

I like your statement that “ignorance is opportunity” and I read recently somewhere that science isn’t about knowledge as much as it is about identifying more precisely what we don’t know. You’ve mentioned dark matter and I wanted to ask about this sister concept to dark energy: Isn’t the argument for dark matter almost entirely circular? In other words, dark matter was originally posited to explain anomalous observations about galactic rotations and other large-scale motions and then this same evidence is often used to support the dark matter concept. Experiments designed to find the stuff that constitutes dark matter have, I believe, turned up empty-handed. Am I being unfair here?

On dark matter I think you are missing some important points. First, there’s no question that you either give up on Newton’s gravity or you need a lot of dark matter on the scale of galaxies to account for their rotation curves.

Next, you find that big clusters of galaxies exhibit the presence of matter three ways – by the motions of the galaxies, by the temperature of the hot gas that falls into the gravitational pit formed by the dark matter, and third by the warping of background images as gravitational lensing. And the same amount of dark matter that you infer from the clusters is just what’s needed, when added to the dark energy density, to account for the flat (Euclidean) geometry of the Universe at large.

Now, the hard part is saying what the dark matter actually is. The experimental searches are only now beginning to probe the most likely regions for some kind of weakly-interacting dark matter particle. So it’s not surprising that they haven’t seen anything. But as the searches get better, and we learn more about the world of particles by experiments at the LHC, so we know what to look for, this subject should get a lot more interesting very soon.

On a different note, cosmology is important to scientists and laymen alike, perhaps most importantly because of what our theories say about the ultimate fate of the universe. The Big Bang cosmology was for some time complemented by the “Big Crunch” vision of the end of the universe resulting from gravitational collapse trillions of years hence. With the new data supporting an accelerating expansion, this vision has given way to a constantly expanding universe that will eventually end in a whimper, not a crunch, what is sometimes described as “heat death” as all motion eventually comes to an end and matter and energy eventually simply dissipate into nothingness. Yet another vision that is still considered possible is the “big bounce” cosmology (one I find appealing), in which universes come and go over unimaginable timescales. How much do you think personal intuitions about the fate of the universe influence cosmologists, if at all? What would your preferred “end of the universe” be, if you could actually choose one regardless of the data?

The present information on the nature of the dark energy does not tell you if the universe will expand forever or end in a big crunch. But, if the dark energy is really the cosmological constant, then the expansion will be literally exponential, so that distant galaxies will have their light stretched to the red and then out of sight. The universe will become a dark, cold, lonely place. If this is right, eventually, after Andromeda smashes into the Milky Way, there will be just one large galaxy in our observable universe: much like the picture Einstein started out with in 1917. This is a reason for funding cosmology now!

The universe doesn’t care what you (or I) think, and it is not required to make any sense to beings who have grown up on Earth, which is far from a typical site in the Universe. Our “common sense” is good for predicting the trajectory of baseballs, but very poor at telling us how the universe should behave.

Last, what do you see as the most pressing issues facing physics today? Quantum gravity? Figuring out what results in the observed accelerating expansion? Making string theory an empirical theory? Or something else entirely?

I think the dark energy points precisely at the unsolved problems that lie right at the junction of quantum physics and gravity. Everybody knows we don’t have a quantum theory of gravity, and if we did, there might be some way to see why the cosmological constant must be so small. I am hopeful such a theory will emerge. At the moment, we have a very wide range of speculations and very few ways to weed this lush garden of ideas. So the path forward is through measurement – trying hard to pin down whether the cosmological constant is or is not a sufficient explanation for cosmic acceleration. So far, all the data are consistent with the dark energy being the cosmological constant, but the precision of the measurements is not very good. I am working on better measurements, using infrared observations with the Hubble Space Telescope.


Please note this login is to submit events or press releases. Use this page here to login for your Independent subscription

Not a member? Sign up here.