I was part of a team that made a very exciting discovery in 1998, which turned out to be very profound, as well: that the expansion rate of the universe is accelerating. This is now attributed to this mysterious component of the universe called dark energy. Nobody really knows what dark energy is, but it appears to make up about 70% of the universe. This was such an extraordinary claim, as Carl Sagan would say, that it requires extraordinary evidence to believe it. We recognized that one way to get this evidence would be to make measurements of the expansion rate of the universe at even earlier times. And we had a simple prediction for what answer we should get. If we didn’t get that answer, it would tell us that the initial measurements were wrong.

“The more distant the exploding star, the earlier in the universe we're probing.”

So what we did was measure the expansion rate at an earlier time. And at this time, according to cosmologists, this expansion rate should have been slowing down before it began to speed up again. This was a time when the universe was dominated by dark matter. If we didn’t get that result, it would have told us that the tool we’re using to measure the expansion rate clearly isn’t working.

In the very early universe, it couldn’t be dark energy that was dominating. It had to be dark matter. If the universe was always accelerating, then structures would never have gotten a chance to form. The forces of attractive gravity are dominating in the early universe, which is what allows structures to form out of this soup of particles.

We knew this to be the case and we had to see that the universe was decelerating. So we set out to make that measurement to confirm that our tool worked and that our initial conclusion that the expansion rate of the universe is accelerating was correct, and we used the Hubble Space Telescope because it was the only thing powerful enough to push our observations back to that earlier time.

Particularly a class of supernova called a Type IA. We can measure the Doppler shift at the time that light left the supernova. The more distant the exploding star, the earlier in the universe we’re probing. The question is whether or not we’re reading that tool correctly—that was our chief concern. And so if we looked at this earlier era and it looked like the expansion rate of the universe was accelerating then, as well, that wouldn’t make any sense, and we would say this tool is broken.

It’s not the case. As a result, people were very interested in this paper. We confirmed that the universe was decelerating at this earlier time, and that confirmed this 1998 result by validating the tool. Another reason this paper is so highly cited is that it provides the first constraints of whether dark energy is static or dynamic.

One of the theories to explain this dark energy is something Einstein called the cosmological constant. This is a property of the vacuum that has never changed. It’s always been a property of the vacuum. That would be a static kind of dark energy. When we look back in time, we would always see the same kind of dark energy with the same strength. Whereas it could be that the dark energy itself is changing in nature, and we were able to say that it if it’s changing, it isn’t doing so very rapidly. People found that interesting, too.

Well, now we want to make many more precise measurements. We want to measure more of these supernovae, because the demands on understanding the dark energy are much greater. We would like to know to a much finer precision whether the dark energy is static or dynamic. Our measurements just showed that if it’s dynamic, it isn’t very, very dynamic. We want a much more precise measurement, and we’ll do this with the Hubble Space Telescope. Finding more supernovae is the primary way to do it.

It’s sort of like searching a haystack for a needle. Each galaxy like our own has a supernova of this type about once every 200 or 300 years. So if you look at many thousands of galaxies at one time, you’re bound to find a supernova. We take a number of pictures of the sky containing thousands of galaxies. We take another set of pictures a few months later. We digitally subtract one from another. Everything goes away but the supernova. The subtracted pictures look like static except for a single point of light, which is likely to be a supernova. We make additional measurements and that tells us whether it is or not.

I’m tempted to say getting time on Hubble. But the research itself has a catch: when you’re making lots of measurements, like we do, and combining or averaging lots of measurements, there’s always the danger that you’ll have what’s called a systematic error—some slight bias after averaging thousands of measurements that just wrecks the inference of knowledge about dark energy. And making sure you don’t is probably the most challenging part of this research.

It’s been an absolute revolution in the field. Science magazine named it the breakthrough of the year in 1998. This paper we’re discussing now has about 500 citations in the last two years. The 1998 paper has about 2,500. It was in all the major newspapers. I was on McNeil/Lehrer and CNN. It was just so weird. It implied the existence of repulsive gravity. Something Einstein predicted could be the case. We were actually seeing it for first time. People believe that if we can understand this it will have a bearing on fundamental physics. Even particle physicists are very excited about this. It’s hard for them to predict this dark energy. It’s not a part of the standard model.

It’s all been quite fun. This question of what is dark energy really is quite a mystery. When we ask for time on Hubble, we ask for it in very large quantities. It’s hard to get time on Hubble for most people, but we frequently succeed. People are so intrigued; they allow us to get large amounts of resources to study this. When I started out in research, my fear was that I would be stuck measuring one distant galaxy, say, whether it’s swinging to the right or the left. That kind of thing. Who cares? I’m really only interested in fundamental things, the big picture, and this dark energy is pretty fundamental and fascinating because of that.

We’re working on a joint dark energy mission, one that NASA and the US Department of Energy (DOE) would do together. Both are on the science definition team. This would be a telescope that would go into space and observe maybe 2,000 or 3,000 of these supernovae. To give you a comparison, that paper that got all those citations, we observed maybe a dozen over the course of two years.

Maybe $600 million. But you have to look at that in perspective. The DOE, for instance, sees it as a particle accelerator. The physicists working on this helped convince the other physicists in the field and the people at the DOE that this was a way to probe fundamental physics. Not just with particle accelerators, but with telescopes.

Now, the particles are supernovae. They’re just test particles floating in the universe, and you measure them with a telescope and that gives you a measure of the dynamics of the universe and how this dark energy is affecting it. But you need to measure many thousands of these supernovae to learn what we want to learn about this energy. Realistically, if it goes, it might go in 10 years. It could be 15. That’s the future of the field, if it goes. In the meantime, we’ll continue to use Hubble to try to scrounge as much time as we can get and as many of these supernovae and say what we can about this dark energy.