Science@Berkeley Lab nameplate Berkeley Lab logo
October 27, 2005
Supernovae: The Stellar Route to Understanding Dark Energy
Part 1: The Fundamental Role of Type Ia Supernovae

The first installment of a new Science@Berkeley Lab series reporting on the recent Supernova Workshop sponsored by SNAP, the SuperNova/Acceleration Probe collaboration, convening leading experts to debate the best ways to study dark energy with exploding stars.

Starting in the late 1980s the Supernova Cosmology Project (SCP), based at Berkeley Lab and led by Saul Perlmutter, pioneered methods of finding enough exceedingly bright, remarkably uniform, very far-off Type Ia supernovae to measure the expansion of our universe. The idea was to compare the brightness of these objects, which is a measure of their distance, with their redshifts, a measure of how much the universe has expanded since they exploded.

At a uniform rate of expansion, redshift and brightness would have a consistent relationship. A supernova brighter than the relationship dictated would also be closer, not having moved as far away in the time since its light was emitted. This would be a sign that expansion was slowing down — just as everyone expected it would be. The question was, by how much?

In 1998 the SCP and their rivals in the High-Z Supernova Search Team (z stands for redshift) announced a startling discovery: the distant Type Ia supernovae they had studied were actually dimmer than expected, meaning they were farther away — a sign that expansion is not slowing, it's accelerating. The mutual gravitational attraction of all the matter in the universe, visible and dark, is insufficient to overcome the negative pressure of a mysterious, invisible something that fills space, something soon to be known as dark energy.

Almost eight years after its discovery, we know that dark energy accounts for more than two-thirds of the density of the universe, but we don't know what it is. Dark energy looms as one of the most puzzling questions — perhaps the most puzzling — facing physical scientists today.

Which is one reason the National Science Foundation, NASA, and the Department of Energy recently assembled a Dark Energy Task Force of astronomers, astrophysicists, and high-energy physicists to advise them on how best to get answers about dark energy, those that can be answered in the short term and those that will require long-term experimental programs.

Theories are many

Theories of dark energy are many but divide into roughly three proposals. Perhaps the least likely: there is no dark energy; gravity itself was significantly weaker in the early universe. Assuming dark energy is real, however, it is either unvarying — what Einstein called the cosmological constant — or it has changed over time, through an agency dubbed quintessence.

To distinguish among these broad alternatives scientists want to measure dark energy's equation of state, the ratio of its pressure to its density, a term labeled w. For quintessence, w changes, with a value more negative than -1/3; for a cosmological constant, its value is always -1. The difference is small, and the best way to detect it is by comparing the distance and redshifts of very far-off objects to high accuracy.

To find and accurately measure the thousands of high-redshift Type Ia supernovae needed to choose among theories of dark energy will require a space-based telescope like SNAP, the SuperNova/Acceleration Probe.

One way to do this may be to test for "baryon acoustic oscillations;" these began as pressure waves in the hot, dense early universe and had lasting effects on the spacing of the galaxies that formed later. Another test is to observe the "weak lensing" of distant galaxies (the gravitational displacement of their light), which betrays the distribution of invisible dark matter in the universe and also measures the growth of large-scale structures (another probe of dark energy). Both these tests may be used to compare the redshifts of far-away objects with their distance and possibly characterize dark energy.

"But Type Ia supernovae have the track record," says Alex Kim of Berkeley Lab's Physics Division. "They are how we discovered dark energy in the first place. They are how we have gotten the highest constraints on dark energy parameters. And they are empirically straightforward."

Yet doing it the way the SCP did it is not good enough to nail down w, says Perlmutter, a member of the Lab's Physics Division and a professor of physics at the University of California at Berkeley. "The problem is that it's very hard to control the details of the changes in the supernova, including dust and other uncertainties, if you work only from the ground or only with the small numbers you can examine with the Hubble Space Telescope. Because we need to study thousands of distant supernovae from space, we developed the concept for SNAP," the SuperNova/Acceleration Probe.

Perlmutter is principal investigator of SNAP and Berkeley Lab's Michael Levi is co-PI. The SNAP collaboration includes more than a hundred scientists from a dozen institutions around the world. So persuasive were the scientific ideas behind SNAP, and so convincing were its design fundamentals, that in 2003 NASA was inspired to join the Department of Energy in announcing a competition for a Joint Dark Energy Mission (JDEM), which stimulated rival designs named Destiny (the Dark Energy Space Telescope) and JEDI (the Joint Efficient Dark Energy Investigation).

Measuring distant Type Ia supernovae to high precision is crucial to all JDEM competitors, although SNAP and JEDI would be equipped to measure baryon oscillation and weak lensing as well. The SNAP collaboration recently convened a three-day workshop, moderated by Kim, which included both SNAP collaboration members and SNAP's JDEM rivals — the world's leading Type Ia researchers.

For three days the workshoppers reported on ground-based searches for "nearby" supernovae, on searches for distant supernovae from the ground and from space, and on theoretical models of how these unique stars explode. They discussed innovative techniques for measuring and comparing Type Ia supernovae, including color magnitude diagrams and spectropolarimetry.

Brighter Type Ia supernovae wax and wane more slowly than fainter ones. When time scales of individual light curves are stretched to fit the norm and brightness is scaled according to the stretch, most light curves match.

Their aim was to pin down the sources of variation in the limited population of Type Ia supernovae known so far and to suggest improved measurement strategies for the future. It is through differences in the expansion rate of the universe over time — presently accelerating, but slowing in the early universe, when matter was still close together and its mutual gravitation strong — that various theories of dark energy can be distinguished. To detect these minor differences at great distances, sources of variation in Type Ia supernovae must be better understood and, to the extent possible, compensated for or eliminated.

"Until recently Type Ia supernovae have been analyzed as a one-parameter family," Kim says, referring to the fact that although some Type Ia supernovae are intrinsically brighter than others, these variations can be corrected simply by correlating maximum brightness with the width of the light curve (the way the exploding star declines from its peak brightness over time). "What is a little worrisome is that there may be more than just that one parameter of variation."

This and subsequent installments of this Science@Berkeley Lab series will touch on a few of the many presentations and discussions that took place during the three-day Supernova Workshop.

When standard candles differ

Identifying a Type Ia supernova depends on its light curve and its spectrum, particularly absorption and emission lines that reveal the presence of specific elements in the evolving atmosphere of the exploding star. Unlike Type II supernovae, which are triggered by the gravitational collapse of the iron cores of very massive stars and have strong hydrogen lines in their spectra, Type I supernovae lack hydrogen lines.

Some Type I's, those designated Ib and Ic, are also core-collapse supernovae, whose precursor stars have lost their outer hydrogen layers. But Type Ia's are in a class by themselves. Not only do their spectra lack hydrogen lines, they show particular silicon absorption lines, a clue to the very different mechanism by which they explode.

It is now widely accepted (although unconfirmed by direct observation, which would require phenomenal luck) that a Type Ia supernova results when a white dwarf star, made of carbon and oxygen, accumulates additional material from a companion star. When the white dwarf is on the verge of reaching the critical Chandrasekhar mass — 1.4 times the mass of our sun packed into a volume about that of the planet Earth — temperature, pressure, and density inside the star instigate thermonuclear burning, triggering a titanic blast.

Since a supernova's spectrum determines its type, one of the first places to look for differences among supposedly similar supernovae is by studying their spectra and correlating these differences to their brightness and light curves. At the workshop, David Branch of the University of Oklahoma reported on spectral variations among a large collection of Type Ia supernovae, which he has divided into different groups including the "cools," the "shallows," and the "broads," distributed around a core group of Type Ia's that meet the classical specifications.

Branch has found lines indicating unusual ions in the spectra of some supernovae, and also familiar lines, like those of ionized iron, in unusual places. Some differences may be due to extrinsic factors like dust between the observer and the supernova, but presumably most of these detailed features are because of differences in the way individual Type Ia's explode.

For example, although the Chandrasekhar mass is always the same, how do ratios of carbon to oxygen in precursor white dwarfs differ? When fusion begins, a nuclear flame proceeds subsonically through the star from the inside out; at some point or points does this deflagration make a transition to supersonic detonation? Does the advancing flame proceed smoothly or in clumps?

Absorption lines of many elements and ions appear, often in several positions, in the spectrum of a normal Type Ia supernova, like this spectrum of SN 1981b taken near maximum brightness. The position of the lines carries information about the chemistry and physics of the explosion. (After Dan Kasen)

Are the elements produced by fusion — which include nickel, iron, magnesium, silicon, sulfur, calcium, and others, a "cocktail of isotopes" — smoothly layered or mixed? Elements at different depths in the rapidly disintegrating supernova apparently travel at different speeds, some relatively slowly, some at very high velocity. The same element can be seen traveling toward the observer faster in some spectra than in others; some iron lines Branch has found suggest detached, high-velocity, doubly-ionized iron (Fe II).

These and other factors affect not only the details of a supernova's spectrum but its overall brightness and its brightness in different colors. Nevertheless, Branch pointed out, the differences among the different groups of Type Ia supernovae — between the shallow and the broad, for example — appear to form a continuum. Understanding the smooth transition from one group to the next allows adjustments to be made; the majority of Type Ia supernovae found so far are indeed suitable for cosmology, and those that are too far off the mark are easily removed from the sample.

Stefano Benetti of the Astronomical Observatory of Padua and his colleagues have also grouped spectra from a set of well-studied Type Ia supernovae into groups, based on the velocities indicated by their silicon II lines. The three groups were dubbed HVG (high velocity gradient) and SVG (slow velocity gradient), the latter including the brightest supernovae, plus FAINT, a group comprised of supernovae that quickly decline in brightness, seem to occur in older galaxies, and indeed are too faint for cosmological purposes.

Benetti suggests that these differences could be due to factors like the density and clumpiness of the explosion, asymmetries in the supernova as a whole — for example, HVG and SVG supernovae could be similar explosions viewed from different angles — and features like intervening dust. Different transitions from deflagration to detonation during the explosions could account for some features in the HVG and SVG groups, and the lack of such a transition might explain the unique FAINT group.

Benetti concluded that once the physics of Type Ia supernovae are better understood, supernova candidates can readily be reconciled to yield consistent data for cosmology.

Future installments in this series will report on innovative analytical methods and theoretical models of Type Ia supernovae, on searches for "nearby" and distant supernovae, and on the discussion of issues raised by the Supernova Workshop.

Additional information

Participants in the SNAP Supernova Workshop, September 2005:

Greg Aldering

Stefano Benetti

David Branch

Ben Dilday

Josh Frieman

Peter Garnavich

Gerson Goldhaber

Peter Hoeflich

Isobel Hook

David Jeffery

Daniel Kasen

Alex Kim

Kevin Krisciunas

Michael Levi

Weidong Li

Peter Nugent

Saul Perlmutter

Mark Phillips

Adam Riess

Fritz Roepke

Roger Romani

Masao Sako

Mark Strovink

Mark Sullivan

Nicholas Suntzeff

Lifan Wang