Scientists at Lawrence Berkeley National
Laboratory, working with colleagues at the European Southern
Observatory (ESO) and the University of Texas at Austin,
have established that the extraordinarily bright and
remarkably similar astronomical “standard candles” known as
Type Ia supernovae do not explode in a perfectly spherical

Led by Lifan Wang, an astronomer and astrophysicist in
Berkeley Lab’s Physics Division, the researchers used the
ESO’s Very Large Telescope (VLT) in Chile to measure the
polarization of light emitted by supernova 2001el as it
brightened and dimmed. This is the first time the intrinsic
polarization of a normal Type Ia supernova has been

The researchers were able to show that at peak brightness
the exploding star was slightly flattened, with one axis
shorter by about 10 percent. By a week later, however, the
visible explosion was virtually spherical.

“For the first time we have actually measured the asymmetry
of a Type Ia supernova,” says Wang. Not only can this
information be used to test models of how Type Ia supernovae
originate and explode, he says, it also helps to underline
“how valid supernovae are for doing cosmology.”

It was by comparing the brightness and redshift of Type Ia
supernovae that the international Supernova Cosmology
Project, based at Berkeley Lab, discovered the accelerating
expansion of the universe; confirmed by other researchers,
their finding was announced in 1998. Accelerating expansion
implies the existence of a “cosmological constant” or other
form of so-called dark energy, now known to constitute some
75 percent of the density of the universe.

Yet while Type Ia supernovae are by far the best standard
candles for measuring cosmological distances, and hence for
investigating dark energy, a small measurement uncertainty

“The asymmetry we have measured in SN 2001el is large enough
to account for a large part of this intrinsic uncertainty,”
says Wang. “If all Type Ia supernovae are like this, it
would account for a lot of the dispersion in brightness
measurements. They may be even more uniform than we

The spectropolarimetry program led by Wang has previously
established that other types of supernovae show considerably
higher degrees of polarization, and therefore asphericity,
than does Type Ia. Not only is the asphericity of Type Ia
supernovae small by comparison, its effect on brightness
measurements may be readily correctable.

Asymmetry and how to measure it:

Wang and his colleagues use the example of a carton of eggs
to explain how asymmetry can affect brightness measurements.
All the eggs in the carton are similar, but the egg shape is
only apparent when they are viewed from the side; viewed
end-on, an egg looks round.

Likewise, if supernovae are not spherically symmetric, they
will shine more brightly in one direction than in others.
Even with a telescope as powerful as the VLT, however,
distant supernovae appear only as point-sources of light, so
asymmetric shapes cannot be seen directly. Instead they must
be inferred from the way the light is polarized.

Polarization refers to the orientation of the plane of the
electric wave component of light and other electromagnetic
radiation. Polaroid sunglasses, for example, “measure”
polarization by blocking or absorbing much of the light
polarized by reflection from horizontal surfaces.

In the light from a spherically symmetric star, however, all
orientations are equally represented, and there is no net
polarization. Not so for an asymmetric star or explosion.
Light emitted along the longer axis shows a net excess of a
particular polarization.

“The differences are very small,” says Dietrich Baade, a
scientist with the European Southern Observatory and a
member of the team that did the spectropolarimetry.
“Measuring them requires an instrument that is very
sensitive and very stable.”

To study SN 2001el, the team used the FORS1
spectropolarimetry instrument in conjunction with the VLT,
the world’s largest optical telescope array. They analyzed
the polarization of various parts of the supernova’s
spectrum, beginning immediately after its discovery in
September 2001. They made repeated measurements as the
supernova grew brighter, reached maximum brightness, and
then slowly faded.

Following the light curve:

“Distance measurements of Type Ia supernovae have typically
been calculated at maximum brightness,” says Wang. “Our
observations of SN 2001el show that asymmetry persists up to
and beyond maximum brightness.”

As spherical symmetry begins to dominate, about a week after
maximum, “it’s not because the supernova is changing shape,
but because we are seeing different layers of it,” says
Wang. Outer layers expanding at thousands of kilometers a
second grow diffuse and become transparent, allowing the
inner layers to become visible. “When it explodes, the outer
part is aspherical, but as we see lower down, the dense
inner core is spherical.”

The spectropolarimetry of SN 2001el suggests that
information from the fading part of Type Ia supernovae light
curves can be used to reduce uncertainties in the relation
between their distance and brightness.

Polarimetry also has much to say about a Type Ia supernova’s
progenitors and the way it burns when it explodes. The
observations of SN 2001el provide evidence for the model in
which a white dwarf star accretes material from an orbiting
companion until it reaches the Chandrasekhar limit, a
critical mass about 1.4 times the mass of our sun.

To fully understand the resulting thermonuclear explosion it
may be essential, as the spectropolarimetry of SN 2001el
shows, to incorporate polarimetric data in three-dimensional
models of the process. Typical one and two-dimensional
computer modeling is inadequate.

In addition to polarimetry, analyzing the spectrum of a Type
Ia supernova during the rising part of its light curve can
reveal specific information about its elemental composition.
“This shows that we need to follow the rise and fall of the
light curve completely,” Wang says.

In the future, very high-precision projects like the
SuperNova/Acceleration Probe (SNAP) satellite, now under
development, will insure that Type Ia supernovae are
compared like to like, thus averaging out any asymmetries.
SNAP is designed to find thousands of new supernovae and
perform the sort of detailed studies of their light curves
and spectra represented by the VLT’s FORS1
spectropolarimetry of SN 2001el. What stands as an
extraordinary achievement for a ground-based telescope will
be standard operating procedure for SNAP.

Berkeley Lab is a U.S. Department of Energy national
laboratory located in Berkeley, California. It conducts
unclassified scientific research and is managed by the
University of California. Visit our website at

Additional information:

“Spectropolarimetry of SN 2001el in NGC 1448: asphericity of
a normal Type Ia supernova,” by Lifan Wang, Dietrich Baade,
Peter Hoeflich, Alexei Khokhlov, J. Craig Wheeler, D. Kasen,
Peter E. Nugent, Saul Perlmutter, Claes Fransson, and Peter
Lundqvist, appeared in the Astrophysical Journal, vol 591, p
1110 (10 July, 2003) and is available online at

Read the European Southern Observatory press release on this

More about the Supernova Cosmology Project

More about the European Southern Observatory

More about the Supernova Research Group at the University of
Texas [].

More about the SNAP satellite [].