The Search for Lensed Supernovae


Type Ia supernovae that have multiple images due to gravitational lensing can provide us with a wealth of information — both about the supernovae themselves and about our surrounding universe. But how can we find these rare explosions?

Clues from Multiple Images

When light from a distant object passes by a massive foreground galaxy, the galaxy’s strong gravitational pull can bend the light, distorting our view of the background object. In severe cases, this process can cause multiple images of the distant object to appear in the foreground lensing galaxy.

gravitational lensing

An illustration of gravitational lensing. Light from the distant supernova is bent as it passes through a giant elliptical galaxy in the foreground, causing multiple images of the supernova to appear to be hosted by the elliptical galaxy. [Adapted from image by NASA/ESA/A. Feild (STScI)]

Observations of multiply-imaged Type Ia supernovae (explosions that occur when white dwarfs in binary systems exceed their maximum allowed mass) could answer a number of astronomical questions. Because Type Ia supernovae are standard candles, distant, lensed Type Ia supernovae can be used to extend the Hubble diagram to high redshifts. Furthermore, the lensing time delays from the multiply-imaged explosion can provide high-precision constraints on cosmological parameters.

The catch? So far, we’ve only found one multiply-imaged Type Ia supernova: iPTF16geu, discovered late last year. We’re going to need a lot more of them to develop a useful sample! So how do we identify the mutiply-imaged Type Ias among the many billions of fleeting events discovered in current and future surveys of transients?

Searching for Anomalies

SNe Ia absolute magnitudes

Absolute magnitudes for Type Ia supernovae in elliptical galaxies. None are expected to be above -20 in the B band, so if we calculate a magnitude for a Type Ia supernova that’s larger than this, it’s probably not hosted by the galaxy we think it is! [Goldstein & Nugent 2017]

Two scientists from University of California, Berkeley and Lawrence Berkeley National Laboratory have a plan. In a recent publication, Daniel Goldstein and Peter Nugent propose the following clever procedure to apply to data from transient surveys:

  1. From the data, select only the supernova candidates that appear to be hosted by quiescent elliptical galaxies.
  2. Use the host galaxies’ photometric redshifts to calculate absolute magnitudes for the supernovae in this sample.
  3. Select from this only the supernovae above the maximum absolute magnitude expected for Type Ia supernovae.

Supernovae selected in this way are likely tricking us: their apparent hosts are probably not their hosts at all! Instead, the supernova is likely behind the galaxy, and the galaxy is just lensing its light. Using this strategy therefore allows us to select supernova candidates that are most likely to be distant, gravitationally lensed Type Ia supernovae.

ZTF and LSST finds

Redshift distribution of the multiply-imaged Type Ia supernovae the authors estimate will be detectable by ZTF and LSST in their respective 3- and 10-year survey durations. [Goldstein & Nugent 2017]

A convenient aspect of Goldstein and Nugent’s technique is that we don’t need to be able to resolve the lensed multiple images for discovery. This is useful, because ground-based optical surveys don’t have the resolution to see the separate images — yet they’ll still be useful for discovering multiply-imaged supernovae.

Future Prospects

How useful? Goldstein and Nugent use Monte Carlo simulations to estimate how many multiply-imaged Type Ia supernovae will be discoverable with future survey projects. They find that the Zwicky Transient Facility (ZTF), which will begin operating this year, should be able to find up to 10 using this technique in a 3-year search. The Large Synoptic Survey Telescope (LSST), which should start operating in 2022, will be able to find around 500 multiply-imaged Type Ia supernovae in a 10-year survey.


Daniel A. Goldstein and Peter E. Nugent 2017 ApJL 834 L5. doi:10.3847/2041-8213/834/1/L5