SN1997bs’s True Identity: Impostor or Not?

Title: LOSS’S First Supernova: New Limits on the “Impostor” SN 1997bs
Authors: S. M. Adams & C. S. Kochanek
First Author’s Institution: Ohio State University
Paper Status: Submitted to MNRAS
Bonus: Video of the first author explaining this paper!

As a scientific field driven by observation, astronomy has a knack for grouping objects together that kinda-sorta look alike, regardless of their underlying physics. Take, for example, our supernova classification scheme. While all supernovae mark the explosive death of a star, they have a wide range of properties. Historically, supernovae are broadly grouped into two categories based on their spectra: Type I’s (which lack signs of hydrogen in their spectra) and Type II’s (which have hydrogen lines). As it turns out, these categories are largely unrelated to the actual physical process driving these explosions. Confusing, right? In today’s paper, the supernova community faces another case of bad taxonomy: objects collectively known as “supernova impostors.”

Supernova Impostors: Theory vs Practice

In theory, supernova impostors are extremely bright explosions of stars which look a lot like supernovae. Unlike supernovae, they leave behind most of the progenitor star surrounded by a shell of newly formed dust. The most famous example of such an object is Eta Carinae’s larger star which has undergone multiple explosions throughout the 1700 and 1800s. While the exact mechanisms of these blasts are uncertain, they appear commonplace in massive Luminous Blue Variable stars (LBVs) like Eta Carinae. Figure 1 shows Eta Carinae as seen by Hubble, surrounded by its homunculus nebula.

Eta Carinae in optical

Fig. 1: Eta Carinae as seen by Hubble. Eta Carinae is a supernova impostor that has undergone several eruption events.

In practice,  objects are often labeled “supernova impostors” without knowledge of their true origin. Supernova impostors, at first glance, look like dim type IIn supernovae. Type IIn supernovae are core-collapse supernovae that have narrow hydrogen emission lines in their spectra. The narrowness of these features are odd because supernovae are fast, energetic events that typically produce broad features; the narrowness implies that the supernova is interacting with dust that has veiled the progenitor. So if an astronomer sees a transient that doesn’t look energetic enough to be a true IIn supernova, they might label it as a “supernova impostor”. But not all so-called impostors seem to be explosions from LBVs like Eta Carinae. In fact, today’s paper argues that one of the archetypes of supernova impostors, SN1997bs, may not be an impostor at all!

SN1997bs: An impostor or the real deal?

As the name suggests, SN1997bs was first discovered in April of 1997 and was classified as a type IIn supernova. The supposed-supernova then faded to be even dimmer than its pre-explosion progenitor star, and appeared to hold steady in the early 2000s. Originally, the flattening of the light curve was interpreted as the steady flux of the explosion’s survivor – making SN1997bs a supernova impostor.

Today’s paper relies on recent photometry to answer the basic question: is the star really still there? The authors monitored SN1997bs with Hubble, Spitzer and the Large Binocular Telescope (LBT) until 2014  nearly two decades after the initial transient! They find that the object has continued to dim far below the progenitor’s initial brightness, as seen in Figure 1. In fact, there is no highly significant detection of the star in any of the images dated between 2013 and 2014. If the star is no longer detectable, what does that mean about SN1997bs? The authors discuss a few theories:

Light curve of SN1997bs

Fig. 1: Optical light curve of SN 1997bs. SN1997bs appears to dim past its progenitor’s original brightness (shown in dotted lines). Here the V (squares) and I (triangles) optical bands are shown, including a V-band upper limit from LBT.  All of the points past 1000 days are all upper limits on SN1997bs’s brightness.

  • Could SN1997bs be a “canonical” supernova impostor? Perhaps SN1997bs is an Eta Carinae-like star and now finds itself in a dirty web of dust. The authors say that this is unlikely because the star would be expected to eventually brighten in redder light (or to “redden”). Instead, the star appears to get bluer and then disappear altogether.
  • Forget the dust…what if SN1997bs is a “tuckered-out” impostor? The star may have erupted and then declined into a much darker state 30 times fainter than the original star! Currently, there is no known mechanism that can dramatically reduce the star’s intrinsic luminosity after a single event. If anything, we would expect a star’s outer envelope to expand during the eruption and lead to an even brighter star. Either some new astrophysics is happening or…
  • SN1997bs may be a true type IIn supernova? The authors argue that the simplest solution to a disappearing star is to assume that it is truly gone. Although SN1997bs would be an extremely subluminous (or low energy) supernova, the authors argue that other faint core collapse supernovae are just now being discovered, and the energy ranges for these events may be broader than we originally thought. Additional factors, such as obscuring dust, may contribute to underestimating the energy of this initial supernova.

In order to be sure that SN1997bs was an authentic supernova, we need better observations to confirm that the star has vanished. Although a compact remnant, like a neutron star, might exist after this explosion, it would be far dimmer than the original star. The uncertainty in SN1997bs’s real identity highlights a fundamental flaw in how we label and study impostors: What if SN1997bs, an archetype of supernova impostors, isn’t an impostor after all? As our knowledge in this field grows, we need to be careful about classifying these eclectic “impostors.”

About Ashley Villar

I am a third year PhD student at Harvard University. I'm generally interested in optical transients, or the dramatic aftermaths of stellar eruptions, collisions and explosions. I'm also broadly interested in how astronomers can efficiently use large datasets produced in future missions. When I'm not working, I bake, exercise and try to enjoy Boston.


  1. Hi Ashley, this is a great article. I would love to understand LBVs a little better. It seems that the reason they are often considered supernovae impostors is because they undergo rare giant outbursts, which makes me wonder about their relationship to other variable stars. I don’t know much about Cepheids, but I have done research on RR Lyrae and my impression is that RR Lyrae variable stars not only undergo oscillations of order ~1day, but also much larger oscillations in size and brightness over a longer period. This being said, it seems unsurprising that LBVs would also experience larger-period oscillations leading to rare, large outbursts. Do you (or anyone else reading) know if there is evidence for a similarity in the mechanisms of other variable stars and LBVs like Eta Carinae?

  2. Wow this is so interesting! I had no idea that this was such a common occurrence. You mentioned that we need better observations to confirm that the star has vanished and I was just wondering if these are intended to be further photometric observations or spectra of some kind to truly be sure that we don’t miss detecting any remnant?

  3. How large of a telescope would be needed to determine is neutron stars indeed do exist in situations like this?

  4. How long following a supernova is it possible to distinguish whether or not there is a remnant, and how much observing time would be needed with, say, Hubble to identify the type of remnant?

  5. Hi, Ashley,
    Nice post! You mention narrowness of the lines as odd because SNe are fast and usually have broad lines; is the mechanism Doppler broadening so that the linewidth reflects the velocity?
    Do we have theory expectations for the minimum energy of a core collapse supernova? If so perhaps one could subtract the implied luminosity from that observed here to set an upper bound on the importance of dust.

    • Hi all – there are many comments and great questions that I see I never responded to 🙂

      1. Zoey: Variables stars like Cepheids or RR Lyraes undergo oscillations which are on a much smaller magnitude than LBV outbursts. Additionally, variables stars often undergo at least pseudo periodic variations, whereas LBV bursts seem haphazard. They are all linked by the typical mechanisms of a star, but likely they are caused by different instabilities in the equations governing these stars.

      2. Jonathan: Mostly photometric these remnants are often extremely dim (or non existent 🙂 ), so we would like to gather the most light possible from the source. Spectra take a much longer time to take because it breaks down the light into its various wavelength components. However, spectral followup could also be extremely useful if we *do* know that remnant is there.

      3.Chris: That’s good a question. It depends on many factors such as the intrinsic brightness of the object and its distance. I looked up the exposure times for this particular case (SN1997bs) using Hubble’s MAST website (which is publicly available!), and it looks like a single observation took ~25 minutes. This may not sound like a lot, but HST time is very, very competitive. Of course, this time would be great reduced in next generation space telescopes like the Webb.

      4. Zach: As you might guess, the narrowness of the line is due to dust in the explosions nearby environment. While it’s a good idea to use the SN properties to learn about this dust, in practice, we *don’t* know of a minimum energy for these core collapse SNe. In fact, cases like SN1997bs are showing us that subluminous SNe seem to be possible, and so the energy distribution we observe in these objects seems wider than expected.


Leave a Reply