This week we have a post by Paul Kuin (Swift UVOT team at MSSL), taking us through the breathtaking roller-coaster ride of discovery as the light from a supernova in the nearby M82 galaxy was observed.
You probably heard by now about the new supernova which was discovered in the M82 galaxy. I thought it would be nice to write about our involvement here at MSSL, and give some background story.
The first I heard about the supernova was through an Astronomical Telegram (ATEL for short) sent Wednesday the 22nd of January by Y. Cao of Caltech and collaborators who took a spectrum and identified this as a Type Ia supernova at 14 days before the peak brightness. I thought that was interesting news. We don’t see many type Ia supernovae in galaxies so nearby. The last one was SN 2011fe, which was seen in M101, the pinwheel galaxy. The supernova was discovered by our colleague Steve Fossey and his students here at UCL. I checked the Swift TOO list (a TOO is a Target Of Opportunity, requested for interesting, unexpected astrophysical events) and saw that Eran Ofek requested a Swift observation, which already had been approved.
Mark Cropper (MSSL) sent an email around alerting everyone in the lab of the discovery, and I responded that Swift was already on it, observing. Ignacio Ferreras then (MSSL) asked for more information. His student Susan Hutton (MSSL) had made a study of M82 using very deep Swift UVOT images that had just been accepted for publication! One of their data consists of sums of images in the UVOT ultraviolet filters, revealing fainter features than before.
I contacted Mike Siegel, head of the Swift UVOT team at Pennsylvania State University, who told me Peter Brown of Texas A&M was taking the lead for our team. Peter was in the process of submitting more TOOs for Swift observations in the six UVOT filters, and with the Swift ultraviolet grism (to take a spectrum of the supernova). We decided on some details via email.
In the meantime the first data had come down from Ofek’s TOO. Swift data are public and available on a quick-look site. The supernova was bright in the optical V, B, and U bands. It was also visible, though fainter, in the UVW1 band which is a filter bluewards of the U band, centered at a wavelength around 252nm, in the near-ultraviolet region (invisible to the eye!). However, the UVM2 and UVW1 bands which are at even shorter wavelengths were not yet available.
Typically, during the early stages of the explosion, the supernova has a very hot expanding photosphere. The expansion increases the brightness while the temperature goes down, but the initial temperatures are very high, putting out much of the emission in the ultraviolet. Therefore the observed faintness of the supernova in M82 in the UVW1 filter is not typical. The most likely reason for such faint UV emission is the large amount of dust that photons have to traverse to leave the dusty galaxy (which is orientated edge-on towards us).
Ignacio had in the meantime plotted the position of the new supernova on his deep image, and on a Hubble ACS optical image with better spatial resolution. I had wondered if there would be any evidence of a progenitor, but nothing special was spotted (a type II supernova would have originated from a very luminous supergiant star, whereas the progenitors of type Ia explosions are much harder to detect, consisting of a binary system, where at least one of the stars is a white dwarf). A bit later we saw other ATELs come by where others did report their searches (ATELs 5789,5794,5795).
I checked a few times, but the grism observations were not yet in the Swift list of “observations done”. By the evening we got the observation in UVW2, which showed the supernova, and later in UVM2 which did not (see the blue images below; the SN position is indicated in the uvw2 image with lines). The UVW2 filter peaks farthest in the ultraviolet of all the UVOT filters, but also has a small sensitivity bump around 420nm of about 0.4% of the peak response, also known as a “red leak”. However, the UVM2 filter response is limited to a small wavelength band around 225nm only. The first conclusion was that the contrast in brightness between the optical and ultraviolet was so large that the UVW2 detection of the supernova was probably due to the high flux of optical photons seeping through the red leak. Note that the UVW2 image was taken first, before the UVM2 one.
The next morning the grism observation was partially available on the Swift quicklook web site. I downloaded the data, and had a look at the first two images. There was a bright zeroth order in line with the spectrum, but it seemed in the wrong place. After getting the published position plotted on the image I was sure. There sat another very bright star right in the same dispersion plane as the supernova. What was going on!?
The grism disperses light into colours, and so spectra of two nearby sources in a grism image can fall over each other, and that is precisely what happened here ! It can be fixed by changing the roll angle of the spacecraft, so I quickly alerted the on-duty scientist that there was a problem. At the daily Swift planning teleconference there was some futher discussion of that issue. Since there was not much time left for the next plan upload to the spacecraft, we worked hard to resolve which angle to use for the next day. Mike Siegel was eventually asked for help and we decided on a new spacecraft roll angle. However, even the early contaminated UVOT spectrum clearly showed that there was not much emission below 290nm, since that area of the spectrum was partially uncontaminated.
The next day I got an email that NASA was going to put out a press release, and that Neil Gehrels, the Swift Project Scientist, thought that Swift should put out an ATEL with our results so far. We decided to ask Peter Brown at Texas A&M to write that. He had already done a lot of the work for that, including making some images of before and after. He also could say, based on the latest data, that there finally was a detection in UVM2. The supernova had either brightened enough, or we had accumulated enough exposure time to get a detection. He also worked with the NASA press people, and soon there was a lot on Twitter with his ATEL and NASA’s press release taken up by many outlets.
By Sunday the 26th of January, 2014, I had downloaded one of the new grism images and extracted the spectrum (plotted above). It shows the characteristic undulations due to a plethora of spectral lines from metals formed and expelled in the SN explosion (we had a blog entry in 2013 on how spectroscopy allows us to understand the composition of galaxies).
It can be seen that the flux drops off quite steeply to the blue. It will be interesting to see if summing exposures will make the spectrum in the UV visible, and what signatures there are of the absorbing material in the UV. Hopefully we will learn something from this nearby supernova Ia that helps us understand them better.