Ryan Lebron - CIERA REU

Background

Type Ia supernovae are the result of carbon-oxygen white dwarf stars exploding. When this occurs, it is possible to determine the nature of their binary companion. If the companion is a non-degenerate star, the explosion will cause a shock after colliding with the companion that will lead to an ultraviolet flash. To detect this signal, very early observations need to be made. Using this knowledge, it was our goal to understand type Ia supernovae by looking at their ultraviolet spectra. More specifically, we wanted to look at very early UV observations in hopes of seeing this flash, which would clarify the natue of the binary companion.

To help us do this, we looked at two type Ia supernovae. The first supernova was SN2011fe, which was selected because it was well observed, and is known to be very normal. This led us to selecting it as our standard model for a supernova. The second was SN2019yvq, which was selected because it is known to display the early time UV flash. Through NASA's high Energy Astrophysics Science Archive Research Center (HEASARC), we were able to obtain ultraviolet data for the two supernovae. We then worked on building a pipeline to reduce this data and retrieve the UV filters that were used to capture the data, the magnitude of the brightness for each filter, and the various times that the measurements were taken. After this, we plotted the supernova's light curves, which confirmed that SN2011fe was normal, and that SN2019yvq showed the early time UV flash.

Methods

To begin we retrieved the UVOT data for both supernovae from from NASA's HEASARC. After this, we began building our pipeline. This included various custom python routines, as well as the HEADAS environment. After being unpacked, the data was run to obtain each individual UVOT epoch and depending on the signal to noise ratio of the image, was processed by two photometry packages. For our photometry, we used an aperture with a 3 arcsecond radius, and a background region of a 15 arcsecond radius. Because a circular aperture will not receive all of the flux from an image, an aperture correction was implemented to account for the missing flux. Along with that, a coincidence loss correction has been calibrated by the instrument team. This calibration corrects for any photons "lost" if the source is too bright to measure correctly. At this point, we now had lists for the two supernovae's fluxes, times that the images were taken, and the respective filters for the image. With these data sets, we began building our light curves. As previously stated, SN2011fe was very well observed, which led to an abundance of data points, which led to a lot of clutter on its light curve. For the sake of this paper, we decided to reduce that clutter by binning some of the observations. To do so, we decided that any images that were of the same filter, and occurred within .75 days of each other, for our purpose, could be considered as one data point. So, we took the weighted average of the times for these images, along with the weighted average of the fluxes. With the newly averaged times, fluxes, along with the filters, the light curves were ready to be built.

Results and Discussion

Light curves for SN2011fe (top) and SN2019yvq (bottom). Each filter has an offset on the vertical axis as follows: V:-2, B:-1, U:0, uw1:+1, uw2:+2, um2:+3. These offsets were applied for the sake of reducing clutter to insure that the difference in the light curves was as visually obvious as possible. The vertical axis is also inverted due to the nature of brightness magnitude, where something that is brighter will have a more negative magnitude. The horizontal axis is measured in time (days) relative to the time of max B-band brightness. The y-axis is measured in absolute magnitude, which is the brightness of an object as it would appear at a standard distance (10 parsecs). Finally, note that downward pointing, open triangles represent upper limits on flux measurements.

In Figure 1, we display the light curves for SN2011fe and SN2019yvq. As previously stated, SN2011fe was a very standard supernova. With this model light curve, we can now compare it to that of SN2019vyq. Looking at SN2011fe, its curve begins with a gradual rise before reaching maximum and slowly declining. Knowing that this is now the standard behavior we see that SN2019yvq does not display it. From around -17 to -13 days, the plot has a short downward spike. before following the standard behavior. That downward spike corresponds the shock post-supernova, which leads to the UV flash.

Conclusion

As shown, SN2019yvq did display the early time UV flash. But, as discussed by Adam Miller, this particular UV flash is not indicative of a single degenerate system, as there was not strong evidence for a companion. This information is part of the motivation to find these events. Searching for the early UV flashes will continue to tell us about the physics of these systems. With these light curves now built, we have a foundation for studying similar supernovae. For the future we will continue to search for candidates through the Zwicky Transient Facility. If one of their candidates meets our criteria, we can submit observations to NASA's Swift satellite, in order to receive the UV data on it. Then, we will run the data through the same pipeline discussed above and build light curves similar to the ones shown here. From this, we can check for the early time signal which will help us understand the supernova's progenitor system. Since type Ia supernovae are responsible for iron-group element buildup in the universe, understanding these systems is important for understanding the chemical enrichment of the universe itself.

About Me

Hello! I'm Ryan Lebron, a senior phyics major from the University of Scranton in Scranton, Pennsylvania. I was an REU student for CIERA at Northwestern University during the summer of 2021. I hope to continue this work throughout my final year of school, and move on to earning my PHD in Astronomy. Feel free to contact me at ryleb27 [at] gmail.com

This material is based upon work supported by the National Science Foundation under Grant No. AST-1757792