About Me!

My name is Sara Eberlin, and I am a junior at the University of Wisconsin-Madison studying physics and astronomy. I atteneded the 2021 Northwestern CIERA REU program, where I conducted research with Professor Mel Ulmer searching Hubble deep field clusters for tranisent events, particularly gamma-ray burst afterglows. Previously, I have done research studying AGNs with Dr. Bart Wakker at the University of Wisconsin.

Outside of research, I love teaching sailing in the summer and, when it gets to cold to go sailing, knitting socks for my friends. I can also be found playing the mandolin, cello, accordion, or violin--badly--on my days off. I also very much enjoy enjoy baking (ask me about my scone recipe), painting with watercolors, and thinking about my kitties Snuffles and Obi Wan back home in Upstate New York.

You can contact me at seberlin [at] wisc.edu

Back to top

Summer research background

Transients are short-lived phenomenon, characterized by sudden increase in brightness followed by a slow or sudden decline. Examples of transient events include supernovae, gamma-ray bursts, tidal disruption events, and gravitational micro-lensing event.

Supernovae and tidal disruption events can last anywhere from a few weeks to over a year. Gamma-ray bursts, on the other hand, can appear and disappear within a matter of milliseconds and can't be seen unless their beam is pointed almost directly at Earth, making them particularly hard to study. Thus, GRBs are more often caught via an 'afterglow', produced by excited interstellar medium surrounding the burst area. A GRB afterglow lasts longer, about a day, and can be observed regardless of emission angle.

Similar to a GRB, there might exist a theoretical class of relativistic burst called a 'dirty fireball' burst. Dirty fireballs are characterized as less relativistic than a GRB, having a higher baryon loading fraction, a wider beam angle, no prompt emission in the gamma ray, and an optical counterpart. Ho et al. [1] analyzed data from the intermediate Palomar Transient Factory (iPTF), a ground based observatory that observes transients in real time, looking for dirty fireballs, but came to no conclusive results. Our search, which is particularly motivated by finding a dirty fireball, has the capability to go deeper and fainter than Ho et al. [1] was able to reach. If dirty fireballs are proven to exist, they could be a more accessible avenue than GRBs to study the processes that produce relativistic bursts, and thus the processes that drive the evolution of the universe.

Within a singular galaxy, a GRB might occur every 100 years and the chances of observing one is relatively low. However, within a large galaxy cluster containing several hundred galaxies, the probability of catching GRBs, GRB afterglows, or dirty fireballs becomes much higher.

The Hubble Frontier Fields mission looked closely at several galaxy clusters with strong gravitational lensing, providing opportunities for observing transients in high redshift lensed galaxies along with foreground galaxies.

We searched data from Frontier Files clusters Abell 2744, Abell 370, and MACS-J0416 for undocumented or unidentified transient events, including phenomenon resembling a dirty fireball. We developed an image subtraction process to investigate a section of Abell 2744, and found 35 variable objects, only 6 of which were documented in the Hubble catalog of Variables. Within archival data of MACS-J0416, we also found an unusual transient that has a similar light curve to a GRB, but appears and disappears within 5 days, as opposed to a couple minutes.

Back to top

Methods

If a transient event were to be caught in an exposure, it might not show up against the light of its host galaxy. We used image subtraction to reveal any objects that varied in brightness between two consecutive exposures and factor out light from unchanging objects, such as an event's host galaxy. The result is a difference image that contains whatever transient events or variability were caught between frames.

Our image subtraction process first uses SWarp, a software program, to align two exposures. Then, the exposures' pixel values are directly subtracted from one another, canceling any pixels that did not change in brightness. Within the resulting image, or the difference image, leftover point sources over a certain threshold, 10σ from the object's original brightness, are flagged by the program SExtractor https://sextractor.readthedocs.io/en/latest/index.html.

Our image subtraction process's first tests looked at a section of the cluster Abell 2744. Figure 1 shows the resulting difference image from a successful test. Red circles show sources that changed by more than 10 simga between exposures. The clusters are distant enough, at z=0.3 for Abell 2744, z=0.42 for MACS-J0416, and z=0.8 for Abell 370, that they fit neatly within Hubble's field of view and the entire cluster can be contained within a singular image. Once we're confident enough in our process, we might be able to apply our pipeline to an entire cluster at once.

We are currently investigating our image subtraction sources and developing light curves for those that were not documented in the Hubble Catalog of Variables. From there, we can look for any transients, specifically ones that match the characteristics of a GRB, GRB afterglow, or dirty fireball. While combing through archival data from the Hubble Frontier Fields, we also aimed to investigate any objects labeled as 'unknown', as opposed to 'star' or 'galaxy', by the SIMBAD or NED databases. This might uncover events that were documented but never identified, or studied but who's nature is inconclusive. From these, we looked for objects that match any transient of interest.

Back to top

Results

Our initial image subtraction test, shown in figure 1, found 35 objects, six of which were in the Hubble Catalog of Variables. The Hubble Catalog of Variables The Hubble Catalog of Variables (HCV) analyzed archival data from a variety of Hubble missions, including the Frontier Fields, looking for variable objects and events over a threshold of 5 standard deviations.

The HCV contains objects like supernova, AGN, and variable stars, but not any GRBs or GRB afterglows. The HCV also did not investigate objects over a magnitude of 27, while we have the opportunity to go beyond 27 magnitude. Throughout the development of our image subtraction process, we used the HCV a good comparison for how well the process could find variability.

While looking into objects labeled as 'unknown' in HCV and Frontier Fields archival data, we found an interesting unknown object in the cluster MACS-J0416 named HFF14Spo. HFF14Spo exhibited a sudden increase in brightness before completely fading within a day.

Figure 2 shows HFF14Spo seen a few days before peak brightness, and figure 3 shows HFF14Spo at peak brightness.

The paper Rodney et al. [2] explored this object, along with a transient event seen about six months later in the same lensed galaxy. Figure 4 shows light curves developed by Rodney et al. for the two events. On the left is a sharp peak and fall of a transient event emitting in the optical, and on the right shows the second event about six months later, seen in the infrared.

HFF14Spo displayed a luminosity of L = 10^41, and faded after peak brightness within a day. Its host galaxy was lensed from a location behind the main cluster, and sits at a redshift of z=1.0054 as determined by Rodney et al.

Back to top

Discussion

We hope our search can complement the Hubble Catalog of Variables (HCV) by using our process of image subtraction, considering fainter objects, and specifically looking for transient events like GRB afterglows and dirty fireballs.

Figure 5 shows where our image subtracted sources are located in the sky compared to HCV sources. The red points are sources found by image subtraction, and the blue points are variables objects found by the HCV team. Objects outlined in pink are shared by both.

Our process found around 30 objects that were not documented in the HCV, yet missed about 20 objects in that region of the sky documented as variable by the HCV.

Figure 6 shows a histogram with the distribution of magnitudes between the image subtracted sources and HCV sources.

Our sources trended to the brighter end, excluding the source at 27 magnitudes on the left which we suspect is from an artifact. However, the magnitudes we found for known objects differed significantly from the magnitudes documented for those objects in the Hubble Frontier Fields and HCV catalogs.

It's likely that our method for determining magnitudes is flawed. Since our sources are not constrained to be under 27 in magnitude, we hope the image subtraction process can uncover fainter objects than the HCV was able to reach.

Discussing the properties of the unknown object HFF14Spo, we can start by ruling transients that its light curve does not match. HFF14Spo faded too fast to be a supernova or tidal disruption event, and is too luminous to be a nova. It showed no emission in the gamma-ray, instead emitting primarily in the optical, and decayed too slowly to be a gamma-ray burst. The light curve peaked too sharply to be an GRB afterglow.

We suspect HFF14Spo might be a gravitational micro-lensing event, as it's light curve matches the a-chromatic and symmetrical properties of a micro-lensing event. Figure 7 shows a light curve of a micro-lensing event observed by GAIA, and Figure 8 on the follwoing page shows a closer look at HFF14Spo's light curve for comparison.

However, it's very unlikely that the geometry needed for a micro-lensing event were met within the cluster. The lensed object and compact object doing the lensing would need to be an equal distance apart as the compact object and the viewer. HFF14Spo's host galaxy's redshift and MACS-J0416's redshift do not correlate in this way, and we determined that HFF14Spo is most likely not a micro-lensing event, but we are still exploring this possibility.

We also suspect HFF14Spo could also match the characteristics of a dirty fireball event. The light curve has a similar shape to that of a gamma-ray burst as well. Figure 10 shows a sample of a dozen different GRB light curves given by NASA. The top left light curve closely matches the shape of HFF14Spo.

However, HFF14Spo emitted over 5 days while the matching GRB light curve only lasted a couple seconds. GRBs commonly display small bumps before or after the burst, which could also explain the small bump at before HFF14Spo's main peak. A dirty fireball lacks emission in the gamma-ray, instead emitting mainly in a lower energy like the optical, which HFF14Spo does.

A dirty fireball is theorized to have a higher baryon loading fraction, meaning it is less relativistic than a GRB. This could explain HFF14Spo's slower rise and decay time.

Rodney et al. concluded that the two transient events observed in the lensed galaxy shown previously originated from the same object, HFF14Spo. However, the events were observed 0.6 arcseconds apart, which corresponds to approximately 5 kpc (determined by Ned wright's cosmology calculator), about half of the way between earth and the center of the Milky Way.

If the events are from two separate objects within the same galaxy, it would make this galaxy exceptional, especially since it has a redshift of z=1.0054. Galaxies that host an unusually high number of transient events are not totally unheard of as well, taking NGC 6946 (NGC 6946 Wikipedia) as an example.

Back to top

Conclusion

From the beginnings of our search it is evident that the potential of large, dense, deep field galaxy clusters is extremely promising. Previous searches for variability within these clusters, like the HCV, may not be wholly comprehensive, and opportunities exist for new discoveries. In the future, we will continue our search within other Frontier Fields galaxy clusters and continue to develop and analyze light curves for the sources we have identified so far. The nature of HFF14Spo, a transient event found by digging into archival data looking for sources categorized as 'unknown', is compelling, and it's properties will continued to be investigated.

Back to top