The volumetric rate of superluminous supernovae from the Dark Energy Survey
Student thesis: Doctoral Thesis
Superluminous supernovae (SLSNe) are a recently discovered class of supernovae (SNe) whose peak luminosities reach more than ten times that of Type Ia SNe. The explosion physics and progenitor systems of SLSNe are poorly understood, but population studies can constrain the possibilities. For example, the volumetric rate of SLSNe should follow the cosmic star formation history if their progenitors are massive stars, and can therefore give a strong indication towards their origin. The volumetric rate is also a critical quantity needed to predict the number of SLSNe that will be detected in the future, informing the observing strategy of upcoming transient surveys.
The Dark Energy Survey (DES) Supernova Survey ran for ∼5:5 monthseach year for five years (2013  2018) with a nominal cadence of 7 days,and had the goal of detecting Type Ia SNe for cosmology. The precise,deep photometry of DES also facilitated the discovery of 22 spectroscopicallyconrmed SLSNe in a redshift range 0:22 ≤ z ≤ 1:99 and absolute magnituderange 19:57 ≥ M_{4000} ≥22:82.
In this thesis I study the volumetric rate of SLSNe from DES. I fit multicolour models to the lightcurves of the DES SLSN sample, and simulate >10^{8} of these representative SLSN lightcurves over the cosmological volume and time observed by DES. By comparing simulated lightcurves to archival information of DES observing conditions, I quantify the probability of detecting SLSN lightcurves in DES as a function of redshift and time. I therebycharacterise the photometric completeness of the DES search for SLSNe.
I devise selection criteria that are informed by the observed properties of SLSNe and their host galaxy environments. I impose these selection criteriaon the DES transient catalogue to recover 30 additional SLSN candidates. Together with the spectroscopic SLSNe, these candidates constitute the largest sample of SLSNe that has yet been constructed.
I use a Monte Carlo simulation to compute the volumetric rate of the total SLSN sample, and nd RV = 190 ^{+26}_{23} (sys) ±52 (stat) SLSNe yr^{1} Gpc^{3}; at a volume weighted mean redshift of z̅ = 0:797. I conduct the most detailed error analysis relative to previous SLSN volumetric rate measurements. The volumetric rate measurement of this work is consistent with previous results and with the shape of the normalised cosmic star formation history, in support of the idea that SLSNe originate from massive stars.
Finally, I use an unsupervised machine learning approach to search foranomalous SN lightcurves. I apply a stationary wavelet transform and principal component analysis to a sample of 28,184 Gaussian processed lightcurvesfrom the DES catalogue. I visualise similarities between the resulting abstractfeatures using the tdistributed stochastic neighbour embedding algorithm and thereby identify six anomalous lightcurves. By inspecting the host galaxies of these, I suggest that two are peculiar core collapse SNe, two are active galactic nuclei, and two are SLSN candidates. I discuss lessons learned for future such analyses.
The Dark Energy Survey (DES) Supernova Survey ran for ∼5:5 monthseach year for five years (2013  2018) with a nominal cadence of 7 days,and had the goal of detecting Type Ia SNe for cosmology. The precise,deep photometry of DES also facilitated the discovery of 22 spectroscopicallyconrmed SLSNe in a redshift range 0:22 ≤ z ≤ 1:99 and absolute magnituderange 19:57 ≥ M_{4000} ≥22:82.
In this thesis I study the volumetric rate of SLSNe from DES. I fit multicolour models to the lightcurves of the DES SLSN sample, and simulate >10^{8} of these representative SLSN lightcurves over the cosmological volume and time observed by DES. By comparing simulated lightcurves to archival information of DES observing conditions, I quantify the probability of detecting SLSN lightcurves in DES as a function of redshift and time. I therebycharacterise the photometric completeness of the DES search for SLSNe.
I devise selection criteria that are informed by the observed properties of SLSNe and their host galaxy environments. I impose these selection criteriaon the DES transient catalogue to recover 30 additional SLSN candidates. Together with the spectroscopic SLSNe, these candidates constitute the largest sample of SLSNe that has yet been constructed.
I use a Monte Carlo simulation to compute the volumetric rate of the total SLSN sample, and nd RV = 190 ^{+26}_{23} (sys) ±52 (stat) SLSNe yr^{1} Gpc^{3}; at a volume weighted mean redshift of z̅ = 0:797. I conduct the most detailed error analysis relative to previous SLSN volumetric rate measurements. The volumetric rate measurement of this work is consistent with previous results and with the shape of the normalised cosmic star formation history, in support of the idea that SLSNe originate from massive stars.
Finally, I use an unsupervised machine learning approach to search foranomalous SN lightcurves. I apply a stationary wavelet transform and principal component analysis to a sample of 28,184 Gaussian processed lightcurvesfrom the DES catalogue. I visualise similarities between the resulting abstractfeatures using the tdistributed stochastic neighbour embedding algorithm and thereby identify six anomalous lightcurves. By inspecting the host galaxies of these, I suggest that two are peculiar core collapse SNe, two are active galactic nuclei, and two are SLSN candidates. I discuss lessons learned for future such analyses.
Original language  English 

Awarding Institution  
Supervisors/Advisors 

Award date  Aug 2019 
Documents
8.97 MB, PDF document
ID: 18257540