Research

Research Interest:

My current research interests are using strong gravitational lensing of background sources to study galaxy clusters. Galaxy clusters are the largest gravitationally-bound objects in the universe. They lend themselves as great cosmic laboratories to study many astronomical, astrophysical, and cosmological phenomena. Strong gravitational lensing occurs when light from a background source goes through the massive potential well created by the galaxy clusters. The light is then deflected, leading to a magnification effect of the background object as well as multiple images of the same source appearing in the field of view.

Complete list of Publications can be found in the following ADS link, which includes arXiv (free) PDFs versions of the articles.

Hubble Frontier Fields

Publication: An Evaluation of 10 Lensing Models of the Frontier Fields Cluster MACSJ0416.1-2403

Accepted for publication in The Astrophysical Journal (ApJ) on June 25, 2018.

My first project explored the robustness of 10 gravitational lensing models from the Hubble Frontier Fields Cluster MACSJ0416.1-2403 (Hubble Space Telescope, HST, optical image on the right). The Hubble Frontier Fields (HFF; Lots et al., 2017) program uses 6 massive galaxy clusters selected for their strong lensing evidence with the objective of studying the background universe (z~5-10). Each of the galaxy clusters has been observed with 140 HST in seven different optical and near-infrared bands.

The Space Telescope Science Institute (STScI) contracted 7 independent lens modeling teams to compute lens models for all of the galaxy clusters. In 2015 the third version of the lens models for MACSJ0416.1-2403 were made available to the public and used all of the available data at the time. Later in 2016, using the Multi Unit Spectroscopic Explorer (MUSE) on the Very Large Telescope (VLT), many new redshifts for the multiple images were measured, confirming the previously known ones, and updating a multiply-lensed system.

Optical image of MACSJ0416.1-2403 from the Hubble Space Telescope. Credit to Digitized Sky Survey (STScI/NASA), and Z. Levay (STScI)

Delta_Models_SPS.pdf

We compared the distribution of the rms source plane scatter of the systems that had spectroscopic measurements at the time when the lens models were computed, training set, agains the distribution of the arcs that have newly measured spectroscopic redshifts. The dotted lines represent lines of constant slope of one. The models that perform the best will be found towards the lower left corner with the lowest rms scatter in the training set as well as in the new set. The symbols represent the different lens models and the errors bars identify the first and third quartiles of the distributions.Credit: Remolina González et al., 2018.

We tested the performance of the lens models by using the source plane root-mean-square (rms) distance between the positions of the multiple images in the source plane as a metric. We compared the models using newly available spectroscopic constraints that were not available at the time the lens models were computed.

We found that the lens models overall are robust and have a high predictability power. The rms scatter increased slightly with the newly added images, which is expected since that information was not used when computing the lens models. The redshift for one of the lensed systems was updated and changed from z=2.1851 to z= 3.2355. We found that in cases like the HFFs where there are hundreds of constraints, having one catastrophic redshift failure does not break the model (this is not the case when only using a low number of constraints as explored by Johnson & Sharon, 2016).

We explored the spatial distribution of the rms scatter with respect to the lens plane. We found that there are models with regions of high rms scatter in particular locations of the galaxy cluster, identifying sections where the lens models have a lower performance. Last we investigated the newly available version 4 lens models for MACSJ0416.1-2403 using all of the current available information. We observe a similar low source plane RMS scatter distribution between all the lens models, demonstrating the progress in the quality of data, better characterization of MACSJ0416.1-2403, and understanding of strong lens modeling systematics.

The Einstein Radius as a First Order Estimation of the Galaxy Cluster Core Mass

In the advent of some current and future surveys, where thousands of galaxy clusters will be discovered and from those hundreds will have strong lensing features, we have to use different resources to analyze the large amount of data. In the case of estimating the mass at the cores of galaxy clusters, the best method is using strong gravitational lensing, which can give you an estimate of the total projected mass including both baryonic and dark matter. To first approximation we can use the mass enclosed by the Einstein Radius. Using this mass estimate is very fast, since it assumes spherical symmetry of the system and no lens models are required. We are studying the uncertainties in the measurement of the mass enclosed by the Einstein Radius as well as how this measurement compares to a very basic lens model.

We are using the state of the art simulation called "Outer Rim" which is a large-volume, high-mass-resolution, gravity-only, N-body simulation run using the Hardware/Hybrid Accelerated Cosmology Code (HACC; Habib et al. 2016) carried out at the Blue Gene/Q (BG/Q) system Mira at Argonne National Laboratory (ANL). This simulation is ideal for statistical analysis as has been shown by Child et al. (2018), where they investigate the concentration-Mass relation for both groups and galaxy clusters.

We are interested in comparing the simulations against the observations from galaxy clusters discovered by the South Pole Telescope (SPT) using the Sunyaev-Zel'dovich (SZ) effect. The SZ effect occurs when photons from the Cosmic Microwave Background (CMB) travel through the hot gas in the intracluster medium (ICM), interact with the fast moving electrons, and get a burst of energy which can then be measured in the sub-mm wavelengths when looking at the CMB. The selection function of the galaxy clusters coming from the SZ is almost redshift-independent and mass-limited. We mimic this selection function in to find the galaxy cluster from the simulation above masses of 10^14 Msun and then from those galaxy clusters identify the strong lensing clusters. Ray-trace images of realistic background sources are made (Li et al., 2016) so as to be used for lensing analysis.

Zoom-in image of the cosmic web of the Outer Rim, large-volume, high resolution, N-body simulation. Credit: Habib et al., 2016.

PICS.pdf

Optical image of a South Pole Telescope on the left and a realistic strong lensing galaxy cluster from the Outer Rim simulation on the right. Credit: Li et al., 2016 and L. Bleem.

Using the ray-traced images, we can then measure an Einstein Radius from the center of the image to the multiply lensed images. Then we compute the mass enclosed by the Einstein Radius using the redshift of the multiple images and the galaxy cluster. This is the first order approximation of the projected mass at the core of the galaxy cluster. From the simulations, we obtain projected mass density profiles for all of the galaxy clusters. We then place the exact same circle located at the center of the image and with a radius equal to the Einstein Radius we had previously measured. From this we can compute the true projected mass enclosed by the circle from the simulation. We then compare these to measurements.

In addition, using the surface density profile we can compute the convergence, shear, and magnification of the galaxy cluster. From this information we can then identify the critical curves, which are lines of a theoretical value of infinite magnification. We want to compare this to the critical curve values from a very simple lens model using only one large dark matter scale halo. From the two critical curves (from the simulation and from modeling) we can make an effective Einstein Ring which is a circle of the same area as the area enclosed by the tangential critical curve. We can then compare the masses enclosed by the effective Einstein Radii.

This is current work in progress; stay tuned for more exciting results as we make additional progress.

(Top) Example of the comparison between the mass enclosed by the Einstein Radius M(<θ_E ), to the mass enclosed by the same circle in the projected mass distribution of the simulations, M_sim. M_sim is the sum of the projected mass density inside the circle (<θ_E ) multiplied by the squared pixel scale (dsx).(Right) Preliminary lens model of one of the ray traced images where the dotted and solid lines represent the critical curves from the model and from the simulations respectively. The red are the tangential while the yellow are the radial critical curves. The green circles are the identified constraints and the magenta are model predictions.

Lens_Model.pdf

Past Research Experience:

Fourier Transform Spectrometer (Physics Department at University of Michigan, Ann Arbor, MI, Summer REU 2015): I participated in the summer REU program at University of Michigan under the supervision of Dr. Jeff McMahon. My project for the summer was to design and begin building a Fourier Transform Spectrometer for the bolometric detector characterization system used to test the bolometers used for the Atacama Cosmological Telescope (ACT). I developed a code to be used in a CNC machine to cut elliptical mirrors at an angle with the objective of reducing the imperfections in the mirrors. I also explored a different approach to build polarization filters.
Interpreting Circumstellar Dust Models around Asymptotic Giant Branch Stars (Physics and Astronomy Department at University of Missouri, Columbia, MO, Summer REU 2014): I participated in the summer REU program at University of Missouri under the supervision of Dr. Angela Speck. I worked with another undergraduate to make improvements to an existing code called ABSCAT developed to compute absorption and scattering coefficients usable for any radiative transfer code. ABSCAT allows for a variety of dust shapes as well as dust species which would then be used in DUSTY, a radiative transfer code.
Projectile Coherence Effects & Few-Body Dynamics (Physics Department at Missouri University of Science and Technology, Rolla, MO, 2012-2016): I was part of the Laboratory of Atomic, Molecular and Optical Research (LAMOR) under Dr. Michael Schultz. The group studies ion-atom collisions to better understand the few-body problem, one of the unsolved fundamental problems in physics. Using a kinematically complete ion-collision experiment you can measure all of the information permitted by quantum mechanics and use this to study scattering cross-sections and their effect on the coherence properties of the projectiles.

Report abuse