Projects on offer for 2017 entry will be added here in November and December. The projects themselves may evolve somewhat before October 2017.
Cold Dust Galaxies - Dr Dave Clements
Recent work using data from the Herschel and Planck missions suggests that galaxies whose far-IR emission is dominated by cold dust, ~20K, may be more common than previously suspected, but the range of properties of such objects is difficult to judge since the surveys that have found them have been either too shallow or have not covered sufficient area. This project aims to select galaxies using the IRAS 100 micron all-sky survey to look for colder dust galaxies, identify them in the optical and near-IR, and to determine what role they play in both the nearby universe and in galaxy evolution.
Far-IR Luminous Quasars - Dr Dave Clements
Quasars are powered by accretion onto supermassive black holes and are among the most luminous objects in the universe. The energy emitted from their central engine can have strong effects on their host galaxy, and such feedback effects are thought to have a strong influence on galaxy evolution. This feedback is usually thought of in negative terms, in that AGN emission turns off star formation. However, a number of quasars are now known to have strong far-IR emission, which is usually indicative of an enhanced star formation rate. This project will begin by identifying further far_IR luminous quasars in Herschel and/or IRAS data and will then examine the nature of these objects both statistically and with detailed study of individual sources. The end result will be a better understanding of the range of possible feedback processes that might be driven by an AGN, and their effects on galaxy evolution.
Cosmology from the CMB - Prof Andrew Jaffe
The Cosmic Microwave Background — light from 400,000 years after the Big Bang — remains the gold standard of cosmological data. Satellites like Planck have mapped the CMB sky, allowing precise determination of parameters like the overall matter density, the expansion rate of the Universe, and the distribution of matter on the largest scales. The next generation of CMB experiments like Polarbear (already taking data) and the Simons Observatory (starting about 2020) will have thousands of detectors dedicated to observing the polarization of the CMB. This will enable even more precise determination of cosmological parameters, indirect measurements of the neutrino mass, and a possible first observation of gravitational waves in the early Universe. This project will enable the student to become part of these observational teams, focusing on the analysis and interpretation of CMB data from these experiments. There will be opportunities to contribute to detailed analysis of low-level data, the creation of advanced statistical tools (e.g., Bayesian hierarchical models), and high-level cosmological modeling.
The first quasars and super-massive black holes - Daniel Mortlock
Quasars, which are powered by material falling into super-massive black holes, have been discovered when the Universe was just 5% of its current age (of 13.8 billion years). The first part of this project would be to use the results from surveys of distant quasars to determine their population and, in particular, how it is evolving with cosmic time. The second part of this project concerns the black holes themselves: it is very difficult to explain the formation of black holes of > 10^9 Solar masses in less than a billion years, but another possibility is that they are not actually so massive - their mass estimates come from extrapolating an imperfect empirical correlation established from a different population of quasars which, typically, have much lower mass black holes, and the current approaches to this problem tend to ignore the various sources of uncertainty. In this project we will go back to the original data and establish what can really be said about the first super-massive black holes and quasars from the available data.
Bayesian Hierarchical Modelling of Cosmology - Alan Heavens & Andrew Jaffe
When we analyse cosmological surveys, whether it is microwave background, galaxy clustering, or weak lensing data, to name some examples, we are performing a statistical analysis. Essentially all that we have learned about cosmology, from the Planck satellite and from surveys such as the CFHTLenS gravitational lensing survey, comes from statistics. To do the analysis properly relies on building as complete a statistical model of the data as possible, and getting this right is vital to extract science from future high-profile surveys such as Euclid, LSST and SKA. To do this, we have been pioneering Bayesian Hierarchical Models for weak lensing, and we will extend this, including to other areas of cosmology. These problems are interestingly large, involving very high dimensional parameter spaces (typically 100,000 to a million), with corresponding computational challenges.
Improving the accuracy and precision of supernova type Ia as cosmological probes - Roberto Trotta
Supernovae type Ia (SNIa) are a particular type of stellar explosion that have the important property of being almost standard candles -- ie, their luminosity is almost the same for all objects, and therefore can be used to measure distances in cosmology. SNIa were instrumental in establishing the accelerated expansion of the Universe (currently ascribed to an unknown form of energy, called dark energy) -- a momentous discovery which was rewarded with the Nobel Prize for Physics 2011.
The number of SNIa observations is now in the several hundreds, and it is set to increase by over a factor of 10 in the next few years. Already today, our inferences about the nature of dark energy are being limited by poorly-understood systematic effects (such as the reddening and dimming introduced by dust, which can be confounded for the dimming due to the expansion of the Universe).
This project will look at one of the frontiers of SNIa cosmology, namely the understanding and reduction of systematic effects arising from our imperfect knowledge of the astrophysical origin of SNIas, instrumental effects (eg., selection effects) and incomplete data (e.g, lack of spectroscopic confirmation). Topics that will be investigated include:
1. The influence of the galactic environment in which the SNIa explodes on its observable properties. If SNIa's have different brightness in different environments, this could compromise or limit their usage as cosmological standard candles. Factors that will be considered are host galaxy mass, star formation rate, metallicity, stellar population age, spectral lines width, host morphology, location within the host.
2. The implementation of a principled method for dealing with observational selection effects into the cosmological data analysis pipeline, building on the BAHAMAS approach (Shariff, Trotta et al, 2016).
3. The development of machine learning and other algorithms to deal with contaminated SNIa data sets. In the absence of spectroscopic data for a majority of SNIa from future surveys, the presence of non-Ia's in the sample can potentially bias our inferences about dark energy. This project will build a complete analysis pipeline to solve this problem.
The project is well suited to a student with interests in computational methods, statistics, cosmology, and messy astrophysics.
Astrophysics and Cosmology with the Square Kilometre Array: Dr Jonathan Pritchard
One of the last frontiers of astrophysics is the first billion years of the Universe, when the formation of the first galaxies and black holes produced the first star light leading to a Cosmic Dawn. Observations of the 21 cm line of neutral hydrogen with HERA and SKA promise to open a window onto this period (z=6-27) for the first time (see e.g. arXiv:1109.6012) and are expected to offer new insights into astrophysics and cosmology. This project will focus on implementing semi-numerical simulations of reionization as a specific realisation of a linear perturbation theory. The aim is to produce a fast algorithm that can be used in parameter estimation studies of the 21cm signal. Alongside this, the project will explore the applications of such simulations for intensity mapping studies in different atomic and molecular lines, both during reionization and for cosmology at lower redshifts. The student will be part of a 21cm research group at Imperial funded by ERC Starting Grant “FIRSTDAWN” and closely involved with preparation for SKA.
Simulations of stellar variability - Yvonne Unruh
High-precision photometric measurements (e.g., with Kepler) are showing a wide range of stellar variability. Apart from shedding light on the underlying phenomena, modelling this variability is also necessary when trying to characterise the atmospheres of exoplanets.
The aim of this project is to build a stellar variability `simulator' for solar-type stars (i.e., stars with masses of approximately 0.4 to 1.4 solar masses) where variability is largely driven by magnetic activity. The variability simulator will need to take into account the sizes, locations and intensities of surface features such as dark starspots
and bright plages/faculae. The project would suit somebody with and interest in solar and stellar physics, and enjoys working with a large range of data sources.
Past projects have included: