PhD Projects

Determining the origins of gas giant exoplanets that exist extremely close to their stars—at only ~1/10 the Mercury–Sun separation—remains a major outstanding question in exoplanet science, even 30 years after the first such planet was discovered. The atmospheres of these “hot Jupiters” hold the clues: their chemical compositions encode where they formed in their natal discs, how they migrated inward, and how they have been shaped by the extreme irradiation they receive. The James Webb Space Telescope (JWST) is transforming our understanding of exoplanets with exquisitely precise infrared spectra, revealing molecules such as water, carbon dioxide and sulphur dioxide in remarkable detail. However, infrared data alone can be misleading: clouds, hazes, temperature structure and even the host star’s own spectrum can all bias the atmospheric abundances inferred from infrared features. Optical (visible) wavelengths provide the information needed to disentangle these effects, but JWST is insensitive to them. Ground-based telescopes, which do have sensitivity at optical wavelengths, therefore play a vital role in completing the spectral coverage. In this project, you will combine new and archival ground-based optical spectra with consistent reanalyses of JWST near-IR data to build the most complete view yet of hot Jupiter atmospheres. This will enable accurate determinations of their metallicities and elemental abundance ratios, allowing robust tests of theories that link atmospheric composition to planet formation and to the effects of stellar irradiation. In doing so, the project will contribute directly to understanding the origin, migration history, and atmospheric evolution of hot Jupiters. Opportunities to participate in ground-based observing may also arise.

Gas giants are the most spectacular outcomes of planet formation, yet their origins are poorly understood. They are massive enough to impact the properties of their parent formation environment, changing its composition and preventing some material from accreting onto their host stars. The GAIA mission will shortly unveil the first significant population of distant gas giant exoplanets, similar to Jupiter and Saturn in our solar system, where signatures of the chemical changes forming planets induce will be encoded in their host star's composition. This project will involve modelling gas giant planet formation to interpret the GAIA exoplanet population. 

We are now entering an era where studying exoplanet atmospheres in detail is technically feasible. Processes such as their climate, dynamics and escape into space are all possible to study for the first time. In this project, you will support observational programs to study exoplanet atmospheres by developing models to match the observations. In particular, you will either support the STELa program on the Hubble Space Telescope, the largest-ever single award of telescope time to study exoplanets, or the BOWIE-Friends program a JWST project to tie atmospheric composition to planet formation. 

Stellar magnetic activity introduces uncertainties and biases in the determination of exoplanet parameters. Being able to disentangle the stellar from the planetary signal is thus critical if we want to study exoplanets. Determining the level of activity is not easy, however, as this has to be inferred from indirect tracers, such as spectroscopic and photometric variability. The aim of this project is to model spectroscopic and photometric variability starting from existing  (magneto-convection) models of stellar surfaces that include the effects of magnetic fields representing a range of activity levels, and to test these models against observations of planet-hosting stars. 

The Simons Observatory (SO) is the latest and best cosmic microwave background experiment so far, and is currently operating in Chile. In addition to producing new results on the cosmic microwave background, it will also produce a millimetre wave survey over >40% of the sky reaching sensitivity levels never before seen. This represents a huge new resource for mm astronomy and will find, among other types of source, high redshift dusty star forming galaxies (DSFGs), lensed DSFGs, groups and clusters of DSFGs, as well as cool and polarized dust in local galaxies. Perhaps the most interesting aspect of this survey is its ability to detect transient mm sources, including gamma ray bursts, tidal disruption events, supernovae, accreting protostars and flaring stars, as well as other poorly understood phenomena. This project will involve working with the SO sources and transients data as it arrives, conducting followup observations of both sources and transients, and comparing these results to relevant models so to understand the various phenomena better, from nearby flaring stars, to accreting AGN, to dusty galaxies across the universe. This work will be undertaken as part of the SO and SO:UK projects.

The role of dusty star forming galaxies (DSFGs) in the overall history of galaxy formation is currently unclear. They may have significant roles in the overall star formation history of the universe and might also play roles in the formation of galaxies in clusters, protoclusters and as precursors to quasars. This project will make use of the deepest far-IR image ever obtained - the Herschel Dark Field - alongside complementary observations and data from other sources potentially including the Simons Observatory, to study the role and evolution of DSFGs over cosmic time. This will include using colour selection to find high redshift DSFGs, conducting followup observations, and comparing these results to theoretical models. We are also in the process of preparing for the next generation of far-IR space missions, including the proposed NASA PRIMA and ESA LETO missions. This project may thus include contributions to the science cases and planning for these missions.

• Theoretical models and computer simulations of the planet-forming environment - Timmy Delage & James Owen
• Identifying the nature of dark matter using the cosmic web - Dr Keir Rogers
• Theoretical Topics in Exoplanet Atmosphere - Dr James Owen
• Finding the most distant quasars with Euclid - Professor Daniel Mortlock
• Disentangling stellar and planetary signals in exoplanet transits - Dr Yvonne Unruh
• Accurate cosmology with the Rubin LSST - Dr Boris Leistedt
• Planet formation in the inner regions of protoplanetary discs - Subhanjoy Mohanty
• Confronting the theory of exoplanet evolution with observations - James Kirk
• Dusty star-forming galaxies near and far - Dr Dave Clements
• Exoplanets origins and evolution - Dr James Owen
• Accretion discs around polluted white dwarfs - Dr Chris Manser and Dr James Owen
• Atmospheres of Habitable Zone Exoplanets around M dwarfs - Dr Subhanjoy Mohanty
• Epoch of Reionization with REACH and SKA - Dr Jonathan Pritchard 
• Cosmology with the CMB - Prof Andrew Jaffe and Prof Alan Heavens
• The most luminous galaxies in the local Universe - Dr Dave Clements
• Astrophysics and cosmology from the 21cm line - Dr Jonathan Pritchard
• Cosmology with the next generation of CMB experiments - Prof Andrew Jaffe
• Planet formation and habitability - Dr Subu Mohanty
• The first quasars and supermassive Black Holes - Dr Daniel Mortlock
• Bayesian Analysis of the dynamic Universe - Dr Florent Leclercq and Prof Alan Heavens
• Bayesian analysis of weak gravitational lensing - Prof Alan Heavens and Prof Andrew Jaffe
• Searching for the most distant quasars - Dr Daniel Mortlock
• Higgs, Dark Matter and the Global Search for Physics beyond the Standard Model - Dr Pat Scott
• Direct Detection of Dark Matter and Global Fits - Prof Roberto Trotta
• Cosmology and Fundamental Physics with Euclid - Prof Roberto Trotta
• Extreme Dusty Star-Forming Galaxies - Dr Dave Clements
• The Nature and Evolution of 70 micron selected galaxies - Dr Dave Clements
• The X-ray-Starburst Connection in the Herschel Era - Dr Dave Clements
• Advanced statistical methods for astrophysical probes of dark energy - Prof Roberto Trotta
• The early Universe and cosmological parameters from the Cosmic Microwave Background, Gravitational Waves, and other observations - Professor Andrew Jaffe
• Determining the topology of the Universe from the Cosmic Microwave Background - Professor Andrew Jaffe
• Accretion Disks, Planet Formation and Habitability Around Red and Brown Dwarfs - Dr Subu Mohanty
• Towards optimal statistics of reionization a5 Emulating radiation from variable stars - Dr Yvonne Unruh nd the 21 cm signal - Dr Jonathan Pritchard
• Cool pre-main sequence stars: their surfaces and circumstellar environments - Dr Yvonne Unruh
• Understanding solar brightness changes on climate-relevant time scales - Dr Yvonne Unruh
• Gravitational lensing, dark matter, and black holes - Professor Steve Warren
•  The most luminous galaxies in the local Universe - Dr Dave Clements
• Pushing the limits of high-z LSS structure cosmology - Dr Boris Leistedt
•  Emulating radiation from variable stars - Dr Yvonne Unruh
• Cosmology with likelihood-free inference - Prof Alan Heavens
•  The highest redshift quasars - Professor Stephen Warren
• Planet Formation in the Inner Disc-The First End to End Model - Dr Subu Mohanty
• Stellar Brightness Variability and Exoplanets – Dr Yvonne Unruh
• The Earliest Stages of Planet Formation - Dr Richard Booth
• Molecules in the Atmosphere of Venus - Dr David Clements and Dr Ingo Mueller

This PhD project will study the formation of planets in the inner regions of protoplanetary disks. Specifically, it will focus on modelling the growth of planetesimals and planetary embryos into full-fledged planets in the inner disk, via both hydrodynamical simulations of dust+gas evolution, and N-body simulations of rocky bodies supplemented with the effects of interactions with gas. The successful applicant will have a strong background in undergraduate physics, including a solid understanding of undergraduate-level hydrodynamics, and good programming skills (e.g., in python). 

Gravitational wave astronomy is entering a transformative era where statistical population studies become possible. Rubin Observatory will provide unprecedented constraints on the astrophysical origins of GW sources through galaxy demographics—stellar mass functions, star formation histories, metallicity distributions, and AGN populations. The timing is critical: Rubin begins operations as pulsar timing arrays probe the nanohertz gravitational wave background from supermassive black hole binaries, while LIGO/Virgo accumulate hundreds of compact binary merger events. This convergence enables joint inference across multiple GW frequency regimes that was previously impossible.

Imperial's active involvement in Rubin LSST DESC, combined with the group's expertise in complex statistical modeling from astrophysical parameters through to pixels, images, and catalogs, positions this project to make transformative contributions. The work spans inferring massive black hole binary populations from AGN demographics, using galaxy evolution data to build priors on cosmic merger rates and test binary evolution physics, predicting gravitational wave background anisotropy from large-scale structure, and hierarchically constraining delay time distributions by combining host galaxy observations with both resolved events and the stochastic background. This requires carefully forward modeling selection effects, detection thresholds, and observational biases at every stage.

Examples of Project Scope or Starting Points

• Use LSST stellar mass functions, star formation rates, and metallicity distributions to build astrophysically-informed priors on compact binary merger histories and test binary evolution models (e.g., common envelope efficiency)

• Constrain massive black hole binary populations using LSST AGN variability and demographics, connecting to pulsar timing array measurements of the nanohertz stochastic background

• Predict and measure anisotropies in the gravitational wave background using LSST large-scale structure data and cross-correlations

• Hierarchically constrain delay time distributions by combining host galaxy properties from Rubin with resolved LIGO events and stochastic background measurements

• Develop Bayesian frameworks for host galaxy identification within GW localization regions and optimize kilonova discovery strategies

Essential Background or Rapid Learning Required

To succeed in these projects, you should either have a solid background in the following areas or be ready to develop these skills quickly:

• Physics: Strong foundation in physical principles with preference for experience in observational astronomy, cosmology, or related fields

• Statistics and machine learning: Familiarity with probability theory, statistical inference, and computational methods

• Programming: Proficiency in efficient vectorized Python and clear data visualization

• Independence and creativity: Ability to work independently, creatively solve problems, and generate new ideas when faced with research challenges

• Oral and written communication: Ability to clearly present technical work, write scientific documents, and engage in collaborative discussions

Particularly Desirable Technical Skills

While not required, the following specific technical skills would be particularly valuable for getting started quickly on these projects. If you have some of these skills you should leverage this as a strength serving your application.

Observational Data Analysis

• Multi-band photometric analysis

• Photometric redshift techniques

• Spectroscopic data reduction and analysis

• Time-domain astronomy and light curve analysis

Large Survey Pipelines

• Image processing pipelines for wide-field surveys

• Source injection and image simulation techniques

• Galaxy image simulation

Statistical Methods & Software

• Bayesian inference frameworks

• Hierarchical modeling

• Large-scale structure statistics and correlation functions

• CMB analysis tools and spherical harmonic methods

• Hierarchical modeling

• Large-scale structure statistics and correlation functions

• CMB analysis tools and spherical harmonic methods

Machine Learning

• Deep learning frameworks

• Simulation-based inference techniques

• Neural density estimation methods

Astrophysical Modeling

• Stellar population synthesis

• Binary population synthesis

• Galaxy formation models

• IGM/CGM modeling and Lyman-alpha radiative transfer

Gravitational Waves

• Gravitational wave data analysis

• GW parameter estimation and signal processing

• Pulsar timing analysis

Lyman Break Galaxies at redshifts 3 < z < 7 trace a critical epoch in cosmic history, spanning galaxy assembly, dust enrichment, and the tail end of reionization. Yet their full potential as both astrophysical laboratories and cosmological probes remains underexploited. The convergence of Rubin's deep multi-band photometry with DESI spectroscopy (and preparing for DESI-II) creates an unprecedented opportunity to simultaneously constrain LBG physical properties, understand galaxy evolution processes, and extract precision cosmological measurements—while carefully accounting for the complex systematics that connect these scales.

Imperial's involvement in many LSST DESC projects on clustering and cross-correlations, including work on LBGs, positions the group to lead comprehensive LBG studies. The project spans three interconnected themes: (1) measuring intrinsic physical properties—stellar masses, ages, star formation histories, dust content—from spectroscopic and photometric modeling; (2) understanding how galaxy evolution processes (dust attenuation, IGM absorption, Lyman-alpha emission, bursty star formation) shape observed properties and selection functions; and (3) leveraging LBG clustering, weak lensing, and cross-correlations with publicly available CMB data from Simons Observatory to constrain cosmological parameters including neutrino masses, dark energy, and growth of structure. Critical to all three themes is developing robust methods for handling observational systematics—Galactic dust extinction, interloper contamination from LAEs and AGN, photometric biases, and non-uniform survey coverage. Image simulation and source injection techniques will be key tools for characterizing selection functions and enabling forward modeling, with potential applications beyond LBGs.

Examples of Project Scope or Starting Points

• Fit stellar population synthesis models to DESI spectra to measure LBG masses, ages, star formation histories, and dust; compare statistical distributions to predictions from cosmological simulations and galaxy formation models

• Characterize how IGM absorption, Lyman-alpha emission (in/out of line of sight), and bursty star formation affect LBG colors, photometric selection completeness, and redshift distributions using forward modeling

• Quantify interloper contamination from LAEs and AGN using narrow-band surveys (Rubin, HSC, IBIS) combined with spectroscopic validation; develop probabilistic membership techniques

• Design and implement source injection strategies to characterize LBG selection functions across magnitude, color, and redshift space, validating photometric measurements and enabling forward modeling (extensible to other populations)

• Measure LBG auto-correlations and cross-correlations with weak lensing and CMB lensing from ACT or Simons Observatory, extracting constraints on cosmological parameters while marginalizing over galaxy bias and population uncertainties

• Build integrated frameworks connecting intrinsic LBG properties to observed clustering and cross-correlation signals, accounting for dust extinction spatial variations and other wide-area survey systematics

Essential Background or Rapid Learning Required

To succeed in these projects, you should either have a solid background in the following areas or be ready to develop these skills quickly:

• Physics: Strong foundation in physical principles with preference for experience in observational astronomy, cosmology, or related fields

• Statistics and machine learning: Familiarity with probability theory, statistical inference, and computational methods

• Programming: Proficiency in efficient vectorized Python and clear data visualization

• Independence and creativity: Ability to work independently, creatively solve problems, and generate new ideas when faced with research challenges

• Oral and written communication: Ability to clearly present technical work, write scientific documents, and engage in collaborative discussions

Particularly Desirable Technical Skills

While not required, the following specific technical skills would be particularly valuable for getting started quickly on these projects. If you have some of these skills you should leverage this as a strength serving your application!

Observational Data Analysis

• Multi-band photometric analysis

• Photometric redshift techniques

• Spectroscopic data reduction and analysis

• Time-domain astronomy and light curve analysis

Large Survey Pipelines

• Image processing pipelines for wide-field surveys

• Source injection and image simulation techniques

• Galaxy image simulation

Statistical Methods & Software

• Bayesian inference frameworks

• Hierarchical modeling

• Large-scale structure statistics and correlation functions

• CMB analysis tools and spherical harmonic methods

• Hierarchical modeling

• Large-scale structure statistics and correlation functions

• CMB analysis tools and spherical harmonic methods

Machine Learning

• Deep learning frameworks

• Simulation-based inference techniques

• Neural density estimation methods

Astrophysical Modeling

• Stellar population synthesis

• Binary population synthesis

• Galaxy formation models

• IGM/CGM modeling and Lyman-alpha radiative transfer

Gravitational Waves

• Gravitational wave data analysis

• GW parameter estimation and signal processing

• Pulsar timing analysis

Simulation-based inference (SBI) has matured significantly in cosmology, enabling powerful parameter estimation for well-controlled settings. However, a critical gap remains: handling the complex observational realities of wide-field photometric surveys—non-uniform depth, spatially varying selection effects, systematic biases like blending and stellar contamination, and astrophysical foregrounds. Traditional approaches rely on simplified parametric models of these effects, which can introduce biases when the true systematics are more complex. This project aims to fill this gap by developing SBI frameworks that operate at the interface between complex cosmological and astrophysical simulations on one side, and realistic data models based on simulated images on the other.

The group's expertise in forward modeling from astrophysical parameters through to pixels and catalogs positions this work at a unique methodological frontier. The key innovation is building SBI pipelines that can forward-model the full chain: starting from cosmological simulations and galaxy formation physics, generating realistic mock images that capture survey-specific systematics (variable depth, blending, masking, photometric scatter), processing these through the same pipelines as real data, and performing inference on summary statistics or compressed representations. This enables marginalizing over complex systematics without requiring tractable analytical likelihoods. The work spans methodology development (implementing neural density estimation, sequential inference, and optimal summary statistics), validation through controlled simulations, and application to extracting unbiased cosmological and galaxy evolution constraints from real survey data. There is natural synergy with Project 2's image simulation components.

Examples of Project Scope or Starting Points

• Build end-to-end forward models connecting cosmological simulations and galaxy formation models to realistic survey images, including blending, stellar contamination, variable depth, and masking

• Develop SBI pipelines using neural density estimators (normalizing flows, mixture density networks) with summary statistics optimized for robustness to observational systematics

• Validate the approach on mock catalogs where ground truth is known, demonstrating unbiased parameter recovery in the presence of complex systematics that would bias traditional methods

• Explore the information content trade-off between summary statistics and direct image-level inference for handling systematic effects

Skills You Will Develop

By the end of your PhD, you will have developed the classic skills that physics PhD students acquire, plus expertise in the following areas specific to Imperial's astrophysics program and this research group:

• End-to-end expertise: From high-level theoretical parameters, models, and science questions, down to low-level practical problems in reducing complex data with advanced algorithms

• Advanced statistical inference and machine learning applied to physics: Bayesian hierarchical modeling, simulation-based inference, neural density estimation, and modern computational statistics for astrophysical problems

• Working in both small groups and large collaborations: Experience collaborating within focused research teams while also contributing to major international survey collaborations (LSST DESC, DESI, Simons Observatory)

• Communication and writing at the highest standards: Publishing in top journals, presenting at international conferences, and communicating complex technical work to diverse audiences

Essential Background or Rapid Learning Required

To succeed in these projects, you should either have a solid background in the following areas or be ready to develop these skills quickly:

• Physics: Strong foundation in physical principles with preference for experience in observational astronomy, cosmology, or related fields

• Statistics and machine learning: Familiarity with probability theory, statistical inference, and computational methods

• Programming: Proficiency in efficient vectorized Python and clear data visualization

• Independence and creativity: Ability to work independently, creatively solve problems, and generate new ideas when faced with research challenges

• Oral and written communication: Ability to clearly present technical work, write scientific documents, and engage in collaborative discussions

Particularly Desirable Technical Skills

While not required, the following specific technical skills would be particularly valuable for getting started quickly on these projects. If you have some of these skills you should leverage this as a strength serving your application!

Observational Data Analysis

• Multi-band photometric analysis

• Photometric redshift techniques

• Spectroscopic data reduction and analysis

• Time-domain astronomy and light curve analysis

Large Survey Pipelines

• Image processing pipelines for wide-field surveys

• Source injection and image simulation techniques

• Galaxy image simulation

Statistical Methods & Software

• Bayesian inference frameworks

• Hierarchical modeling

• Large-scale structure statistics and correlation functions

• CMB analysis tools and spherical harmonic methods

Machine Learning

• Deep learning frameworks

• Simulation-based inference techniques

• Neural density estimation methods

Astrophysical Modeling

• Stellar population synthesis

• Binary population synthesis

• Galaxy formation models

• IGM/CGM modeling and Lyman-alpha radiative transfer

Gravitational Waves

• Gravitational wave data analysis

• GW parameter estimation and signal processing

• Pulsar timing analysis