Oxford Astrophysics will run a summer research programme for undergraduate physics students again in summer 2024. The list of available projects and application forms can be found below. Please ensure you meet the eligibility requirements before applying. 

The selected students will work with a supervisor in the department, usually a postdoctoral researcher or lecturer, on a self-contained research project. The programme will also include lectures/seminars on current astrophysics topics, and on academic careers in astro. Students will be encouraged to take part in department life , joining researchers for coffee, discussions and seminars.

The projects run for typically 8 weeks, nominally from 1 July to 23 August. The duration may be adjusted slightly to be shorter or longer, or to accommodate summer travel. Students will be paid as employees of the University, receiving a payment of £12.06 per hour (subject to tax and National Insurance deductions). Students are normally expected to work full time on their project, but hours can be discussed with your supervisor. We will make sure that all computing resources needed for the selected projects are available to all selected candidates.

There are also a number of astrophysics projects available for the UNIQ+ scheme, which can be found on the UNIQ+ website -- please note that applications for UNIQ+ have different criteria and are managed separately and not by the department. 

Application instructions, and the list of projects offered for 2024 can be found below. 

Eligibility criteria

Students currently in third year of a relevant undergraduate degree are eligible to apply. Students who have completed a 3-year undergraduate degree and are now taking a taught Masters course are also eligible, as long as they are not in their final year. Applications are welcome from institutes outside of Oxford. Unfortunately, due to UK visa regulations, we are only able to accept applications from candidates who do not require a visa to work in the UK. EU students currently in the UK who have been granted Pre-Settled status are also welcome to apply along with current students in the UK on a Tier 4 visa that allows vacation employment. If you have queries about your personal circumstances, please get in touch with ashling.gordon@physics.ox.ac.uk.

How to apply

To apply, submit a CV and a 1-page cover letter (in a single PDF file) while filling out the Oxford Astrophysics Summer Research Programme 2024 application form.

The cover letter should summarise your academic accomplishments to date, your motivation for participating in the programme, and mention which of the projects advertised you might be interested in. You will also be asked to contact one academic referee who can provide a short letter in support of your application. Referees should submit their letter by filling out the Oxford Astrophysics Summer Research Programme 2024 referee form before the same deadline.

Deadline for applications: March 15th 2024.

Available projects

 

Does Scorpius X-1 produce fast jets? 

Supervisors: Dr Lauren Rhodes and Dr Alex Cooper

Very long baseline observations of the black hole X-ray binary MAXI J1820 have demonstrated the existence of both fast (~c) and slow (~0.3c) ejecta. We have radio observations of a neutron star X-ray binary (Scorpius X-1) that show slow ejecta (https://public.nrao.edu/gallery/sco-x1-the-movie/) but not evidence of the corresponding fast jets. We are searching for a student that is interested in working with radio observations (but previous experience is not required) to help us determine whether the Scorpius X-1 system could produce fast jets that we simply do not observe because of relativistic beaming. Depending on the progress of the researcher, there is the possibility to expand this work to other X-ray binary systems. 

Recommend skills/experience: Python programming

 

Testing the galaxy bias coevolution relations with current observations

Supervisor: Dr Carlos Garcia-Garcia

The relation between the distribution of matter and galaxies is biased. Galaxies only account for the visible matter in the Universe, a portion of the total amount of matter. At very large scales, the relation is linear. However, at smaller scales, galaxies are subject to very complicated formation and evolution mechanisms that make the relation between the total matter and galaxies distributions non linear. However, there are hints from simulations that the evolution of the different galaxy bias terms can be written in terms of the galaxy linear bias. If proven true, it would dramatically simplify current cosmological analysis. The goal of this project is to validate this prediction with galaxy clustering observations from current surveys such as DESI Legacy Survey, eBOSS or 2MPZ.

Recommend skills/experience: Knowledge of physics and statistics and coding (or willingness to learn to code) in Python.

 

The growth of supermassive black holes 

Supervisor: Dr Becky Smethurst

It has long been thought that galaxy mergers lead to the majority of supermassive black hole (SMBH) growth, however cosmological simulations have recently shown that this is not the case. The majority of SMBH growth seems to occur through calmer processes internal to the galaxy itself, such as gas being funnelled to the central SMBH down a galaxy's spiral arms. However, these processes are understudied, and the assumption that mergers dominate supermassive black hole growth still pervades astrophysics. A summer student would join our research group tackling this problem from many different perspectives. An interesting avenue of research would be to look at the differences between the motion of stars and gas in galaxies with growing supermassive black holes, and determine if they are offset from each other. If they are offset, this suggests something like a merger has disturbed the galaxy and most likely led to the growth of the black hole. If they are not offset, then a calmer process internal to the galaxy has most likely led to the growth of the black hole. A successful student would determine if there are correlations between the supermassive black hole growth rate and the disturbance of the galaxy in a large sample of galaxies from the SDSS MaNGA survey. This would be an observational project requiring a student to work with real astrophysical data. 

Recommend skills/experience: Some knowledge of the Python coding language would be beneficial. 

 

The correlation between light and dark matter across cosmic time

Supervisors: Dr Anastasia Ponomareva, Dr Tariq Yasin

In the standard cosmological model, each galaxy resides at the centre of a vast dark matter halo. A key problem in astrophysics is explaining the existence of tight scaling relationships between properties of the luminous galaxy (such as its mass) and dynamical proxies for its invisible dark matter halo (such as a galaxy’s rotation speed). These are unexpected given the stochastic process of galaxy formation under hierarchical structure formation.

The Oxford MIGHTEE team, with pioneering new radio observations of neutral hydrogen (the most direct probe of galaxy dynamics), have now measured some of these relationships over the past billion years. However this is just the beginning: over the next few years the Square Kilometre Array will come online, extending our knowledge of galaxy dynamics back to when the universe was just a third of its current age. This radio data will be further bolstered by constraints on the mass distribution around galaxies from weak lensing observations with instruments such as Euclid.

There is now a pressing need for theoretical predictions for the relationships between galaxies and their host halos over cosmic time. Developing such relations is the goal of this project. We will achieve this by combining numerical simulations of dark matter with statistical models for the distribution of galaxies within halos. These predictions will then be used to test the consistency of our current understanding of cosmology and galaxy formation against the unprecedented forthcoming trove of data extending deep into the universe's past, a regime in which the dark matter paradigm has yet to be tested at the galaxy scale.

The student will gain skills in statistics, Bayesian inference, dark matter models and galaxy dynamics.

Recommended experience: Ability to code in Python, statistics

 

Optimising the superconducting quantum frequency converter for quantum computing and astronomical applications

Supervisors: Dr Nikita Klimovich, Dr Boon Kok Tan

In recent decades, there has been remarkable growth in superconducting quantum electronic technologies, playing a crucial role in quantum information science, astrophysics, and beyond. Breakthroughs include quantum-noise-limited parametric amplifiers, revolutionary imaging technology featuring MKID detectors, ultra-sensitive heterodyne mixers, and microscopic spectrometers operating at frequencies in the GHz range. These advancements find applications in diverse fields, from dark matter searches and black hole imaging to various quantum technology platforms like quantum communications and computation. A recent trend involves extending similar technologies to higher frequencies, reaching into the hundreds of GHz to the 1 THz regime. This expansion aims to improve qubit readout in quantum computers at higher bath temperatures, explore uncharted regions in axion-like dark matter searches, and study the cosmology of the early universe.

The millimetre/sub-millimetre range, largely unexplored outside astronomical purposes, holds significant novelty and potential for ground-breaking research. At Oxford, we are at the forefront of this progress, drawing on our extensive experience in mm/sub-mm astronomical instrumentation. We actively address challenges in transitioning to higher frequencies. A major challenge in this regime is generating low-noise, high-purity, and high-power tunable broadband tones at such frequencies. Conventional techniques, such as semiconductor frequency multipliers based on Schottky diodes, have moderate conversion efficiencies and substantial heat dissipation. These issues hinder their deployment for ultra-sensitive applications like quantum computing or large pixel count astronomical heterodyne arrays.

Our group is addressing this challenge by developing novel techniques and prototyping new superconducting devices that can efficiently translate signals to higher frequencies. However, the exploration of the parameter space is vast, and the preliminary framework needs refinement to answer crucial questions for optimizing device operation. We are confident that this technology will not only work in principle but also in experimental applications. Still, there is currently no systematic method for thoroughly optimizing the design of such devices with specific desired performance parameters.

In this project, the student will learn to use and build upon our existing Matlab and Python code for simulating such parametric frequency conversion devices. The focus will be on investigating the impact of various design parameters on the performance of these frequency converters. This project is ideal for students interested in quantum devices for fundamental physics experiments and those who enjoy developing theoretical/computational software. For students interested in experimentation, there will be ample opportunities to test these devices using state-of-the-art cryogenic facilities and top-end laboratory equipment. The student will be supported by several DPhil students and technicians, in addition to their supervisors.

 

How smooth is my galaxy?

Supervisor: Dr Thomas Williams

Observations of nearby star-forming galaxies reveal that gas emission across them is not smooth, but rather clustered into sources and evacuated in holes. Different phases of the interstellar medium look extremely different, from the smooth atomic hydrogen discs to the clumpy, cloud-dominated molecular hydrogen. Hydrodynamical simulations of galaxies show that the smoothness of the various gas phases is extremely sensitive to the various physical recipes and strength of various processes, and so quantifying clumpiness is vital to inform models. This project will quantify the smoothness of different gas phases within galaxies, using VLA/MeerKAT HI, ALMA CO, and new JWST observations of 19 galaxies taken as part of the PHANGS-JWST survey to see how sensitive each tracer is to, e.g., feedback from supernovae and regions of a stronger interstellar radiation field, and determine how best to map out the various distinctive sources and holes in galaxies.

Recommended skills/experience: Python, statistical analysis and interpretation

 

A Radio Study of the Black Hole X-ray Binary Swift J1727.8−1613 with the Allen Telescope Array 

Supervisor: Dr Joe Bright

Our team (including members of Oxford Astrophysics, the SETI Institute, and Breakthrough Listen) have been carrying out an extensive monitoring campaign with the Allen Telescope Array (ATA) on the luminous black hole X-ray binary (BHXRB) Swift J1727.8−1613, which has been in outburst since Aug. 2023. Radio observations allow us to track the jets and ejections produced in these systems, and to gain insight into how accretion is connected to their launch. The observing campaign is ongoing, however we already have 10s of epochs in hand where we clearly detect the source, each epoch containing multiple frequencies thanks to the wide observing bandwidth of the ATA. A robust and well tested calibration and imaging pipeline for the ATA is under active development at Oxford. The student will use this pipeline to process these data and measure the evolving multi-frequency flux density from Swift J1727.8−1613. Combined with publicly available X-ray data, this will allow us to understand the evolving outburst properties from the target, and place it in context with other accreting binary systems containing a compact object. 

At the end of the summer project, the student will have produced a publication ready radio dataset, whilst learning about radio and X-ray astronomy, accretion onto compact objects, and will have had the opportunity to run their own observations of the target with the ATA. This project will contribute significantly to a publication, and the student will be involved in writing the manuscript. This will give them first hand experience in paper writing, as well as authorship on a peer reviewed piece of work, which will boost the quality of any future applications they make. 

Recommended skills/experience: We would require the student to be familiar with the Python programming language. Familiarity with Linux operating systems would be beneficial but is not required.

 

Dark matter and galaxy dynamics: enduring puzzles

Supervisors: Dr Tariq Yasin, Dr Harry Desmond

In the 1980s, Vera Rubin’s observations of rotation curves around spiral galaxies provided key evidence that most of the universe’s mass is non-baryonic, paving the way for the widespread acceptance of the dark matter paradigm. However, despite decades of further study and great observational advances, many facets of rotation curves remain mysterious.

This project will investigate one such facet, "Renzo’s rule" which states that “for any feature in a galaxy’s luminosity profile there is a corresponding feature in the rotation curve and vice versa.”  This implies a strong correlation between luminous and dark matter, which is very strange given the assumed collisionless nature of the latter. However, despite its validity being widely acknowledged, Renzo’s rule is entirely informal, based largely on visual inspection of rotation curves.

The goal of the project is to develop a rigorous mathematical description of Renzo's rule (and perhaps also similar heuristics of galaxy dynamics). We will do this by quantifying the features of state-of-the-art rotation curves to establish the statistical significance of the coupling of luminous and dark matter. This will enable new tests of the composition and dynamics of the universe.

The student will gain skills in statistics, galaxy dynamics and dark matter modelling.

Recommended skills/experience: Ability to code in Python, statistics

 

Sulfur and argon abundances in early galaxies with JWST/NIRSpec

Supervisor: Dr Alex Cameron

Chemical elements are formed by a range of processes in different types of dying stars. Since different elements arise from different processes, relative abundances of different elements can be used to understand the properties of the stellar populations in galaxies, and how these galaxies have grown across time.

Sulfur and argon are two elements of particular interest for understanding early chemical enrichment. It is thought that significant fractions of the present-day abundances of these elements arise from Type Ia supernovae. Since this type of supernova only occurs on long timescales after the onset of star formation, it is predicted that abundances of these elements will be lower in the early Universe compared to elements such as oxygen and neon which enrich very quickly.

The goal of this project will be to use James Webb Space Telescope (JWST) spectroscopy to measure sulfur and argon abundances in high-redshift galaxies, providing constraints on how rapidly the enrichment of these elements proceeded within the first 3 billion years of cosmic history. These abundance measurements will then be compared to existing chemical evolution models to explore the impact of Type Ia supernova on the chemical enrichment histories of these galaxies. This will provide valuable insight into the properties of early stellar populations.

 

Machine-learning the halo bias model

Supervisors: Prof. Pedro Ferreira, Dr Deaglan Bartlett, Dr Harry Desmond

According to the Lambda-CDM model of cosmology, galaxies form in huge spheres of dark matter called halos, which form the basic building blocks of the universe. Much of cosmology relies on the relation, called "bias", between the spatial distribution of halos and the underlying dark matter density field [1]. This relation cannot be determined a priori, necessitating assumptions about its functional form.

The aim of this project is to apply a new technique in machine learning, symbolic regression (SR), to uncover the true form of halo biasing. This will be achieved by training state-of-the-art SR algorithms on cosmological simulations of the formation and evolution of dark matter halos, and evaluating possible functional forms using novel information-theoretic metrics [2]. Besides being of fundamental interest in cosmology, an improved bias model is desperately needed for cutting-edge, field-level cosmological analyses (e.g. [3]) and thus the model the student will create should become an integral part of such analyses.

The student will gain expertise in cosmology, machine learning, statistics and Python programming.

[1] https://arxiv.org/abs/1611.09787
[2] https://arxiv.org/abs/2211.11461
[3] https://arxiv.org/abs/1909.06396

 

Searching for Technosignatures with the Murriyang (Parkes) Radio Telescope 

Supervisors: Dr Joe Bright, Dr Steve Croft

The University of Oxford has recently become the headquarters of Breakthrough Listen, a project dedicated to the discovery of intelligent life beyond our Solar System. One aspect of the program is the search for radio technosig- natures: narrow bandwidth Doppler drifting signals indicative of communication. One of the main Breakthrough Listen observing facilities is the 64-m Murriyang radio telescope located at the Parkes Observatory in New South Wales, Australia. The Breakthrough Listen program has collected a wealth of data from Murriyang on nearby stars and the galactic center using the ultra-wideband and multibeam receivers (operating in the L- and S-bands). High time and frequency data products are available from these observations which can be searched using Doppler drift searching algorithms such as turboSETI (and updated codes which are currently under development). The aims of this project will be i) searching for technosignature candidates in archival Murriyang data, and ii) characterising the radio frequency interference (RFI) environment at the Parkes Observatory. RFI is the leading contaminant when searching for narrow band signals from beyond our Solar System. The student will learn and employ techniques to separate RFI from candidate signals using e.g. off-source comparison data (signals should only be present when pointed at the target) and/or anomaly detection, a powerful machine learning tool allowing for unusual data to be discovered. The final products of this project will be a database of technosignature candidates (and associated analysis to determine their origin), and insights into the properties of RFI at the Parkes site (occurrence rate at times, frequencies, telescope positions, etc.). 

There will be the opportunity for students to become trained observers on Murriyang, allowing them to perform observations as part of the Breakthrough Listen initiative. The student will work in collaboration with a summer student at Berkeley University who will be performing a complementary analysis on data from the Green Bank telescope. The analysis performed by the student will contribute to a publication in a peer reviewed journal, giving the student valuable insight into paper writing and enhancing any subsequent applications for PhD positions.

We acknowledge the Wiradjuri People as the traditional owners of the Parkes Observatory. 

Recommended Skills/Experience: We would require the student to be familiar with the Python programming language. Familiarity with Linux operating systems would be beneficial but is not required. 

 

Anomaly detection with ZTF

Supervisors: Prof. Chris Lintott, Dr Steve Croft

Finding unusual, interesting objects is a task of increasing urgency as large surveys such as the Vera C Rubin Observatory’s LSST approach; this project aims to develop tools for finding unusual light curves both for technosignature searches and for studies of astronomical oddities. 

We will make use of data from the Zwicky Transient Facility (ZTF), a multi-filter survey of the northern sky with high (~2 day) cadence, which has produced millions of lightcurves that are available in a public archive. The data are being used to search for transient sources, categorise large classes of variable sources, and to identify unusual objects.

One unusual object that has attracted much interest over the last few years is Boyajian’s Star (Boyajian et al. 2016), a star that exhibits non-periodic dips in brightness of over 20%. The most favoured explanation for these dips is the presence of dust in the system, but the behaviour also has some characteristics expected of artificial megastructures in orbit around the star (Wright et al. 2016). A larger sample of similar objects is needed in order to better understand the underlying physical mechanisms in these systems.

Drawing on existing code for accessing ZTF data via the Lasair broker developed for LSST, the student will search the lightcurves for Boyajian’s Star analogues, as well as other unusual behaviour. Possible tools to be used for classification include existing anomaly detection frameworks, classical statistical tools, and machine learning algorithms. The project would suit someone with either strong coding skills, or a willingness to learn as they go. The necessary computing resources can be accessed remotely from a laptop or desktop computer.

 

Recommendation systems for radio SETI 

Supervisors: Prof. Chris Lintott, Dr Steve Croft

An increasingly common mode of interaction with machine learning and other expert systems is via a recommendation system, which manipulate often abstracted properties of the data being studied to present objects of interest to an expert user. This project will apply this approach to the task of identifying signals of interest in a large body of SETI data from the Green Bank Telescope (GBT).

An example of such a system is Astronomaly, described by Lochner & Bassett (2020); we have recently applied these tools to galaxy classification (see e.g. Walmsley et al. 2023). This project will initially make use of an existing, trained reverse image search pipeline, and present data to citizen scientists via the Zooniverse.org platform. 

The dataset consists of a large number of radio spectrograms of a variety of targets (mostly nearby stars) from the GBT, which can be loaded into Python using publicly available libraries. Large numbers of signals are present in the data, the vast majority of which are due to human-generated radio frequency interference (RFI), making up the “haystack” in which we search for the “needles” of candidate technosignatures or other interesting anomalies. 

This project would suit a student who wants to get experience in using the outputs of sophisticated machine learning, and possible extensions if time allows include either characterising any unusual signals found or developing more specialised ML. The project also has the potential to help characterise the variety of transmitters making up the RFI haystack at GBT, which is of interest to a variety of other users of the telescope.

Given the involvement of citizen scientists, there will also be some opportunity for online outreach, either through writing or via presentations. The necessary computing resources can be accessed remotely from a laptop or desktop computer.

 

Constraining baryonic physics in cosmology

Supervisor: Dr. Tassia Ferreira

Weak lensing observations find that the dark matter today is more smoothly distributed than expected from the analysis of early-Universe measurements of the Cosmic Microwave Background. One explanation could be a missing ingredient to the weak lensing analysis: baryonic effects. The cosmic baryonic component is one of the leading systematics in weak lensing analyses. In large cosmological surveys like the Dark Energy Survey, about 40% of the cosmic shear data points are discarded in the analysis due to our inability to appropriately account for baryonic effects. The recent detection of the cross-correlation between X-rays and weak lensing has emerged as a promising avenue to complement cosmic shear analyses by providing important information on baryonic physics.

In this project, you will do exciting work to validate the results from this cross-correlation by comparing them with data from the Cosmic Microwave Background that contains imprints of the large-scale structure.

Recommend skills/experience: Knowledge of physics and statistics and coding (or willingness to learn to code) in Python.

 

Understanding and characterising natural and artificial signatures of coherent radio emission from nearby stars

Supervisors: Dr. Alex Cooper, Dr. Vishal Gajjar

The Breakthrough Listen initiative (https://breakthroughinitiatives.org/initiative/1 ) is currently undertaking the largest ever scientific research program aimed at searching the skies for signatures of extra-terrestrial life. Extremely sensitive radio observations, coupled with new data analysis techniques, will probe the nearest million stars to hunt for techno signatures expected from advanced civilisations. These observations will also probe new classes of coherent radio transients; and may serendipitously detect mysterious, extragalactic fast radio bursts.

In this project, the student will make new theoretical predictions to determine the nature of both natural and artificial radio signatures these state-of-the-art observations will discover. They will use novel computational tools to examine how the propagation of coherent radio signals will shape the signal detected in Breakthrough Listen datasets. These predictions will allow us to identify extra-terrestrial signals more readily in the data, but also work backwards to derive their true nature at source. Furthermore, the student will investigate what kinds of new, natural radio signatures are accessible by these observations. Specifically, they will perform numerical calculations to predict observable radio emission from the magnetic interaction between exoplanets and their stars/moons, by analogy to observed radio emission stemming from binary interactions in our local solar system and as predicted in compact object mergers.

There will be extensive opportunities to collaborate with students and scientists within Oxford Astrophysics, but also at University of California, Berkeley including with Dr. Steve Croft.

Recommended skills/experience: The student should be comfortable using the Python programming language and should have an interest in extra-terrestrial searches and/or radiative processes. 

 

Lifting the Galactic veil to find gravitational waves from the Big Bang

Supervisors: Dr Kevin Wolz, Dr Adrien La Posta

Our model of the Big Bang is incomplete. The Universe may have undergone an early phase of acceleration called cosmic inflation. Inflation sends ripples through spacetime, known as primordial gravitational waves. The Cosmic Microwave Background (CMB) should contain a record of these gravitational waves through a specific pattern in its polarisation signal. However, the Milky Way emits polarised light orders of magnitude brighter. Distinguishing the two requires statistical models tailored to the complexity of Galactic patterns in the sky. The University of Oxford is involved in on-going searches for this elusive signal from the Big Bang as a member institution of the Simons Observatory (SO). 

The goal of this project is to make SO more robust against Galactic contamination. To do so, the student will get the chance to familiarise themselves with the SO parameter inference software, and implement a statistical tool called the Jeffreys prior.

Recommended skills/experience: Knowledge of statistics, Python programming