2016 Highlights

On February 11 2016, the LIGO collaboration announced the first direct detection of gravitational waves and the first observaton of binary black holes. Accurate theoretical models of the signal were needed to find it and, more importantly, to decode the signal to work out what the source was. These models rely on large numbers of numerial solutions of Einstein’s equations for the last orbits and merger of two black holes, for a variety of binary configurations. The DiRAC Data Centric system, COSMA5, was used by researchers at Cardiff University to perform these simuations. With these results, along with international collaborators, they constructed the generic-binary model that was used to measure the masses of the two black holes that were detected, the mass of the final black hole, and to glean some basic information about how fast the black holes were spinning. Their model was crucial in measuring the properties of the gravitational-wave signal, and The DiRAC Data Centric system COSMA5 was crucial in producing that model. More information on the detection of gravitational waves can be found at the LIGO collaboration website.

In the figure to the right, the top plot shows the signal of gravitational waves detected by the LIGO observatory located in Hanford, USA whist the middle plot shows the waveforms predicted by general relativity. The X-axis plots time and the Y-axis plots the strain, which is the fractional amount by which distances are distorted by the passing gravitational wave. The bottom plot shows the LIGO data matches the predications very closely. (Adapted from Fig. 1 in Physics Review Letters 116, 061102 (2016))

If you could see deep into the subatomic world with slow-motion glasses, you would see empty space teeming with particles, appearing and disappearing in a tiny fraction of a second. If you were very lucky you might glimpse a particle so heavy and rare as to have escaped detection so far at CERN’s Large Hadron Collider. Such particles would be harbingers of new physics going beyond our existing Standard Model, but are proving very hard to find.

One way to demonstrate their existence indirectly is through the impact they have, as part of the teeming vacuum, on the properties of particles we can study directly. A new experiment will start shortly at Fermilab near Chicago to measure the magnetic moment of the muon (a heavier cousin of the electron) to an astonishing accuracy of 1 part in 1010. If we can calculate the effect of all the particles in the Standard Model on this magnetic moment, then a discrepancy with experiment would indicate the existence of new particles. An improved theoretical uncertainty from the Standard Model needs numerical calculations for the theory of the strong interaction, QCD. This is because the largest source of uncertainty comes from quarks, which are strongly interacting (see Figure 1); this contribution is known as the hadronic vacuum polarization (HVP). HPQCD has developed a new lattice QCD method using DiRAC to determine the HVP with improved accuracy

We used our method to determine the strange and charm quark contributions to the HVP (arXiv:1403:1778) for the first time from lattice QCD, and then the very small bottom quark contribution (1408.5768). In the past year have determined the up and down quark contribution (1601.03071) and estimated the effect of disconnected diagrams (lower picture in Figure 1) (1512.03270). Our final result, summing all of these contributions, has a total uncertainty of 2% (see Figure 2). The error is dominated by possible systematic effects from ignoring the electric charge of the quarks and taking the masses of the up and down quarks to be the same. Work is now underway to include both of these effects in the calculation to reduce the uncertainty further, ahead of the new experimental results in 2018.

A Tidal Disruption Event (TDE), which occurs when a star is destroyed by the gravitational field of a supermassive black hole, produces a stream of debris, the evolution of which ultimately determines the observational properties of the event. Typically, investigations of this process resort to predicting future accretion of material assuming the debris orbits are Keplerian or simulating restricted parameter regimes where the computational cost is a small fraction of that for more realistic parameters. In a series of papers that took advantage of Complexity@DiRAC, we simulated the long-term evolution of the debris stream for realistic TDE parameters, revealing new effects that can significantly affect the observable properties of these events (Coughlin & Nixon 2015; Coughlin, Nixon et al. 2016a,b). The figure to the right is taken from Coughlin, Nixon et al. (2016a: MNRAS 455, 4, 3612). In this work, we showed that a post-periapsis caustic – a location where the locus of gas parcels comprising the stream would collapse into a two-dimensional surface if they evolved solely in the gravitational field of the hole – occurs when the pericentre distance of the star is of the order of the tidal radius of the hole. We showed that this ‘pancake’ induces significant density perturbations in the debris stream, and these fluctuations can be sufficient to gravitationally destabilize the stream, resulting in its fragmentation into bound clumps.

As these clumps fall back towards the black hole their accretion can result in strong bursts of radiation, while the low-density regions between clumps can yield relatively little. This makes the luminosity of the event highly time variable, and may be an explanation for the rapid time variability of systems such as Swift J1644. However, the detailed behaviour of the clumps is not yet well understood, as it depends on their internal thermodynamics and how quickly they can cool and condense. If they can cool fast enough, they may be able to survive a second pericentre passage and potentially form planets orbiting the black hole, otherwise they will be re-disrupted at pericentre and join the accretion flow. These processes will be the subject of future DiRAC simulations. Over the next few years new observing missions will provide a substantial increase in the number of events for which we have data. Determining how the luminosity of a TDE varies with time is critical to interpreting these events and discerning the progenitor properties, such as black hole mass and spin.

To find out what remote planets orbiting other stars are made of, astronomers analyse the way in which their atmospheres absorb starlight of different colours and compare it to a model, or ‘spectrum’, to identify different molecules. One of such objects is Kepler-69c, which is 70 percent larger than the size of Earth. Its orbit of 242 days around a sun-like star resembles that of our neighboring planet Venus, however the composition of Kepler-69c is still unknown.

In collaboration with NASA AMES we have developed molecular data to model (hot) super-Venuses, yet to be detected class of object with the sulphur dominated chemistry based on Volcanic activity. According to the atmospheric chemical models, among the main constituents of the atmospheres of the rocky super-Earth or super-Venus are sulphur dioxide (SO₂) and sulphur trioxide (SO₃). These spectroscopically important molecules are generally not included in terrestrial exoplanet atmospheric models due to the lack of the laboratory data. We anticipate our new data will have a big impact on the future study of rocky super-Earths and super-Venuses. The recently discovered system of exoplanets TRAPIST-1 is one of the existing examples where our data will play important role in future atmospheric retrievals.

Our SO₃ data set, named UYT2, is one of the biggest line lists in the ExoMol database. It contains 21 billion vibration-rotation transitions and can be used to model transit IR spectra of volcanic atmospheres for temperatures up to 800 K. In order to accomplish this data intensive task, we used some of the UK’s most advanced supercomputers, provided by the Distributed Research utilising Advanced Computing (DiRAC) project and run by the University of Cambridge. Our calculations required millions CPU hours, hundreds thousands GPU hours, and up to 5 TB of RAM, the processing power only accessible to us through the DiRAC project.

COSMOS consortium researchers exploiting DiRAC HPC Facilities have made progress towards ambitious milestones in five key inter-related areas: (i) extreme gravity, (ii) inflation and the early universe, (iii) cosmic microwave sky, (iv) dark energy and (iv) galaxy archaeology. In the limited space available, we highlight recent breakthroughs in general relativistic simulations of black holes and inflation.

On 11 February 2016, the LIGO Scientific Collaboration announced the ground-breaking discovery of a gravitational wave signal (GW150914) representing the first incontrovertible observation of a black hole binary system, as well as the most energetic event ever observed in the Universe. Given recent advances such as our innovative GRChombo AMR code, COSMOS Consortium researchers (several also LSC members) are well-placed to exploit these new scientific opportunities in gravitational wave astronomy. We have continued our long-standing industrial collaboration through the COSMOS Intel Parallel Computing Centre (IPCC) to improve our research codes in cosmology and general relativity; we have particularly focussed on GRchombo and improving performance on multi- and many-core (Xeon Phi) systems, following in the wake of our 2015 HPCwire Award. Intel has also responded to the demands of ever increasing dataset size by developing a powerful software defined visualisation framework, including the OpenSWR software rasteriser and the OSPRay parallel ray tracer and volume renderer, a program in which COSMOS IPCC is directly involved. These tools hves allowed us to create unprecedented real-time visualisations of our science results using several TB datasets. In 2016 we demonstrated the adaptive mesh capabilities of ParaView powered by OSPRay and OpenSWR by creating videos of GRChombo black hole mergers for two of the largest HPC conferences, International Supercomputing 2016 and Supercomputing 2016, as publicised in insideHPC

Inflation is now the paradigmatic theory of the early universe, providing a dynamical mechanism to explain the many features of our current cosmology. In particular, it explains why the universe is homogenous over cosmological scales. Nevertheless, can inflation begin in the first place, if the initial conditions are not homogenous? To answer this question requires a numerical approach, as solving the full GR equations is not possible for generic inhomogenous initial conditions. We attacked this problem with the state-of-the-art numerical relativity code GRChombo and the results are reported in arXiv:1608.04408. We found that high scale inflation is generically robust – it will inflate despite large inhomogeneities. More interestingly, we found that low scale inflation – preferred by many string theory phenomenological models – is not robust, failing to inflate in the presence of large scale but subdominant inhomogeneities.

One of the biggest unresolved problems in Einstein’s theory of gravity is the weak cosmic censorship conjecture, which posits that spacetime singularities are always hidden inside a black hole horizon. Since nothing can escape a black hole, the breakdown of Einstein’s theory at singularities would be ‘censored’ from the outside world. In our numerical work with GRChombo, we find that the conjecture appears to be false if space has four or more dimensions. For example, the late stages in the evolution of a rapidly spinning black hole in six spacetime dimensions is shown (left); this is a disk-like structure with a great deal of fractal microstructure – the thinner rings are connected by ever thinner membranes that ultimately lead to a ‘naked singularity’. Given these results in extra dimensions, it is clear that, if cosmic censorship is true, then it must be a property specific to the three-dimensional space that we live in; without this property, Einstein’s theory of general relativity could lose its predictive power.

The fundamental constituents of the strong force are quarks and gluons, which themselves bind together to form the familiar building blocks of nuclear physics, protons and neutrons. The two most common forms of quarks are the up quark and the down quark. The electric charge of the up quark is +2/3, whereas the down quark carries charge -1/3. A proton is composed of two up quarks and one down quark, adding to a net charge of +1, whereas the neutron has two down and one up quark, producing a chargeneutral object. Lattice QCD simulations have reached a precision now, where isospin breaking effects become important. This has two sources, the mass difference of up and down quarks, and the electromagnetic interactions. Both effects are of the same order of magnitude, so a direct calculation from QCD and QED is necessary.

Most important for our understanding of the salient features of QCD, such as quark confinement and spontaneous chiral symmetry breaking, is the understanding of the properties of the vacuum. Simulations of QCD and QED allow us to study the effect of dynamical quarks on the vacuum. A snapshot of the vacuum fields is shown in Figure 1, in form of a three-dimensional slice of space-time. The topological charge density of the QCD fields is rendered with the magnetic field of QED. The positive topological charge lump at the lower left of the image is accompanied by large magnetic field strength, presenting an opportunity to observe the chiral magnetic effect which separates right- and left-handed quarks. It has been predicted theoretically, but never observed.

Isospin breaking effects are crucial for our existence. The Universe would not exist in the present form if the neutron — proton mass difference were only slightly — — different. If it would be larger than the binding energy of the deuteron, no fusion would take place. If it would be a little smaller, all hydrogen would have been burned to helium. Knowing the mass of neutron and proton and how it depends on the mass and charge of the individual quarks, we can express the allowed region in terms of the u and d quark masses and the electromagnetic coupling, as shown in the Figure 2. It turns out that both α_EM and the ratio of light quark masses must be finely tuned.

Bibliography:
[1] R. Horsley, Y. Nakamura, H. Perlt, D. Pleiter, P.E.L. Rakow, G. Schierholz, A. Schiller, R. Stokes, H. Stüben, R.D. Young and J.M. Zanotti, J. Phys. G43 (2016) 10LT02.
[2] R. Horsley, Y. Nakamura, H. Perlt, D. Pleiter, P.E.L. Rakow, G. Schierholz, A. Schiller, R. Stokes, H. Stüben, R.D. Young and J.M. Zanotti, JHEP1604 (2016) 093

Galaxies fall into two clearly distinct types: active, blue-sequence galaxies that are rapidly forming young stars, and passive red-sequence galaxies in which star formation has almost completely ceased. These sequences are closely related to the visual galaxy classification system, based on the prominence of spiral arms first suggested by Edwin Hubble.

Numerical simulations by the EAGLE collaboration shows that these sequences, and thus the Hubble sequence of galaxies, are created by a competition between star formation-driven outflows and gas accretion onto the supermassive black hole at a galaxy’s centre.

In galaxies containing fewer than 30 billion solar-like stars, young stars, exploding as supernovae, are able to drive buoyant outflows that prevents high density gas building up around the black hole. As the galaxies increase in stellar mass, they convert a larger fraction of the inflowing gas into stars and grow along the blue sequence.

More massive galaxies, however, are surrounded by a hot corona and this has important ramifications. The supernova-driven outflow is no longer buoyant and star formation is unable to prevent the build up of gas in the galaxy, and around the black hole in particular. This triggers a strongly non-linear response from the black hole. Its accretion rate rises rapidly, heating the galaxy’s corona and disrupting the incoming supply of cool gas. The galaxy is starved of the fuel for star formation and makes a transition to the red sequence, and further growth predominantly occurs through galaxy mergers. Interestingly, our own Milky Way galaxy lies at the boundary between these two galaxy sequences suggesting that it will undergo a transition to a red and `dead’ early-type galaxy in the near future.

This model has deep implications for understanding the Universe. Far from being exotic predictions of General Relativity, the growth of supermassive black holes sets a fundamental limit to the mass of star forming galaxies.

The figure on the left shows formation timescales of galaxies as a function of stellar mass in the Eagle simulation (points) and observational data (Ilbert 2015). The separation of galaxies into blue (rapidly star forming) and red (passive) sequences is clearly seen. Most low mass galaxies follow the star forming blue galaxy sequence, doubling their stellar mass every 3 billion years, but more massive galaxies have much longer star formation growth timescales (a horizontal dotted line shows the present day age of the Universe; galaxies with longer star formation timescales are shown above the line). The transition between the sequences occurs at a stellar mass of round 30 billion solar masses. While it is not possible to reliably measure the masses of black holes in observed galaxies, it is possible to do this for galaxies in the EAGLE cosmological simulation. The simulated galaxies follow the observed data closely and points are coloured by the mass of the black hole relative to that of the host halo. In red-sequence galaxies, black holes account for around 0.01% of the halo mass, but the fraction is a factor of 100 smaller in blue-sequence galaxies. Around a galaxy mass of 30 billion solar masses, there is considerable scatter in the relative mass of the black hole and in the star formation growth timescale. However, at a fixed galaxy mass, systems with a higher black hole mass have substantially longer growth timescales, implying the existence of an intimate connection between black hole mass and galaxy type.

State-of-the-art observations show that within approximately 1 billion years after the Big Bang, all of the primordial hydrogen that permeated the immense space between galaxies had been destroyed by high energy photons in a process known as reionization. This time period in the history of the Universe, also known as ‘the dawn of galaxies’ remains one of the most active frontiers in modern astrophysics from both an observational and theoretical point of view.

Much remains unknown about the sources which provided the high energy photons needed for reionization. The standard picture suggests it was likely the first generation of galaxies with a possible additional contribution coming from radiating supermassive black holes. More exotic solutions have been proposed as well, including annihilating dark matter. Identifying the sources responsible for reionization as well as understanding their properties is key to constraining the conditions which governed all subsequent galaxy formation in the Universe.

Unfortunately, our current telescopes only probe the tail end of this epoch. For this reason, theoretical and computer models are relied upon for insight into the physics during the reionization. Simulating this regime is both complicated and computationally expensive; however, over the past two years, researchers at the Institute for Astronomy (IoA) and Kavli Institute for Cosmology Cambridge (KICC), have designed a new, state-of-the-art computer algorithm which they are exploiting on DiRAC supercomputers at the Universities of Cambridge, Leicester and Durham to better understand the epoch of reionization. This new algorithm (presented in a recently submitted paper by Katz et al. 2016) follows the flow of high energy photons which are directly and on-the-fly coupled to the chemical and thermodynamic state of the gas at different speeds depending on the density of the medium. This has overcome many issues relating to standard algorithms in the field at much reduced computational cost, allowing us to simulate significantly more complex systems than previously possible.

The team at IoA and KICC has thus for the first time generated a multi-scale view of galaxy formation during this epoch – from the large scale cosmic web all the way down to individual star forming regions inside of primordial galaxies. Because the flow of radiation is followed explicitly, the researchers have used an inhomogeneous radiation field to compute the distribution of emission lines coming from ionized metals (such as [CII] and [OIII]) which can be observed by some of our most sensitive telescopes. Likewise, they have self-consistently tracked the location and mass of ionized, neutral, and molecular gas to make testable predictions about these quantities which control star formation in the Early Universe.

In the coming years, new observations from the James Webb Space Telescope, Atacama Large Millimeter Array, and Square Kilometer Array will shed light on this exciting frontier and be compared to these simulated models to give insight into the physical processes which drive ‘the dawn of galaxy’ formation.

In Cheung et al [JHEP 1612 (2016) 089, arXiv:1610.01073] we computed spectra of highly excited charmonia and charmed mesons with light (up and down) quark masses, corresponding to m_π ≈ 240 MeV, closer to their physical masses than in our earlier study where m_π ≈ 390 MeV. A highlight of the results was the presence of hybrid mesons (where the gluonic field is excited), some of which had exotic quantum numbers (cannot be solely a quark-antiquark pair). The resulting phenomenology of hybrid mesons gave insight into the effective low-energy gluonic degrees of freedom. One main conclusion was that the spectra showed only mild dependence on the light-quark mass, and the pattern of states, in particular the hybrid mesons, did not change significantly as the light-quark mass was reduced. This is an important systematic check and suggests that the interesting phenomenology of excited and hybrid mesons we found is not an artefact of working with unphysically-heavy light-quark masses.

In Moir et al [JHEP 1610 (2016) 011, arXiv:1607.07093], we presented the first ab-initio study of coupled-channel scattering involving charm mesons (and the first lattice QCD study of three coupled scattering channels). Working with light-quark masses corresponding to m_π ≈ 390 MeV, from computations of finite-volume spectra we determined infinite-volume D_π,D_η,D_s K isospin-1/2 scattering amplitudes. Figure 2 shows the various S-wave (orbital angular momentum = 0) cross sections: upper (lower) panel shows quantities proportional to the diagonal (off-diagonal) cross sections. The singularity structure of the amplitudes showed a near-threshold bound state in J^P=0⁺ corresponding to the physical D₀^*(2400) resonance, a deeply bound state with J^P=1^– corresponding to the D^*, and evidence for a narrow J^P=2⁺ resonance. This study represents a significant step toward addressing some puzzles in the spectroscopy of charmed mesons.

Initial conditions for galaxy formation simulations are specified within the Λ-CDM scenario by a Gaussian random field. This leads to a troubling lack of control: we often cannot ask precisely what leads to certain outcomes for a galaxy. What is it about a galaxy’s mass, environment or history that gives it its shape or its colour? Imagine we could control the initial conditions of a galaxy in cosmological simulations and tweak them to alter one aspect of its formation while keeping the other aspects the same.

Thanks to the DiRAC facility, our group has run the first generation of high resolution cosmological simulations that alters the “DNA” of a galaxy (Roth et al. 2016, MNRAS, 455, 974; Pontzen et al., arXiv:1608.02507). Roth et al. 2016 sets out the “genetic modification” approach and Pontzen et al. showcases its power (see figure) to understand what quenches star formation in a galaxy. By minimally altering the linear overdensity in the cosmological initial conditions, Pontzen et al changes the merger history of a galaxy while leaving its final halo mass, large scale structure and local environment unchanged. The changing merger history influences the impact of feedback from Active Galactic Nuclei leading to three very different fates for the galaxies: star forming, temporarily quenched, permanently quenched.

A recent international collaboration between the UK, France, US and Canada have performed systematic studies of both nuclear radii and binding energies in (even) oxygen isotopes from the valley of stability to the neutron drip line. Both charge and matter radii are compared to state-of-the-art ab initio calculations along with binding energy systematics. Experimental matter radii are obtained through a complete evaluation of the available elastic proton scattering data of oxygen isotopes. The theoretical ab initio calculations were possible thanks to exploiting High Performance Computing resources and showed that, in spite of a good reproduction of binding energies, conventional nuclear interactions derived within chiral effective field theory fail to provide a realistic description of charge and matter radii. A novel version of two- and three-nucleon forces leads to considerable improvement of the simultaneous description of the three observables for stable isotopes but shows deficiencies for the most neutron-rich systems. Thus, crucial challenges related to the development of nuclear interactions remain. Phys. Rev. Lett. 117, 052501 (2016) – Editor’s suggestion.