Recent research

Here are some brief descriptions of research I have been working on, together with links to the relevant published articles or preprints.

A star of the beat generation

posted Dec 2, 2017, 1:50 PM by Jeremy Drake

Stars like the Sun generate copious magnetic fields that induce all sorts of interesting behaviour at the stellar surface.  Sunspots betraying bundles of emerging magnetic field, and the diffuse white halo of light scattered off the magnetically-driven solar wind seen around the Sun during a total eclipse, are good examples of this. The magnetic field is generated by the combination of the Sun's rotation and the convection that characterises the outer 30% or so of its radius. The magnetic field is always changing and goes through a cycle in which the polarity of the fieldthe "north" and "south" of the magnetreverses and changes back again over a period of about 22 years.  Similar cyclic magnetic behaviour is seen in about 60% of Sun-like stars.

Iota Horologii is a star like the Sun but much younger, with an age of "only" about 600 million years. It came to prominence when found to host a planet with a mass of 2.5 Jupiters in an orbit similar to that of Earth.  ι Hor rotates four times faster than the Sun because it has not had time to lose as much angular momentum through its wind. This faster rotation drives a stronger magnetic dynamo. Earlier studies had found a much shorter magnetic cycle than the Sun's with a period of only 1.6 yearsin fact the shortest known to date. As part of a much wider investigation of this intriguing exoplanet host, SAO Postdoctoral Scholar Julian Alvarado-Goméz studied the light from ionized calcium atoms in the star's upper atmosphere, or "chromosphere."  The chromosphere is heated by energy dissipated from the surface magnetic field and so can be used to trace the magnetic cycle. In a paper published in the 2017 October 10 edition of the Monthly Notices of the Royal Astronomical Society, Alvarado-Goméz and colleagues find the cycle is actually two cycles, with periods of 1.4 and 2 years, superimposed and "beating" against each other to make an apparent 1.6 year average. Still to be done is to understand exactly how and why magnetic cycles on the Sun and stars arise at all.

The icing on the snake

posted Nov 9, 2017, 7:37 AM by Jeremy Drake   [ updated Nov 9, 2017, 7:37 AM ]

Planets are born in the residual gas and dust from the star formation process that is left orbiting a nascent star in the form of a disk. The closest example of such a planet hatchery is in the constellation of Hydra, the water snake. The TW Hydrae system, whose Atacama Large Millimeter Array image famously shows a disk face-on to us carved out with dark lanes that could possibly be gaps swept out by newly-born planets.  The gas disk lasts for a few million years, eventually being dissipated by evaporation or loss to a disk wind, or by spiraling onto the parent star under gravity.

The inner edge of the disk is truncated at a few stellar radii by the star's strong  magnetic field.  Gas then follows the magnetic field lines, accreting onto the star in narrow columns down the magnetic poles. In free-fall, velocities reach up to 500 km per second and most of this energy of motion is converted into heating the gas to 2-3 million degreeshot enough to emit at X-ray wavelengthsin a shock as it hits the stellar surface. The shocked gas is slowed by a factor of about 4, so just before merging into the stellar surface it can still be moving at about 100 km per second.  Italian astrophysicist Costanza Argiroffi, from the Palermo Astronomical Observatory in Sicily, lead a study published in the November 2017 edition of Astronomy & Astrophysics to try and detect this motion of the shocked gas accreting onto TW Hydrae by its Doppler shift using the spectrometers onboard the Chandra X-ray Observatory. By combining many observations of the star and carefully comparing the spectrum with older stars long-bereft of accreting gas, the Doppler shift was detected. But the velocity was much smaller than expected - "only" 40 km per second.  This indicates that the accretion streams that provide the last 1% or so of the star's final massthe icing on the cakelie at an angle to our line-of-sight. The dominant magnetic pole of TW Hydrae is then not at the geographical stellar pole, but instead closer to the equator at a latitude of 10-30 degrees.

Monster CME from the Demon Star

posted Oct 30, 2017, 9:10 PM by Jeremy Drake   [ updated Nov 1, 2017, 8:06 AM ]

The Arabic name "Algol", of the eclipsing close binary star sometimes known as the Demon Star, derives from the "head of the ghoul"in Greek tradition the head of the venomous snake-haired Medusa severed by the demigod fellow Perseus. It comprises a B8 dwarf and a K0 subgiant in a 2.86 day orbit.  When the fainter K star passes in front of the brighter B star, Algol dims significantlybehaviour thought of as somewhat spooky long before its binary nature was known.  The stars are tidally-locked, and both rotate with the same period as their orbit, much like the Moon is tidally-locked to its orbit about the Earth.  The consequent rapid rotation of the K star, in concert with its outer convection zone, engenders vigorous and violent magnetic activity.  The hot B8 dwarf is instead bereft of an outer convection zone and appears magnetically quiescent.  This is handy because when the B star passes in front of the K star the eclipse signature in X-rays can be used to infer the spatial structure of the K star's X-ray emitting corona.

Back in late August of 1997, the joint Italian-Dutch BeppoSAX X-ray satellite caught Medusa's head having a bad hair day. An immensely energetic magnetic flare was observed that lasted  for nearly three days, and released a total X-ray energy of 1037 ergequivalent to the entire visible light output of the Sun for almost an hour and a million times more energetic than the largest solar flares. The flare was eclipsed by the B star, enabling the scientists who first studied it to estimate its size and location. Part of the X-ray signal from the flare was also observed to suddenly dim, then recover.  It was suggested that this dimming was caused by a coronal mass ejection like those that commonly accompany large solar flares.  SAO Postdoctoral Scholar Sofia Moschou took up this idea and analysed the ejection using a geometrical CME description developed for solar CMEs called the "ice cream cone" model.  She found the event had a mass of two to twenty thousand trillion tonnes, similar to that of the fifth largest asteroid in the solar system, Interamnia, and moved at a speed of about 3 million kph.  This work, lead by Dr. Moschou, is accepted for publication in the Astrophysical Journal.


posted Aug 28, 2017, 9:19 PM by Jeremy Drake   [ updated Aug 28, 2017, 9:26 PM ]

It was the first time I saw the solar corona with my own eyes.  Having worked in the field of stellar coronae for more than two decades, the eclipse of 2017 August 21 was quite special to me. 
In 1991, Wilton Barnhardt—a friend from Oxford days—and I travelled to Mexico on an extensive Kerouac-esque (well, perhaps at times...), road trip designed to take in the 1991 July 11 eclipse. The southern end of the Baja Peninsular, close to La Paz, was predicted to be the best spot to see it from. But it's quite a long drive down—900 miles or so from Tijuana—and we would have missed some places in northern Mexico we wanted to stop at. While Baja's Valle De Los Cirios was appealing, the ferry to the mainland would have taken an additional day, and considerable cash that was not in great supply. So, we opted for Mazatlan on the mainland that also made a good stopping point for continuing the trip.
The morning was clear, with blue skies and a smattering of clouds toward the horizon.  The early partial phase was nicely visible, starting around noon, through my number 14 welders' glass held more precariously than I'd like to admit by a complex but not entirely convincing web of elastic bands to my binoculars.  But, as noon turned to afternoon, we watched with mounting dismay, like Macbeth beholding the rise of Birnam Wood to Dunsinane, as clouds rolled in to blot out the remainder of the eclipse.  Totality under clouds is horribly disappointing—it got darker, street lights came on, and then the light gradually returned again. That's about it.  We hit the road and moved on, me as the astronomer feeling a bit self-conscious about having earlier talked up the damp squib eclipse so much to my friend.  I would not get a good chance to see another one until last week.

Of course, it's an utterly fatuous exercise to try and photograph the eclipse—hundreds, thousands, of eclipse experts with telescopes and fantastic professional-grade gear will do an infinitely superior job to a naive tourist with a consumer-level camera, and will plaster their beautiful images all over the web for anyone to get hold of.  What's the point of me doing it? I suppose its a bit like wearing that old "Beckham" number 7 shirt while playing football (soccer, if you must!) rather poorly at lunchtimes, before I did my knee in.  Or muddling through Agustin Barios' La Catedral on classical guitar, making a hash of it in the same places every single time, and the other side of the Universe from the brilliance of Ana Vidovic's rendition that is so dazzlingly beautiful it brings tears to my eyes.  The pictures are indeed a bit rubbish in the grand scheme of eclipse photography, and at the time of writing I've not had time to do anything but download a couple of them straight from the camera. But they are mine, and I took them at that place and time when I actually got to see the solar corona.

Gone with the wind

posted Jul 25, 2017, 5:21 AM by Jeremy Drake   [ updated Nov 6, 2017, 8:16 AM ]

The discovery of a planet orbiting our nearest star, Proxima Centuri, is a jewel in the crown of exoplanetary science. Further gilded by its location in the "habitable" zone, where an Earth-like planet would not freeze or boil its oceans, Proxima b represents a chance that the nearest exoplanet to us could actually harbour life. 

Surface liquid water requires the presence of a substantial atmosphere - without atmospheric pressure, water simply boils to the gaseous state. Close proximity to Proxima poses problems in this regard: as noted previously, the intense stellar wind where Proxima b resides could strip away a planetary atmosphere.  A strong planetary magnetic field might provide some protection, but atmospheric escape can still occur through the magnetic poles.  This "polar wind" occurs on Earth, but is not strong enough to endanger our atmosphere.  It occurs because air molecules at high altitude get ionized by solar extreme ultraviolet and X-ray radiation, and the light electrons liberated in the process can rise to sufficient altitudes to begin to escape Earth's gravity.  The rise of electrons generates an electric field that pulls the positively charged heavier ions upward, essentially resulting in a flow into space.  

Proxima b experiences ionizing radiation levels hundreds of times higher than the Earth, and we expected its polar wind to be commensurately stronger.  In a study lead by Katie Garcia-Sage of the NASA Goddard Space Flight Center and the Catholic University of America, featured in a NASA release, a Forbes article, and published in the July 20th edition of the Astrophysical Journal Letters, we used supercomputer models to compute the ionospheric outflow that the Earth would experience were it in the location of Proxima b. A best guess for the thermospheric temperature suggests a polar wind escape rate sufficient to remove an Earth atmosphere in a few hundred million years - a short timescale compared with the planet's likely age of nearly 5 billion years.  While the atmosphere could plausibly be replenished by geological processes, such as volcanic activity, the loss rates we estimate are conservative and do not include other erosion effects from the stellar wind and coronal mass ejections.  In all likelihood, Proxima b is bare, devoid of an atmosphere that has gone with the wind.

Check the space weather before packing

posted Jul 17, 2017, 7:36 AM by Jeremy Drake   [ updated Jul 17, 2017, 7:36 AM ]

The remarkable TRAPPIST-1 system is now know to host at least seven planets, three of which reside in the so-called temperate or "habitable zone", where surface temperatures should in principle be able to sustain water in liquid form.  But, as we have argued in earlier postings, there is a big difference between a planet being in the habitable zone and actually being habitable. Liquid water, in particular, can only exist within a relatively dense atmosphere. 

TRAPPIST-1 is an "utracool" red dwarf - a tiny star one tenth the size of the Sun with a surface temperature of about 2600 degrees.  It is so faint that only planets that orbit really close in can have temperatures warm enough for liquid water. The problem for the TRAPPIST-1 planets is that their star is not so weak at ultraviolet to X-ray wavelengths and also in regard to its magnetic field and wind. TRAPPIST-1 rotates once every 3.3 days, and this rapid rotation has built up its surface magnetic field to an immense strength of 100 times our Sun's.  Despite the star being so small, this magnetic field drives a stellar wind that ends up being of a similar strength to the solar wind.  Because the wind strength experienced by a planet decreases with the square of its distance from its parent star, the TRAPPIST-1 planets in their close-in habitable zone get blasted by a wind up to ten thousand times stronger than the Earth experiences.

On Earth, we are protected from most of the ravages of the solar wind by the Earth's magnetic field. However, any magnetic fields the TRAPPIST-1 planets have will be largely overwhelmed by the magnetism of their star.  The star-planet magnetic fields can directly connect over much of each planet, allowing the full force of the stellar wind to stream straight onto their surfaces.  The concept of atmospheric protection by a planetary magnetic field does not hold here. These conditions will likely strip off any atmospheres - and liquid water - the TRAPPIST-1 planets might have had on timescales of a hundred million years or so. Most habitable zone planets around similar low mass M dwarfs are likely to end up dry and barren - uninhabitable in the habitable zone. 

Your new vacation home on Proxima b just got cheaper!

posted Jan 11, 2017, 8:48 AM by Jeremy Drake   [ updated Nov 6, 2017, 8:13 AM ]

The term "habitable zone" used to be an esoteric concept on the fringes of astrophysics, describing the region around a star in which a planet could support liquid water.  The discovery over the last decade or so that planets are actually very common in the Universe has thrust habitable zones into the common lexicons of both astrophysics and popular science. While the classical definition of a habitable zone remains useful, it is not obvious that all planets in habitable zones will actually be habitable.

In several past works, we have argued that the space weather environment of a planet - the plasma and magnetic field conditions resulting from the host stellar wind - is likely to be crucial to habitability.  The link to habitability is through the atmosphere that is continually eroded by the action of the host stellar wind.  The key question is whether a planet can hold on to its atmosphere over billion year timescales. The Earth managed it, but evidence suggests that Mars, probably because of its lack of a strong protective magnetic field, has long since lost most of its surface water due to solar wind scouring.

The recently discovered planet in the habitable zone of our closest neighboring star, Proxima Centauri, has been one of the most exciting findings of the last decade. The planet is close enough that next generation telescopes will be able to observe it directly:  Proxima b could become the first habitable planet studied in detail. But Proxima is a host star quite different to the Sun - cool and faint, shining with only 1/600th of the Sun's power. The habitable zone of Proxima is then very close to the star, and Proxima b orbits twenty times closer to Proxima than the Earth to the Sun. SAO scientist Cecilia Garraffo has lead a new study of the impact of this proximity to Proxima. Supercomputer model simulations of the wind from Proxima indicate it blows with similar strength to that of the Sun. Being so close though, Proxima b gets blasted by this wind, experiencing a wind pressure several hundred times that experienced by Earth. Twice each 11 day orbit it passes through more dense wind streams that raise the pressure to 2000 times that at Earth.  If Proxima b has a magnetic field it might serve as some protection for its atmosphere, but it will be a highly dynamic magnetospheric environment, expanding and compressing like a bellows on timescales of less than a day. The much less severe dynamics of the Earth's magnetosphere are thought to be a factor in the Earth's atmospheric loss, as plasma is released by the forced opening and closing of the magnetic field.  It seems doubtful under such conditions that Proxima b will have retained any sort of atmosphere capable of sustaining life. 

Detailed atmosphere calculations are needed to assess the true habitability of Proxima b, but the property does not appear quite as appealing as it did in the first glossy brochures.  This work was published in the 2016 December 10 edition of the Astrophysical Journal Letters, and has featured a press release or two and an article on Forbes.

Sculpting a nova explosion using a supercomputer

posted Jan 6, 2017, 7:52 AM by Jeremy Drake   [ updated Sep 18, 2017, 11:19 AM ]

Several postings within these pages have discussed new observations and insights into nova explosions. The term "nova" derives from 16th century astronomer Tycho Brahe's nova stella - "new star" - description of supernova SN 1572 in Cassiopeia.  Unable to see the progenitor of the explosion, from his perspective it appeared like a bright new star where before there was none. Novae seemed similar - new stars appearing where nothing was visible to the naked eye before.  Classical novae are actually like mini supernova
explosions, but rather than the detonation of an entire star the explosion originates on the surface of a white dwarf that has accreted material from a close companion star - see the postings on U Scorpii and V407 Cygni for further details.  

The "Fastest Nova in Town",  V745 Sco, is a special member of this class.  The white dwarf orbits within the dense wind of its red giant companion from which its strong gravity scavenges the hydrogen fuel that powers thermonuclear outbursts at intervals of about 25 years. One persistent problem in understanding observations of nova explosions has been evidence for distinctly aspherical blasts. Why would an explosion over the surface of a very spherical white dwarf be aspherical? We have pursued this problem for several recent nova events using supercomputer hydrodynamical simulations. 

The key to the form and evolution of a nova explosion turns out to lie in its immediate environment.  The expanding blast wave propagates at different speeds through gas of different density.  The denser the gas the slower it moves, but more dense gas also looks brighter when it has been heated by the blast.  Salvo Orlando, of the Palermo Astronomical Observatory in Sicily, has lead a study based on simulations of the V745 Sco 2014 event to try to understand Doppler shift observations that revealed a very aspherical explosion. Earlier studies have indicated that a white dwarf orbiting in a dense wind tends to attract higher density gas in the plane of its orbit.  The supercomputer simulations confirmed that we cannot reproduce a sufficiently aspherical explosion without this.  The result is a blast and ejecta that shoot out poleward, expanding rapidly northward and southward of the white dwarf. The dense equatorial gas instead lights up more brightly in X-rays and UV light, producing a ring-like emission structure. Careful comparison of different simulations with X-ray observations indicate the explosion threw off about 1/10 of an Earth mass, with an energy equivalent to about 1,000,000,000,000,000,000,000,000,000 tonnes of TNT (4x1043 erg).  This work was published in the 2017 February 1 edition of Monthly Notices of the Royal Astronomical Society and has been featured in a Chandra X-ray Center press release.

Plus ça change...

posted Jan 5, 2017, 9:28 AM by Jeremy Drake   [ updated Jan 5, 2017, 9:28 AM ]

Our nearest stellar neighbour, the aptly monickered Proxima (actually dubbed "Proxima Centurus" by its discoverer, Robert Innes, of the Union Observatory in Johannesburg), shot to stardom recently after having been found to harbour a habitable zone planet with a mass only 30% larger than that of the Earth.  Before this jaunt down the red dwarf carpet, Proxima was merely a conveniently nearby M6 dwarf - a faint, diminutive star with only an eighth of the mass of the Sun and a seventh of its radius. Proxima is cool, in the literal sense - just 3000 K at its surface compared with the Sun's 5800 K.  So cool, in fact, that its internal structure is quite different to that of the Sun.  

The Sun has an outer convection zone taking up the top 30% or so of its radius.  Inside that, its structure is stable and "purely radiative".  It has been thought that the interface between the radiative and convective zones on the Sun is key for its dynamo that generates magnetic fields visible at the surface in the form of sunpots and energetic UV to X-ray emission. Proxima does not have a central radiative zone - it is convective all the way to the centre. So, logically, its dynamo should operate quite differently.  

An earlier posting presented new X-ray evidence that the dynamos of fully convective stars, like Proxima, are instead surprisingly solar-like.  Further evidence has now emerged showing common dynamo action in another way: cyclic behaviour. The solar dynamo has a well-known cycle that results in the surface magnetic field waxing and waning and reversing polarity every 11 years.  SAO scientist Brad Wargelin has lead a team analysing long-term optical, UV and X-ray observations of Proxima that have probed its long-term magnetic behaviour. The visible light data reveal a slow modulation in brightness with a period of seven years. This brightening and dimming is the result of changes in the number of starspots on the stellar surface stemming from a magnetic cycle - in Proxima's case a 7 year one instead of 11.  X-ray emission measured by a collection of different satellites over the years also shows a sympathetic secular variation in phase with the starspots.  The implication of these new observations is that the magnetic fields of all stars with outer convection zones, including the Sun, are generated in the same way within convection zone and do not depend on the presence of a central stable radiative zone.  This work was published in the 2017 January 21 edition of Monthly Notices of the Royal Astronomical Society.

The actually quite large red CME that couldn't

posted Nov 5, 2016, 6:19 AM by Jeremy Drake   [ updated Nov 5, 2016, 6:53 AM ]

Stars like the Sun have strong surface magnetic fields that harbour substantial amounts of energy.  The energy builds up as the fields are twisted and stretched by the turbulence and flows beneath the stellar surface where they are anchored. This energy is released from time to time as the fields interact and snap back into less stressed states.  The energy is dissipated in bursts of X-ray and ultraviolet light and by launching plasma into space at hundreds of kilometers per second.  These phenomena are referred to as flares and coronal mass ejections (CMEs).

A 2013 posting on research examining how much mass stars can lose in coronal mass ejections (CMEs) highlighted a mass and energy "catastrophe" if relations between 
flares and CMEs on the Sun are extrapolated to much more magnetically active stars: the implied energies and masses were simply too high to be reasonable. How is the catastrophe to be averted? 
A failed supercomputer  model of a CME on the vey active young star AB Doradus might provide a clue.  Magnetically active stars have large-scale magnetic fields up to 100 times that of the Sun.  To escape the star, the CME must be ejected with sufficient force to break through the overlying magnetic field canopy.  The CME we simulated had the energy of the largest CMEs seen on the Sun - equivalent to about a hundred quadrillion tons of TNT or about 2 billion of the most powerful nuclear warheads ever made - but it failed to break through the magnetic canopy of AB Dor.  The rather unspectacular movie of the simulation appears to the right. The crisis could be averted because very active stars hang onto their weaker CMEs and recycle the energy in the corona, reducing the mass and energy budgets to more reasonable values.  This work was published in the Proceedings of IAU Symposium 320, held in Hawaii 2015 August.

1-10 of 56