How do they find new planets?
A few days ago, the WASP project (Wide Angle Search for Planets) announced the discovery of their 17th planet called WASP-17b (the star is called WASP-17, WASP-17b is the first planet to be discovered around this star). WASP discovers new planets by using  the "transit method". With this method hundreds of images of the same patch of sky are collected over several nights and by advanced analysis methods slight changes in the brighness in any of the stars are found. The slight dimming of a star could be due to a large planet (Jupiter- or Saturn-sized) passing in front of the star. But qute often the dimming is caused by the transit of a background double star - or if the star is a binary star itself and the eclipse is only partial: a gracing eclipse. The dimming is typically of the order 1-2 percent.

How do they know it's a planet?
To confirm that such a transit is due to a planet it is necessary to make additional measurements to determine the mass of the transiting object. This is done by taking a series of spectrograms (by sending light from the telescope through a prism or a grating). In each spectrogram hundreds of spectral absorption lines are visible. Each line is caused by the absorption of light in the atmosphere of the star due to electronic transitions in the atoms that make up the hot plasma that comprise the atmosphere. Each atomic line has a certain wavelength corresponding to the transition energy. The exact same wavelengths can be measured accurately in laboratories on Earth. However, when studying stellar spectrograms the lines are shifted due to the movement of the star relative to the telescope situated on Earth. After correcting for Earth's rotation and the movement of Earth around the Sun, the movement of the distant star is measured; this is realtive to the barycenter of our solar system. In this way the velocity curve of the star is measured and it typically consists of many individual spectrograms collected on several nights.

The WASP-17b planet has an orbital period of about 3.7 days. By combining the light curve (from images, showing the dimming) with an accurate velocity curve (from spectrograms, showing the movement) scientists can determine the orbit of the planet and the radius and mass of the planet: in this case it is an eccentric orbit (it is not a circular orbit, but an ellipse), the radius is about twice that of Jupiter, and the mass is only half that of Jupiter. In other words it has a much lower density than Jupiter, probably be cause it is relatively hot since it is so close to its host star.

WASP-17b is in a retrograde orbit
The most exciting thing about WASP-17b is that the orbit around its host star is retrograde. That means it is travelling in the opposite direction as the rotation of the central star. This phenomenon is also known for several of the moons in our Solar system. It is thought that this is due to collisions or close encounters of massive objects (other moons or other planets). Therefore scientists have suggested that the retrograde orbit of WASP-17b is due to a relatively recent close encounter. This would also explain the eccentric orbit of the planet, since tidal forces will tend to circularize such orbits, especially when the planet is very close to the star.

How do they know it's a retrograde orbit?
This is done from a detailed analysis of the radial velocity curve. The change in velocity is due to the pull of the planet causing the much larger star to move - but only slightly. This explains the main part of the velocity change, but there is also a more subtle effect changing the velocity, the so-called Rossiter-McLaughlin effect. As the planet moves in front of the star, part of the light will not be visible. Let us imagine that this is the right part of the stellar disc and that 2 percent of the light is therefore not being measured. Let us say the star is rotating from left to right, meaning the left part is coming towards us and the right part is receding. The Doppler effect will cause the light on the left part to be blue shifted (the right part will be red shifted; the light near the centre of the disc will not be shifted). Since 2 percent of the red-shifted light is not being observed the measurements will indicate that the star is coming towards us (relative to the situation outside the eclipse). This Dopper-shift will change during the eclipse - from a blue shift to a red shift. This is exactly the opposite of what is observed in all other known exo-planets and this is one of the things that makes WASP-17b special.

Who are "they" ?
The paper presenting the results can be found here: Anderson et al. 2009.


(0)

Inden for den sidste uge har tre forskellige forskergrupper offentliggjort direkte billeder af planeter omkring andre stjerner. I over 10 år har man indirekte kunnet måle hvordan store Jupiter-lignende planeter trækker i deres "moderstjerne", men i år er det for første gang lykkedes at tage direkte billeder. Dels fra Hubble Rumteleskopet og fra Jorden ved at bruge adaptiv optik med nogle af verdens største optiske teleskoper: VLT og Gemini. Jeg har skrevet lidt mere om det her.

 

Billedet her viser området omkring den sydlige stjerne beta Pictoris (i stjernetegnet "Kunstmaleren") taget med ESOs Very Large Telescope (Copyright ESO 2008). Stjernen beta Picoris findes midt i billedet - og planeten er den lille lysende plet lidt over stjernen til venstre for midten. På det infrarøde billede kan man også se den store skive af støv omkring stjernen, som er resterne fra skabelsen for bare 12 millioner år siden.

(0)

The Japanese Aerospace Exploration Agency (JAXA) recently launched SELENE, a new mission to study the surface and geology of the moon. The Japanese nickname is Kaguya, named after the lunar princess in a Japanese fairytale. It was launched on 14 September 2007 and the first images have now been released, check them out here. More detailed info here.


(0)

Words of wisdom or famous last words? I have used IDL's curvefit.pro in the past, but have run into problems when I want to fix the value of a certain variable--or constrain the value of a variable. Also, the uncertainties of the fitted parameters quoted by curvefit are often very misleading (you may argue I don't know what I'm doing). I have found that assigning point weights seems to sometimes cause problems for curvefit.pro.

Marquardt wrote mpfit.pro, which resolves all these issues. I thought you might need it one day, so here's the programs and a nice tutorial: http://cow.physics.wisc.edu/~craigm/idl/idl.html

(0)

After twelve nights at the Anglo-Australian Telescope Laszlo and I have now surrendered the control room to Paul Butler, who will be searching for Saturn-sized planets for the coming 48 nights. During the past six nights it was clear, but we were not able to observe Procyon right after twilight, so we pointed the telescope at the known binary star Delta Orionis. It is one of the bright stars in Orions belt, which is also visible from the northern hemisphere.  In the picture you can see me pointing at the star in the control room of the AAT.  Note that Orion really appears to be upside down! The data will be combined with photometry from the WIRE satellite to produce the most accurate determination of the physical properties of the two stars comprising the system.



(0)

The weather has finally cleared up at Siding Spring observatory. We have been observing for about eight hours each night from the Anglo-Australian telescope (AAT) near Coonabrabran. The AAT is a 3.9-m telescope and is equipped with the University College of London high-resolution Echelle Spectrograph (UCLES). Using the spectra from UCLES we can measure the tiny oscillations on the surface of the solar-like star Procyon. We are collecting spectra of Procyon from Japan, USA, France, Spain, Chile and Australia. The aim is to get almost continuous observations of the star for a week or more. The measurements of the oscillations in the star will be used to probe the interior of the star in great detail. It is very much like geologists measuring earthquakes to learn about the layers below us. We call it asteroseisomlogy.


(0)

The weather has been horrible for the past five nights. Apparently, out of nowhere a massive system of clouds is forming to the west of the Siding Spring observatory. Although the system is slowly moving towards the coast and sometimes seems to disperse and getting less dense we cannot acquire our target Procyon despite it being a very bright star (V=1). The forecast for the next couple of days seems to be improving.

(0)

We are now at the Anglo-Australian Telescope (AAT) in Coonabarabran. The AAT is reached after a six hours drive north-west of Sydney. We have spent new year's eve up here with Laszlo Kiss and our night assistent Shaun James. We'll be here for about two weeks enjoying the starry nights and beautiful landscape. There are lots of kangaroos and wallabies jumping around the Siding Spring observatory. Yesterday we went bush walking to look for koalas, but didn't see any. However, we saw two huge emus walking slowly down the road.



(0)

In the spring of 2005 Frank Grundahl and myself attended a workshop in Uppsala where we learned how to reduce Échelle spectra. The REDUCE code is written mostly in IDL by Nikolai Piskunov (Uppsala) and Jeff Valenti (STScI). I have successfully reduced spectra from UVES@UVES and achieved S/N ratios that were 30% higher than a reduction that Jens Viggo Clausen made with IRAF.

This week I have been reducing new data from the FEROS@ESO-2.2m spectrograph. Everything has run smoothly except for the wavelength calibration. I thought I'd share some information here that other people may use in their reduction of FEROS spectra.

The first thing I did was to find the approximate wavelength range of one Échelle order. I did this by identifying the water vapor lines in the spectra. There are several groups but I picked the ones around 6275 ångström in relative order number 25 (the order just next to the one containing H-alpha at 6562.79 ångström). There are also two strong Fe lines in the order that I used to find the approximate scale (ångström/pixel) in order to infer the wavelength range in that spectral order. In the program wavecal.pro I could not immediately recognize the lines in Thorium/Argon (Th/Ar) atlas. I then swapped the input array so wavelength increased from left to right instead, and then I could recognize the Th/Ar lines. I was then able to improve on the wavelength scale and go on to the next Échelle order.

My next problem was that I couldn't get the 2D solution to work even after fitting 10 orders manually with a 1D solution. The orders that I had fitted 1D solutions to manually all had residuals of exactly 0.0 m/s in the 2D fit. I think the problem was that I didn't supply wavecal.pro with the right absolute Échelle order. However, I was able to deduce it (within ±1 order) from the documentation on FEROS.

The image below is a screenshot from the wavecal.pro widget (right click to get full resolution image). I have indicated the absolute order with yellow text and also the wavelength range it covers (accurate within ±0.5 ångström). The red arrows mark the boxes where you have to type the absolute Échelle order of the base order (ie. relative order = 0). I may add, as a footnote, that there are more than 37 Échelle orders on the CCD, any you may be able to extract up to 39 or even 40 (partial) orders with the REDUCE program. However, the S/N in the orders at the top/bottom of the CCD will be quite low, so I don't recommend to bother trying to include them.




(0)

As you may now I spend most of my time analysing data from the star tracker on the WIRE satellite. This is the latest discovery with WIRE:

In June 2004 I discovered that the bright southern star Psi Centauri was a new eclipsing binary star. The period of the two orbiting stars was 38.8 days. You may read the scientific paper for more details.

Just a few months ago WIRE observed Psi Cen again at the time when we predicted that the primary eclipse would occur again. In the picture below you can see the data from 2004 (blue points) and 2006 (yellow points). The eclipse occurs when the faintest of the two stars passes in front of the brightest star thus blocking out about 30% of the total light. The slight offset in time (or phase) between the two light curves is there because we have a small uncertainty on the period of the system of about 50 seconds (relative to the 38.8 day period that's not a lot). With the new data we can improve the period to within just 3 seconds.

More interestingly we have just received some new spectroscopic data. This will enable us to measure the movement of the two stars relative to each other. From this we can determine the absolute masses and radii of the stars with very high accuracy - perhaps as good as half a percent. This result will put constrains on the theoretical models of the two late B / early A-type stars (around B9 and A2) that make up the Psi Cen binary system.



(0)

Today I got the exciting news that I will be observing with the Anglo-Australian Telescope. It is the largest telescope on the Australian continent with a 3.9 meter mirror. It is found at the Siding Springs observatory about 500 km north of Sydney. I will be observing the solar-like star Procyon. It is a star slightly more evolved that our Sun, i.e. it is an evolutionary stage between the hydrogen core burning phase (as our Sun) and a red giant star.

Laszlo L. Kiss and I will carry out the observations during 12 nights starting on the 29th of December 2006. That means we will be celebrating new years eve at the telescope! Up to five other telescope around the world will be pointed towards Procyon at the same time so will hopefully be able to measure the pulsations of the star with nearly 24 hour coverage every day for about two weeks.

The Anglo-Australian telescope

(0)

The planet Mercury is the smallest planet in our solar system. It is also the planet that is closest to the Sun with an orbital period of 88 Earth days. The diameter of Mercury (4880 km) is about 40% that of the Earth and it is only 40% larger than our Moon. About 10-15 times every century Mercury passes in front of the Sun and we can see a partial eclipse (a transit) of the Sun. The transit is very tiny since the Sun is so immense in comparison, and thus it requires a small telescope to see it.


(0)

In 2001-2 I was a PhD student in Vienna in the astronomy group of Werner Weiss. At that time I started developing a program for semi-automatic abundance analysis of the stellar spectra of A-G type main sequence stars. During the past four years the program has been improved and today you can download it from the VWA webpage.

About two years ago I promised Conny Aerts (Leuven, Belgium) to perform an abundance analysis of a sample of Gamma Doradus stars. They are A-F type stars found to be slightly cooler than the "red edge" of the classical instability strip. In other words, typical surface temperatures on these stars are 6500-7000 Kelvin, ie. somewhat hotter than the Sun. The first paper on the analysis of the spectra has been completed by Peter De Cat. At the moment I am writing the second paper on the abundance analysis of the Gamma Dor stars.

So why am I doing this? Well, the surface abundance is an important parameter when doing detailed modelling of pulsating variable stars.


(0)

I am currently attending the workshop "the Future of Asteroseismology" in Vienna, Austria. Around 120 astronomers, mainly interested in pulsating stars, are here for three days. The most recent results on various classes of variable stars have been presented during the past days from quakes in the helium rich atmospheres of white dwarfs to 5-minute solar-like oscillations in "normal" (=G-type MS) stars.

This afternoon I presented results from the star tracker on the WIRE satellite. The high-precision photometry from WIRE has been used to probe delta scuti stars (about twice the mass of the sun) and more recently to take the measure of the size and masses of stars in detached eclipsing binaries. Some of the stars in these binary systems are also pulsating (ie. essentially "star quakes") which can in principle be used measure what goes on inside the stars.

The future of asteroseismology will be exciting. New space missions like COROT (launch in December 2006) and Kepler (October 2008?) will reveal unprecedented details on the interior of the stars.


(1)

Malene and I are preparing our relocation from Copenhagen (Denmark) to Sydney (Australia). Today I was looking through a lot of papers and books at my office at the Niels Bohr Institute. A lot of ancient knowledge is buried in these dusty papers and a lot of wise thoughts have been written down in these forgotten tomes. For example, I found the notes I made for a lecture on "the twin paradox" in 1993. I think I discarded that. I found my notes for the MONS workshop held in Aarhus in 2000 (a proposal for a Danish asteroseismology satellite) and the first COROT science workshop (a French-European satellite). I kept these notes.

Also, I came across the master's thesis by Jakob Vinter from Aarhus. On the first page was a very intersting quote from one Harlow Shapley (1885-1972): "A hypothesis or theory is clear, decisive and positive, but it is believed by no one but the man who created it. Experimental findings, on the other hand, are messy, inexact things which are believed by everyone except the man who did that work." It really sums up how I sometimes feel when working as an observational astrophysisist.



(0)

Yesterday, around 500 astronomers at the 26th General Assembly in Prague voted for a new definition of the celestial bodies in our solar system. There are three distinct kinds of objects in our solar system (excluding the satellites or moons). The three kinds of objects are:

1. Planets
2. Dwarf planets
3. Small solar-system bodies

In group (1) we find the four terristial planets (they have a solid surface: Mercury, Venus, Earth, and Mars) and the four gas planets (Jupiter, Saturn, Uranus, and Neptune). In group (2) we find dwarf planets like Ceres, Pluto (with its moon Charon), and 2003 UB₃₁₃. Group (3) comprises the comets, asteroids, Trojans (mainly sharing the orbit of Jupiter but also four are currently known in the orbit of Neptune), and a multitude of Kuiper belt objects (some are very likely "dwarf planets").

As a new member of the IAU I voted for the new resolution 5A+B, thus degrading Pluto to being a dwarf planet. I am sure that in the near future we will have a new resolution that will comprise all known planets -- not just the planets in our solar system. Today almost 200 such exo-planets are known with masses from several Jupiters to a few times the Earth. Already in November 2006 the French/European satellite mission COROT will be launched. The aim of COROT is to discover new "super-Earths" around nearby stars. In 2008 the NASA Kepler satellite mission will fly and it is expected that several true Earth-like planets will be discovered.

I am at the International Astronomical Union's 26th General Assembly in Prague. I have attended sessions on convection, binary stars and a joint discussion on asteroseismology. It's very inspiring to hear fellow researchers talk about their favorite stars. I fear my poster on photometry of binary stars from the WIRE satellite has only attracted little attention. At least I can see that there are only few copies left of my latest paper on the newly discovered binary star Psi Centauri. I have discussed writing up a paper on the North star (Polaris) with Nancy Evans (her pet Cepheid star) so that is likely to happen soon. More on the status of our solar system tomorrow... the fate of Pluto will be decided on the general assembly meeting in the afternoon.



(0)

I have managed to produce a nice A0 poster for the International Astronomical Union (IAU) meeting in Prague next week. The title is "Eclipsing Binaries from Space". I have used photometry from WIRE to monitor three bright detached eclipsing binaries (dEBs). These "double stars" are interesting because they can be used as a way to determine the relative radii of the stars with high accuracy (better than 1% -- that's difficult to achieve for single stars). Furthermore, my fellow astronomer John Southworth will collect new spectroscopy of the systems and we will then be able to determine absolute radii and masses. These values can be compared directly to theoretical models for single stars. Such models are important in many key areas of astronomy. For example they are used constrain the age of stellar clusters and the properties of distant galaxies. In total we expect to collect WIRE data (thanks to Derek L. Buzasi at USAFA and the WIRE team!) for 10-12 dEBs during the coming year.

(0)

The Planet Definition Commitee of the IAU has proposed a new definition of what a planet is: it is an object orbiting a star (but is not itself a star) that has a strong enough gravity that it will be shaped like a sphere. A moon is an object orbiting a planet, and the center of gravity must lie inside the host planet. Today we know 12 planets in our solar system that agree with the new definition. They are Mercury, Venus, Mars, Earth, Ceres (located in the asteroid belt), Jupiter, Saturn, Uranus, Neptune, Pluto+Charon (a double planet system), and "2003 UB313" (found in the Kuiper belt). More planets in the Kuiper belt surrounding our solar system are likely to be discovered in the coming years.


(0)