HD 80606b transit detected

I’m very pleased to be able to announce that HD 80606b is a transiting planet!

It looks like priority of discovery goes Claire Moutou and the French and Swiss team, who beat at least one other team to submission by a matter of hours. I’m attaching a draft of the French and Swiss Team’s paper that was just sent to me. Congratulations to Everyone involved!

Here’s the preprint.

Details to follow…

go

Image Source: Mearth Live.

Update 4 : Feb. 14 2009, 07:12:00 UT

The first reports are coming in. Gregor Srdoc in Croatia got a lightcurve through most of the night for HD 80606 combined with HD 80607. No sign of a transit, but the data is relatively noisy due to imperfect weather.

Veli-Pekka Hentunen reports that weather conditions in Finland were bad generally, and were specifically bad in Varkaus.

At least four sets of observations from various locations in Arizona are currently underway, including both the 40” and the 1.3m at USNO Flagstaff under the able command of Paul Shankland.

Jonathan Irwin reports that data from Mearth through 5 UT shows no sign of an egress.

Ohio State Grad Student Jason Eastman reports on his remote Demonex observations (from the comments page):

Halfway through the night…

We started observing at UT 02:30 in the V band. No sign of an egress at the ~0.005 mag level.

http://www.astronomy.ohio-state.edu/~jdeast/demonex/HD80606b.R.2009-02-14.jpg

That link will be updated with the entire night’s data in the morning.

So it’s not looking particularly good for a transit, but I’m really happy that data is coming in. We’ll have a definitive answer sometime tomorrow.

Thanks to everyone who observed. It’s really cool how a planet 190 light years away can bring observers all over the globe into a common mission.

Update 3 : Feb. 13 2009, 23:29:00 UT

We’re now closing in on the moment of inferior conjunction, which hopefully will wind up being the midpoint of a central transit. The current weather in Europe looks like it’s clear for observers in Finland and Northern Italy, so it’s now quite likely that we’ll get a definitive answer from the campaign.

No word yet on whether an ingress was observed, but Jonathan Irwin did send a nice light curve from last night’s baseline run with Mearth. He writes:

Here’s our entire night of data (about 11 hours) from one telescope, using 80607 as the comparison star. Raw and binned x12 (about 5 minutes per bin). We are getting rms scatter of about 1.6 times Poisson with this fairly quick reduction.

There is usually a slight offset when the target crosses the meridian (data point 777) due to flat-fielding error, that I have not removed in this – over the ~20 arcsec separation of the pair it’s pretty small. There is also a bit of a blip there as my guide loop recovers its lock after crossing – still needs a little tuning :)

Fingers crossed for tonight!

Update: Clear Skies in Arizona. Dave Charbonneau writes:

http://mearth.sao.arizona.edu/live/

Clear skies. You can even watch the images in real time, and see how many
MEarth scopes are on ‘606…

Update 2 : Feb. 13 2009, 17:04:00 UT

It’s now the middle of the night in the Far East, and the transit window has opened. The weather in Japan looks a little spotty, but Southern China is in the clear.

Observers in Arizona reported good weather last night, but the forecast is a little iffy for tonight.

In addition, I just got an e-mail (UT 17:48) from Gregor Srdoc in Croatia, who is on the sky under quite good conditions just after nightfall…

Update 1 : Feb. 13 2009, 06:03:03 UT

There’s about a half-day left until the possible start of the ingress. On the map above, I’ve marked the locations of confirmed observers with small red dots. HD 80606b is 190 light years above the spot labeled with the orange circle.

Observers in the US are currently taking data of both HD 80606 and its binary companion, HD 80607. It’s always good to have an out-of-transit baseline photometric time series.

Dave Charbonneau checked in with a status report:

MEarth is ready. You can watch us in real time at
http://mearth.sao.arizona.edu/live/

If the roof is closed, it is cloudy.

The up-to-the-minute stop-action animations showing the disconcertingly reptilian movements of the telescopes are completely mesmerizing. Mearth (pronounced “mirth”) is located at the Fred Lawrence Whipple Observatory on Mt. Hopkins in Arizona, and spends most of its nights searching for potentially habitable terrestrial planets transiting nearby M dwarfs. The telescopes have a list of ~2000 nearby red dwarf stars. Each star is subjected to repeated visits of ~30-45 minute duration. The idea is to catch transiting planets in progress and to broadcast the information to larger telescopes that can obtain immediate real-time photometric confirmation of a discovery. (For a more detailed overview of Mearth, see Irwin, Charbonneau, Nutzmann & Falco 2008.)

Update 0 : Feb. 12 2009, 22:47:40 UT

I’ll be posting updates on the global HD 80606b transit campaign as I get them, with newer updates going to the top of this post.

A number of observers have indicated that they’ll be on the sky. Right now, it looks like telescopes are confirmed for Finland, Israel, Italy, Japan and the US. Given the vagaries of the weather, however, it would be great if we can get as much coverage as possible. As Vince Lombardi would have put it, “We’re looking at 15%, so if you can get 1%, get out there and give 110%!”

Everyone is encouraged to comment as the campaign progresses (click the number next to the post title to access the discussion page). I’ve lifted the restriction that only allows registered oklo users to comment, but all comments are now held for moderation, in order to keep the Viagra contingent off the air.

ready set…

Even as I write, HD 80606b is closing in fast on its inferior conjunction. It’s basically a roll of a die, a roughly one in 6 chance, that the orbital alignment is good enough for a primary transit to be observable. (The odds are boosted from the a-priori geometric probability of 11% to ~15% by the fact that the secondary transit was fully consistent with an uninclined passage directly behind the star.)

Here’s the situation:

A central transit will last roughly 16 hours, with the ingress best suited for observers in the Far East, and the egress best suited for observers in North America. Europe is the place to be for transits that are closer to grazing. HD 80606 itself is favorably sited in Ursa Majoris, and is at low air mass for basically the entire night, especially at higher latitudes.

Good luck to everyone who’ll be observing!

What’s your angle?

P. Diddy flossing his '606 pose.
Happy ‘606 day!

HD 80606b swung through periastron at about 01:40 UT this morning (Feb. 8, 2009) and will spend the balance of the week spinning out toward inferior conjunction, which will occur at 00:50 UT on Valentine’s day (Feb. 14th).

Proposals for GO-6, the first general observing cycle of the forthcoming Spitzer Warm Mission, were due on Friday. Jonathan and Drake and I worked right down to the 5 PM PST wire, polishing our request to complement the Nov. 2007 8-micron periastron observations with a pair of additional photometric time series at 4.5 microns (Warm Spitzer’s longest IR wavelength). Two HD 80606b events are observable during GO-6; the first at the very start of the warm mission on May 30, 2009, and the second on Jan 08, 2010. We’re keen to watch the planet ring down from its maximum brightness, so we’ve proposed for a window that runs from 10 hours before periastron to 30 hours after periastron. In the 4.5 micron bandpass, we’re predicting a maximum planet-to-star flux ratio of a bit more than one part in a thousand — easily within Spitzer’s sensitivity.

Here’s a diagram showing the portion of the orbit that we’re proposing to observe. Even though the orbital period is 111.43 days, our forty-hour proposed observation encompasses more than 200 degrees of true anomaly. A planet with e=0.932 is quite truly anomalous.

In the near term, though, I’m very eager to see what shows up in my inbox on Valentine’s day, when observers across the Northern Hemisphere will be monitoring HD 80606 to ascertain whether a primary transit for the planet can be observed.

Here’s the geometric situation. If HD 80606’s orbit were inclined only negligibly to the sky plane, then Earth’s view of the system would be a simple reflection of the standard diagram. At inferior conjunction, six days after periastron, the planet is heading away from the star and slightly toward Earth:

The occurrence of the secondary transit tells us, however, that the orbital inclination relative to the sky plane is in reality close to 90 degrees. Using the Illustrator scale tool to compress along the north-south direction, we can see the result of increasing the inclination.


“Sooner can a camel thread the eye of a needle…”

CoRoT-7b

From a cell phone picture transmitted by an oklo.org agent

The photo above is grainy, but what’s truly remarkable is that the depth of the dip is only 0.03%. Earth transiting in front of the Sun as seen from afar blocks roughly 0.01% of the Sun’s light. Look at the signal-to-noise of the bottom composite-average curve.

I can sure empathize with the CoRoT team. Their symposium date was set up long ago. Kepler is launching in a few weeks. The results of the Doppler surveys are suggesting that super-Earths with orbital periods of 50 days or less (with correspondingly high transit probabilities) are present around 30% of solar-type stars. Ground-based photometry is pushing below 0.5 millimagnitudes at 1-minute cadence. The pressure is on. And there’s an absolutely fascinating candidate planet that isn’t quite yet out of the oven, due to a paucity of high-precision radial velocities that would pin down the mass. What do you do?

I agree! You go ahead and announce.

Everything about CoRoT-7b reemphasizes the fact that planets are wont to turn up in every corner of parameter space to which observations are sensitive. In this case, a V=11 K0V star in the direction of the galactic anti-center displays 176 individual 1.5-hour 0.3 mmag photometric dips with a strict 0.854 day periodicity. These measurements suggest a 1.7 Earth-radii planet with a 20-hour year — a world that makes 51 Peg b look like Fargo North Dakota.

The abstract for Daniel Rouan’s talk at the meeting (transcribed from the cell phone photograph) describes the procedures that the CoRoT team has implemented to rule out the various false positives that can plague transit surveys. This gives a sense of the amount of follow-up work that needs to be done in order to secure a planet as small as this one (also, see comments section for this post, for many additional details):

To qualify/falsify the interpretation of the observed transits, we have considered different alternative interpretations: (1) transit of a main sequence star in front of a giant star — rejected by the measured log(g) of the target; (2) a grazing eclipse by a stellar companion — rejected by the radial velocity measurements (3) a weak Background Eclipsing Binary that would be inside the target mask — partially rejected by on/off transit photometric observations performed from the ground at angular distance from the target larger than 2 arcsec, and by high-resolution imaging at distances larger than 0.3 arcsec. (4) a triple system made of the target star (K0V) and a faint star (M5V) eclipsed by a giant planet or a dark stellar companion — rejected by the study of the transit colours which are the same as those of the main target.

Exoplanet.eu is quoting a significantly uncertain mass of 0.035 Jupiter masses (11 Mearth) for the planet, a figure that could have been arrived at via assumptions about the density and/or limits on the radial velocity detection. An 11 Earth-mass planet would induce an eminently detectable K=8 m/s RV signal, so it’s a bit odd that a firmer estimate of the mass isn’t available yet. The CoRoT “galactic anticenter” field is located in Monoceros, at RA~06h 45m, DEC~+0d, meaning that the candidate star is currently visible to HARPS (at air mass <2) all through the first half of the night. Exoplanet.eu also states an age of 1.1 Gyr for the star, so youth, with its attendant stellar activity, could possibly be making it tough to get good velocity precision.

In any case, it’s a remarkable detection, and will be hugely influential as soon as the mass is confirmed. The planet is orbiting at only four stellar radii — with the star filling nearly a thousand square degrees of sky…

The Big Swing

Image from computer modeling by J. Langton and D. Kasen.

HD 80606b — everyone’s second-favorite planet — is in the news! Our article describing the Spitzer Space Telescope’s 8-micron observations of the planet’s periastron passage made the cover of this week’s issue of Nature, and JPL has issued a press release on the results.

The planet has been a long-running topic here at oklo.org, with the storyline developing over a series of posts during the past few years. A incomplete list might include:

Post one (older), two, three, four, five, though six (newer).

The outsize eccentricity of HD 80606b’s orbit leads to very brief, very intense encounters every 111.4 days as the planet swings through periastron. On the Nov. 20, 2007 encounter, we used Spitzer to monitor the 8-micron emission of the star and planet for a thirty hour period. The observations spanned the time leading up to superior conjunction and periastron, and continued for several hours thereafter:

The resulting time series looks like this:

The most remarkable feature of the light curve is the dip at time 2454424.72. The alignment of the planetary orbit turns out to be close enough to edge-on that a secondary eclipse occurs. The a-priori chance of observing the eclipse was only about 15%, and so we were lucky. Our interpretation of the light curve is that we’re seeing the planet heat up rapidly, from a temperature of roughly 800K to a temperature of about 1500K over a time period lasting roughly five or six hours. This indicates that the starlight is being absorbed at quite a high level in the atmosphere, where the air is thin and the heat capacity is low.

The details are all in the Nature paper. I’ll be posting it on astro-ph shortly, but in the meantime, a .pdf draft of the article is here, along with the (quite extensive) supplemental information section, and the figures (one and two) from the article.

The information that comes directly from Spitzer amounts to a 30-hour, one-pixel grayscale movie of a storm that was brewing on the planet back in the Monroe Administration. Hydrodynamical modeling, however, can flesh out the details, and the goal over the coming years will be to compute simulations that are as detailed and as physically correct as possible. In the next post, I’ll go into more detail, but here’s an advance look at the results of a “synthetic mission” in which a probe has been inserted into orbit around the planet 2.2 days prior to periastron. The resulting footage runs through 8.9 days after periastron. The orbital dynamics and the illumination are all self-consistent…

Footage from a synthetic probe.

WASP-12b

WASP-12b. Now there’s an unpleasant travel destination.

Nevertheless, this particular planet, whose transits were recently announced by the SuperWASP collaboration, is quite a remarkable world. For starters, inveterate bottle-poppers can celebrate a WASP-12b New Year on literally nine out of every ten days — the orbital period is a mere 26 hours and 11 minutes. The temperature of the planetary photosphere at the substellar point likely exceeds 2500K. Cherry orange, to be exact.

Because of its ultra-short orbital period, WASP-12b is attracting quite a bit of interest. The planet has a radius 1.8x larger than Jupiter, which should make it eminently feasible to detect secondary transits from the ground in either the optical or near-infrared. One expects, furthermore, that a planet with an orbital period just a shade over a day should have long since damped out its eccentricity, but (to better than 2-sigma) the orbit appears to be non-circular, with e=0.049 +/- 0.015. Even if another planet exists in the system, there should long since have been evolution to a tidal fixed point, followed by circularization. If the orbit really is eccentric, then GR precession of the periastron amounts to a whopping 0.2 degrees per year, nearly 2000x faster than Mercury’s stately 43” per century.

I got an opportunity to visit Harvard this month, and while I was there, David Latham remarked that he had used a remotely operated telescope in Arizona to get a high-precision light curve of a WASP-12b transit. Latham is a meticulous observer, and so, in order to get the best possible baseline, he had cued up the telescope a number of hours prior to the predicted ingress. He related that he’d been completely startled to find, upon analyzing his photometry, that the transit had occurred several hours ahead of schedule. Without a doubt, transit timing variations are going to be one of the big exoplanet stories of 2009, but they’re going to be measured in seconds, not hours. Imagine the commotion that would result if the Sun rises a few hours late tomorrow morning!

The WASP-12 mystery was solved by the amateur astronomers Veli-Pekka Hentunen and Markku Nissinen of Taurus Hill Observatory near Varkaus, Finland. Bruce Gary, who runs the Amateur Exoplanet Archive forwarded the news of their work:

AXA contributors and TransitingPlanets members,

I just received two data files for WASP-12 as observed by Veli-Pekka Hentunen and Markku Nissinen (Finland) which suggest that the discovery paper for this exoplanet has a misprint for the ephemeris. Their observations on January 1 was a “no show” (attached) whereas their observations on January 4 had a nice transit (attached). According to the discovery paper’s ephemeris there should have been a transit on January 1 but not on January 4. However, the discovery paper has a discrepancy between the stated ephemeris and the stated HJD for WASP survey observations. The Hentunen and Nissinen observations can be explained if the discovery paper’s stated WASP survey HJD is correct and their HJDo has a number transposition, such that HJDo = 4506.7961 (instead of 4506.9761). This is described on the AXA web page for WASP-12: http://brucegary.net/AXA/WASP12/wasp12.htm

[…]

We amateurs have to keep the pro’s honest! Nice work, Veli-Pekka Hentunen and Markku Nissinen.

Bruce L. Gary, webmaster
Amateur Exoplanet Archive

Indeed! The typographical error in the discovery ephemeris has now been corrected, and with it, the puzzling “early” transit was revealed to be a completely separate event in the unending sequence of near-daily occultations. It seems somehow fitting that a seemingly alarming discrepancy for the hottest planet known was resolved by a pair of dedicated amateur observers during the long, dark, and frozen Finnish nights.

HAT found a Neptune,

and at 880K it’s close to ten times hotter (but likely the same color) as the original edition.

In the twenty months following Gillon et al.’s startling discovery that Gliese 436b is observable in transit, literally dozens of additional transiting planets have been found. New transiting hot Jupiters are now routine enough that they’re generally trotted out in batches. Reported cases of transit fever have also been on the decline, with symptoms often amounting to little more than a passing distraction.

That said, it’s been been a very long dry spell waiting for a second example of a transiting Neptune-mass planet, which makes HAT-P-11b both exciting and newsworthy. In a preprint that muscled its way to the top of today’s astro-ph mailing, Gaspar Bakos and collaborators have produced a admirably solid analysis of what’s definitely the toughest ground-based detection to date.

HAT-P-11b’s transit depth is 4.2 millimag, which is the smallest planet-produced dip yet detected by a photometric survey. (HD 149026b has a smaller transit depth, but it was discovered via the Doppler velocity method and then followed up photometrically for the transits during the time windows predicted by the orbital solution.) The HAT-P-11b analysis was further confounded by a photometrically variable parent star and ~5m/s stellar jitter on the radial velocity observations. The paper is definitely worth reading carefully.

HAT-P-11b is quite similar in mass and radius to Gliese 436b, and it’s actually somewhat larger than Neptune on both counts. When the mass and radius are compared to theoretical models, it’s clear that, like Gliese 436, it’s mostly made of heavy elements (that is, some combination of metal, rock and “ice”) with an envelope of roughly 3 Earth masses of hydrogen and helium). It’s completely dwarfed when placed next to an inflated hot Jupiter, HAT-P-9b, for instance:

Interestingly, HAT-P-11b seems to have a significant eccentricity, on the order of e=0.2. Drawn to scale with the parent star, the orbit looks like this:

The dots demarcating the orbit are not to scale. With 500 pixels of resolution, you can just barely see the planet. (I put one in front of the star, and tacked a copy onto the orbit for good measure.)

The e=0.15 eccentricity of Gliese 436b has caused a lot of consternation. For any reasonable value of the so-called tidal quality factor, Q, the circularization timescale for Gliese 436b’s orbit is considerably shorter than the age of the system. This has led to attempts (to date unfulfilled) to locate Gliese 436c. HAT-P-11b doesn’t have this problem. For a given Q, it’s circularization timescale is a full thirty times longer than that of 436b. The orbit will still be measurably eccentric even when the 0.8 solar mass primary starts to turn into a red giant.

an interesting development

Last night, I received a mysterious e-mail from Gaspar Bakos of the Harvard-Smithsonian CfA. It consisted of a single line:

A phaeton tuned fun

Now I certainly wouldn’t want to detune anyone’s fun, so I’m turning the comments page off…