VB 10 b - disproven astrometric planet detection claim

Infrared radial velocities of vB 10
http://arxiv.org/abs/0908.3751

Abstract wrote:

We present radial velocities of the M8V-type, very low-mass star vB 10 that have been obtained at four different epochs of observations between 2001 and 2008. We use high-resolution (R ~ 20,000) near-infrared (J-band) spectra taken with the NIRSPEC instrument on the Keck II telescope. Our data suggest that vB 10 shows radial velocity variability with an amplitude of ~1 km/s, a result that is consistent with the recent finding of a massive planet companion around the star by Pravdo & Shaklan (2009). More velocity measurements and a better sampling of the orbital phase are required to precisely constrain the orbital parameters and the individual masses of the pair.

Infrared radial velocities? Nice stuff... if the infrared radial velocity measurements can reach comparable precision to the visible wavelengths stuff, things start to look promising for detecting planets around late-type stars of spectral types M and L.

For example the star 55 Cancri B has a V-band magnitude of 13.15, which is very dim, not good for making radial velocity measurements. In the J-band it is a much more promising 8.560, and in the K-band is 7.666. And with 55 Cancri A harbouring so many massive planets, you've got to wonder what's in orbit around the companion.

I've always wondered about 55 Cnc B myself as well. Maybe 55 Cnc B will make as much a mockery of planet formation theories as Gliese 876.

The CRIRES Search for Planets Around the Lowest-Mass Stars. II. No Giant Planet Orbiting VB10
Authors:
Jacob L. Bean,
Andreas Seifahrt,
Henrik Hartman,
Hampus Nilsson,
Ansgar Reiners,
Stefan Dreizler,
Todd J. Henry,
Guenter Wiedemann

http://arxiv.org/abs/0912.0003

Quote :

We present radial velocities of the very low-mass star VB10 obtained over a
time span of 0.61 yr as part of an ongoing search for planets around stars at
the end of the main sequence. The radial velocities were measured from
high-resolution near-infrared spectra obtained using the CRIRES instrument on
the VLT with an ammonia gas cell. The typical internal precision of the
measurements is 10 m/s. These data do not exhibit significant variability and
are essentially constant at a level consistent with the measurement
uncertainties. Therefore, we do not detect the radial velocity variations of
VB10 expected due to the presence of an orbiting giant planet similar to that
recently proposed by Pravdo and Shaklan based on apparent astrometric
perturbations. In addition, we do not confirm the ~1 km/s radial velocity
variability of the star tentatively detected by Zapatero Osorio and colleagues
with lower precision measurements. Our measurements rule out planets with M > 3
M_Jup and the orbital period and inclination suggested by Pravdo and Shaklan at
better than 5 sigma confidence. Planets with masses down to 1 M_Jup would need
to have unusually large orbital eccentricities (e > 0.7) and be phased in a
very specific way to have eluded detection with our data. We conclude that the
planet detection claimed by Pravdo and Shaklan is spurious on the basis of
these results. Although the outcome of this work is a non-detection, it
illustrates the potential of using ammonia cell radial velocities to detect
planets around very low-mass stars.

The first astrometric planet lived only six months Take off your hats

But if seriously, this big blow on astrometric project STEPS (Stellar Planet Survey).

Ahh darn.

Borislav wrote:

The first astrometric planet lived only six months Take off your hats

Haha ouch. That does hurt. Astrometry's just cursed I guess. Lay VB 10 b to rest right next to Barnard's Star b.

Curse van de Kamp ?

Borislav wrote:

Curse van de Kamp ?

Probably
The Extrasolar Planets Encyclopaedia did not move VB10 b to unconfirmed. Perhaps this is because the authors of the paper leave open the (slight) possibility that the planet's orbit is modulated to the phase coverage and highly eccentric, so as to avoid detection through incomplete coverage of the orbit.

The curse of van de Kamp... I like that

if the planet orbits in the plan of the sky? She would not have been able to be discerned by this method. Why to imagine an eccentric orbit while the incline is sufficient?

Stalker: the astrometric measurements suggested an inclination close to edge-on (96.9 degrees).

Isn't Lalande 21185 still in doubt.

Not quite dis-proven.

The first rule of planet discovery by astrometry: there are no planets discovered by astrometry.
The second rule of planet discovery by astrometry: there ARE NO PLANETS DISCOVERED BY ASTROMETRY.
The third rule of planet discovery by astrometry: if someone says "let me try to confirm that planet" or points a telescope in that general direction, the planet's existence is over.
The fourth rule of planet discovery by astrometry: only two teams of astronomers are needed.
The fifth rule of planet discovery by astrometry: one team to make the discovery, one team to disprove it.
The sixth rule of planet discovery by astrometry: no gas giants, no terrestrials.
The seventh rule of planet discovery by astrometry: planet disproof will take as long as it has to.
The eighth rule of planet discovery by astrometry: if this is the first planet discovered by astrometry, IT WILL BE DISPROVEN.

Haha that is awesome Ctrl+S'ed that to my hard drive.

Moved to retracted planets. Sob!

Never enough nails in the coffin.

Strong Constraints to the Putative Planet Candidate around VB 10 using Doppler spectroscopy
http://arxiv.org/abs/1001.0043

Abstract wrote:

We present new radial velocity measurements of the ultra-cool dwarf VB 10, which was recently announced to host a giant planet detected with astrometry. The new observations were obtained using optical spectrographs(MIKE/Magellan and ESPaDOnS/CHFT) and cover a 63% of the reported period of 270 days. We apply Least-squares periodograms to identify the most significant signals and evaluate their corresponding False Alarm Probabilities. We show that this method is the proper generalization to astrometric data because (1) it mitigates the coupling of the orbital parameters with the parallax and proper motion, and (2) it permits a direct generalization to include non-linear Keplerian parameters in a combined fit to astrometry and radial velocity data. In fact, our analysis of the astrometry alone uncovers the reported 270 d period and an even stronger signal at 50 days. We estimate the uncertainties in the parameters using a Markov Chain Monte Carlo approach. The nominal precision of the new Doppler measurements is about 150 s$^{-1}$ while their standard deviation is 250 ms$^{-1}$. However, the best fit solutions still have RMS of 200 ms$^{-1}$ indicating that the excess in variability is due to uncontrolled systematic errors rather than the candidate companions detected in the astrometry. Although the new data alone cannot rule-out the presence of a candidate, when combined with published radial velocity measurements, the False Alarm Probabilities of the best solutions grow to unacceptable levels strongly suggesting that the observed astrometric wobble is not due to an unseen companion.

And now if using radial velocities to rule out VB 10b wasn't enough, here's some astrometry that also doesn't support its existence.

Astrometric search for a planet around VB 10

Quote :

Taking all the uncertainties into account and using Monte-Carlo resampling of the data, we are able to reject the existence of VB 10b with the announced mass of 6.4 M_J with the false alarm probability of only 0.0005. A 3.2 M_J planet is also rejected with a false alarm probability of 0.023.