Home > worlds > HAT-P-13: good news and bad news

HAT-P-13: good news and bad news

March 31st, 2010

So when presented with that particular formulation, I generally prefer to get the bad news first:

Stefano Meschiari and I have investigated how the new radial velocity data for the HAT-P-13 system affect the possibility of measuring transit timing variations for the short-period planet “b” as the heavy, long-period planet “c” rumbles through its periastron passage later this spring.

First, recall the overall set-up. HAT-P-13 was discovered in transit by Gaspar Bakos and his HAT Net collaborators last summer. HAT-P-13 “b” is a standard-issue hot Jupiter with 0.85 Jupiter masses and a fleeting 2.916-day orbital period. The radial velocity follow-up indicated that the system also contains an Msin(i)~14.5 Jupiter mass object on an eccentric orbit with a P~430 day period. If the two planets are close to coplanar, then the system should have tidally evolved to an eccentricity fixed point — a configuration that allows one to extract Juno-mission style interior information from the inner planet for free.

System Version 1.0 for HAT-P-13 generates significant transit timing variations for the inner planet during the weeks surrounding the periastron passage of the outer planet. In a post two weeks ago, I showed some invigorating calculations by Matthew Payne and Eric Ford, which charted the details of the timing variations. Here’s a figure inspired by the Payne-Ford analysis that uses the systemic console’s TTV routines to zoom in on the imminent HAT-P-13 periastron:

The above picture is quite rosy, at least as far as the outlook for TTVs is concerned. With orbital models that are based on the Bakos et al. discovery data for the system, the transit-to-transit time intervals for planet b veer from ~17 seconds shorter than average to ~17 seconds longer than average (relative to the long-term mean) as planet c runs through its periastron and exerts its maximum perturbing influence. This shift from a compressed period to an expanded period occurs rather abruptly over a span of about 2 weeks. Most provocatively, there are significant and feasibly observable differences between the TTV profiles produced by the coplanar configuration and by the configurations with 45-degree mutual inclinations. And finally, all the action was predicted to occur just before the end of HAT-P-13′s yearly observing season (see Bruce Gary’s revived AXA page for wealth of additional detail). It’s not hard to revel in the thought of all the ground-based observers pooling their results (in the spirit of 1761 and 1769) and emerging with a big-picture result!

The new Winn et al. data, however, definitely rain on the TTV parade. The augmented (out-of-transit) data set now shows that the period of planet c is about 20 days longer than previously believed, and c’s eccentricity also drops slightly, from e_c=0.69 to e_c=0.666. With the new orbital model, the differences in the TTVs generated by the co-planar and mutually inclined configurations are considerably smaller. The overall amplitude of the variations is cut nearly in half, and the excitement is pushed far more precariously against the end of the observing season:

And the good news? As described in the last post, the Winn et al. data show that the orbital plane of planet b is probably aligned with the equator of the parent star, which, in turn, means that it’s quite likely that the b-c system is indeed coplanar.

If we assume coplanarity, then the system should be at an eccentricity fixed point in which the apsides of the two planets are aligned. A measurement of the eccentricity of planet b then allows the interior structure and the tidal dissipation of planet b to be measured.

The augmented radial velocity data set permits a better measurement of planet b’s orbital eccentricity. Figure 5 of the Winn et al. paper has the relevant plot, which shows the distribution of Markov-Chain models for the eccentricity and apsidal angle of planet b. If the orbits are aligned, then the true model needs to fall within the red dotted lines, which mark the position of the (much better determined) apsidal line for planet c. From looking at the figure, the apsidally aligned configurations seem to have e_b ~ 0.01±0.005.

I asked Josh if he could send a histogram that shows the distribution of eccentricities for planet b for the subset of models that satisfy the alignment criterion. He got back to me very quickly with the following plot:

The result is: e_b = 0.0106 ± 0.0040, which implies a best-guess planetary structure that has (1) a small core, (2) a Love number k_2~0.34, and (3) a tidal dissipation quality factor Q~10,000 (see our paper, Batygin, Bodenheimer & Laughlin 2009 for details).

Categories: worlds Tags:
  1. darin
    March 31st, 2010 at 14:24 | #1

    Bad news indeed. We’ll need some good times to clear all this up. On the plus side, it looks like good transit times will help determine c’s orbital parameters (not just inclination).

    Do the new data from Winn also shift the results of Batygin, Bodenheimer, & Laughlin?

  2. March 31st, 2010 at 17:13 | #2

    Hi Greg,
    Given this issue with the star being to close to the sun to properly observe, is there any change you can get the EPOXI people to turn their spacecraft back on and watch this one from space? I know a few people on the impact half of that mission, but not the transit half, if you need a contact…

  3. March 31st, 2010 at 19:27 | #3

    So, essentially, what we have learned in 20 plus years of studying extra solar systems is this: “You would really not want to live there!”

    I expect that that is about to change soon.

  4. awhoward
    April 2nd, 2010 at 11:53 | #4

    Don’t despair, photometrists. I wouldn’t be surprised if the orbit of ‘c’ needs further refinement after we’ve observed its upcoming peri passage at Keck. The third body ‘d’ announced in Winn et al. is currently only modeled as a linear trend in the RVs. After we nail the parameters of ‘c’ with the upcoming peri passage, this could well bring ‘d’ into focus and the parameters of all will be better determined.

    On the other hand, I don’t think the best-fit eccentricity of ‘b’ is going to change much with additional measurements. The RV phase curve of ‘b’ is already very well sampled.

    Take home message: continue to observe HAT-13! It’s still one of the most exciting games in town.

  5. May 13th, 2010 at 05:37 | #5


Comments are closed.