arXiv subjects

Michael M. Zhang

Publications and source records attributed to Michael M. Zhang.

3 recordsLinked to original sources

arXiv2023

Transmission spectroscopy of the lowest-density gas giant: metals and a potential extended outflow in HAT-P-67b

Extremely low-density exoplanets are tantalizing targets for atmospheric characterization because of their promisingly large signals in transmission spectroscopy. We present the first analysis of the atmosphere of the lowest-density gas giant currently known, HAT-P-67 b. This inflated Saturn-mass exoplanet sits at the boundary between hot and ultrahot gas giants, where thermal dissociation of molecules begins to dominate atmospheric composition. We observed a transit of HAT-P-67 b at high spectral resolution with CARMENES and searched for atomic and molecular species using cross-correlation and likelihood mapping. Furthermore, we explored potential atmospheric escape by targeting H$α$ and the metastable helium line. We detect Ca II and Na I with significances of 13.2$σ$ and 4.6$σ$, respectively. Unlike in several ultrahot Jupiters, we do not measure a day-to-night wind. The large line depths of Ca II suggest that the upper atmosphere may be more ionized than models predict. We detect strong variability in H$α$ and the helium triplet during the observations. These signals suggest the possible presence of an extended planetary outflow that causes an early ingress and late egress. In the averaged transmission spectrum, we measure redshifted absorption at the $\sim 3.8\%$ and $\sim 4.5\%$ level in the H$α$ and He I triplet lines, respectively. From an isothermal Parker wind model, we derive a mass loss rate of $\dot{M} \sim 10^{13}~\rm{g/s}$ and an outflow temperature of $T \sim 9900~\rm{K}$. However, due to the lack of a longer out-of-transit baseline in our data, additional observations are needed to rule out stellar variability as the source of the H$α$ and He signals.

astro-ph.EP↗

arXiv2019

The Orbit of WASP-12b is Decaying

WASP-12b is a transiting hot Jupiter on a 1.09-day orbit around a late-F star. Since the planet's discovery in 2008, the time interval between transits has been decreasing by $29\pm 2$ msec year$^{-1}$. This is a possible sign of orbital decay, although the previously available data left open the possibility that the planet's orbit is slightly eccentric and is undergoing apsidal precession. Here, we present new transit and occultation observations that provide more decisive evidence for orbital decay, which is favored over apsidal precession by a $Δ\mathrm{BIC}$ of 22.3 or Bayes factor of 70,000. We also present new radial-velocity data that rule out the Rømer effect as the cause of the period change. This makes WASP-12 the first planetary system for which we can be confident that the orbit is decaying. The decay timescale for the orbit is $P/\dot{P} = 3.25\pm 0.23$ Myr. Interpreting the decay as the result of tidal dissipation, the modified stellar tidal quality factor is $Q'_\star = 1.8 \times10^{5}$.

astro-ph.EP↗

arXiv2017

Parallel Markov Chain Monte Carlo for the Indian Buffet Process

Indian Buffet Process based models are an elegant way for discovering underlying features within a data set, but inference in such models can be slow. Inferring underlying features using Markov chain Monte Carlo either relies on an uncollapsed representation, which leads to poor mixing, or on a collapsed representation, which leads to a quadratic increase in computational complexity. Existing attempts at distributing inference have introduced additional approximation within the inference procedure. In this paper we present a novel algorithm to perform asymptotically exact parallel Markov chain Monte Carlo inference for Indian Buffet Process models. We take advantage of the fact that the features are conditionally independent under the beta-Bernoulli process. Because of this conditional independence, we can partition the features into two parts: one part containing only the finitely many instantiated features and the other part containing the infinite tail of uninstantiated features. For the finite partition, parallel inference is simple given the instantiation of features. But for the infinite tail, performing uncollapsed MCMC leads to poor mixing and hence we collapse out the features. The resulting hybrid sampler, while being parallel, produces samples asymptotically from the true posterior.

stat.ML↗