The dedication of the Magellan Adaptive Optics (MagAO) on the 6.5 m Clay Telescope has opened a new era in high-contrast imaging. MagAO is a large-aperture AO system that can reach moderate Strehl ratios in the visible wavelengths (0.6 to 1 mum). This unique wavelength coverage provides a new avenue for studying planet atmospheres and the environments where planets are born, because visible light is much more sensitive to dust extinction compared to near-infrared, and also contains the strongest gas accretion indicator, Halpha. The presence of cirucumplanetary disks (CPDs) can be inferred by measuring the amount of dust reddening as well as by searching for Halpha emission. Since 2012, I have been observing wide-orbit planet-mass companions with MagAO to search for the evidence of disks. I found that CT Cha B and 1RXS 1609 B have a significant dust extinction of AV ~ 3 to 5 mag, so they likely have their own dust disks. This suggests that CPDs may survive after 10 Myr, longer than the average lifetime of circumstellar disks (~5 Myr). For GQ Lup B, I showed that it is actively accreting material from its disk and emitting strong Halpha emission. MagAO observations demonstrate that CPDs could be common among young gas giants.

Since our knowledge of CPDs and exomoons has been mostly from theories and simulations, direct imaging of CPDs is of paramount importance to understand how planets and satellites actually form. I used the Atacama Large Millimeter/submillimeter Array (ALMA) to search for CPDs around 7 planetary-mass companions at 1.3 mm and 12CO (2--1), with a beam size of ~0.1'' and a 3sigma flux limit of ~0.1 mJy. However, except for FW Tau C, which was shown to be a low-mass star from the dynamical mass measurement, no CPDs were found around other companions. This surprising null result implies that CPDs are much more compact and denser than expected. CPDs probably have radii of <1000 RJup, so they are faint and optically thick in the radio wavelengths, and mid- to far-infrared may be more favorable to characterize disk properties. The MIRI imager on the James Webb Space Telescope can test this compact disk hypothesis by probing disk thermal emission between 5 and 25 mum.

I also carried out a few research projects broadly related to stellar physics, including (1) MagAO Halpha imaging of the binary proplyd LV1 and the brightest protostellar jet HH 508 in the Orion Nebula. I found that HH 508 has complex structure, implying that it could be a wide-angle outflow or two separate jets, (2) MagAO Halpha imaging of the stellar wind of the very massive star eta Carinae. I found that the Halpha line-forming region has a characteristic radius of 12 mas, or ~30 AU, in good agreement with model predictions, and (3) Submillimeter Array 12CO(2--1) mapping of the spiral galaxy NGC 6946 to understand the statistical properties of giant molecular clouds.