What clears protoplanetary disks after a few million years?
In plain words
The fraction of young stars that still have disks drops sharply over a few million years, which sets the deadline for forming gas giants. Ultraviolet and X-ray light from the star, magnetized winds and radiation from nearby massive stars could each remove the gas.
Precise statement
The fraction of stars with inner-disk infrared excess declines with an e-folding time of roughly 2-3 Myr. Determine the dominant dispersal agent: internal photoevaporation by stellar X-ray, EUV or FUV radiation, magnetized disk winds, or external photoevaporation by massive stars in the birth cluster, by matching measured wind mass-loss rates (forbidden-line profiles such as [Ne II] 12.8 micron and [O I] 6300 Angstrom), transition-disk statistics and disk lifetimes versus environment. The answer is the mass-loss rate per channel versus stellar mass and environment.
What would settle it
A model that reproduces disk fractions versus age and environment, the transition-disk population and measured wind mass-loss rates for a large sample.