Not all stars are the Sun: rethinking the Mixing Length

Characterizing heat transport in stars is notoriously complicated, and the task of reproducing the physics involved with high precision on stellar evolutionary timescales is a long-standing problem in stellar modeling. For this reason, convection in 1-D stellar evolution codes is addressed primarily through a framework known as mixing length theory (MLT). In this simplified picture, the efficiency of convective heat transport is encapsulated by the mixing length parameter alpha_{MLT}, measured in pressure scale heights.
While it is well understood that our nearest star is not a valid representation of stars in general, it remains the standard procedure to calibrate alpha_{MLT} according to solar specifications, and then to apply this value in any stellar model, regardless of mass or composition. However, there has been growing evidence that the use of a solar-calibrated mixing length is not always appropriate.
In this study, we investigate the scope of validity of a solar-calibrated mixing length over a range of evolutionary phases and input physics for very metal-poor stars, and find that the solar-calibrated mixing length is widely ineffective at reproducing the observed properties of such stars.