We present a theoretical model embedding the essential physics of early galaxy formation ( z \simeq 5 - 12 ) based on the single premise that any galaxy can form stars with a maximal limiting efficiency that provides enough energy to expel all the remaining gas , quenching further star formation .
This simple idea is implemented into a merger-tree based semi-analytical model that utilises two mass and redshift-independent parameters to capture the key physics of supernova feedback in ejecting gas from low-mass halos , and tracks the resulting impact on the subsequent growth of more massive systems via halo mergers and gas accretion .
Our model shows that : ( i ) the smallest halos ( halo mass M _ { h } \leq 10 ^ { 10 } M _ { \odot } ) build up their gas mass by accretion from the intergalactic medium ; ( ii ) the bulk of the gas powering star formation in larger halos ( M _ { h } \geq 10 ^ { 11.5 } M _ { \odot } ) is brought in by merging progenitors ; ( iii ) the faint-end UV luminosity function slope evolves according to \alpha = -1.75 \log z - 0.52 .
In addition , ( iv ) the stellar mass-to-light ratio is well fit by the functional form \log M _ { * } = -0.38 M _ { UV } -0.13 z + 2.4 , which we use to build the evolving stellar mass function to compare to observations .
We end with a census of the cosmic stellar mass density ( SMD ) across galaxies with UV magnitudes over the range -23 \leq M _ { UV } \leq - 11 spanning redshifts 5 < z < 12 : ( v ) while currently detected LBGs contain \approx 50 % ( 10 % ) of the total SMD at z = 5 ( 8 ) , the JWST will detect up to 25 % of the SMD at z \simeq 9.5 .