We report on deep near-infrared F125W ( J ) and F160W ( H ) Hubble Space Telescope Wide Field Camera 3 images of the z = { 6.42 } quasar J1148+5251 to attempt to detect rest-frame near-ultraviolet emission from the host galaxy .
These observations included contemporaneous observations of a nearby star of similar near-infrared colors to measure temporal variations in the telescope and instrument point spread function ( PSF ) .
We subtract the quasar point source using both this direct PSF and a model PSF .

Using direct subtraction , we measure an upper limit for the quasar host galaxy of m _ { J } > 22.8 , m _ { H } > 23.0 AB mag ( 2 \sigma ) .
After subtracting our best model PSF , we measure a limiting surface brightness from 0 \farcs 3 - 0 \farcs 5 radius of \mu _ { J } > 23.5 , \mu _ { H } > 23.7 AB mag arcsec ^ { -2 } ( 2 \sigma ) .
We test the ability of the model subtraction method to recover the host galaxy flux by simulating host galaxies with varying integrated magnitude , effective radius , and Sérsic index , and conducting the same analysis .
These models indicate that the surface brightness limit ( \mu _ { J } > 23.5 AB mag arcsec ^ { -2 } ) corresponds to an integrated upper limit of m _ { J } > 22 - 23 AB mag , consistent with the direct subtraction method .
Combined with existing far-infrared observations , this gives an infrared excess \log ( IRX ) > 1.0 and corresponding ultraviolet spectral slope \beta > -1.2 \pm 0.2 .
These values match those of most local luminous infrared galaxies , but are redder than those of almost all local star-forming galaxies and z \simeq { 6 } Lyman break galaxies .