A DEEP<i>HUBBLE SPACE TELESCOPE</i>SEARCH FOR ESCAPING LYMAN CONTINUUM FLUX AT<i>z</i>∼ 1.3: EVIDENCE FOR AN EVOLVING IONIZING EMISSIVITY

Abstract

We have obtained deep Hubble Space Telescope far-UV images of 15 starburst\ngalaxies at z~1.3 in the GOODS fields to search for escaping Lyman continuum\nphotons. These are the deepest far-UV images m_{AB}=28.7, 3\\sigma, 1" diameter)\nover this large an area (4.83 arcmin^2) and provide the best escape fraction\nconstraints for any galaxy at any redshift. We do not detect any individual\ngalaxies, with 3\\sigma limits to the Lyman Continuum (~700 \\AA) flux 50--149\ntimes fainter (in f_nu) than the rest-frame UV (1500 \\AA) continuum fluxes.\nCorrecting for the mean IGM attenuation (factor ~2), as well as an intrinsic\nstellar Lyman Break (~3), these limits translate to relative escape fraction\nlimits of f_{esc,rel}&lt;[0.03,0.21]. The stacked limit is\nf_{esc,rel}(3\\sigma)&lt;0.02. We use a Monte Carlo simulation to properly account\nfor the expected distribution of IGM opacities. When including constraints from\nprevious surveys at z~1.3 we find that, at the 95% confidence level, no more\nthan 8% of star--forming galaxies at z~1.3 can have relative escape fractions\ngreater than 0.50. Alternatively, if the majority of galaxies have low, but\nnon-zero, escaping Lyman Continuum, the escape fraction can not be more than\n0.04. Both the stacked limits, and the limits from the Monte Carlo simulation\nsuggest that the average ionizing emissivity (relative to non-ionizing UV\nemissivity) at z~1.3 is significantly lower than has been observed in Lyman\nBreak Galaxies (LBGs) at z~3. If the ionizing emissivity of star-forming\ngalaxies is in fact increasing with redshift, it would help to explain the high\nphotoionization rates seen in the IGM at z&gt;4 and reionization of the\nintergalactic medium at z&gt;6. [Abridged]\n