First measurement of the Mg II forest correlation function in the Epoch of Reionization

Abstract: In the process of producing the roughly three ionizing photons per atom required to reionize the IGM, the same massive stars explode and eject metals into their surroundings. While the overly sensitive Lya transition makes Gunn-Peterson absorption of background quasar light an ineffective probe of reionization at z > 6, strong low-ionization transitions like the MgII doublet will give rise to a detectable "metal-line forest", if metals pollute the neutral IGM. We measure the auto-correlation of the MgII forest transmission using a sample of ten ground based z >= 6.80 quasar spectra probing the redshift range 5.96 < z_MgII < 7.42 (z_MgII,median = 6.47). The correlation function exhibits strong small-scale clustering and a pronounced peak at the doublet velocity (768 km/s) arising from strong absorbers in the CGM of galaxies. After these strong absorbers are identified and masked the signal is consistent with noise. Our measurements are compared to a suite of models generated by combining a large hydrodynamical simulation with a semi-numerical reionization topology, assuming a simple uniform enrichment model. We obtain a 95% credibility upper limit of [Mg/H] < -3.73 at z_MgII,median = 6.47, assuming uninformative priors on [Mg/H] and the IGM neutral fraction x_HI. Splitting the data into low-z (5.96 < z_MgII < 6.47; z_MgII,median = 6.235) and high-z (6.47 < z_MgII < 7.42; z_MgII,median = 6.72) subsamples again yields null-detections and 95% upper limits of [Mg/H] < -3.75 and [Mg/H] < -3.45, respectively. These first measurements set the stage for an approved JWST Cycle 2 program (GO 3526) targeting a similar number of quasars that will be an order of magnitude more sensitive, making the Mgii forest an emerging powerful tool to deliver precision constraints on the reionization and enrichment history of the Universe.