The enrichment history of the intergalactic medium -: Measuring the C iv/H i ratio in the Lyα forest

We have obtained an exceptionally high signal-to-noise, high-resolution spectrum of the gravitationally lensed quasar Q1422+231. A total of 34 C Iv systems are identified, several of which had not been seen in previous spectra. Voigt profiles are fitted to these C rv systems and to the entire Ly alpha forest in order to determine column densities, b-values, and redshifts for each absorption component. The column density distribution for C DI is found to be a power law with index alpha = 1.44 +/- 0.05, down to at least logN(C rv) = 12.3. We use simulations to estimate the incompleteness correction and find that there is in fact no evidence for flattening of the power law down to logN(C rv)= 11.7-a factor of 10 lower than previous measurements. In order to determine whether the C Iv enrichment extends to even lower column density H I clouds, we utilize two analysis techniques to probe the low column density regime in the Ly alpha forest. First, a composite stacked spectrum is produced by combining the data for Q1422 + 231 and another bright QSO, APM 08279 + 5255. The signal-to-noise ratio of the stacked spectrum is 1250, and yet no resultant C rv absorption is detected. We discuss the various problems that affect the stacking technique and focus in particular on a random velocity offset between H I and its associated C rv which we measure to have a dispersion of sigma(nu) = 17 km s(-1). It is concluded that, in our data, this offset results in an underestimate of the amount of C Iv present by a factor of about 2, and this technique is therefore not sufficiently sensitive to probe the low column density Ly alpha clouds to meaningful metallicities. Second, we use measurements of individual pixel optical depths of Ly alpha and corresponding C rv lines. We compare the results obtained from this optical depth method with analyses of simulated spectra enriched with varying C rv enrichment recipes. From these simulations, we conclude that more C Iv than is currently directly detected in Q1422+231 is required to reproduce the optical depths determined from the data, consistent with the conclusions drawn from consideration of the power-law distribution.