Observations of the redshift z=7.085 quasar J1120+0641 have been used to search for variations of the fine structure constant, alpha, over the redshift range 5.5 to 7.1. Observations at z=7.1 probe the physics of the universe when it was only 0.8 billion years old. These are the most distant direct measurements of alpha to date and the first measurements made with a near-IR spectrograph. A new AI analysis method has been employed. Four measurements from the X-SHOOTER spectrograph on the European Southern Observatory's Very Large Telescope (VLT) directly constrain any changes in alpha relative to the value measured on Earth (alpha_0). The weighted mean strength of the electromagnetic force over this redshift range in this location in the universe is da/a = (alpha_z - alpha_0)/alpha_0 = (-2.18 +/- 7.27) X 10^{-5}, i.e. we find no evidence for a temporal change from the 4 new very high redshift measurements. When the 4 new measurements are combined with a large existing sample of lower redshift measurements, a new limit on possible spatial variation of da/a is marginally preferred over a no-variation model at the 3.7 sigma level.

We present an analysis of 23 absorption systems along the lines of sight towards 18 quasars in the redshift range of $0.4 \leq z_{abs} \leq 2.3$ observed on the Very Large Telescope (VLT) using the Ultraviolet and Visual Echelle Spectrograph (UVES). Considering both statistical and systematic error contributions we find a robust estimate of the weighted mean deviation of the fine-structure constant from its current, laboratory value of $\Delta\alpha/\alpha=\left(0.22\pm0.23\right)\times10^{-5}$, consistent with the dipole variation reported in Webb et al. and King et al. This paper also examines modelling methodologies and systematic effects. In particular we focus on the consequences of fitting quasar absorption systems with too few absorbing components and of selectively fitting only the stronger components in an absorption complex. We show that using insufficient continuum regions around an absorption complex causes a significant increase in the scatter of a sample of $\Delta\alpha/\alpha$ measurements, thus unnecessarily reducing the overall precision. We further show that fitting absorption systems with too few velocity components also results in a significant increase in the scatter of $\Delta\alpha/\alpha$ measurements, and in addition causes $\Delta\alpha/\alpha$ error estimates to be systematically underestimated. These results thus identify some of the potential pitfalls in analysis techniques and provide a guide for future analyses.

(abridged) We present a new analysis of a large sample of quasar absorption-line spectra obtained using UVES (the Ultraviolet and Visual Echelle Spectrograph) on the VLT (Very Large Telescope) in Chile. In the VLT sample (154 absorbers), we find evidence that alpha increases with increasing cosmological distance from Earth. However, as previously shown, the Keck sample (141 absorbers) provided evidence for a smaller alpha in the distant absorption clouds. Upon combining the samples an apparent variation of alpha across the sky emerges which is well represented by an angular dipole model pointing in the direction RA=(17.3 +/- 1.0) hr, dec. = (-61 +/- 10) deg, with amplitude (0.97 +0.22/-0.20) x 10^(-5). The dipole model is required at the 4.1 sigma statistical significance level over a simple monopole model where alpha is the same across the sky (but possibly different to the current laboratory value). The data sets reveal a number of remarkable consistencies: various data cuts are consistent and there is consistency in the overlap region of the Keck and VLT samples. Assuming a dipole-only (i.e. no-monopole) model whose amplitude grows proportionally with `lookback-time distance' (r=ct, where t is the lookback time), the amplitude is (1.1 +/- 0.2) x 10^(-6) GLyr^(-1) and the model is significant at the 4.2 sigma confidence level over the null model [Delta alpha]/alpha = 0). We apply robustness checks and demonstrate that the dipole effect does not originate from a small subset of the absorbers or spectra. We present an analysis of systematic effects, and are unable to identify any single systematic effect which can emulate the observed variation in alpha.