A new era of fine structure constant measurements at high redshift

TL;DR

This paper introduces advanced calibration, AI-driven spectral modeling, and component-specific measurements to improve the precision of high-redshift fine structure constant (α) measurements, finding no significant variation across observed scales.

Contribution

It presents a novel combination of laser frequency comb calibration, AI-based spectral analysis, and component-wise α measurement for more robust quasar absorption spectra analysis.

Findings

Achieved six times smaller wavelength calibration residuals with LFC.

Collected 40 new α measurements constraining spatial fluctuations.

Found no significant variation in α, with Δα/α = -0.27 ± 2.41 × 10^{-6}.

Abstract

New observations of the quasar HE0515$-$4414 have been made using the HARPS spectrograph on the ESO 3.6m telescope, aided by the Laser Frequency Comb (LFC). We present three important advances for $\alpha$ measurements in quasar absorption spectra from these observations. Firstly, the data have been wavelength calibrated using LFC and ThAr methods. The LFC wavelength calibration residuals are six times smaller than when using the standard ThAr calibration. We give a direct comparison between $\alpha$ measurements made using the two methods. Secondly, spectral modelling was performed using Artificial Intelligence (fully automated, all human bias eliminated), including a temperature parameter for each absorption component. Thirdly, in contrast to previous work, additional model parameters were assigned to measure $\alpha$ for each individual absorption component. The increase in statistical uncertainty from the larger number of model parameters is small and the method allows a substantial advantage; outliers that would otherwise contribute a significant systematic, possibly corrupting the entire measurement, are identified and removed, permitting a more robust overall result. The $z_{abs} = 1.15$ absorption system along the HE0515$-$4414 sightline yields 40 new $\alpha$ measurements. We constrain spatial fluctuations in $\alpha$ to be $\Delta\alpha/\alpha \leq 9 \times 10^{-5}$ on scales $\approx 20\;{\rm km\,s}^{-1}$, corresponding to $\approx25\;$kpc if the $z_{abs} = 1.15$ system arises in a $1\;$Mpc cluster. Collectively, the 40 measurements yield $\Delta\alpha/\alpha=-0.27\pm2.41\times10^{-6}$, consistent with no variation.