The spectrum of the z _ { em } = 2.63 quasar Q1009 + 2956 has been observed extensively on the Keck telescope .
The Lyman limit absorption system z _ { abs } = 2.504 was previously used to measure D/H by \citeauthor Burles1998b using a spectrum with signal to noise approximately 60 per pixel in the continuum near Ly \alpha at z _ { abs } = 2.504 .
The larger dataset now available combines to form an exceptionally high signal to noise spectrum , around 147 per pixel .
Several heavy element absorption lines are detected in this LLS , providing strong constraints on the kinematic structure .
We explore a suite of absorption system models and find that the deuterium feature is likely to be contaminated by weak interloping Ly \alpha absorption from a low column density H i cloud , reducing the expected D/H precision .
We find D/H = 2.48 ^ { +0.41 } _ { -0.35 } \times 10 ^ { -5 } for this system .
Combining this new measurement with others from the literature and applying the method of Least Trimmed Squares to a statistical sample of 15 D/H measurements results in a ‘ ‘ reliable ’ ’ sample of 13 values .
This sample yields a primordial deuterium abundance of ( \text { D / H } ) _ { p } = ( 2.545 \pm 0.025 ) \times 10 ^ { -5 } .
The corresponding mean baryonic density of the Universe is \Omega _ { b } h ^ { 2 } = 0.02174 \pm 0.00025 .
The quasar absorption data is of the same precision as , and marginally inconsistent with , the 2015 CMB Planck ( TT+lowP+lensing ) measurement , \Omega _ { b } h ^ { 2 } = 0.02226 \pm 0.00023 .
Further quasar and more precise nuclear data are required to establish whether this is a random fluctuation .