We present precise H i 21 cm absorption line redshifts observed in multiple epochs to directly constrain the secular redshift drift \dot { z } or the cosmic acceleration , \Delta v / \Delta t _ { \circ } .
A comparison of literature analog spectra to contemporary digital spectra shows significant acceleration likely attributable to systematic instrumental errors .
However , we obtain robust constraints using primarily Green Bank Telescope digital data .
Ten objects spanning z = 0.09 –0.69 observed over 13.5 years show \dot { z } = ( -2.3 \pm 0.8 ) \times 10 ^ { -8 } yr ^ { -1 } or \Delta v / \Delta t _ { \circ } = -5.5 \pm 2.2 m s ^ { -1 } yr ^ { -1 } .
The best constraint from a single object , 3C286 at \langle z \rangle = 0.692153275 ( 85 ) , is \dot { z } = ( 1.6 \pm 4.7 ) \times 10 ^ { -8 } yr ^ { -1 } or \Delta v / \Delta t _ { \circ } = 2.8 \pm 8.4 m s ^ { -1 } yr ^ { -1 } .
These measurements are three orders of magnitude larger than the theoretically expected acceleration at z = 0.5 , \dot { z } = 2 \times 10 ^ { -11 } yr ^ { -1 } or \Delta v / \Delta t _ { \circ } = 0.3 cm s ^ { -1 } yr ^ { -1 } , but they demonstrate the lack of peculiar acceleration in absorption line systems and the long-term frequency stability of modern radio telescopes .
A comparison of UV metal absorption lines to the 21 cm line improves constraints on the cosmic variation of physical constants : \Delta ( \alpha ^ { 2 } g _ { p } \mu ) / \alpha ^ { 2 } g _ { p } \mu = ( -1.2 \pm 1.4 ) \times 10 ^ { -6 } in the redshift range z = 0.24 –2.04 .
The linear evolution over the last 10.4 Gyr is ( -0.2 \pm 2.7 ) \times 10 ^ { -16 } yr ^ { -1 } , consistent with no variation .
The cosmic acceleration could be directly measured in \sim 125 years using current telescopes or in \sim 5 years using a Square Kilometer Array , but systematic effects will arise at the 1 cm s ^ { -1 } yr ^ { -1 } level .