A linear correlation has been proposed between the CO luminosity ( L ^ { \prime } _ { CO } ) and full-width at half maximum ( FWHM ) for high-redshift ( z > 1 ) submillimeter galaxies .
However , the controversy concerning the L ^ { \prime } _ { CO } -FWHM correlation seems to have been caused by the use of heterogeneous samples ( e.g. , different transition lines ) and/or data with large measurement uncertainties .
In order to avoid the uncertainty caused by using different rotational transitions , in this work we make an extensive effort to select only CO ( J = 1 - 0 ) data from the literature .
We separate these wide-ranging redshift data into two samples : the low-redshift ( z < 1 ) and high-redshift ( z > 1 ) samples .
The samples are corrected for lensing magnification factors if gravitational-lensing effects appeared in the observations .
The correlation analysis shows that there exists significant L ^ { \prime } _ { CO } -FWHM correlations for both the low-redshift and high-redshift samples .
A comparison of the low- and high-redshift L ^ { \prime } _ { CO } -FWHM correlations does not show strong evolution with redshift .
Assuming that there is no evolution , we can use this relation to determine the model-independent distances of high-redshift galaxies .
We then constrain cosmological models with the calibrated high-redshift CO data and the sample of Type Ia supernovae in the Union 2.1 compilation .
In the constraint for w CDM with our samples , the derived values are w _ { 0 } = -1.02 \pm { 0.17 } , \Omega _ { m 0 } = 0.30 \pm { 0.02 } , and H _ { 0 } = 70.00 \pm { 0.60 } { km s ^ { -1 } Mpc ^ { -1 } } .