Palomar 5 ( Pal 5 ) is a faint halo globular cluster associated with narrow tidal tails .
It is a useful system to understand the process of tidal dissolution , as well as to constrain the potential of the Milky Way .
A well-determined orbit for Pal 5 would enable detailed study of these open questions .
We present here the first CCD-based proper motion measurement of Pal 5 obtained using SDSS as a first epoch and new LBT/LBC images as a second , giving a baseline of 15 years .
We perform relative astrometry , using SDSS as a distortion-free reference , and images of the cluster and also of the Pal 5 stream for the derivation of the distortion correction for LBC .
The reference frame is made up of background galaxies .
We correct for differential chromatic refraction using relations obtained from SDSS colors as well as from flux-calibrated spectra , finding that the correction relations for stars and for galaxies are different .
We obtain \mu _ { \alpha } = -2.296 \pm 0.186 mas/yr and \mu _ { \delta } = -2.257 \pm 0.181 mas/yr for the proper motion of Pal 5 .
We use this motion , and the publicly available code galpy , to model the disruption of Pal 5 in different Milky Way models consisting of a bulge , a disk and a spherical dark matter halo .
Our fits to the observed stream properties ( streak and radial velocity gradient ) result in a preference for a relatively large Pal 5 distance of around 24 kpc .
A slightly larger absolute proper motion than what we measure also results in better matches but the best solutions need a change in distance .
We find that a spherical Milky Way model , with V _ { 0 } = 220 km/s and V _ { \mathrm { 20 kpc } } , i.e. , approximately at the apocenter of Pal 5 , of 218 km/s , can match the data well , at least for our choice of disk and bulge parametrization .