This paper can be heplyou: Estimating Divergence Times in Phylogenetic Trees Without a Molecular Clock.

"estimation of divergence times without assuming a constant molecular clock, where inference is based on DNA (or amino acid or protein) sequences from the species of interest. “Time” could here either be relative time, i.e., all divergence times are relative to the unknown age of the root of the tree, or absolute time if some fossil dating(s) relating the relative times to absolute time are available".

Syst Biol (2005) 54 (3): 500-507. doi: 10.1080/10635150590947311

http://sysbio.oxfordjournals.org/content/54/3/500.long

It is not enough to have just phylogeny on specie level. If you have phylogenetic tree with each branch corresponding to single species, and you can determine divergence time for lineages by means of node dating with molecular clock calibrated with fossil record, but this dating does not give you direct information on species age. The problem is that inside each linage there could be some extinct or non-discovered species, so the age of closest (most recent) node that separates any specie from the rest of the tree is not the same as specie age. This node age could be interpreted instead as maximum for range of possibilities for specie age. The real specie age could be much younger than this maximum limit value.

So you should analyze population from given specie, determine variability and calibrate molecular clock within specie limits. Potentially it is possible to determine age of last common ancestor of all populations that belongs to specie - that will be value close to the specie age, considering that there were no "bottleneck" events in specie evolution.

You can review methods used in Dick et al. (2013) though it seems that there are some issues with incomplete population sampling in this work.

Dick et al. 2013. Neogene origins and implied warmth tolerance of Amazon tree species. Ecology and Evolution 2013; 3(1): 162–169. doi: 10.1002/ece3.441