This is a tricky issue. As for most microbes, in the absence of a fossil record and/or good biogeographic data, it is very difficult to put absolute times on yeast divergence events. Though I have not seen a paper doing so, one way to do this is to transform observed molecular divergences among these species into absolute times using empirical mutation rate estimates, such as those from Lynch et al. (2008).

Nevertheless, there have been some attempts to estimate divergence times in yeast that you can cite, e.g. from Beltrao & Serrano (2005) and references therein:

The estimated divergence times of the other yeast species from S. cerevisiae were as follows: C. glabrata, 300 My; D. hansenii, 800 My; K. lactis, 400 My; Y. lipolytica, 900 My; C. albicans, 800 My; S. paradoxus, 50 My; S. bayanus, 50 My; S. mikatae, 50 My; N. crassa, 1,000 My; and Sch. pombe, 1,100 My. These values were based on phylogenetic studies found in the literature [32,42,43].