As Cyril mentioned, a screenshot would be nice, I also cant figure out how to get a z-score....
But in general: you should be cautious with largely different sample sizes in your groups. In combination with inhomogeneity in the variances between the groups, the ANOVA is no longer robust.
Yes, whith large df a t distribution approximates a normal distribution, but I think it doesn't help in this case, because Bruna asked how to read the SPSS output of the ANOVA and Mehmet, Cyril and me are unable to reproduce a SPSS ANOVA output with z-values, although you are right that there is a relationship between z, t and F.