When f0 becomes zero, then there is no void "seed" any more in the model. This is why f0 will keep zero when tension is applied later.

I guess, If you set a threshold: namely under compression, when f0 decreases to a very small value, e.g. 10E-4, then keep f0 be this value and don't let f0 further decrease. Then you do tensile test, I guess now you will find void growth in the pre-compressed specimen.

Don't spend time waiting for the perfect situation, something that is not very likely to come