I feel thiourea is better. Though we have used Na2S and thiourea both in different synthesis, we have not studied the relative merit or demerit. However, the possibility of Na getting incorporated in to the product cannot be ruled out. Though no noticeable Na was found ( I don't remember the limits now as it was done years ago), still it can be in smaller concentrations. We have also used the method of passing H2S through the metal precursor ( eg. cadmium nitrate) solution. This may be the best source.