ADF test

[sunmoon]

Hello.

I was making an unit root tests to my variables in order to test test stationarity (including intercept in test equation and with test for unit root in level,) Although, for two of my six variables i needed to make first differences in order to be stationary.
Because of this my regression should be changed, right?
How should i right this two variables that just became stationary at first differences in my model?
One of the variables is reerulc i should generate a series for difference as: dreerulc=d(reerulc)?
Because when i do this for my two stationary variables ate first differences and then run the regression again i have weaker results.
When do we introduce a logarithm in a serie? Because if i do so my model becomes more robust.