Log-linearization of Tobin's Q from Gali2007

Hi Dynare Forum

I am trying to log-linearize the following equation in a NK model with time-varying depreciation:

So far I have the following derivations:

I am unsure on how to get to terms I have derived to look like the ones presented in Gali2007:
Screenshot 2023-11-21 at 13.16.13

And then I am specifically concerned about the new delta*delta_bar i get. If this should indeed be positive or if this should be negative…

Can anyone help?