Thank you for your suggestions..
For the last 24 hours I have been thinking about this opamp and I want to share what I found, and also what I could not. First, I think that I can model this opamp by the model given below. Using open circuit time constant method, I found that dominant pole after compensation arises due to C2 and given by 1/[C2.(gmn5.gmn3.rA.rB.rout)]. However, using open circuit time constant method I could not find where the pole the output moved to. this is because, open circuiting all other caps other than Cout gives me the same output pole expression as without compensation. I also tried to derive the whole transfer function (Vout/Vin), but it was too tedious to come up with poles and zeros.
Does anybody have an idea to show that indeed nondominant poles moved to higher frequencies after compensation?
Thank you all..