Diffusion capacitance of PN junction diode

Hi ,

I have problem understanding the concept of Diffusion capacitance of PN junction diode .

For Transition capacitance (reverse bias ) this is calculated as c =dq/dv , where dq is change in charge and dv is change in voltage . here the dq is any one charge ( either +ve or -ve ) exactly similar to the parallel plate capacitor , where c is calculated as c=q/v (as dipole) , where q is charge on any one plate . so , this is fairly consistent with our understanding of basic capacitor .

however for diffusion capacitance , c is calculated as c = [ dq ( excees hole pumped in N type from P type) + dq ( execc electron pumped in P type from N type 0 ] /dv .

Here why we are taking both charges ?? unlike the normal capacitor where charge on only plate r is considered .
Doesn't it mean the diffusion capacitance will be zero when both ( excess minority in each type ) is equal , moreover the calculation of C for diffusion cases seems to be breaking the rule ( +q-q)/v .
what is the concept behind this , Kindly help me
Regards
carol