Explain in deep-detailed the backpropagation step for pooling layer

I do know there are a lot of webpages trying to go deep down in how neural networks learn using gradient decent and the derivatives. What I need a really deep explanation of how gradient decent and derivatives are done in VGG16 by using all the formulas and all the derivatives and specifically for pooling layer(average pooling). Mathematics equations and explanations in one epoch. I do not need code just mathematics and explanation.
Budget is $20. deadline is 2 days.