The following equations would be required to execute backward propagation:
These equations will give you an idea of what is going on behind the scenes. Here, a suffix, d, has been added, which is a representation of the partial derivative that acts during backward propagation:
