This seems to be a sensitive process since we are taking measurements from two different photodiodes and combining them to output an image signal. Is this combination process affected by low light? I would assume that it may become less reliable and could negatively affect the image quality. How does this compare with contrast detection that is used for auto-focus when there are low light conditions?

For my camera, it uses both both phase detection autofocus and contrast detection autofocus.And it indeed performs poorly in low light environments.