Given that the irradiance and collector current for a (random) photo-transistor is linearly correlated like below.
Then I am wondering why the output voltage graph of the transimpedence amplifier from the application notes from Sharp is not completely linear with irradiance.