Reason for wide efficiency gap for little change in distance between particle gun and detector


I am curious why I am having a very unrealistic gap in the efficiency of deposited energy on my detector for very little change in the photon origin. The detector size is 50mmx50mmx50mm (i.e. each side has have a dimension: -25mm < x < 25mm.

The origin of the incoming photons of x0 and y0 was set randomly but between -25mm < x < 25mm, while z0 was fixed at -12.5mm, -25mm, 27.5mm, and -50mm for four different simulations. The deposited energy away from the surface is much smaller than that on the surface for very little change of 25mm (70%) to 27.5mm (30%) even though there is nothing (vacuum) obstructing the incoming photons. However, the efficiency for 27.5mm and 50mm appear to be similar.

What could be the cause of this efficiency gap, could it be that something is wrong with my simulation.

I am patiently waiting for insights from the experts in this forum. Kindly give a helping hand. Thanks in advance.