With MATLAB it is almost always better to loop over indices rather than over data values. That makes it easier to save the data in arrays. For example:
Eg_vec = 1.55:0.1:3.6;
Eg_num = numel(Eg_vec);
output = nan(1,Eg_num); % preallocate the output array.
for k = 1:Eg_num
Eg = Eg_vec(k);
.. the rest of your code in the loop
output(k) = max(efficiency);
end
plot(Eg_vec,output)
Avoid i and j as loop iterators, they are already defined as the imaginary unit.
