Because the  wavelength accuracy on UV-Vis measurements is located between (0.2 - 0.8) , then when we dissolve 22 mg from DPPH in 50ml Methanol (stock solution ) and prepare work solution by adding 6 ml from stock solution to 94ml methanol , the optical density of negative blank ( DPPH +Methanol  ) should be  between   0.700 and  0.800  at 515 nm

if you have material with strong antioxidant capacity  you can extract small amount or make serial dilutions to have proper results ( optical densities ) within range 0.2-0.8

Dear Baso Didik Hikmawan

Om P. Sharma,  and Tej K. Bhat have published this article and it might answer your queston.

http://dx.doi.org/10.1016/j.foodchem.2008.08.008

This paper may be helpful ;   doi:10.1016/j.foodchem.2003.05.007

Basically, it is due to the absorbance of dpph at these concentrations. If it is too concentrated, the spectrophotometer would not be able to give any reading. If it is too diluted, there won't be enough dpph radical to be scavenged by your sample, thus giving 100% scavenging all the time.

due the interval of analysis on the UV-Vis. if you work with results using the IC 50, you only can compare your results with works used the same concentration of DPPH or you can to use the IAA (index of antioxidant activity) used to correct the difference on the concentration of DPPH.

https://www.researchgate.net/publication/222209132_Antioxidant_activity_index_%28AAI%29_by_the_22-diphenyl-1-picrylhydrazyl_method

good luck

Article Antioxidant activity index (AAI) by the 2,2-diphenyl-1-picry...