when i add graphite at three levels to my metal matrix composites, i get the density values initially increases and then decreases with graphite addition.
based on rule of mixture, the density of composites certainly reduced with increase of graphite percentage because of low density of graphite (2.25 g/cc)
About density measurement of graphite flakes you may want to check: https://www.researchgate.net/post/How_do_we_find_the_bulk_density_and_particle_density_of_exfoliated_graphite_flaks_in_Lab
It is very difficult to homogeneously mix a small amount of graphite into copper matrix due to the significant difference in density of both constituents. If the amount is too small, the graphite will be most probably acts as lubricant, reducing friction during compaction, as mentioned by Dorota Nagolska. As a result, the density of the composite is increased. As more graphite is added, the mixing homogeneity will increase, thus the density of the composite will decreased as mentioned by Raj Kumar. Based on experience, it is very difficult to add more than 5 vol% of graphite into the metal matrix composites because of the passive nature of graphite. So I think your "3 levels" of graphite contents were not significantly difference. My suggestion is to adjust the graphite content for each level. In order to do so, you might want to introduce surface modification on the graphite.
I have investigated in that field about previous month. Graphite particles usually act as a solid lubricant which facilitates movement and rearrangement of matrix/reinforcement particles causing higher densification. what is the particle size of graphite? what is the amount of addition? I mean what weight percentage or volume percentage.
what percentage of graphite did you add? I have added varying percentage of Graphite up to 10 wt.%. At10 wt.% I have seen particle agglomeration and properties deterioration.