You may find meaning in the attached file. please explore

Normally one would not apply such a huge field when measuring warmup after ZFC. What was your motivation to do so?

What was the experimental prescription to obtain the "emu/g" values (this comes down to knowing whether you have a bulk sample with no sample holder contribution or maybe a thin film with a substrate signal that needs to be accounted for etc. It takes some knowledge on specimens and (reasonable) expectations to answer your question). Material purity is one of them, highlighted in the paper Joseph proposed.

The crucial is the size of your material. For nanoparticles or thin films there can be two main extrinsic contributions to the magnetization:

i) surface spins

ii) absence of spin compensation in the spin cycloid. BiFeO3 is G-type antiferromagnet with superimposed cycloidal ordering. The period of the cycloid is about 62nm. If you have nanomaterials with size smaller than 62nm spins in the cycloid are no longer compensated leading to enhanced magnetization.

From ZFC-FC dependence I guess that you have ceramic or powder-like sample (i.e. bulk materials). Substitution with Sr can destroy spin cycloid resulting in enhanced magnetization (desired behavior from the point of view of multiferroic properties) and in weak ferromagnetic component. This ferromagnetic signal is reflected in your ZFC/FC dependence on temperature (you can try to fit the M(T) data with the Bloch law i.e. M~(T-Tc)^3/2 where Tc is the Curie temperature). The difference between ZFC and FC line probably is due to weak ferromagnetic domain reorientations. Thus you observed net intrinsic magnetization. Enclosed I am sending you a few papers devoted to this problem.

Article Synthesis and Properties of Bismuth Ferrite Multiferroic Nanoflowers

Article Dielectric and magnetic properties of (Bi1-xLaxFeO3)0.5(PbTi...

Article XRD, Raman and magnetic studies of Bi 1− x La x FeO 3 solid ...