Abstract
A compensated magnetic state in an ideally homogeneous system with long range magnetic order is characterized by a net zero magnetization (M) throughout the sample (macroscopic). In the pristine state of the sample (i.e. with no external field, H = 0), this implies that at the magnetic compensation temperature (T-comp) we must have M = 0 at H = 0 irrespective of any thermal and magnetic history of the sample and any underlying physics. This simple fact voids the usual identification (and interpretation) of M-H loop parameters at and in the vicinity of magnetic compensation temperature, specifically the coercivity, the remanence, and the exchange bias characteristics. The physics of coercivity and exchange bias continues to be fully relevant, but its manifestation in an M-H loop would get camouflaged at (and near) a magnetic compensation because M -> 0 at H = 0. We present an experimental elucidation of the above premise through a case study in the spin ferromagnetic Sm1-xGdxAl2 alloys [x = 0.01-0.06], where the specimens with x <= 0.03 show compensation below the Curie temperature T-c, while those with x >= 0.03 have rather small magnetization due to near cancellation of opposing contributions, but are otherwise devoid of compensation. The experiments comprised low field (near zero) as well as high field (70 kOe) magnetization measurements from the paramagnetic state down to 5 K in the ordered regime (T-c similar to 125 K) and isothermal M-H loop studies on the remnant magnetic state of polycrystalline samples.