Thermodynamically, the phase separation process leads to structured/organized state of molecules which relates to a decrease in system entropy according to Gibbs free energy equation del(G) = del(H) - Tdel(S) W and correspondingly the process would be exothermic for spontaneous phase separation. 

Generally, phase separation is initiated by fluctuations in the local phase density which leads to formation of nuclei with defined interface. This requires some activation energy which depends on the available interfacial energy.  For a stable phase separation process, this interfacial energy is consumed to reach a minimum Gibbs energy state.

A good example of this phenomenon is preparing a low surface energy surface coating (superhydrophobic surface) from binary ceramic system such as TiO2–Cu2O composite. (Ref. http://m.iopscience.iop.org/0957-4484/25/24/245601/pdf/0957-4484_25_24_245601.pdf)

Additionally, the following reference may be useful:

Metallopolymer Nanocomposites By A.D. Pomogailo, V.N. Kestelman   

Hope it helps,

Since phase separation is a way to reduce entropy, it should go down right!

The surface with phase seggregation will have fluctuations in surface energy as mentioned by Dr. Mitra. Depending on the effect you are looking for, your surface will have an anisotropic response. I guess you are interested in characterize some of the properties? Surface potential 

thank you for your responses....

I have found out that in ceria-alumina system a heterogeneous/mixed Al-rich and Ce-rich phase is formed beyond a certain concentration of ceria loading in alumina. The crystallisation of Ceria in the system also increases with higher loading of Ce. Can these two things be correlated with each other?

It would be interesting to find the binary phase diagram of ceo2 vs al2o3.. or a ternary one of ce al and o. They have to be reported somewhere as ceria protective oxides onto al aleations are very used.