Among the two classes of geometry optimization/minimization methods i.e., first order derivative (steepest descent and conjugate gradient) and second order derivatives (Newton-Raphson), Can anyone tell me that what are the merits and demerits of each of the class and under what conditions which one to choose ?