Implemented different Δx inside and outside in function "solve_D_adaptedGrid". Keeping solve_D separate for alpha analysis.

Plotting m/m0 for different N, for 1D case and 3D case. Spherical mass conservation comparable to 1D case, but oscillating.

Problem with this: it doesn't cause the anticipated slow-down of
dynamics (interfacial resistance), since the mobility will now be
enhanced at the boundary.

interesting for interfacial resistance.