# Divergence of laplacian

It seems to me that, in some derivations on fluid dynamic books I am reading, the identity $$\nabla \cdot (\nabla^2 u) = 0$$ where $u$ is a vector field, is used.

Does this identity exist? Is it true?

Note: $\nabla^2 u= \nabla(\nabla \cdot u) - \nabla \times (\nabla \times u)$. Taking the divergence gives: $\nabla \cdot(\nabla^2 u)= \nabla^2 (\nabla \cdot u)$. The reason the statement holds in your case is likely that $u$ is assumed to be an incompressible flow field, i.e. $\nabla \cdot u=0$, which is the incompressible continuity equation. Else, it may not hold.
The definition of vector Laplacian is $$\nabla^2 \mathbf{u} = \nabla(\nabla \cdot \mathbf{u}) - \nabla \times (\nabla \times \mathbf{u})$$ Since curl is always solenoidal, the divergence of the second term is $\mathbf{0}$, so we are left with $$\nabla \cdot (\nabla^2 \mathbf{u}) = \nabla^2(\nabla \cdot \mathbf{u})$$ and it may not be $\mathbf{0}$.