Brownian dynamics simulations of single-wall carbon nanotube separation by type using dielectrophoresis.

We theoretically investigate the separation of individualized metallic and semiconducting single-wall carbon nanotubes (SWNTs) in a dielectrophoretic (DEP) flow device. The SWNT motion is simulated by a Brownian dynamics (BD) algorithm, which includes the translational and rotational effects of hydrodynamic, Brownian, dielectrophoretic, and electrophoretic forces. The device geometry is chosen to be a coaxial cylinder because it yields effective flow throughput, the DEP and flow fields are orthogonal to each other, and all the fields can be described analytically everywhere. We construct a flow-DEP phase map showing different regimes, depending on the relative magnitudes of the forces in play. The BD code is combined with an optimization algorithm that searches for the conditions that maximize the separation performance. The optimization results show that a 99% sorting performance can be achieved with typical SWNT parameters by operating in a region of the phase map where the metallic SWNTs completely orient with the field, whereas the semiconducting SWNTs partially flow-align.