if it is SWCNT , there are many techniques.. simplest of which is centrifugatioin in amine/THF [DOI: 10.1021/ja051774o] or Milo Shaffer's alkali based method ACS Nano 2012, 6 (1), 54-62.

there are some methods how to separate them, but no one is working on 100%

I agree with Sridhar. There are several publication where density gradient based ultracentrifugation can give you 99.9% purity of semiconducting carbon nanotube. However there are reports that one can selectively grow the semiconducting carbon nanotube using sulfur/ceria by CVD methods.

In my opinion, the separation technique developed by Kataura's group has the benefit of being a quick, scalable, continuous process with higher resolution separation with respects to density gradient centrifugation and selective functionalization.

The process is based the structure-dependent interaction strength of SWCNTs with an agarose gel. The dispersion of SWCNT free of large bundles is laoded on the top column where the nanotubes that exhibit the strongest interaction with the gel (the semiconducting ones) are adsorbed by the sites of the column, while the unbound nanotubes (the metallic ones) flow through the column. The supernatant acts both as a solubilizing agent and as a mediator for the interaction between SWCNTs and the agarose gel.

The setup is simple as well: all you need is an agarose gel, a column for chromatographic separation, an aqueous solution of sodium dodecyl sulfate and one of sodium deoxycholate, a bath sonicator, a centrifuge (to remove large bundles prior to separation) and SWCNTs.

EXPERIMENTAL SECTION

According to "Continuous Separation of Metallic and Semiconducting Carbon Nanotubes

Using Agarose Gel" by Tanaka et al. on Applied Physics Express 2 (2009) 125002:

"To prepare the column, the outlet of a 2.5 ml plastic disposable syringe was plugged with cotton, and the syringe was filled with agarose beads (GE Healthcare Sepharose 2B, bead size range, 60 – 200 m). The column was washed with water, and then equilibrated with a 1% SDS solution. As shown in Fig. 2, when the SWCNT dispersion and successive 1% SDS solution were applied to the column, color separation was observed. Greenish SWCNTs were trapped in the upper region of the column, while reddish SWCNTs were not adsorbed to the column. These colors correspond to those of semiconducting and metallic HiPco-SWCNTs, respectively. The greenish SWCNTs were tightly bound to the gel, and were not eluted when the column was washed with 1% SDS solution (until 70 2000). By the addition of other kinds of surfactant solution, 1% sodium deoxycholate (DOC; Wako Pure Chemical Industries), the bound SWCNTs were successfully eluted as a gel-free solution fraction ."

I might suggest to look for Kataura's more recent patents and papers for updated procedures.

thanks, are these methods be used if the CNTs on silicon substrate?