posted on 2021-04-09, 00:04authored byBrendan Mirka, Nicole A. Rice, Phillip Williams, Mathieu N. Tousignant, Nicholas T. Boileau, William J. Bodnaryk, Darryl Fong, Alex Adronov, Benoît H. Lessard
Ultrapure semiconducting
single-walled carbon nanotube (sc-SWNT)
dispersions produced through conjugated polymer sorting are ideal
candidates for the fabrication of solution-processed organic electronic
devices on a commercial scale. Protocols for sorting and dispersing
ultrapure sc-SWNTs with conjugated polymers for thin-film transistor
(TFT) applications have been well refined. Conventional wisdom dictates
that removal of excess unbound polymer through filtration or centrifugation
is necessary to produce high-performance TFTs. However, this is time-consuming,
wasteful, and resource-intensive. In this report, we challenge this
paradigm and demonstrate that excess unbound polymer during semiconductor
film fabrication is not necessarily detrimental to device performance.
Over 1200 TFT devices were fabricated from 30 unique polymer-sorted
SWNT dispersions, prepared using two different alternating copolymers.
Detailed Raman spectroscopy, x-ray photoelectron spectroscopy (XPS),
and atomic force microscopy (AFM) studies of the random-network semiconductor
films demonstrated that a simple solvent rinse during TFT fabrication
was sufficient to remove unbound polymer from the sc-SWNT films, thus
eliminating laborious polymer removal before TFT fabrication. Furthermore,
below a threshold polymer concentration, the presence of excess polymer
during fabrication did not significantly impede TFT performance. Preeminent
performance was achieved for devices prepared from native polymer-sorted
SWNT dispersions containing the “original” amount of
excess unbound polymer (immediately following enrichment). Lastly,
we developed an open-source Machine Learning algorithm to quantitatively
analyze AFM images of SWNT films for surface coverage, number of tubes,
and tube alignment.