Polymer-Free Carbon Nanotube Thermoelectrics with Improved Charge Carrier Transport and Power Factor

Abstract

Semiconducting single-walled carbon nanotubes (s-SWCNTs) have recently attracted attention for their promise as active components in a variety of optical and electronic applications, including thermoelectricity generation. Here we demonstrate that removing the wrapping polymer from the highly enriched s-SWCNT network leads to substantial improvements in charge carrier transport and thermoelectric power factor. These improvements arise primarily from an increase in charge carrier mobility within the s-SWCNT networks because of removal of the insulating polymer and control of the level of nanotube bundling in the network, which enables higher thin-film conductivity for a given carrier density. Ultimately, these studies demonstrate that highly enriched s-SWCNT thin films, in the complete absence of any accompanying semiconducting polymer, can attain thermoelectric power factors in the range of ∼400 μW m −1 K −2 , which is on par with that of some of the best single-component organic thermoelectrics demonstrated to date.

Thermoelectric properties of sorted semiconducting single-walled carbon nanotube sheets

“…They revealed that the Seebeck coefficient was increased as the purity of s-SWNT increased [29], which was also systematically studied by Piao et al [30]. For the s-SWNT sheets, it was demonstrated that Seebeck coefficient was further increased by optimizing the doping level chemically [14,16,31,32] or electrochemically [33][34][35] and, quite importantly, the higher purity of s-SWNT leads to a larger increase of Seebeck coefficient [29], thus the value reached to 2000 μV K −1 [14].…”

“…On the other hand, the electrical conductivity of s-SWNT was much lower than that of m-SWNT, and the electrical conductivity decreased as the s-SWNTs purity increased [36]. Such a negative correlation between Seebeck coefficient and electrical conductivity cannot be ruled out even under changing the carrier density; namely, the increase of electrical conductivity of the s-SWNT network from 10 3 to 10 6 S m −1 by chemical doping resulted in the large decrease of the Seebeck coefficient [14,31,32]. Thus, the power factor (PF) of the s-SWNT network, defined as σS 2 , has a maximum value when the electrical conductivity is around 10 4 -10 5 S m −1 [14,31,32].…”

“…Such a negative correlation between Seebeck coefficient and electrical conductivity cannot be ruled out even under changing the carrier density; namely, the increase of electrical conductivity of the s-SWNT network from 10 3 to 10 6 S m −1 by chemical doping resulted in the large decrease of the Seebeck coefficient [14,31,32]. Thus, the power factor (PF) of the s-SWNT network, defined as σS 2 , has a maximum value when the electrical conductivity is around 10 4 -10 5 S m −1 [14,31,32]. It has been revealed that the maximum PF can be improved by optimizing 1) the diameter of SWNT [14], 2) the morphology of the SWNT network including the bundle size, SWNT length [37][38][39], and 3) removal of dispersant [31,32].…”

“…Thus, the power factor (PF) of the s-SWNT network, defined as σS 2 , has a maximum value when the electrical conductivity is around 10 4 -10 5 S m −1 [14,31,32]. It has been revealed that the maximum PF can be improved by optimizing 1) the diameter of SWNT [14], 2) the morphology of the SWNT network including the bundle size, SWNT length [37][38][39], and 3) removal of dispersant [31,32]. Thus, the maximum PF of s-SWNT network can exceed 500 μW m −1 K −2 [31], which was much greater than that of m-SWNT network (<10 μW m −1 K −2 ) [29].…”

Semiconducting Single‐Walled Carbon Nanotubes or Very Rigid Conjugated Polymers: A Comparison

“…In the ultimate case all nanotubes directly bridge the channel if it is short enough. [235,236] It is still unclear how much this resistance is influenced by the angle of overlap for two nanotubes, [236] by residual surfactant or wrapping polymer [250,251] and how it changes with doping/ gating. [81,110] The density of a nanotube network (see Section 2.4) is an important parameter for FETs.…”