Thermoelectric composites of poly(3-hexylthiophene) and carbon nanotubes with a large power factor

Abstract

Composite films of poly(3-hexylthiophene) and single-as well as multi-walled carbon nanotubes are demonstrated to offer a competitive thermoelectric performance. The power factor significantly exceeds values obtained with either constituent alone provided that the conjugated polymer is sufficiently pdoped. The use of single-walled carbon nanotubes consistently results in a higher electrical conductivity with a maximum value above 10 3 S cm À1 and thus gives rise to a power factor of 25 AE 6 mW m À1 K À2 for a filler content of only 8 wt% and a maximum 95 AE 12 mW m À1 K À2 for 42-81 wt%. Moreover, a carbon nanotube content of 8-10 wt% does not compromise the low bulk thermal conductivity of the polymer matrix, which promises a high figure of merit of at least ZT > 10 À2 at room-temperature. All samples are cast on plastic substrates, emphasising their suitability for large-area, flexible thermoelectric applications.

Photoinduced p‐ to n‐type Switching in Thermoelectric Polymer‐Carbon Nanotube Composites

“…Macroscopically, drop cast fi lms on PET substrates present an electrically percolating behavior as a function of CNT content, with a percolation threshold around c p ≈ 3.5 wt% ( Figure 2 a), similar to earlier results. [ 24 ] This threshold points to relatively well dispersed carbon nanotubes, in agreement with the SEM, AFM, and TXM data shown in Figure 1 . The macroscopic σ also increases sharply, in this case by fi ve orders of magnitude, when comparing the neat polymer and composites with CNT weight fractions above percolation.…”

“…The macroscopic σ also increases sharply, in this case by fi ve orders of magnitude, when comparing the neat polymer and composites with CNT weight fractions above percolation. Interestingly, the Seebeck coeffi cient varies from that of the neat polymer (≈1000 µV K −1 ) [ 24 ] to that of the CNTs (≈−10 µV K −1 ) and correlates well with the percolation threshold observed for σ (Figure 2 b). Strikingly, at around c s ≈ 40 wt% CNT content, the Seebeck coeffi cient changes sign.…”

“…[ 12,13 ] Interestingly, the thermal conductivity of bulk polymers as well as conjugated macromolecules can be tuned by controlling molecular orientation. [14][15][16] In order to increase the modest electrical conductivity of polymers, a number of strategies have been proposed, including careful doping, [ 4,11,[17][18][19][20][21] making composites of polymers with conductive fi llers such as CNTs, [22][23][24] or fabricating multilayer composite A broad range of organic electronic applications rely on the availability of both p-and n-type organic semiconductors, and the possibility to deposit them as sequential layers or to form spatial patterns. Examples include transport layers in diodes (OLEDs, photovoltaics, etc.…”

“…Composites of regular, unmodifi ed CNTs and doped P3HT have been shown to deliver power factors around 95 µW m −1 K −2 . [ 24 ] Here, composites with different stoichiometries were prepared by dispersing nitrogen doped CNTs (hereafter referred to simply as CNT) in o -dichlorobenzene (oDCB) and dissolving P3HT in chloroform, followed by mixing, sonicating, and drop-casting on fl exible PET substrates (see the Experimental Section for details). Scanning electron microscopy (SEM) images of composites of P3HT and nitrogen doped multiwalled CNTs with low CNT content, c, appear well dispersed, while samples containing large fractions of CNTs exhibit agglomeration ( Figure 1 a).…”

“…And indeed, the Seebeck coeffi cient can be changed to a positive value by doping the composites with FeCl 3 (data not shown). [ 24 ] The comparison between composites made from the two CNT batches suggests that the thermoelectric behavior of these composites strongly depends on the electronic properties of the matrix polymer as well as the specifi c interactions between P3HT and CNTs. So next we attempted to verify this hypothesis.…”