Morphology Determines Conductivity and Seebeck Coefficient in Conjugated Polymer Blends

Abstract

The impact of nanoscale morphology on conductivity and Seebeck coefficient in p-type doped all-polymer blend systems is investigated. For a strongly phase separated system (P3HT:PTB7), we achieve a Seebeck coefficient that peaks at S ∼ 1100 μV/K with conductivity σ ∼ 3 × 10 S/cm for 90% PTB7. In marked contrast, for well-mixed systems (P3HT:PTB7 with 5% diiodooctane (DIO), P3HT:PCPDTBT), we find an almost constant S ∼ 140 μV/K and σ ∼ 1 S/cm despite the energy levels being (virtually) identical in both cases. The results are interpreted in terms of a variable range hopping (VRH) model where a peak in S and a minimum in σ arise when the percolation pathway contains both host and guest sites, in which the latter acts as energetic trap. For well-mixed blends of the investigated compositions, VRH enables percolation pathways that only involve isolated guest sites, whereas the large distance between guest clusters in phase-separated blends enforces (energetically unfavorable) hops via the host. The experimentally observed trends are in good agreement with the results of atomistic kinetic Monte Carlo simulations accounting for the differences in nanoscale morphology.

Conjugated Polymer Blends for Organic Thermoelectrics

“…Despite the fact that the length scale of phase separation in the model was different from that in actual devices where the clusters are around several hundreds of nanometer across, the model was argued to be relevant to the actual system because the percolation problem is largely invariant to absolute length scales. [66] The trends from kMC simulations in conductivity and Seebeck coefficient agree well with those observed in experiments, both for phaseseparated (P3HT:PTB7 w/o DIO) and well-mixed (P3HT:PTB7 w/ DIO) morphologies.…”

“…[75][76][77][78][79][80] In Ref. [66] Zuo et al investigate the effect of the processing agent DIO on the morphology and thermoelectric properties of blended systems. Whereas in section 3.1 a clear peak in thermopower versus composition curve was shown for P3HT:PTB7 blends, the peak disappeared upon adding 5 vol% DIO to the thermoelectrically optimized blend, as shown in Materials A and B are indicated by red and black colors, respectively.…”

“…[81] To explain the relation of morphology with thermopower in OSC blends, the mechanism shown in Figure 7 was proposed. [66] In contrast to the nearest-neighbor hopping framework originally used to explain the composition dependence of thermoelectric blends, c.f. The blend morphologies, as measured by AFM in Figure 6(b,c), can be mapped on the schematic morphologies in the top panels of Figure 7.…”

“…In this scenario, both transport energy and Fermi energy are located in the DOS of material A, decreasing both thermopower and power factor. [66] [66] Copyright 2018, American Chemical Society.…”

“…Clustered morphologies were generated by the numerical annealing procedure developed by Peumans et al [82] and later used by Marsh et al [83] More details of this model can be found in Ref. [66] The calculated thermoelectric properties and corresponding blend morphologies are shown in Figure 8. Despite the fact that the length scale of phase separation in the model was different from that in actual devices where the clusters are around several hundreds of nanometer across, the model was argued to be relevant to the actual system because the percolation problem is largely invariant to absolute length scales.…”