Origin of Open Circuit Voltage in Solar Cells Based on Polymer Wrapped Carbon Nanotubes and Fullerenes

Thin films of semiconducting p-type single-walled carbon nanotubes (SWCNTs) have been recently used both for photo-absorption and for hole transport in the quest for all-carbon solar cells ^[1-6]. Quantum efficiency as high as 43% at the peak absorption wavelength of the nanotubes has been reported using single-chirality tubes ^[3]. These devices were of bilayer type, with efficiencies around 1%. More recently, bulk heterojunction (BHJ) type solar cells were also reported with a certified efficiency of 3.1% ^[2]. With the major focus on the interface between SWCNTs and fullerenes where the excitons generated by photo-absorption get dissociated, varying values of open circuit voltage (V_oc) have been obtained. Common to all these devices is the use of semiconducting polymers such as polyfluorene and polythiophene along with the nanotubes. During the separation process for single chirality tubes, the excess polymer is usually removed to obtain a polymer-nanotube mass ratio of 1:1. However, along with the SWCNT interfaces, there will still exist polymer-fullerene and polymer-electrode interfaces, which can influence the device performance. For example, S. Na et al ^[7] have shown that a polyfluorene derivative interfacial layer on the cathode side can in a BHJ solar cell with poly(3-hexylthiophene) (P3HT) and 1-(3-methoxycarbonyl)-propyl-1-phenyl-(6,6)C₆₁(PCBM) can increase its efficiency.

We report the effect of polymers on the performance of two single chirality SWCNT-C60 bilayer solar cells using: (a) poly(9,9-dioctylfluorene-2,7-diyl) (PFO) separated (7,5) tubes and (b) 9,9-di-octylfluorenyl-2,7-diyl and bipyridine copolymer (PFO-BPy) separated (6,5) tubes. We show that along with polymer-SWCNT weight ratio, the absolute weight of the polymer is also a critical parameter affecting the device operation.  Performance of devices with and without SWCNTs both having the same polymer content is compared. In devices without SWCNTs, the polymer alone acts as interfacial layer between C60 and the anode. Polymer concentration dependence is also studied in these devices. It is found that the polymer layer, even at very low concentrations, increases the device efficiency by nearly an order of magnitude. While the reference cell with only C60 gave an efficiency of 0.07%, the devices with PFO interfacial layer had efficiencies up to 0.6% with an increase in all three efficiency determining parameters. The V_oc of devices with PFO increased in the range of 0.2 -0.5 V over that of the C60 reference device, increasing with PFO layer thickness. In addition, while the PFO-SWCNT-C60 device gave a V_oc of 0.40 V, PFO-C60 device gave a V_oc of 0.56 V even though both devices contain equal concentration of PFO. These results indicate that the origin of V_oc may be the PFO-C60 interface and that the presence of SWCNTs has a negative effect on achievable V_oc. With PFO-SWCNT-C60 device working like a ternary system with PFO concentration dependent efficiency, optimization of this system should take into account the PFO-C60 interface as well.

References

[1] A. Dowgialo et al, ACS Nano, 8, 8573 (2014)

[2] M. Gong et al, Nano Lett., 14, 5308 (2014)

[3] M.J. Shea and M.S. Arnold, 102, 243101 (2013)

[4] M.P. Ramuz et al, ACS Nano, 6, 10384 (2012)

[5] D.J. Bindl et al, Chem. Phys., 413, 29 (2013)

[6] D.J. Bindl and M.S. Arnold, J. Phys. Chem. C, 117, 2390 (2013)

[7] S. Na et al, Org. Elec., 10, 496 (2009)