One of the unavoidable realities of our era is that fossil fuels are not infinite. Someday, and sooner rather than later, available stores will be exhausted. To overcome the supply and other limitations of this source of energy, scientists are actively searching for alternatives. Solar energy derived from organic (i.e., plastic) photovoltaic (OPV) cells is a promising replacement for fossil fuels, not least because it is potentially affordable. A current drawback to the technology, however, is that OPVs are less efficient than their non-organic silicon counterparts. Among organic solar cells, those based on poly(3-hexylthiophene):fullerene (P3HT:PCBM) are the best known at converting sunlight to electricity. To make an OPV device, P3HT and PCBM are blended in a solvent and coated on a surface, or substrate, usually glass covered with indium tin oxide (ITO). When the device is irradiated, pairs of positive and negative charge carriers, called excitons, are created in the P3HT regions of the substrate and become dissociated at the interface of P3HT and PCBM. Depending on their individual charges, the carriers are then propelled by built-in potential through networks either to ITO or aluminum (Al) electrode layers. This type of blended structure is known as a bulk heterojunction.
The interpenetration of these networks is generally believed to be key to achieving high power efficiency. It is governed by the structure (morphology) of the organic active layer, which is influenced by many factors, including the regioregularity (akin to purity) of the polymer, the composition of the blending materials, the solvent, and a heating process known as thermal annealing. Among all these factors, thermal annealing has the greatest effect on the morphology of P3HT:PCBM—and thus proper formation of the charge transport networks—and the device performance. For this reason, OPV cells usually undergo a post-annealing step to ensure a certain level of power-conversion efficiency.1–7
Figure 1. (a) X-ray reflectivity curves during thermal annealing of an OPV cell measured in an in-situ chamber. (b) Mass-density profiles of an aluminum (Al) layer (black filled squares) and interlayer (red filled circles), which were obtained by model fitting of the x-ray reflectivity curves. (c) Roughness of the Al surface (black filled squares) and poly(3-hexylthiophene):fullerene (P3HT:PCBM) surface (red filled circles) during thermal annealing. (d) The change in thickness of the Al layer (black filled squares) and interlayer (red filled circles) with increasing temperature. Theta: The angle between the incident x-rays and the sample.
The structural changes brought about by the annealing process have been investigated using a variety of methods, such as UV-absorption spectra measurements, atomic force microscopy, transmission electron microscopy (TEM), x-ray diffraction measurements, and variable angle spectroscopic ellipsometry. TEM has shown large P3HT and PCBM domain sizes in the annealed film, which indicates improved phase separation (i.e., of the two materials). Moreover, a sharp increase in P3HT peak intensity revealed by x-ray diffraction suggests that an increase in P3HT crystals induced by the thermal process may also be an advantage. All the same, none of these results fully explains the link between device performance and structural properties. Indeed, larger crystal size was recently reported to cause poor device performance.8 The main problem with these studies is that most of them omit the Al electrode because it is difficult to measure the polymer structures beneath it. However, in real device fabrication, the annealing process is generally carried out after depositing the Al electrode. Taking it into account is therefore essential to understanding device behavior.
We have carried out real-time measurement of structural changes during the annealing process using synchrotron x-rays. We prepared an in situ annealing chamber equipped with liquid nitrogen cooling lines and a vacuum pump. P3HT:PCBM (1:0.65) thin films with a 40nm-thick Al layer were annealed in the chamber under vacuum conditions. We increased the temperature from 30 to 180°C in steps, measuring the reflectivity curve and 2D diffraction pattern at each step. Figure 1(a) shows the variation in the x-ray reflectivity curves during thermal annealing measured by our experiments. Figure 1(b–d) presents the results obtained by model fitting the reflectivity data. We assumed a gradually decreasing mass-density profile, which we found to be a better fit than an abruptly separated interface.9 Figure 1(b) shows the variation in the mass densities of the Al layer (black filled squares) and the interlayer (red filled circles) with a rise in temperature. Thermal annealing increases the surface roughness of Al: see Figure 1(c). On the other hand, the roughness of the interface where the interlayer and P3HT:PCBM meet remains constant up to 120°C. Thereafter, it grows and saturates at 150°C. Figure 1(d) shows the thickness changes in the Al layer and the interlayer during annealing. The interlayer measured ∼3nm for the pristine sample and was 4.8nm at 180°C. The interlayer thickness began to increase at 120°C, which is similar to the results for mass density and roughness of the P3HT:PCBM. Alteration in P3HT crystals was investigated using grazing incident wide-angle scattering. Thermal annealing increased the size of crystals of P3HT in films both with and without the Al layer. In the latter, annealing enhanced the preferred orientation of the crystals, while in the former, it had the opposite effect.
We have directly measured the real-time variation at the interface of an Al layer and a P3HT:PCBM organic layer during thermal annealing using synchrotron x-rays. These measurements demonstrate that Al atoms diffuse into the organic layer during the annealing process and disturb the alignment of P3HT crystals toward the surface normal. The current-density voltage characteristics show that the Al interdiffusion lowers the contact resistance between the Al electrode and the P3HT:PCBM layer. Consequently, more current is generated in the device and the power conversion efficiency increases.
As next steps, we intend to confirm the interdiffusion of the Al atoms more directly, for example, by TEM. We also plan to apply our methods to other devices with different active-layer thicknesses.
Hyo Jung Kim
Seoul National University
Department of Materials Science and Engineering
Seoul National University
Hyun Hwi Lee
Pohang Light Source