To the Editor

— The paper by Wong et al.1 in the July 2007 issue of Nature Materials presents a new platinum metallopolyyne donor polymer (P1) with a bandgap of 1.85 eV that provides a photovoltaic power-conversion efficiency, η, of up to 4.93% in combination with a C60 fullerene derivative (PCBM) as acceptor. This high efficiency represents an important step towards the development of more efficient plastic solar cells. We argue, however, that the optical properties of the new polymer presented in the paper are incompatible with such high efficiency and that — based on the optical data — the efficiency is unlikely to exceed 2%.

Wong et al. report that blends of P1 and PCBM (1:4 by weight) provide a short-circuit current of 15.43 mA cm−2 under simulated solar light and a maximum monochromatic external quantum efficiency (EQE) of 87% (at 580 nm) for a 70–75-nm thin film. Both values are extremely high for polymer solar cells. As the EQE represents the number of charges collected in the external circuit per incident photon, an EQE of 87% implies that at least 87% of the photons are absorbed by the active layer. Given the fact that only the polymer contributes significantly to the absorption at 580 nm, 87% photon absorption seems improbable for a 70-nm thin film containing only 20% polymer by weight.

Quantitative insight on the amount of absorbed photons can be obtained from the real and imaginary parts of the complex refractive index reported by Wong et al. for the P1:PCBM blend, in combination with optical modelling using the transfer-matrix formalism that enables determination of the optical electric field in the active layer at each wavelength in the device2. This method takes into account the absorption and reflection at each individual layer and interface, and is commonly used in combination with spectroscopic ellipsometry to determine light absorption in polymer solar cells3,4,5. For the calculation we used the device architecture as reported by Wong et al: glass was taken as the substrate, followed by ITO (132 nm), PEDOT:PSS (27 nm), the active layer P1:PCBM (70 nm), and evaporated Al (100 nm). The optical constants for the substrate and electrode materials are available in the literature6,7, and in the software used for the calculations (Essential Macleod Software, Thin Film Center).

Figure 1 shows the EQE reported by Wong et al. and the fraction of absorbed photons from the optical modelling. Assuming that the internal quantum efficiency (IQE, number of charges per absorbed photon) is 100%, the maximum obtainable EQE equals the fraction of absorbed photons, which is 45% in the visible range. Figure 1 demonstrates that the reported EQE is about two times higher than the fraction of photons absorbed, which is physically unrealistic. Clearly, the reported EQE and η are inconsistent with the refractive index data of the blend.

Figure 1
figure 1

Experimental EQE1 compared with percentage of photons absorbed in a 70-nm film of P1 and PCBM.

Full size image

By integrating the number of absorbed photons with the AM1.5G solar spectrum, the upper limit for the short-circuit current is found to be 6.8 mA cm−2 for P1:PCBM layers. Together with the open-circuit voltage of 0.82 V and fill factor of 0.39 (ref. 1), this implies an upper limit to the power-conversion efficiency of η = 2.2% for 70-nm P1:PCBM layers. The real value may even be less than 2% when we take into account that the IQE is unlikely to be 100% over the whole wavelength range, and that the short-circuit current is usually sublinear with light intensity, which overestimates the spectral integration.

Even with η = 2.2%, the new organometallic polymer P1 represents an interesting step forward in the field of organic photovoltaics, but it still has to close a substantial gap with the best materials that are available at present8.