AFM Measures Photocurrent Distribution in Solar Cells

Lauren I. Rugani

The power conversion efficiency of organic solar cells remains below desired levels, in part because variations in processing conditions alter film morphologies. Current microscopy techniques can provide information on these structural effects; however, they can’t provide information to correlate individual nanostructures and overall device performance. David S. Ginger and a research team from the University of Washington in Seattle have employed photoconductive atomic force microscopy (AFM) to plot high-resolution photocurrent distributions in organic thin-film blends.

Photoconductive AFM was used to measure height (left) and photocurrent (right) of an MDMO-PPV:PCBM film spin-coated from toluene. The active regions of the device show FWHM widths of ~20 nm for 0-V applied bias and ∼40 nm for –1-V applied bias. Reprinted with permission of Nano Letters.

The devices studied were fabricated by spin-coating a 40-nm-thick layer of PEDOT:PSS, followed by an 80-nm active blend layer of poly[2-methoxy-5-(3',7'-dimethyloctyl-oxy)-1,4-phenylene vinylene]:(6,6)-phenyl-C₆₁-butyric acid methyl ester (MDMO- PPV:PCBM) onto indium tin oxide glass-coated substrates. Aluminum top contacts 40 nm thick were evaporated through a shadow mask to create 1.5-mm² active devices. The film blends were spin-coated from chlorobenzene, xylenes or toluene to vary the scale of phase separation in the active layer.

The researchers performed photoconductive AFM in the spaces between the aluminum contacts. An atomic force microscope from Asylum Research was positioned above an inverted optical microscope to image topography and photocurrent simultaneously. For illumination, an attenuated 5-mW, 532-nm laser was focused onto the substrate and aligned with the platinum-coated contact-mode AFM tip.

The microscopy data revealed that charge collection was most efficient at domain interfaces. Although an original bias of +0.25 V showed current to be dominated by only a fraction of the film, charge was collected more efficiently as the applied voltage became more negative. The team found each domain to have a unique short-circuit current, open-circuit voltage and fill factor, all of which contribute to bulk device efficiency. These values varied from 1.4 to 26.4 pA, 0.35 to 0.53 V and 0.42 to 0.58, respectively.

The researchers also observed a range of photocurrents within topographically similar regions, indicative of heterogeneities in vertical film structure. They suggest that the device performance could be improved if the entire film were fabricated to match the best domains. They also found variations in performance on a 200- to 600-nm-length scale that might have arisen from variations in interfacial resistance between the film and the underlying transparent conducting electrode.

Investigations of the intensity dependence of the photocurrent revealed that the dependence at low power remains linear up to 40 times solar intensity before becoming sublinear. Ginger notes the importance of this because the group’s measurements were taken under illumination intensities roughly 10 times that of sunlight — well within the linear regime — to obtain a good signal-to-noise ratio.

A good relative agreement between photocurrent measurements of the millimeter-size devices and of local 2-μm² areas demonstrates that local photocurrents are relevant to overall device performance. The team is working on several aspects to improve the mechanism, including using surface chemistry to guide phase separation to control the film morphology.

Although the technique is targeted toward testing organic solar cells, Ginger notes its potential in studying devices such as LEDs or transistors. “We hope to use photoconductive AFM along with time-resolved EFM to use optical excitation to map recombination and trapping phenomena in these devices as well,” he said.

Nano Letters, published March 2007.