Photos: Left: National Renewable Energy Laboratory; Right: Ken Bennett/Wake Forest University

IN DISPUTE: David Carroll [right] of Wake Forest University, led a team that last year reported having developed an organic solar cell with a record-setting 6.1 percent energy-conversion efficiency. Such technology could eventually be used to make power-generating coatings for cars, tents, and roof shingles. But some solar-cell experts, including Keith Emery of the National Renewable Energy Laboratory [left], contend that Carroll and other organic PV researchers are sidestepping proper verification of their devices.

Power output is determined by exposing the cell to the calibrated light source and measuring the current produced as a function of the voltage on the circuit. The optimal combination of voltage and current—corrected for spectral mismatch—represents the cell's maximum power output. Divide by the energy of illumination the cell received and you've got the cell's conversion efficiency.

Although such standard verification procedures have long been used for inorganic PV, it's easy to introduce errors when applying them to organic solar cells. That's because the unconventional semiconductors that make organic technology so exciting are also tricky to manipulate and measure. Most of the organic PV being worked on today—and drawing the greatest interest from investors and researchers—are called bulk-heterojunction cells, which Heeger and his collaborators invented in the early 1990s. They are composed of conducting polymers and carbon nanostructures called buckyballs that, in the right combinations, mimic the light-absorbing p-n junction of inorganic photovoltaics. (Another type of organic PV technology known as dye-sensitized or Grätzel cells displays higher efficiencies in the lab but has limited commercial appeal because of its liquid components.)

Most experimental organic PV cells are produced by pouring a solution of polymers and buckyballs onto a glass plate, spinning the plate to spread the solution into a film, heating it to drive off the water, and then sandwiching the resulting film between electrodes. Under illumination, the conducting polymer absorbs photons, kicking off electrons that are then attracted by the buckyballs and routed to an electrode. To optimize the transfer of charges from plastic to buckyball—and thus the device's efficiency—researchers continually seek improvements in materials, heat treatments, and other processing tricks. Another tack is to stack PV cells on top of one another. Heeger's group created such an architecture for its claimed 6.5 percent efficient device.

Organic cells behave differently under illumination than their inorganic cousins. Organic cells absorb mostly short-wavelength light at the blue end of the spectrum, while inorganic cells—including the reference cells used to calibrate the test lamps—absorb mostly red and infrared light. That makes spectral mismatch more pronounced when testing organic PV, because the reference cell and experimental cell have widely divergent spectral responses.

At the same time, organic semiconductors are fragile, making it hard to get a precise fix on the spectral mismatch. To measure a cell's spectral response accurately can take hours and thousands of readings across the spectrum, during which oxygen- and moisture-sensitive semiconductors, coatings, and electrodes can degrade.

Finally, organic PV researchers tend to produce very small cells—some less than a millimeter on a side. The larger the cell, the greater the chance of its containing uneven layers of film, which degrades performance. But diminutive cells are more likely to exhibit what solar-cell experts call perimeter effects—extra illumination that creeps in from beyond the edges of the test cell, thereby exaggerating its performance. Certification labs will test any size cell (Emery says NREL has tested cells as small as 0.008 cm2), but they give greater credence to cells that are 1 cm2 or larger, a size that Emery says “is sort of the transition point where perimeter effects become less important.”

Shaheen and Emery estimate that uncorrected spectral mismatch and extra illumination can exaggerate the conversion efficiency of test cells by as much as 50 percent. And their rapid degradation under testing conditions makes the organic PV a moving target. The perceived inattention to these measurement artifacts is what's fueling skepticism. “The community may think that there's 6 percent organic PV out there. I don't believe there is,” Emery says.

In a letter published in the December issue of Applied Physics Letters, Emery directly challenged the 6.1 percent efficiency claim from Wake Forest. The Wake Forest paper had described the group's use of temperature cycling to create polymer filaments in the bulk-heterojunction cells that enhanced their charge conduction and enabled thicker absorbing layers. The process purportedly boosted the cell's efficiency by nearly a fifth. Wake Forest's news release quoted project leader David Carroll, director of the university's nanotechnology center, as hoping to reach 10 percent efficiency in the coming year [see photo, “In Dispute”]. Emery's letter, however, argued that the true efficiency of Carroll's cell might be closer to 3 percent.

Emery's critique hinges on a mathematical cross-check that PV researchers rely on to spot weird results. It repurposes the measurements taken to map a cell's spectral response, using them to predict the current expected from the cell under sunlight. In essence, researchers multiply the values measured under monochromatic light by that wavelength's contribution to the solar spectrum and then sum them.

Carroll and his colleagues did not present that cross-check in their paper, so Emery used the spectral-response data provided and made the calculation himself. To square the resulting predicted currents with the 6.1 percent efficiency claimed, the cells would need to be generating more than one charged electron for each photon absorbed—a result that Emery dismisses as “not physically possible.” Noting that Carroll accounted for spectral mismatch, Emery concludes that perimeter effects inflated the results. “The mistake is probably in the device area—which is not unexpected for a device with an area less than 10 mm2,” Emery wrote.

Carroll, however, stands by his results. He says Emery's cross-check is flawed because the spectral response presented in the paper is that of a related but nonidentical cell. Why present power measurements for one cell and spectral response for another? According to Carroll, the temperature cycling that produced the 6.1 percent cells also left them incapable of withstanding the spectral-response tests. “They came apart,” he says.

Carroll charges that his critics' complaints are just the politically motivated sniping of scientific competitors. He expresses the utmost respect for Emery as a scientist, but Carroll also considers Emery to be the collaborator of a competitor who, like Carroll, is vying for grant money from the U.S. Air Force Office of Scientific Research, which underwrites much of the basic research in organic PV. “People are very aggressive,” Carroll says, “and they're aggressive because of money.”

Emery insists that he simply calibrated test equipment for the researcher in question. “We had absolutely nothing to do with designing the cells,” Emery says. “I won't let anyone in my lab get involved in fabricating PV technology.” He adds that doing so would violate his lab's certification.