Researchers at the National Institute of Standards and Technology (NIST) have for the first time examined, with nanometer-scale precision, the variations in chemical composition and defects of widely used solar cells. The new techniques, which were used to investigate a common type of solar cell made of the semiconductor material cadmium telluride, promise to aid scientists to better understand the microscopic structure of solar cells and may ultimately suggest ways to boost the efficiency with which they convert sunlight to electricity.
Even though standard methods to characterize solar cells have long proven useful in guiding their fabrication and design, the available diagnostic tools give only a limited understanding of why the devices operate at sub-optimal efficiency. For instance, although a method known as electron-beam induced current, which analyzes samples using the beam of an electron microscope, provides data on nanoscale variations in solar cell efficiency, it gives little information on the underlying crystal defects and impurities that degrade the efficiency. Two other methods, photoluminescence and cathodoluminescence, which induce light emission from the samples, provide only insufficient or indirect information on the mechanisms of efficiency losses.
In their study, the scientists used two complementary methods that rely on an atomic force microscope (AFM). Photothermal induced resonance (PTIR) provides information on the solar cell's composition and defects at the nanometer-scale by measuring how much light the sample absorbs over a broad range of wavelengths, from visible light to the mid-infrared. The other method, known as direct transmission near-field scanning optical microscopy (dt-NSOM), creates detailed nanoscale images that capture variations in the composition of the solar cells and defects in their structure, by recording how much light is transmitted at specific sites within the cell. That method produces sharper images than PTIR.
The setup for PTIR resembles a finely tuned version of a Rube Goldberg contraption. First, light pulses from a laser illuminate a sample of cadmium telluride. When the sample absorbs the laser light, it heats up and expands. The expansion nudges the sharp tip of an AFM that is in contact with the sample. The tip converts the heat-induced expansion into mechanical motion, causing the cantilever on which it is mounted to vibrate. Finally, the vibration is detected by bouncing light from another laser off the cantilever into the AFM detector.
Because the amplitude of the cantilever's vibrations is proportional to the energy absorbed by the cadmium telluride sample, PTIR measurements provide key information about the material. For instance, when the tip is held at one location and the wavelength of the pulsed laser light is varied, information is generated about the spectra of radiation absorbed at different points along the sample, with nanoscale resolution.
When the AFM tip moves over the sample but the laser's wavelength remains fixed, PTIR yields an absorption map of the material, which reveals variations in chemical composition from one part of the sample to the other. The small size of the probe tip provides absorption information with a spatial resolution smaller than the laser wavelength.
In the dt-NSOM technique, light from the sharp tip of an AFM probe illuminates a small part of the sample. A photodetector in contact with the sample then measures the amount of light transmitted through the material as the probe scans over it.
The experiments showed that defects in the crystal arrangement of the material are related to impurities in chemical composition, propagated along and from the boundaries between adjoining crystal grains. The team also demonstrated techniques to measure the spatial variation of so-called deep defects in the cadmium telluride samples. These defects, which cause electrons and holes (positively charged particles) in cadmium telluride and other semiconductors to recombine instead of generating electricity, are one of the key reasons that solar cells do not perform as well as the theoretical models predict.
According to the researchers, these findings will aid solar cell research, leading to a better understanding of a variety of photovoltaic materials, and consequently, engineering them for greater efficiency.