Despite the saying, sometimes you can’t believe it when you see it. Researchers at Université de Technologie de Troyes in France, at the University of Illinois at Urbana-Champaign and at Argonne National Laboratory in Illinois have shown that this is true for apertureless scanning near-field optical microscopy, a technique used for subwavelength optical imaging. They found an artifact arising from imperfections in the tapping tip used in the technique. According to Marc Lamy de la Chapelle, an associate physics professor at Troyes, the investigation originated with imaging of surface phenomena on gold nanoparticles. “We wanted to observe it in the near field. As the signal is very low, we observed this artifact.”This schematic shows the setup of an apertureless scanning near-field microscope. Researchers discovered that the vibration of the tip — when in tapping mode — causes avoidable image artifacts. Images reused with permission from L. Billot, Applied Physics Letters. ©2006, American Institute of Physics.Classically, diffraction restricts resolution to about half a wavelength. Consequently, optical resolution is a few hundred nanometers at best, too large for nanoparticles. In apertureless scanning near-field optical microscopy, a sharp tip enables imaging far below the diffraction limit. When close to the sample, the tip’s near-field effect concentrates electromagnetic waves, permitting imaging at very small resolution.This subwavelength imaging is performed using an atomic force microscope. In operation, the tip is forced up and down in a tapping mode as it scans across the sample’s surface. A light source, such as a laser, illuminates the tip, and a detector collects the varying light intensity. Because the tip is vibrating while the background remains constant, the desired signal is extracted by locking into the tip’s frequency.Topographical images of gold nanoparticles (A and D) are shown in comparison with error signals generated by an atomic force microscope tip in tapping mode (B and E) as well as with images acquired with the use of a 620-nm laser as the light source (C and F). Arrows show the path of the scanning tip, and the scale bars = 400 nm.In practice, the researchers found that the technique does not always work as designed. They looked at two kinds of samples: 50-nm-high gold ellipses manufactured via electron beam lithography and 500-nm-wide × 450-nm-deep gold nanowells crafted with nanoimprint lithography.They used a custom-built atomic force microscope, with either a dye laser operating at 580 to 620 nm or a krypton-ion laser at 647.4 nm — both from Coherent Inc. — as a light source. For a detector, they used a photomultiplier tube made by Hamamatsu. With this setup, they scanned the tip across the samples, setting the tip to maintain a constant vibration amplitude between 10 and 100 nm.Because the tip encountered hills and valleys among the raised ellipses or sunken wells, it took time for it to recover to the set point. That difference from the ideal, explained Lamy de la Chapelle, is the source of the error. “The artifact originates from the real displacement and vibration of the tip in tapping mode.”The effect adds a blip to the microscopic image that is dependent on scan direction and that occurs at edges. Given the artifact’s origin, there are several possible countermeasures. The error comes from reflection off the tip, so one countermeasure is to focus the laser at a point below the surface and away from the tip. Another is to reduce tip vibration amplitude variation. It also might be possible to reduce the error by reducing the scan rate.Finally, Lamy de La Chapelle noted that optical images should be published with topography, with error signal and in two scanning directions. “This suggestion has been done to be sure that the optical images are not artifactual.”Applied Physics Letters, Vol. 89, 023105, 2006.