Molecular fingerprints acquired on a small scale

Hank Hogan

At Max Planck Institut für Biochemie in Martinsried, Germany, a team has succeeded in taking some very small fingerprints, using scattering-type near-field scanning optical microscopy (s-NSOM) to measure the infrared spectra of individual nanobeads and viruses. The particles were as small as 18 nm.

The near-field technique complements other methods of detecting nanoparticles, such as light microscopy, by providing a means to identify the makeup of particles without the need to label them first. Researcher Markus Brehm noted that this could be helpful when characterizing engineered nanodevices or when researching biological specimens. “The technique will be useful whenever chemical identification of material is needed at nanoscale resolution,” he said.

In using s-NSOM, he exploited the fact that a sharp, conductive probe tip confines nearby electromagnetic fields and enhances them. These near-field augmentations make it possible for the tip to sense local properties of particles and materials.

The optical properties of a sample near the probe are determinable by measuring the light scattering from the tip. In addition, the sharpness of the tip — not the wavelength of the scattered light — determines the lateral resolution. Thus, resolutions of less than 20 nm are possible using visible or infrared illumination. In their experiments, the researchers used tips with a diameter of roughly 30 nm, as measured by electron microscopy.

The investigators report in the July issue of Nano Letters that they used a line-tunable carbon monoxide laser from Invivo GmbH of Adelzhausen, Germany, that produced a beam of ∋6 μm wavelength. They also used cantilevered, platinum-covered silicon tips from MikroMasch Inc. of Wilsonville, Ore., which they mounted in a custom-built scanning force microscope. They operated the microscope in a tapping mode, forcing the tips up and down about 25 nm at 33 kHz. When capturing infrared spectra, they moved the sample beneath the tip.

To create a scattering near-field effect, they focused the laser onto the moving tips. Some of the light reflected back, and they superimposed this onto a reference beam in a Michelson interferometer. They captured the resulting signal using a mercury-cadmium-telluride infrared detector from Kolmar Technologies Inc. of Newburyport, Mass. By regularly shifting the reference beam’s phase, they separated amplitude and phase from each other in the interferometric signal. After demodulating this signal, the researchers had the s-NSOM output.

In a series of studies, they first looked at spherical beads of polymethyl methacrylate (PMMA) that varied in diameter from 30 to 70 nm. They also examined tobacco mosaic viruses, which were cylindrical and had a diameter of 18 nm. They chose these subjects for the demonstration in part because they had distinct vibrational resonances in the spectral region covered by the laser that was being used. They put the particles into suspension, dispersed that across a silicon or gold-coated silicon substrate, and then dried it. They mapped the substrate and the particles using their s-NSOM setup.

Brehm reported that the researchers did find something unexpected when they looked at their results. “We were quite surprised by the strength of the spectral signature,” he said.

He added that this outcome was predicted when they used a simple dipole model on the various electromagnetic interactions. Whether in theory or practice, the enhancement from near-field effects was pronounced, upping the signal scores of times. In conventional infrared transmission spectrometers, a sample thickness of ~1 μm would be needed to generate a signature of similar strength.

Near-field scanning optical microscopy was used to visualize tobacco mosaic virus, with an additional component that measured the light scattered from polymethyl methacrylate (PMMA) particles placed on the virus. Numbers indicate the height of the PMMA particles in nanometers. Courtesy of the American Chemical Society.

As might be predicted given the strength of the signal, the researchers were successful in acquiring telltale infrared signatures. For example, they found a phase contrast peak with the substrate near 1730 cm^—1 for a 68-nm PMMA bead, which agrees with known absorbance spectra. They also found a phase contrast peak near 1660 cm^—1 for virus particles, again in agreement with known absorbance spectra. By comparing the s-NSOM results with the size of the particles as determined by other methods, the researchers estimated that the spatial resolution of their instrument was ~30 nm. They also calculated that the amount of sample needed for a measurement was about 10^—20 l.

Although the demonstration was promising, some improvements must be made before the technique can be widely applied. The clear difference between the substrate, which has no spectral features in the frequency region of the probe beam, and the virus particles varied according to the infrared frequency. Thus, to build up a complete spectrum of a single virus, the investigators had to repeat the scan using a different frequency. Each pass took 10 minutes, along with some laser configuration time. Consequently, extracting the spectrum of a single particle took a day — too long for many applications. It also is difficult to maintain the setup for that long.

Brehm reported that the researchers were working on enhancements to the technique that would eliminate this problem. “The key is to move from our sequential way of taking spectra — taking images repeatedly at different wavelengths — to a parallel method that takes a complete spectrum at each pixel during scanning,” he said.

He noted that progress has been made by combining infrared frequency-comb spectroscopy with near-field microscopy. The group also is working with theoreticians to improve the modeling of the scattering from the tip-particle geometry substrate, something that will be necessary to more completely interpret the measurements.