Getting to the (Very Small) Point of Imaging

Hank Hogan

Sometimes ideas pan out better than expected. Such is the case with an imaging innovation from the University of Maryland, Baltimore County. Under the direction of assistant professor of chemistry and biochemistry Brian M. Cullum, researchers there devised fiber optic nanoprobes for high-resolution, nonscanning chemical imaging. Using the probes to perform surface-enhanced Raman scattering (SERS), the technique could have such applications as materials science and cellular biology.

Researchers have developed a nanoprobe that enables high-resolution, nonscanning chemical imaging through the use of surface-enhanced Raman scattering effects. Here, it is shown positioned on an optical microscope for the investigation of cellular surfaces. Once in place, the untapered end of the probe can be imaged with the multispectral system.

Raman scattering enables label-free chemical identification. Although the scattering is weak and capturing the signal difficult, especially for trace chemicals, surface effects generated by photons near a metallic surface boost the signal a millionfold or more. The technique has been combined with subdiffraction-limited, scan-based imaging to enable 100-nm spatial resolution, but such techniques require hours or days, too long for many applications.

In developing the probes, the scientists tried etching a fiber optic bundle to create needed structures. “When we did this, it turned out — to our surprise — that we generated a series of uniform spikes around each element,” Cullum said.

Those spikes ringed the fiber elements used by the researchers to collect signals. As a result, he said, the probes achieved much greater signal throughput and better resolution than they would have done otherwise.

They employed a micropipette puller to taper the tips of an optical fiber imaging bundle comprising 30,000 individual elements. The 4-μm-diameter cores were drawn down to diameters ranging from 100 to 1100 nm, the final size a result of conditions and expertise. “With a little practice, it is possible to reproduce the tips of these fibers very accurately,” Cullum said.

Next, the researchers etched the tips with hydrofluoric acid, which attacked the fiber core faster than the surrounding cladding. The result was a recessed core surrounded by six peaks for each element. They then deposited silver on the tips, tilting the probe at an angle, creating six silver islands evenly spaced around the nanowell rim. The islands served as the highly uniform tips needed for surface-enhanced Raman scattering imaging. The scientists, therefore, could image a large area with high resolution without scanning.

For an imaging demonstration, they used the 632.8-nm emission from a laser made by JDS Uniphase of Milpitas, Calif., to illuminate a sample. They placed the nanoprobe in contact with the sample and collected the resulting scattered photons. They selected the desired imaging wavelength with an acousto-optic tunable filter from Brimrose Corporation of America of Baltimore and captured images using an intensified CCD made by Princeton Instruments of Trenton, N.J.

With this setup, they imaged benzoic acid and brilliant cresyl blue and showed that the images were the result of surface-enhanced effects, not spontaneous Raman scattering. They also showed that the resolution, as expected, was twice the diameter of the fiber elements.

Cullum foresees applications in the imaging of lipid rafts in cellular membranes and in the monitoring of impurities in silicon wafers. On the technology side, he envisions the development of a nanoprobe that can be easily connected to a CCD chip or camera for rapid imaging at high spatial resolution.

He is working on solving a disadvantage of the technique, the loss of sensitivity, caused by silver oxidation. “We may have a long-term solution to this problem, with the generation of multilayer gold SERS substrates,” he said.

Analytical Chemistry, Nov. 1, 2006, pp. 7535-7546.