When you think of thin or fine strands of material, human hair and spider silk may be two examples that spring to mind. But these structures are several orders of magnitude larger than the latest optical waveguide.

An international team of scientists has created silica nanowires - ultrafine strands of glass that guide light and have a diameter of about one-thousandth of a human hair.

Despite their small size, low-loss silica nanowires look set to have a big future. For example, they could become connections for miniature photonic circuits, optical probes capable of detecting biological particles or even the basis of photonic devices such as tiny wavelength splitters.

Fabricating silica fibres that are smaller than 1 µm in diameter is much easier said than done. But last December a team from the US, China and Japan succeeded. Eric Mazur and colleagues from Harvard University, Zhejiang University and Tohoku University claim to have developed a fabrication method that is both repeatable and reliable.

"About two years ago Limin Tong, a researcher from China, joined my group and he was interested in making very-small-scale fibres," explained Mazur. "While in my group he developed this two-step technique to pull these incredibly thin fibres."

Tiny wires, major achievement

The process creates glass wires, which have diameters as small as 50 nm and boast an optical loss of less than 0.1 dB/mm for visible and infrared light. According to the research team, this level of transmission is good enough for use in optoelectronic devices.

"Many attempts have already been made to create thin fibres, but none have had the length and smoothness that we have accomplished," Mazur explained. "All the other processes have resulted in fibres with much rougher sidewalls, which makes them useless at guiding light because it gets scattered. The walls of our nanowires are almost atomically smooth."

These days, scientists can reliably make glass wires with diameters on the order of a few microns by heating a conventional 125 µm diameter optical fibre in a flame then pulling it. When it comes to thinner structures, Mazur says that the temperature distribution of the flame causes problems.

"The flame tends to be very turbulent and the temperature around its edge varies continuously," he said. "If you pull the fibre too fast, or it is too hot, then the wire breaks or the diameter is not uniform."

The solution to the problem, and the secret behind Mazur's uniform nanowires, is a surprisingly simple technique. A flame-heated fibre is drawn into a 1 µm wide wire and wrapped in a spool around a sapphire taper before it is drawn a second time.

"The sapphire taper is then placed in the flame keeping the spool just outside it," explained Mazur. "The taper serves as a buffer to even out the temperature. This is why we are able to draw such reliably uniform wires." Mazur and colleagues have been able to pull uniform nanowires around 2 cm in length.

Another advantage of these ultrafine wires is that they are incredibly strong and flexible, and can be twisted and tied into tight knots. For example, Mazur's team have bent a 280 nm wide wire into a radius of 2.7 µm without breaking, and fracture tests show that the wires have a typical tensile strength of 5.5 GPa. What's more, calculations suggest that a 450 nm diameter wire could route red light around a 90° bend (5 µm radius) with a bending loss of just 0.3 dB.

But perhaps the most amazing result is that, despite having a diameter that is much smaller than the wavelength of visible light, the nanowires act as a waveguide. "We have shown that most of the light is actually guided around the wire rather than inside it," explained Mazur. "The wire serves more as a rail than a funnel."

According to Mazur, a 300 nm diameter nanowire guides about the same amount of red (633 nm) light outside the wire as inside. This makes the nanowires very sensitive to their surroundings and opens up a vast range of applications.

Top of his list is using the nanowires as optical sensors. "One of the things we are doing right now is to coat the outside of these wires with biological receptor molecules," he explained. The idea is that when a specific pathogen binds to the receptor, it interacts with the propagating light and changes the transmission spectrum of the nanowire. "We can't detect single molecules - that would be really hard. But we can detect tens or hundreds of molecules, making this a very sensitive sensor."

The researchers are now collaborating with chemists and biologists at Harvard to construct and test such sensors.

The nanowires are also destined for use in photonic devices, ranging from simple low-loss connections to more complex devices, such as wavesplitters. As much of the light is guided around the wire, it is easy to couple light between two touching wires, through a process called evanescent field coupling. The ease with which light can be injected or removed from nanowires makes them particularly attractive for use in miniature photonic circuits.

Mazur claims that a contact region about 10 µm long is enough to transfer the light from one nanowire to another.

"It is very easy to slide the light from one wire to another and this is good for photonic circuits," said Mazur. "You can for example change the transmission of one fibre by having it on top of a grating. This is another thing that we are actively engaged in."

Turning to photonic devices, Mazur's team has already studied a simple nanowire knot, which also acts as a ring cavity, or in other words a miniature interferometer. Light travels around the ring of the knot then undergoes constructive or destructive interference, depending on the length of the knot and the wavelength of the light.

"Just by making a knot you have an extremely simple photonic device," said Mazur. "We have measured the Q factor of this and it is quite encouraging. If you were using normal optical fibre, you would not have a high Q factor and the ring would not be a good resonator. The reason our device works so well is that a lot of light travels outside the ring."

Mazur hopes that the applications being tested in his lab will cross over into industry within the next few years. Having patented the technique, he is also keen to persuade others to join him in developing new applications and devices.

Promise and potential

It seems obvious that the best is yet to come in this area of research, which is very much in its infancy. "Who would have thought that you could bend light around a radius of just a few times its wavelength?" said Mazur. "We've seen light with a wavelength of about a micron travel around a bend of around several microns. It's extraordinary."

And for Mazur, there is the added satisfaction that these amazing results have come out of research that is mounted on a simple breadboard. "On the one hand the guiding of light and things like nanotechnology and photonics are high-tech terms, but on the other the tools that we used to achieve all of this were exceedingly low-tech," he said. "It's simple, reliable and reproducible."