“There is a great need for materials and substrates that can be used to make infrared detectors directly grown on low cost silicon (Si) electronic circuitry, but such a technology does not exist yet,” says team leader Ray LaPierre. “In our work, we describe how indium arsenide antimonide (InAsSb) nanowires can be used to build a new type of multi-spectral camera that can indeed be directly integrated with existing Si technology.”

For IR cameras, the infrared windows of greatest interest lie between 1–2.5 μm (short-wave infrared, or SWIR), 3–5 μm (mid-wave infrared, MWIR) and 8–12 μm (long-wave infrared, or LWIR), he explains. There are two main IR technologies available today. The first is based on mercury-cadmium-telluride (MCT) and the second on group III-V elements of the periodic table (that is, indium gallium arsenide (InGaAs) and indium antimonide (InSb).

MCT is the only technology that can access all of the SWIR, MWIR and LWIR ranges and MCT-devices that dominate the market today. Imaging arrays based on MCT mostly make use of small-area and expensive CdZnTe substrates, however, which means they are mainly restricted to military applications. The fact that MCT detectors also contain mercury (Hg), which is toxic, makes them far from ideal. SWIR detectors based on group III-V materials, such as InGaAs grown on InP substrates or InSb on InSb, are obviously better in this respect, but they are also expensive to manufacture and restricted to small areas.

Lattice mismatch means IR device materials cannot be directly grown on Si

There is another big problem in that current technologies cannot be easily integrated with Si, something that is needed for image read-out and processing. “This is because IR device materials cannot be directly grown on Si due to lattice mismatch,” explains LaPierre. “To get around this problem, researchers generally grow them instead on expensive II-VI (for example, CdZnTe) or III-V (InP and InSb) substrates that can later be integrated with Si electronic circuitry using indium bump bonding or interconnects. These additional steps slow down the overall manufacturing process and make it costlier too.”

The problems do not end there. “Most current IR cameras cannot image over a wide range of wavelengths, but such a multi-colour capability would be very useful for advanced IR imaging systems,” he says. “Existing multi-spectral cameras are fabricated by growing layers with different material bandgaps on top of one another to absorb different regions of the infrared light spectrum. These bandgaps are generally restricted to lattice-matched materials, which restrict these cameras to only a few layers with specific absorption wavelengths.”

Multiple resonances in a single material system

The McMaster researchers have now found that semiconductor nanowire arrays support optical resonant modes thanks to an antenna effect. Indeed, they allow the nanowires to act as very effective waveguides that concentrate and absorb light over a length of just microns.

“At these resonant wavelengths, the nanowire arrays can absorb light much more than the equivalent thickness of a thin film thanks to this effect,” explains LaPierre. “This enables more efficient photodetection than existing technology and with relatively little material.”

And that is not all: “this resonant light absorption shows wavelength selectivity that can be tuned continuously across the visible and IR wavelengths by adjusting the nanowire diameter,” he tells nanotechweb.org. “Large diameter nanowires absorb long wavelengths of light while smaller diameter ones absorb short wavelengths. By tuning the diameter, we can select which wavelengths are optimally absorbed in a detector or a camera. This allows for multiple resonances in a single material system (with different nanowire diameters being placed on the same silicon substrate, or chip) and this principle can be exploited as a new concept for multi-spectral imaging with improved wavelength selection compared to existing detectors.”

“Relieving” lattice mismatch

The researchers also succeeded in “relieving” the lattice mismatch when growing the nanowires on Si thanks to elastic relaxation near the nanowire surface. “This approach allows high lattice-mismatched InAsSb nanowires to be grown directly on Si sensors,” says LaPierre.

The team says that it will now be fabricating electrodes on the nanowire arrays to make a camera pixel that is sensitive to specific wavelengths. “This work is being performed in collaboration with Teledyne Dalsa and Lockheed Martin Canada,” he reveals. “Credit goes to my PhD student Mitchell Robson, who fabricates all the nanowires. PhD student Khalifa Azizur-Rahman is in charge of the theoretical work and two undergraduates, Daniel Parent and Peter Wojdylo, the optical measurements.”

Full details of the study are reported in Nano Futures 1 3.