About a decade ago Dr. Eric Mazur, professor of physics at Harvard University, and several of his graduate students discovered a remarkable material known as "black silicon." Curious about what would happen if they irradiated silicon with a femtosecond laser, the group placed an ordinary silicon wafer in a vacuum chamber, filled it with chalcogen-containing gas, and blasted the silicon with ultra-short, super-intense laser pulses. The result was a blackened surface covered with a vast array of microscopic spikes that proved to be up to 500 times more sensitive to light than a standard silicon chip (Fig. 1). SiOnyx, a spin-off company, was formed in 2006 to commercialize black silicon-based photonic devices for a variety of industries.
"We're producing detectors with APD-like response (>100A/W) at low operation bias (<3 V) that are capable of scaling to imaging arrays that will address important market applications where ambient or artificially produced illumination is limited," says Stephen Saylor, CEO of SiOnyx. "Our detectors also exhibit extended wavelength response, with proven performance in short-wave infrared frequencies (SWIR). This further enhances black silicon's applicability to applications in medical imaging, machine vision, night vision, and solar energy."
Figure 1. Responsivity of a black silicon photodetector as compared to commercial silicon, germanium, and InGaAs photodiodes. Responsivity is the amount of current detected for a given amount of light power. Images courtesy of SiOnyx.
The Silicon Conundrum
Silicon detectors are commonly used in photonic systems. Although it is effective for detecting visible light, silicon is virtually useless for detecting other wavelengths, such as SWIR. In fact, between one-third and one-half of solar energy that passes directly through silicon cannot be captured. To overcome this limitation, scientists have made detectors from more exotic materials such as indium, gallium, arsenide, lead, and cadmium to capture infrared light. Although these do the job, some are very expensive to produce and others are extremely toxic to humans and the environment.
"Another limitation with silicon is that it cannot be doped to high levels because its saturation point for most elements is very low," says Mazur. "However, by irradiating silicon with high-intensity laser pulses as short as one billionth of a millionth of a second, in the presence of dopant-containing gas (typically H2S or SF6), dopant concentrations in the silicon can be increased by 10,000 times or more."
Once the chemical structure of the silicon is disrupted by the laser, dopant compounds enter the structure and become "locked in" as the substrate cools and recrystallizes. The result is a highly doped (nanocrystalline silicon containing 1.6% sulfur), 200- to 300-nm-thick, shallow-junction interface that is thousands of times more sensitive to light than conventional semiconductor materials.
Figure 2, above: An array of small black silicon pixels. Arrays of photodetectors are used to take pictures, as in a digital camera. Figures 3 and 4, below: Scanning electron micrograph of black silicon surface with micrometer-scale spikes.
"In a normal semiconductor, impurities (dopants) take very particular energy states, usually inside the band gap (the range of energies in a semiconductor that are forbidden for any native electron to take)," says Mazur. "With black silicon, we have introduced so many impurities that they no longer take individual states and instead begin to form a band, thus changing the band structure of the material." Because a material's band structure determines most of its interesting device properties, this technique may allow SiOnyx to redesign materials to behave in different (and potentially more desirable) ways.
Black silicon results in near-unity absorption from the ultraviolet to the short wave infrared. It also exhibits large photoconductive gain at room temperature, with a quantum efficiency greater than one. Because all this action occurs in the 200-300 nm shallow-junction interface, it is possible to make detectors that are 100 to 1000 times thinner than a standard silicon device. "Silicon is known as an indirect absorber of light, which means that photons cannot be absorbed on their own, but must be assisted by a vibration of atoms within the material, which we call a phonon (rather than photon)," says Mazur. "On average, it takes hundreds of microns of thickness of silicon to ensure that most photons will find a phonon and be absorbed. Black silicon does not share this characteristic: all photons can be absorbed in the first few hundreds of nanometers (versus the first few hundreds of microns for standard silicon)."
Black silicon's unique physical characteristics will revolutionize photonic device architecture.SiOnyx has successfully incorporated black silicon into new silicon devices that show high efficiency, room-temperature photoconductive gain, broad-spectral silicon photodetection, and enhanced near-infrared photovoltaic response.
Figure 5. A microscopic picture of black silicon test circuits.
Point-detector photodiodes made from black silicon absorb about 90% of light at visible and infrared wavelengths ranging from 400 to 1550 nm. The measured optical responsivity of 100 A/W at 950 nm is 100 times more sensitive than standard detection methodologies-an external quantum efficiency of 10,000%. This gain is achieved at a mere 3V of operational bias, enabling direct integration with hybrid and digital circuitry. "The extension of silicon's spectral sensitivity out to 1550 nm has profound implications for laser sensing applications at the 1064, 1330, and 1550 nm nodes," says Saylor.
Black silicon will also redefine silicon based imaging. The shallow-junction laser processenhances silicon's detector response by a factor of 100 or more, allowing the creation of 1 µm2 of black silicon pixels that produce more signal than 36 µm2 of traditional silicon pixels. The improved photoconductive gain results in each photon producing nearly 100 electrons. The ability of black silicon to capture visible and infrared photons within a thin half-micron layer also solves the problem of crosstalk and red/infrared sensitivity issues.
SiOnyx has entered several partnerships to expand its research on black silicon. MIT professor Tonio Buonassisi is using the synchroton at Lawrence Berkeley National Laboratory to determine where the dopants are located in the lattice structure. Professor Alberto Salleo at Stanford University is conducting photothermal deflection spectroscopy experiments. "We don't have any results yet, but I'm excited that so many other groups are helping us understand the origin of these interesting properties in heavily doped semiconductors," says Mazur.
"We have a material that performs like an exotic semiconductor but enjoys the economic benefits of scale," says Saylor."We also have theability to leverage the manufacturing scale of the standard CMOS fabrication infrastructure.The SiOnyx process is conformable to a variety of fabrication environments and device specific architectures.For the first time in the history of photonics, we can produce silicon devices that are competitive with the unique opto-electronic characteristics of exotic materials, in an environmentally safe way, with results that exceed even the most advanced implementations of this legacy technology."
The Grad Student Experience: Mark Winkler
Mark Winkler, a 2009 Ph.D. candidate at Harvard University, is working with Professor Eric Mazur to further understand why non-equilibrium doping creates so many interesting opto-electronic properties in silicon. "My experiments use optical and electronic spectroscopy to determine the energy states of the dopants we introduce into the material," he says.
Winkler is excited about the how black silicon will benefit society. "Our technique is potentially a route to take materials nature has provided for us and engineer them to behave optically in ways that are more desirable or technologically convenient. The impact-from cameras to solar cells-could be enormous."
Regarding the most important evolution in the team's understanding of black silicon, Winkler believes it was "separating the formation of the interesting surfaces (the spikes that you see in SEM images related to our work) from the infrared properties, which is purely a product of the high-level of doping we achieve. This has allowed us to engineer the process (for example, producing flat, infrared-absorbing samples) in ways that isolate the interesting properties and pave the way for future experiments that will unlock all the material's secrets."
Mark Crawford is a freelance science writer based in Madison, Wisconsin.