Tracing gases

The ability to sniff out trace amounts of potentially dangerous gases is critical in many contexts. In the home, for example, detecting buildup of carbon monoxide can save lives. Various manufacturing environments rely on early detection of toxic effluents such as sulfur dioxide, methane, and nitrous oxide to ensure safe operation. Sensors that can pick up trace gases can be enormously important in meteorology, environmental protection, and bio-hazardous material identification.

IR sensors make up a significant part of the gas-sensor instrumentation market, and IR spectroscopy sensors remain the most accurate and reliable because they don't rely on catalytic or electrochemical interactions. The sensor elements in many other types of devices are prone to react with chemical species similar to the target and can become contaminated or depleted with prolonged use. Since IR gas-sensor elements do not touch the gas, they are not poisoned by contact with the environment.

Conventional IR gas sensors, unfortunately, are big, complex, and expensive. A non-dispersive, IR spectroscopic gas sensor, for example, includes a broadband IR light source, one or more IR detectors, each with a unique filter to select gas-specific absorption bands, and instrumentation electronics.

The IR detector is filtered to monitor a specific part of the IR spectrum. As the gas of interest passes between the powered IR source and the detectors, the absorption spectrum changes, along with the detector output. The instrument electronics process this data to determine the concentration of the gas of interest, potentially setting off an alarm, triggering an action, or simply providing the information for monitoring purposes.

We have developed a novel, thermally stimulated, narrow-band IR source/detector by combining a photonic crystal consisting of an array of holes etched into a dielectric substrate (silicon, for example)¹ and a periodically perforated metallic (typically gold or aluminum) thin film. The structure is formed as a layer stack—silicon and metal—and then we etch holes through the metal layer and into the silicon to a certain depth. The dielectric photonic-crystal structure is passive and exhibits a strong broad absorption around a resonance. This acts as a radiation reservoir for the conductive array, which plays an active role through plasmon interactions and is opaque at all wavelengths except those at which coupling occurs.² The metal film, perforated with apertures approximately the size of the wavelength of the light, permits resonant coupling of the incident radiation from the underlying silicon with surface plasmons at the metal surface. This coupling yields unusually high optical transmission efficiencies where, again, the periodicity of the structure comes into play.

The emission process begins in the bulk silicon, where thermal stimulation produces blackbody-like radiation. The silicon/air pattern of the photonic crystal reshapes this spectrum of radiation and centers it around a specific wavelength of resonance defined by the lattice spacing of the crystal.³

The photons cannot penetrate through the thin metal film on the top surface. Instead, they excite surface plasmon modes at the lower substrate/gold interface, where vertical trenches create a boundary condition. There is a resonant interaction of the incident light with the surface plasmons on both surfaces of the metal film. The interaction occurs by coupling of the photon momentum to the plasmon dispersion through the allowed reciprocal wavevectors of the periodic grating. The resulting equation is:

where ε₁ is the dielectric constant of the interface medium, ε₂ is that of the metal, a is the lattice periodicity, λ is the wavelength, and i and j are integer numbers. The result of this equation is that the dielectric layer in contact with the metal, along with the lattice constant of the array, will influence the spectral position of the resonant wavelength. The surface plasmons then decay into photons, which are emitted from the surface.

We prepared samples with square and hexagonal arrays having lattice constants between 2 and 9 µm. We observed a linear dependence between the wavelength of resonance and the periodicity of the array.

It is important to note that this behavior only works when all of these processes occur together in one continuous sequence. If we separate the system into its individual components—a bulk silicon wafer, a silicon photonic crystal, and a perforated metal film—we do not observe the same effects. We cannot, therefore, describe this interaction as a filtering-type mechanism but rather as an enhanced emission mechanism.

Our SensorChip is a wavelength-tuned, micro-electro-mechanical-systems-based micro-bridge element. Using photonic-crystal and surface-plasmon interactions, the surface emits and absorbs efficiently in a narrow waveband centered on the signature wavelength of the target gas. Our ability to tune the IR emission and absorption allows us to integrate the optical filter, detector, and IR source onto a single device. A simple lithographic mask change allows us to change the spectral tuning of the micro-bridge in production. Thus we are literally printing sensors on silicon-on-insulator wafers.

Figure 1. IR photons from the chip reflect off a mirror (top) and travel back to the chip or detection (center), with the flux changing depending on the amount of gas in between (bottom).

We created this structure in a micro-bridge geometry to enable power-efficient heating of the device and to enhance its sensitivity to gas concentration. The marriage of these concepts reduces power-consumption times, size, and price compared to discrete IR gas sensors.

In practice, we electrically heat the micro-bridge, causing it to emit radiation. The IR energy from the chip passes through the sample gas, reflects off the mirror at the far end of the sample chamber, passes again through the gas, and is reabsorbed by the silicon micro-bridge that also serves as a tuned IR detector (see figure 1).

Figure 2. Fractional absorption depends on the concentration of carbon dioxide between the sensor and the mirror, with shutter representing complete blockage of light.

In this arrangement the silicon bridge receives energy from both the electrical stimulation and the reflected light and comes to thermal equilibrium. If the target gas is introduced, the light reabsorbed by the silicon micro-bridge will decrease, causing the system to reside at a lower temperature. The drop in temperature is proportional to the concentration of the target gas and is measured via a change in resistance or voltage across the bridge (see figure 2).

Our optical technology platform allows us to build all the optical components—emitter, filter, detector—onto a single silicon chip, reducing system size, power consumption, complexity, and cost. Since our chip is designed according to standard microelectronics design practices, we can fabricate it in standard semiconductor foundries. This could emerge as the first commercially available application of photonic crystals.

We collected raw data using the sensor chip with various concentrations of carbon dioxide (CO₂) gas introduced into the chamber between device and mirror. With 100% attenuation defined as the signal resulting when the entire reflected signal is blocked by a shutter, the gas absorption measurements are then normalized to this complete attenuation measurement. The system is sensitive to the presence of CO₂, and the various concentrations can be distinguished. Further refinement to the surface patterning design will tighten the spectral emission of the bridge and enhance the system's sensitivity to gas.

By incorporating photonic-crystal technology to control the spectral emission of light from surfaces and building both the emitting and sensing capabilities on a chip, we hope to bring a new realm of science closer to vital commercial applications. oe

References

1. J. Joannopoulos, R. Meade, and J. Winn, Photonic Crystals (Princeton University Press, Princeton, 1995).

2. H. Raether, Surface Plasmons (Springer, Berlin, 1988).

3. M. Pralle et al., Applied Physics Letters 81[23], 2002.

Described as the City of Contrasts or "the Little Paris," Bucharest is a modern crucible of creation. At this ancient crossroads of Eastern Europe, centuries have stirred tradition, bucolic beauty, and modern civilization into a rich patisserie that feeds the imagination. It's no surprise, then, that a young Irina Puscasu would walk Bucharest's winding streets and find her inspiration in all the colors of the rainbow. "I've always been interested in color. That's what's so funny?now I work in the infrared, which you can't see," she says with a laugh. "But I also needed to know how and why things worked."

Born with an artist's disposition, Puscasu likely was destined to travel the more disciplined road of scientific exploration. "I came from a family of engineers," she says. "My mother's a chemist, my father a civil engineer, and my brother is a computer engineer." Mentoring and role models continued to be important to Puscasu's development, even as she chose the University of Central Florida's Center for Research and Education in Optics and Lasers (CREOL; Orlando, FL) to pursue her postgraduate work in physics. At CREOL, Puscasu received the school's first Student of the Year award for her role as president of the SPIE student chapter, vice president of the OSA student chapter, and participation in women-mentoring programs and even local science fairs, in addition to her scientific work. Today, Puscasu continues to support younger minds on SPIE's membership and Women in Optics committees.

Puscasu continues her work on IR sensing at Ion Optics Inc. (Waltham, MA), where she's using her experience with lithography to create gas sensors on a chip, among other projects. "This is the high point of my career because I'm finally getting the chance to make something that's useful," she says. "I haven't had a chance to do that going all the way back to my thesis!" -Winn Hardin