Chemical species can be detected optically using optical correlation spectroscopy: broadband light passing through an unknown gas sample picks up spectral features of the sample; by correlating the light with known spectra, species in the sample can be identified. Work is underway to develop programmable diffraction gratings that can synthesize molecular spectra for use in an optical correlation spectrometer. These consist of micromachined diffraction elements whose "weighting" can be varied. When broadband light is incident on the array, light reflected from the individual elements interferes to produce a desired spectrum.
|Programmable Diffraction Grating|
There are many instances in which it would be desirable to be able to look through an optical device and determine spectroscopic information about chemicals which may be in the atmosphere at a remote location. These include: battlefield conditions where the target may be chemical agents, biological agents, conventional explosives, fuel-air bombs, tactical nuclear weapons, vehicle emissions, safety monitoring of storage areas for hazardous materials, environmental monitoring and control, or industrial process control. Correlation spectroscopy is an attractive technique in this regard, since it combines the attributes of high selectivity and good sensitivity. In a conventional correlation spectrometer, broadband infrared radiation passes through the sample volume where the chemical species present absorb energy in characteristic spectral lines. The infrared radiation the traverses a reference cell containing a known target compound, whose spectral features are modulated (for example, by pressure modulation). If the spectral features of the sample match (or correlate) with those of the target compound, an overall intensity modulation is observed. Otherwise, the output intensity is not modulated. The primary drawback associated with correlation spectroscopy is the need for reference cells containing real compounds. These cells end to be bulky and fragile, and different modulation schemes are required for different compounds. As a result, it is difficult to envision a correlation spectrometer which is capable of detecting more than a few compounds. Further, there are many compounds of interest, such as transient species, highly toxic species, or corrosive species, for which construction of a reference cell is difficult or impossible.
These limitations can be overcome by encoding the desired spectral information in computer generated holographic optical elements which can be used to replace the cumbersome reference cells of conventional correlation spectroscopy. The elements can be fabricated on silicon wafers using standard microlithographic of the target compound, an "mask and etch" techniques. Storage of a large number of holograms in CD-like format would enable the development of a compact "holographic correlation spectrometer" which is capable of detecting multiple species (including toxic, corrosive and transient compounds).
A numerical procedure for encoding infrared spectra in holographic elements has been developed. This method utilizes scalar domain diffraction theory, in combination with an iterative Fourier transform algorithm, to generate a surface relief profile which contains the desired spectral information. Using this procedure, a reflection mode hologram was designed to synthesize the HF spectrum at a diffraction angle of 15°. It consists of 4096 lines, each of width 4.5 μm, for an overall length of 18.4 mm. The surface relief profile was fabricated on a silicon wafer using a four-mask-level process and anisotropic reactive ion etching. After the etching process is complete, a thin gold layer was sputter deposited onto the surface profile to enhance the surface reflectivity. Figure 1 is a photograph of a wafer containing an HF element, and Figure 2 shows a scanning electron micrograph of a portion of the diffractive element.
|Figure 1. A photograph of a wafer containing the finished HF element (the large rectangular region on the bottom left).||Figure 2. A scanning electron micrograph of a portion of the holographic optical element designed to synthesize the HF spectrum.|
The infrared spectrum of HF is shown in Figure 3a. It consists of a series of sharp rotational lines contained within a vibrational envelope. The theoretical synthetic spectrum predicted by the design algorithm (Figure 3b) closely matches the real spectrum. The infrared spectrum recorded for the HF hologram is shown in Figure 3c. Apart from an increase in the widths of the rotational lines, the spectrum shown in Fig. 3c closely corresponds to the real HF spectrum. The increased linewidth is due to a combination of the finite spectral resolution of the 18 mm element, and the excess linewidth due to optical aberrations in the read-out apparatus. Both of these factors can easily be improved to yield narrower linewidths. The data of Figure 3 represent the first demonstration of synthetic infrared spectra.
|Figure 3. a) The infrared spectrum of HF. |
b) The theoretical spectrum predicted by the design algorithm.
c) The measured spectrum.g
Over the past decade, diffractive optics have been developed for use inspatial domain optical processes, such as image correlation which is used for target recognition and tracking. The ability to encode accurate spectral information in diffractive elements now extends these concepts into the spectral domain. The holographic correlation spectrometer, which can be viewed as the spectral domain analog of an image correlator, is one application of this new capability. It is anticipated that other sensing (and imaging) techniques can take advantage of the spectral processing functions that can be performed by these optical elements.