Proc. SPIE. 6549, Terahertz for Military and Security Applications V
KEYWORDS: Signal to noise ratio, Detection and tracking algorithms, Spectroscopy, Control systems, Heterodyning, Laser stabilization, Terahertz radiation, Chemical detection, Toxic industrial chemicals, Absorption
Goodrich has been developing a high resolution, broad band spectrometer that operates in the Terahertz (THz) region of the spectrum with the intent of performing chemical detection. THz spectroscopy exploits rotational resonances for detection of gas phase compounds. High resolution THz spectroscopy can improve detection and identification through increased probability of detection and reduced false alarms. The Goodrich THz spectrometer is based upon CW photomixer technology in a heterodyne configuration. The current Goodrich design offers continuous tunability across a 0.1 THz to 1.2 THz frequency range. One of the unique aspects of the Goodrich spectrometer is laser system control that has demonstrated difference frequency line widths on the order of 1.5 MHz with stability measured over long time scales. Absolute frequency accuracy is of the order of 4 MHz. The spectrometer design enables high THz energy densities with narrow line widths tunable over a broad spectrum. The system has demonstrated SNR better than a cryogenically cooled hot electron bolometer. This capability allows the Goodrich system to accurately determine absorption signatures of multiple chemicals with exceptional performance. Goodrich has completed initial system testing and verified performance. Initial tests were completed to determine SNR of the heterodyne photomixer transceiver. System performance was also verified for laser line width, stability, and repeatability. The spectrometer was tested against various toxic industrial chemicals. Preliminary data for HCN, HCl, NH3, and SO2 is presented.
Collaboration with the University of Virginia (UVa) and the University of California, Santa Barbara (UCSB) has resulted in the collection of signature data in the THz region of the spectrum for ovalbumin, Bacillus Subtilis (BG) and RNA from MS2 phage. Two independent experimental measurement systems were used to characterize the bio-simulants. Prior to our efforts, only a limited signature database existed. The goal was to evaluate a larger ensemble of biological agent simulants (BG, MS2 and ovalbumin) by measuring their THz absorption spectra. UCSB used a photomixer spectrometer and UVa a Fourier Transform spectrometer to measure absorption spectra. Each group used different sample preparation techniques and made multiple measurements to provide reliable statistics. Data processing culminated in applying proprietary algorithms to develop detection filters for each simulant. Through a covariance matrix approach, the detection filters extract signatures over regions with strong absorption and ignore regions with large signature variation (noise). The discrimination capability of these filters was also tested. The probability of detection and false alarm for each simulant was analyzed by each simulant specific filter. We analyzed a limited set of Bacillus thuringiensis (BT) data (a near neighbor to BG) and were capable of discriminating between BT and BG. The signal processing and filter construction demonstrates signature specificity and filter discrimination capabilities.
This work presents spectroscopic characterization results for biological simulant materials measured in the terahertz gap. Signature data have been collected between 3 cm-1 and 10 cm-1 for toxin Ovalbumin, bacteria Erwinia herbicola, Bacillus Subtilis lyophilized cells and RNA MS2 phage, BioGene. Measurements were conducted on a modified Bruker FTIR spectrometer equipped with the noise source developed in the NRAL. The noise source provides two orders of magnitude higher power in comparison with a conventional mercury lamp. Photometric characterization of the instrument performance demonstrates that the expected error for sample characterization inside the interval from 3 to 9.5 cm-1 is less then 1%.