A Passive, Real-time, Terahertz Camera for Security Screening, using Superconducting Microbolometers E.N. Grossman, C.R. Dietlein, M. Leivo, A. Rautiainen, and A. Luukanen 1 Optoelectronics Division, National Institute of Standards and Technology, Boulder, CO 80304, USA 2 VTT Technical Research Center of Finland, Espoo, Finland Abstract — We describe a broadband terahertz camera based on a modular 64-element linear array of hot-spot microbolometers Unlike many superconducting sensor arrays, the readout for this array is performed by entirely uncooled electronics; no SQUIDs or cryogenic HEMTs are employed. The operating principles for the microbolometer and the readout scheme are described and compared with those of similar superconducting sensors. Index Terms — Microbolometer, superconducting, terahertz, camera I. INTRODUCTION Terahertz (0.1-1 THz) imaging has been used for decades, mainly in radio astronomy. However, it is presently of intense interest for security screening of personnel[1], which requires a completely new level of user-friendliness, reliability, and robustness. The capability that THz frequency offers is a combination of good penetration through obscurants (clothing in particular) with good diffraction-limited spatial resolution. Other things being equal, passive imaging is preferred over actively illuminating the target, partly because of public acceptance issues with irradiating people, and partly because of the limited capabilities and high cost of THz sources. However, the sensitivity levels required for useful passive screening of people are highly challenging, particularly indoors, where background temperature contrasts are low. Superconducting sensors offer a realistic means of attaining this sensitivity. They can be incorporated into the necessary array sizes much more straightforwardly than uncooled coherent (heterodyne) sensors, which are the only realistic uncooled alternative. Fig. 1 illustrates this sensitivity challenge. Passive radiometer sensitivity is ordinarily specified in terms of noise- equivalent temperature difference (NETD), the minimum contrast in brightness temperature that can be detected with unity signal-to-noise ration (SNR) in a specified integration time, conventionally 1/30 s. The radiation being imaged is simply the Rayleigh-Jeans tail of the blackbody emission from the target, along with reflected blackbody emission from the background. Indoors, the resulting contrasts are therefore generally <20K. The images in Fig. 1 are absolutely calibrated in terms of brightness temperature (a calibrator lies to the right of the subject's head). Each contains the same broadband passive image of a seated male subject concealing a ceramic knife beneath a synthetic fleece. The original image was taken with a single-pixel superconducting sensor, with an NETD of ~200mK. Varying amounts of Gaussianly- distributed random noise were then added in each different panel in order to simulate the same scene, as observed with sensors of progressively poorer sensitivity. Fig. 1 Images of a typical THz concealed weapons detection scenario, with noise added to simulate observation by sensors of NETD=0.5K, 1K, 2K, and 5K. This sequence of images illustrates several points regarding the application requirements. First, threat items can have signal strengths up to 5-10K. Second, realistic THz scenes contain substantial clutter however, both in the background and on the target itself, at levels up to 1-2 K (an example is the folds of the clothing,) that could easily prevent detection of weaker threat signatures. Third, the eye can distinguish and reject this clutter when the sensitivity and spatial resolution of this image are adequate. Thus, the required NETD for this type of screening is approximately 0.5 K. This corresponds, in a 100 GHz bandwidth, to an NEP of 0.7 pW/Hz 1/2 , well beyond the present state-of-the-art for incoherent, uncooled THz sensors. The heart of the camera is a linear 1x64 array of superconducting microbolometers. The array is organized into 8-element modules, each containing a 5x24mm device die in an overall 25x25mm package layout (See Fig. 2). The modularity is important (a) to mitigate the effects of II. SENSOR DESIGN, PRINCIPLE OF OPERATION, AND PERFORMANCE 2 2 1 1 2 U.S. Government work not protected by U.S. copyright IMS 2009 1453