Detection of the mm-wave radiation using a low-cost LWIR microbolometer camera from a multiplied Schottky diode based source Basak Kebapci 1, Firat Tankut 2, Hakan Altan 3, and Tayfun Akin 1,2,4 1 METU-MEMS Research and Applications Center, Middle East Technical Univ., Ankara, Turkey 2 MikroSens, Ankara, Turkey 3 Dept. of Physics, Middle East Technical Univ., Ankara, Turkey 4 Dept. of Electrical and Electronics Engineering, Middle East Technical Univ., Ankara, Turkey ABSTRACT This paper presents the measurement and analysis method for detection of the mm-wave signal at 96 GHz, by using a low-cost microbolometer infrared (IR) camera with 70 µm pixel-pitch optimized for detection in the 8-12 µm (LWIR) range. The mm-wave beam derived from a multiplied Schottky diode based source is detected within ~ 65 % of the whole area of a 160 120 pixel focal plane array microbolometer sensor. Under ~73 mw incident power, responsivity is measured as 7.3 V/W, and the average noise for the measurement is determined as 12 µv, which includes both detector and readout electronics contribution. From the measured parameters, the integrated Noise Equivalent Power (NEP) is calculated as 1.63 µw within the 7.8 khz readout bandwidth. By using a simple setup, it is shown that a low-cost microbolometer camera which is designed for LWIR range can detect a distinct mm-wave beam at 96 GHz. Keywords: Low-cost microbolometers, mm-wave, noise equivalent power (NEP). 1. INTRODUCTION Millimeter waves (mm-wave) are between 30 GHz and 300 GHz in the electromagnetic spectrum, and they are longer than the IR radiation wavelength. They pass through clothes and reflect off the body, therefore they are very useful for the security screening [1]. Also, harmless, non-ionizing properties of the mm-waves make them preferable for biological applications. Antenna coupled bolometers are widely used cost effective sensors for detection of the mm-waves [2, 3]. In these kind of bolometers, there is an antenna structure that can resonate with the incoming mm-wave radiation and enhance the detector s sensitivity in the desired ranges. NEP is a common parameter for the imaging sensors to quantify the detector sensitivity. It is the amount of radiation flux (Ø) required to make the signal to noise ratio (SNR) equal to 1 (S/N=1) (see Equation 1.1.). NEP can also be explained with the responsivity (R) (see Equation 1.2.) of the detector. Equation 1.3 defines the NEP with the voltage responsivity (R v) relationship [4, 5]. NEP = R v = S N = Ø NEP signal output (V) received power Ø (W) signal output (V)/R v signal output (V)/ V noise (V) = V noise (V) R v ( V W ) (1.1) (1.2) (1.3) As seen from Equation 1.3, a low NEP means high responsivity and low voltage noise in a detection system. In the mm-wave range, low NEP values in the microbolometers can be achieved by using antenna structures. M. Abdel Rahman Millimetre Wave and Terahertz Sensors and Technology VIII, edited by Neil A. Salmon, Eddie L. Jacobs, Proc. of SPIE Vol. 9651, 96510L 2015 SPIE CCC code: 0277-786X/15/$18 doi: 10.1117/12.2199450 Proc. of SPIE Vol. 9651 96510L-1
et al. gives the NEP as 10.2 10 9 W/ Hz in their antenna coupled microbolometer detector working at 94 GHz [6]. A. Rahman et al. achieves 8.3 10 11 W/ Hz NEP with the single element uncooled Nb microbolometer [7]. At 95 GHz, 12 10 12 W NEP value is introduced by the Raytheon group [3]. There have been also many studies with the microbolometer cameras in the shorter wavelength ranges (THz range: 0.3-10 THz) [8, 9, 10], and some of these THz groups have tried to get image in the THz range by using the IR microbolometer cameras which are optimized for the IR region. In one of these studies [11], by using the 160 120 BAE Systems microbolometer LWIR camera, imaging of the obscured razor blade is achieved. This paper also presents a similar work; here a multiplied Schottky diode based mm-wave source output at 96 GHz is detected by an IR microbolometer camera developed at our facilities which is low-cost due to the absence of complex process steps during the production [12], and which is optimized for the LWIR range. From the detected mm-wave clear beam profile, average responsivity, noise, SNR, and the NEP values are extracted. The measurement results demonstrate the feasibility of developing a low cost THz imager based on CMOS and MEMS production technologies. The performance demonstrated here can be further improved by optimizing the pixels for mm-wavelengths. 2. EXPERIMANTAL SETUP AND THE MEASUREMENT The setup is built on an optical table along a single axis. Figure 1 shows the experimental setup used for the measurement. The multiplied Schottky diode based millimeter wave source working at 96 GHz and the 70 µm pixel pitch, 160 120 FPA, a low-cost microbolometer camera developed by MikroSens are used in the course of the measurement. Non-modulated mm-wave signal is transmitted through the microbolometer camera by the aid of two Teflon lenses. Both lenses have 10 cm focal length. One lens is placed 10 cm before the source, the other one is placed 10 cm away from the microbolometer. Computer Screen Microbolometer Teflon Lenses Camera (MikroSens) / 'N mm -Wave Source @96 GHz Waveform Generator 10cm 60cm 10cm Figure 1. The setup used for the detection of the mm-wave beam at 96 GHz by the IR microbolometer camera. The output of the source is amplitude modulated via a function generator in order to measure the power of the emitted radiation with a Golay Cell (Tydex TC-1T). The microbolometer FPA which is sealed under a silicon cap which is also optimized for the LWIR range was placed at the focal length of the second Teflon lens. During the measurement, the signal is not modulated in order to get the maximum power from the mm-wave source. The working frequency and the power of the source are 96 GHz and 90 mw respectively. The incident power on the microbolometer detector is calculated as 73 mw. Figure 2 gives the noise and the beam profile at 96 GHz detected on the microbolometer by aid of the camera electronics and software. The image is filtered using a median filter in order to improve the measurement quality by removing the spatial noise. Since the wavelength of the radiation is much larger than individual pixels, the data was binned on the order of the wavelength of the mm-wave beam at the focus of the Teflon lens. Figure 2.c shows the 3 3 median filtered beam profile. The bright area observed on the screen covers ~65% of the 160 120 microbolometer array. By using the readout of the MikroSens camera, we can extract the relative powers on the pixels. Figure 3 gives the 3D beam profile with different relative power intensities. Figure 4 shows a better image taken by the MikroSens camera at a different time, with a better focus at a higher source power (~100 mw). Proc. of SPIE Vol. 9651 96510L-2
o Ê 2 E o 4- á 6- Noise Profile Median Filtered Beam Profile 8- X Position (mm) X Position (mm) XPosition (mm) (a) (b) (c) Figure 2. The noise and the beam profiles of the detection at 96 GHz by the microbolometer camera. (a) The noise profile of the measurement, while there is no source power, (b) The raw beam profile at 96 GHz, (c) 3 3 median filtered data to remove the noise of the incident signal given in (b). Bean Profile Y Posilion (mm) o o X Posilion (mm) Figure 3. 3D view of the mm-wave beam (given in Figure 2.b) of the multiplied Schottky diode based source with 90 mw working power, detected by the microbolometer camera, at 96 GHz. Figure 4. The mm-wave beam profile of the multiplied Schottky diode based source with ~100 mw working power, detected by the microbolometer camera. A better focus of the beam is achieved in this measurement. As seen from the beam profiles, the incident mm-wave is visible in the 70 µm pixel-pitch, 160 120 FPA microbolometer. The fringes that are visible in the images are thought to be due to reflections due to the high degree of coherence of the source. Bigger and comparable size pixels with the incident wave and optimized sensor packaging in a strong vacuum environment will improve the detection further. Proc. of SPIE Vol. 9651 96510L-3
3. RESULTS AND DISCUSSION Noise, SNR, responsivity, and NEP characteristics of the detector are understood by looking at the voltage outputs in the readout. Average noise and SNR in the detector and readout electronics are extracted as 12 µv and 4.3, respectively. The incident power on each pixel on the illuminated area is unknown, therefore, the incident power on the array is divided by the illuminated area, and an average responsivity of the measurement is estimated as 7.3 V/W from the different output voltages. Figure 5 and 6 give the histograms showing the noise and NEP characteristics of the detection. The average NEP is calculated as 1.63 µw within 7.8 khz bandwidth. Average spectral NEP value is calculated as 18.6 10 9 W/ Hz by dividing the average NEP with the readout bandwidth. Table 1 summarizes the results of the measurement. The average NEP and the spectral NEP values of the measurement is comparable with the studies done about this field as given before. 7000 Noise Histogram 6000 5000 á? 4000 ó É 3000 0 z 2000 1000 0 15 20 25 30 Noise (MV) 35 40 45 50 Figure 5. The noise histogram of the mm-wave measurement at 96 GHz, with a 90 mw source by the low-cost microbolometer camera. The average noise throughout the illuminated pixels is calculated as 12 µv. Integrated NEP Distribution (W) Figure 6. The NEP histogram of the mm-wave measurement at 96 GHz, with a 90 mw source, by the low-cost microbolometer camera. The average NEP throughout the illuminated pixels is calculated as 1.63 10 6 W. Proc. of SPIE Vol. 9651 96510L-4
Table 1. A summary for the results of the mm-wave beam measurement done at 96 GHz with the low-cost microbolometer IR camera. Figure of Merits Results Responsivity 7.3 V/W Noise 12 µv SNR 4.3 NEP 1.63 µw Readout Bandwidth 7.8 khz Spectral NEP 18.6 10 9 W/ Hz An antenna-coupled structure will improve the performance of the detection much better as talked in Section 1. On the other hand, the cap wafer of the microbolometer decreases the transmission of the mm-wave 30%, therefore the cap wafer of the microbolometer should also be optimized up to the desired mm-wave ranges in the future. 4. CONCLUSION In this study, a low-cost microbolometer IR camera is used for the mm-wave detection. We have shown that the 96 GHz mm-wave beam can be detected by using a low-cost LWIR microbolometer camera that is designed for only LWIR (8-12 µm) range. The calculated average NEP is 1.63 µw which is comparable with the literature. This low NEP value shows that even the low-cost microbolometer cameras are appropriate for the mm-wave or THz measurements. However, to improve the image, antenna-coupled structures can be integrated with these microbolometers and optimized packaging in an optimized vacuum environment can also aid in the measurements. REFERENCES [1] Sheen, D. M., Bernacki, B, and McMakin, D., Advanced millimeter-wave imaging enhances security screening, SPIE Newsroom, (2012). [2] Luukanen, A., Miller, A. J., and Grossman, E. N., Active millimeter-wave video rate imaging with a staring 120-element microbolometer array, Proc. of SPIE, 5410, 195-201 (2004). [3] Anderson, C. et al., A 24 x 28 bolometer based passıve millimeter wave imager, Military Sensing Symposium (MSS), (2002). [4] Vincent, J. D., [Fundamentals of Infrared Detector Operation and Testing], Wiley, (1990). [5] Daniels, A., [Field Guide to Infrared Systems], SPIE Press, (2007). [6] Abdel-Rahman, M, Al-Khalli, N, Kusuma, A, and Debbar, N, A slot antenna-coupled microbolometer for detection at 94 GHz, Progress in Electromagnetics Research Letters, 32, 137-143 (2012). [7] Rahman, A., Duerr, E., de Lange, G., and Hu, Q., Micromachined room-temperature microbolometer for mm-wave detection and focal-plane imaging arrays, Proc. of SPIE, 3064, 122-133 (1997). [8] Grant, J. et al., A monolithic resonant terahertz sensor element comprising a metamaterial absorber and micro-bolometer, Laser Photonics Rev., 7(6), 1043 1048 (2013). [9] Marchese, L. et al., A microbolometer-based THz imager, Proc. of SPIE, 7671 (76710Z), 1-8 (2010). [10] Oda, N., Lee, A.W. M., Ishi, T., Hosako, I., and Hu, Q., Proposal for real-time terahertz imaging system, with palm-size terahertz camera and compact quantum cascade laser, Proc. of SPIE, 8363 (83630A), 1-13 (2012). [11] Lee, A.W. M. and Hu, Q., Real-time, continuous-wave terahertz imaging by use of a microbolometer focal-plane array, Optics Letters, 30(19), 2563-2565 (2005). [12] Tepegoz, M., Kucukkomurler, A., Tankut, F., Eminoglu, S., and Akin, T., A miniature low-cost LWIR camera with a 160x120 microbolometer FPA, Proc. of SPIE, 9070 (90701O), 1-8 (2014). Proc. of SPIE Vol. 9651 96510L-5