Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.cplire.ru/html/lab234/pubs/2011_27.pdf Äàòà èçìåíåíèÿ: Wed Sep 26 18:28:43 2012 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 08:51:51 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: universe

IEEE TR ANSAC TIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 21, NO. 6, DECEMBER 2011

3635

Optical Response of a Cold-Electron Bolometer Array Integrated in a 345-GHz Cross-Slot Antenna
Mikhail A. Tarasov, Leonid S. Kuzmin, Valerian S. Edelman, Sumedh Mahashabde, and Paolo de Bernardis
Abstract--Two series/parallel arrays of ten cold-electron bolometers with superconductor­insulator­normal tunnel junctions were integrated in orthogonal ports of a cross-slot antenna. To increase the dynamic range of the receiver, all single bolometers in an array are connected in parallel for the microwave signal by capacitive coupling. To increase the output response, bolometers are connected in series for dc bias. With the measured voltage-to-temperature response of 8.8 µV/mK, absorber volume of 0.08 µm3 , and output noise of about 10 nV/Hz1/ 2 , we estimated the dark electrical noise equivalent power (NEP) as NEP = 6 10-18 W/Hz1/ 2 . The optical response down to NEP = 2 10-17 W/Hz1/ 2 was measured using a hot/cold load as a radiation source and a sample temperature down to 100 mK. The fluctuation sensitivity to the radiation source temperature is 1.3 10-4 K/Hz1/ 2 . A dynamic range over 43 dB was measured using a backward-wave oscillator, a variable polarization grid attenuator, and cold filters/attenuators. Index Terms--Bolometer arrays, cold-electron bolometers (CEBs), cross-slot antennas, superconducting integrated circuits.

I. I NTRODUCTION HE COLD-electron bolometer (CEB) array has been proposed for use as the detector for the 345-GHz channel of BOOMERanG [1]. The requirement is to develop a CEB array with a junction field-effect transistor (JFET) readout for 90 channels. The noise equivalent power (NEP) of the CEB should be lower than photon noise for an optical power load of 5 pW, and polarization resolution should be better than 20 dB for observations of cosmic microwave background (CMB) foregrounds. The radiation power in the CEB

T

Manuscript received July 29, 2011; revised September 9, 2011; accepted September 19, 2011. Date of publication October 27, 2011; date of current version December 2, 2011. This paper was recommended by Associate Editor M. Mueck. This work was supported in part by Swedish agencies, namely, the Swedish Research Council (VetenskapsrÅdet), the Swedish National Space Board (Rymdstyrelsen), the Swedish Foundation for International Cooperation in Research and Higher Education (STINT), and the Swedish Institute (Svenska Institutet); by the Russian Foundation for Basic Research under Grant OFIM12145; and by the Ministry of Education and Science of the Russian Federation under Grant "Invited Principal Investigator." M. A. Tarasov is with the Kotelnikov Institute of Radio Engineering and Electronics, Russian Academy of Sciences, Moscow 125009, Russia (e-mail: tarasov@ hitech.cplire.ru). L. S. Kuzmin and S. Mahashabde are with Chalmers University of Technology, 41296 Gothenburg, Sweden. V. S. Edelman is with the P. L. Kapitza Institute for Physical Problems, Russian Academy of Sciences, 119334 Moscow, Russia. P. de Bernardis is with the University of Rome "La Sapienza," 00185 Rome, Italy. Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TASC.2011.2169793

is absorbed in a thin normal metal film connected to two superconductor­insulator­normal (SIN) metal tunnel junctions. These SIN junctions serve for electron cooling (similarly to the Peltier effect in semiconductors), and the output signal proportional to the absorbed power is measured on them. Electron cooling makes it possible not only to improve the sensitivity but also to expand the dynamic range due to an increase in the saturation power, because the absorbed power is removed from the absorber by a cooling current [2]. Contrary to heating by dc bias in a hot-electron bolometer, in the CEB, dc bias leads to direct electron cooling. As a result, the noise properties of this device are considerably improved by decreasing the electron temperature. However, for applications in atmospheric radio astronomy such as the BOOMERanG project, the power of microwave background radiation is usually higher than the saturation power of a single bolometer. Our previous theoretical and experimental studies on single CEBs show quite promising NEP down to 2 10-18 W/Hz1/2 [3] for dark measurements. Nevertheless, simulations show that it is impossible to satisfy power load requirements of 5 pW with a JFET readout for a single CEB, for both current- and voltage-biased modes. A novel concept of a series/parallel array of CEBs in a currentbiased mode has been proposed to effectively match a JFET amplifier readout [4]. The main advantage of the CEB array in comparison with a single CEB is the distribution of incoming power between N CEBs; summing the output signals results in an increased response from the array. An effective distribution of power is achieved by a parallel RF connection of CEBs, which couple to the RF signal through additional capacitance values. The total response is increased because the voltage response of each CEB improves for lower background power, and this is increased by a factor of N in the array configuration, with a corresponding decrease in absorber overheating and saturation. The voltage responsivity in the current-biased mode is SV = (Varr / Ts )/Gsum = N ((Vsing / Ts )/Gsum ), where Varr and Vsing are voltages across an array and a single bolometer, respectively; Te is the electron temperature; and Gsum is the total heat conductance. The amplifier noise related to the array is also proportionally reduced to array responsivity SV . The amplifier impact to NEP 2 NEP2 = (V 2 +(NR I )2 )/SV , where V and I are the amp voltage and current noise spectral densities, respectively, and Te is the electron temperature. On the other hand, an increase in N leads to an increase in electron­phonon noise. In our design, we found an optimal number of CEBs, i.e., around ten; in this case, the total noise of the detector becomes less than the photon noise of the incoming signal power load of 5 pW [4].

1051-8223/$26.00 © 2011 IEEE


3636

IEEE TR A N S AC T I O N S O N AP P L I E D SU P E R C ONDUCTIVITY, VOL. 21, NO. 6, DECEMBER 2011

Fig. 1. Optical image of the cross-slot antenna. Arrays of CEBs are connected to four ports of the antenna. Diagonal slots are capacitive shunted by an additional Al layer deposited above a SiO insulator.

Fig. 2. SEM image of half of an array consisting of five absorbers and ten SIN tunnel junctions. One can see five absorbers of bolometers as narrow strips in the center. SIN tunnel junctions (horizontal, center) are connected to the antenna terminal at its edges through a planar capacitor that provides a parallel connection of bolometers to the antenna port.

II. E XPERIMENTAL T ECHNIQ UE The layout design of a CEB array is optimized for polarization measurements in a 345-GHz frequency band in order to measure the CMB and foreground polarization with balloonborne experiment BOOMERanG. Bolometers are integrated in a cross-slot antenna that is placed in the center of a 7 mm â 7 mm chip on an oxidized Si substrate. Antenna design is similar to [5]. Each orthogonal array consists of ten CEBs connected in series for dc bias and readout. A photo of the antenna is presented in Fig. 1. Dark narrow slots are covered with an Al oxide capacitive layer. In each port of the antenna, there are placed five CEBs that are connected in series for each polarization, producing an array of ten CEBs for the vertical and ten CEBs for the horizontal components. The SIN tunnel junctions of the CEBs are made of a CrAl/AlOx/Al trilayer with a nonsuperconducting CrAl bilayer as a normal layer. An advanced shadow-evaporation technique was used for fabrication of the CEB. A detailed view of half of an array with five absorbers and ten tunnel junctions is presented by a SEM image shown in Fig. 2. Such a chip with the antenna is attached by the back side to an extended hyperhemispheric Si lens with an antireflection coating at 345 GHz. For dynamic range measurements, the lens faces the optical window of the cryostat and is protected against overheating by two low-pass filters (LPFs). These multimesh filters are produced by QMC Instruments and provide more than 10-dB attenuation above the cutoff frequency of 3 THz for LPF W97s and above 1 THz for LPF B694. The filters were placed at the windows in the radiation shields, at the 70- and 3-K temperature stages. To improve the thermal performance, we placed neutral density filters (NDFs) with attenuation of about 6 dB in front of each LPF. As a result, IR radiation is suppressed; the temperature of the radiation background is the same as that of the radiation shield, and by this, we are able to avoid any visible overheating of the cold stage or reduction in the hold time for the cryostat.

III. D C M EASUREMENT R ESULTS The IV characteristic of an array of ten SINIS CEBs clearly demonstrates the sum gap voltage of 20 SIN junctions. A ratio of dynamic resistance at zero voltage to normal resistance of this array is over 1000, which is close to the theoretical estimation for the operating temperature of 280 mK. We measured the voltage at 0.1 nA bias current as a function of temperature. The maximum measured voltage response to temperature is 8.8 µV/mK. For NEP estimation, we measured the output noise of the array using a MOSFET input instrumentation amplifier (OPA111) as the first amplification stage. For theoretical estimations of the performance of such a bolometer array, we can use the power flow determined by electron­phonon interaction 5 (T 5 - T0 ) so that G = dP /dT = 5 T 4 . In this P= case, the responsivity is S = dV /dP = (dV /dT )/(dT /dP ) = (dV /dT )/G. The volume of the absorber for our array of ten bolometers is = 10-19 m3 , and the material parameter for aluminium is = 1.2 109 W · m-3 · K-5 ; the thermal conductivity is thus G = 3.6 10-12 W/K at 280 mK phonon temperature. Using the measured bolometer output voltage noise (vn = 10 nV/Hz1/2 in the white noise region) and the temperature response of dV /dT = 8.8 10-3 V/K, we can estimate the minimum dark electrical NEP = vn /S = 6 10-18 W/Hz1/2 . For the actual power load of 5 pW at 345 GHz, the photon contribution to the total NEP can be estimated as NEPphot = (2P0 hf )1/2 = 4.8 10-17 W/Hz1/2 . It means that our bolometer will operate in a background phonon noise limit. IV. O PTICAL R ESPONSE We measured the response of this array to the microwave radiation emitted by a cryogenic blackbody radiation source. The radiation source was mounted on the 0.4-K stage; it consists of a constantan foil equipped with a heater and a thermometer. This foil covers the radiation pattern of the antenna and lens. Using


TA RASOV et al.: OPTICAL RE SPONSE O F A CEB ARRAY INTEGRATED IN A 345-GHz CRO S S-SL OT ANTENNA

3637

Fig. 3. Voltage response of the bolometer array to (open circles) bath temperature changes 0.118­0.2 K; (solid lines) variations of blackbody source temperature 2.7, 3.8, 4.7, 5.6, and 6.5 K; and (dashed lines) 345-GHz radiation from a BWO with additional attenuation of -21, -18, and -15 dB.

Fig. 4. Voltage response to (triangles) fixed radiation power dV /dP and (open circles) temperature response dV /dTph versus phonon temperature of the bolometer array.

a backward-wave oscillator (BWO) spectrometer/reflectometer, we measured reflection of the foil R = 0.70 ± 0.05 at 345 GHz. This value is different from the zero reflectivity of a blackbody, and the actual emissivity of such source is = 0.30 ± 0.05. The response to heating of the emitter is presented in Fig. 3. The measured voltage response to temperature variations of the emitter is 25 µV/K. Taking into account the emissivity of foil and the root-mean-square (rms) voltage noise 0.38 µV in the frequency range 0­40 Hz (mainly determined by external interferences), one can obtain the temperature sensitivity, which is 5 mK rms. Taking the experimentally measured voltage noise spectral density of the amplifier of 10 nV/Hz1/2 , which dominates in the total noise, we obtain a temperature sensitivity of 1.3 10-4 K/Hz1/2 . We can also calculate the power emitted by the heated foil using Planck's formula for central frequency f0 = 345 GHz and for bandwidth f = 100 GHz of a cross-slot antenna. At a temperature of 3 K, we get P = · hf · f = 3 10- exp(hf /k T ) - 1
14

Such phonon temperature saturation of voltage can be explained as a balance between electron cooling and overheating due to external radiation and Joule heating via the leakage resistance of tunnel junctions. Estimating the cooling power at 0.1 nA and 1 mV gives Pcool I (V - V ) 10-13 W. Joule heating for V = 1 mV bias and zero-bias resistance that we assume as leak resistance R0 = 10 M give Pheat = V 2 /R = 10-13 W. The voltage of saturation for a given current can be obtained from the simple expression IV - IV = V 2 /R: we get Vs = 0.5IR0 [(1 + 4V /I R0 )1/2 - 1] = 1.1 mV, which is close to the observed value.

V. DYNAMIC R ANGE M EASUREMENTS The effectiveness of connecting bolometers in an array and of electron cooling is illustrated by optical measurements of the dynamic range. For this purpose, we used a BWO that operates in the frequency range 250­380 GHz. A calibrated polarization grid attenuator was used for ramping the incident power on the detector. Inside the cryostat, in addition to a cold 20-dB NDF, we also used a cold rotatable stage with a can switch between a 10-dB NDF and an open aperture by an external magnetic field. The measured dependence of the output voltage versus the attenuation of the signal is presented in Fig. 5. Assuming that the weakest detectable signal is determined by amplifier noise (10 nV/Hz1/2 ) and that the strongest is determined by the saturation level, as presented in Fig. 5, at 200 µV, we find that the full dynamic range of this bolometer array is over 43 dB. With a better readout amplifier, this value can increase.

W

where h = 6.626 10-34 J s is Planck's constant, k = 1.38 10-23 J/K is Boltzmann's constant, f is the frequency, and is the emissivity of the radiation source. The voltage response to incoming power is thus dV /dP = 8 108 V/W. For the experimentally measured noise of 10 nV/Hz1/2 , this corresponds to an optical NEP = 2 10-17 W/Hz1/2 . In Fig. 3, we show that the responses of the detector to variations in the power from a thermal radiation source and from a BWO are very similar, whereas the response to changes in the physical temperature of the sample is clearly very different. This difference can be due to suppression of the energy gap due to thermal heating. Voltage saturation to phonon temperature V (Tph ) below 200 mK with its derivative dV /dTph approaching zero at low bath temperatures does not lead to a decrease in optical response dV /dP (see Fig. 4). This means that the electron system is still sensitive to the incoming radiation, and this even makes measurements more stable in the region of saturation to phonon temperature.

VI. P OLARIZAT ION S ENSITIVITY We also measured the sensitivity to the polarization degree of the incoming signal by rotating the polarized signal source in front of the optical window. In this experiment, we used a 115-GHz impact ionization avalanche transit time (IMPATT) oscillator and frequency tripler. The voltage dependence of response for polarization angle = 0 , 15 , 30 , 45 , 60 is presented in Fig. 6. The maximum response scales as sin .


3638

IEEE TR A N S AC T I O N S O N AP P L I E D SU P E R C ONDUCTIVITY, VOL. 21, NO. 6, DECEMBER 2011

R EFERENCES
[1] BOOMERANG--Balloon Telescope: Measurements of CMB Polarization. [Online]. Available: http://en.wikipedia.org/wiki/BOOMERanG_ experiment [2] L. Kuzmin, "Ultimate cold-electron bolometer with strong electrothermal feedback," in Proc. SPIE Conf., 2004, vol. 5498, pp. 349­361. [3] A. Agulo, L. Kuzmin, and M. Tarasov, "Attowatt sensitivity of the cold electron bolometer," in Proc. 16th ISSTT , Gothenburg, Sweden, May 2­4, 2005, pp. 147­152. [4] L. Kuzmin, "Array of cold-electron bolometers with SIN tunnel junctions for cosmology experiments," J. Phys., Conf. Ser., vol. 97, p. 012310, 2008. [5] G. Chattopadhayay, F. Rice, D. Miller, H. G. LeDuc, and J. Zmuidzinas, "A 530-GHz balanced mixer," IEEE Microw. Guided Wave Lett., vol. 9, no. 11, pp. 467­469, Nov. 1999.

Fig. 5. Dependence of the output voltage on the attenuation of the incoming signal for the 345-GHz radiation from a BWO. Signal attenuation values for curve B1 are directly taken from the calibrated attenuator. Curve B3 is taken from an additional 10-dB cold attenuator. Mikhail A. Tarasov received the degree from M. Lomonosov Moscow State University, Moscow, Russia, in 1977, the Ph.D. degree in 1983, and the Doctor of Sciences (Habilitation) degree in 1997. Since then, he has been with V.Kotel'nikov Institute of Radio Engineering and Electronics of Russian Academy of Sciences, where he is currently a Principal Investigator. For the last 20 years, he has been spending up to six months every year with Chalmers University of Technology, Gothenburg, Sweden, collaborating in development of SIS mixers, Josephson detectors, and cold electron bolometers. The list of his publications includes over 200 papers in scientific journals and proceedings of international conferences. His research interests are mainly in superconducting electronics, microwave spectroscopy, and noise in superconducting devices.

Fig. 6. Output voltage dependence on bias voltage for a rotation of the polarized signal source in steps of 15 .

We were not resolution due tions dominant signal by more

able to measure the ultimate polarization to vibrations that make instabilities and reflecwhen the rotation of polarization reduces the than 10 dB.

VII. C ONCLUSION The CEB array integrated in a cross-slot antenna was measured in a dilution refrigerator with an optical window in the temperature range 0.3­0.1 K. The optical response with NEP = 2 10-17 W/Hz1/2 and fluctuation sensitivity to the radiation source temperature of 1.3 10-4 K/Hz1/2 were measured using a cryogenic blackbody radiation source. The dynamic range over 43 dB and the sensitivity to polarization of the incoming 345-GHz radiation were measured through an optical window using a BWO and an IMPATT diode with frequency tripler. Measured characteristics satisfy requirements for balloon-borne experiment BOOMERanG, and CEBs could be considered for future balloon- and ground-based radio telescope experiments.

Leonid S. Kuzmin was born in Moscow, Russia, in 1946. He received the Ph.D. degree in physics (Degree of Candidate of Science in Physics and Mathematics) from Moscow State University, Moscow, in 1977, with the thesis topic "Nondegenerate single-frequency parametric amplification using Josephson junctions with self-pumping" defended in 1997 and "Correlated Tunneling of Electrons and Cooper Pairs in Ultrasmall Tunnel Junctions." In 2000, he was a Docent with Chalmers/GÆteborg University. Since 2009, he has been a Professor with Chalmers University of Technology, Gothenburg, Sweden. He is the Chairman of 13 international workshops and three international schools for young scientists with a general title "From Andreev Reflection to the Earliest Universe." He has authored more than 200 publications (including 114 in referred journals; citation index: 1305, h-index: 19, average citations per item: 11.45).

Valerian S. Edelman was born in 1940, he received the degree from Moscow Physical and Technical Institute, Moscow, Russia, in 1963. He defended the Ph.D. degree in 1968 and the Doctor of Sciences (Habilitation) degree in 1975. Since 1963, he has been with the staff of P.Kapitza Institute for Physical Problems of the Russian Academy of Sciences, Moscow. The main fields of scientific activity are low-temperature physics, electronic properties of metals and 2-D systems, scanning tunneling microscopy and spectroscopy, and lowtemperature detectors. Since 1990, Dr. Edelman has been the Editor in Chief of the journal "Instruments and Experimental Techniques."


TA RASOV et al.: OPTICAL RE SPONSE O F A CEB ARRAY INTEGRATED IN A 345-GHz CRO S S-SL OT ANTENNA

3639

Sumedh Mahashabde was born in Shahad, Maharashtra, India, in 1985. He received the B.E degree in instrumentation engineering in 2007 from Mumbai University, Mumbai, India, and the M.Sc. degree in microtechnology in 2010 from Chalmers University of Technology, Gothenburg, Sweden, where he is currently working toward the Ph.D. degree in physics. His research interests include terahertz bolometers, superconducting tunnel junctions, and thermoelectric effects in superconductors.

Paolo de Bernardis was born in Florence in 1959. He is a Professor of physics with the University La Sapienza, Rome, Italy. He was the Principal Investigator of the Italian international experiment BOOMERANG stratospheric balloon. During the 1998 Antarctic flight, BOOMERANG measured, for the first time, the fluctuations of the primordial plasma and demonstrated the "lack of curvature" of the universe, thus estimating "the density" of total mass and energy. In 2003, BOOMERANG/B2K was launched again to measure the state of polarization of the radiation background microwave. He was the State Coordinator of the Italian international MAXIMA experiments on cosmic background radiation and as co-investigator of the HighFrequency Instrument on the Planck satellite European Space Agency. Prof. de Bernardis was a recipient of the Balzan Prize for Observational Astronomy and Astrophysics and the Premio Feltrinelli of the Accademia dei Lincei. He was a Member of the Astronomy Working Group of the European Space Agency and the State Coordinator of the Study on Issues and Models in Cosmology and Fundamental Physics Space of the Agenzia Spaziale Italiana. He has numerous editorial activities.