The stability of a terahertz receiver based on a superconducting integrated receiver

Abstract

We present the results of stability testing of a terahertz radiometer based on a superconducting receiver with a SIS tunnel junction as the mixer and a flux-flow oscillator as the local oscillator. In the continuum mode, the receiver with a noise temperature of 95 K at 510 GHz measured over the intermediate frequency (IF) passband of 4-8 GHz offered a noise equivalent temperature difference of 10 ± 1 mK at an integration time of 1 s. We offer a method to significantly increase the integration time without the use of complex measurement equipment. The receiver observed a strong signal over a final detection bandwidth of 4 GHz and offered an Allan time of 5 s.

Low-Frequency Noise and Passive Imaging With 670 GHz HEMT Low-Noise Amplifiers

“…Imaging radars operating as high as 650 GHz [5] are now being adapted to exploit them. Passive submm imaging on the other hand, has primarily been addressed with cryogenic sensor arrays [6]- [11]. The present work demonstrates that the new generation of HEMT amplifiers does indeed provide adequate sensitivity for passive submm imaging in terrestrial (indoor or outdoor) applications.…”

Ultra-low-noise magnetic sensing with long Josephson tunnel junctions

“…A flux-flow oscillator (FFO) is an LJTJ with relatively high dumping in which a unidirectional viscous flow of mutually repulsive fluxons occurs and coherent electromagnetic radiation is emitted from one of its ends where the fluxon chain collides with the boundary [19,20]; 1 µW of power can be emitted in the frequency range 100-700 GHz within a 10 MHz linewidth [21]. The flux-flow in LJTJs is a very well studied phenomenon that has been exploited for a long time for the realization of voltage-controlled local oscillators in low-noise integrated THz receivers [22]. The Eck step voltage, V E , gives the number of fluxons passing per unit time across the LJTJ and is determined by two external independent stationary currents: one is referred to as the control current, I ctl , injected into one of the junction electrodes to create the magnetic field at the two ends of the LJTJ, while the bias current, I b , applied through the tunneling barrier, accelerates the fluxons and moves them from one junction extremity to the opposite one.…”

“…For comparison, a standard dc-SQUID magnetometer [4] having the same size as our demonstrator would have an intrinsic voltage noise S We stress that our design was far from being optimal and so the magnetic sensitivity is subject to further improvements, as it is essentially determined by the ratios L p /A p and R d /R m . In particular, R d can be decreased by using uniformly biased and properly tailored LJTJs [22,47], while R m increases with reduced constriction width, w c , but is limited ultimately by fringing field [13] and quenching [48] effects. As the dynamic resistance of the Eck step is inversely proportional to the Josephson current densities, the proposed sensor would drastically benefit from high-J c junctions as those used for the realization of FFOs.…”

Superconducting flux-flow type ultra-low-noise magnetic sensors - An alternative to dc-SQUIDs

“…A flux-flow oscillator (FFO) is a LJTJ with relatively high dumping in which a unidirectional viscous flow of mutually repulsive fluxons occurs and coherent electromagnetic radiation is emitted from one of its ends where the fluxon chain collides with the boundary [18,19]. The flux-flow in LJTJs is a very well studied phenomenon that since long is being exploited for the realization of voltage-controlled local oscillators in low noise integrated T Hz receivers [20]. The Eck step voltage, V , gives the number of fluxons passing per unit time across the LJTJ and is determined by two external independent stationary currents: one is referred to as the control current, I ctl , injected into one of the junction electrodes to create the magnetic field at the two ends of the LJTJ, while the bias current, I b , applied through the tunneling barrier, accelerates the fluxons and moves them from one junction extremity to the opposite one.…”

“…We stress that our design was far from being optimal and so the magnetic sensitivity is subject to further improvements, being it essentially determined by the ratios L p /A p and R d /R m . In specie, R d can be decreased by using uniformly biased and properly tailored LJTJs [20,25], while R m increases with reduced constriction width, w c , but is limited ultimately by fringing field [11] and quenching [40] effects. For this sample the thermally originated, loop magnetic (integral) noise, B n ≈ 20 pT , was negligible (S 1/2 B = O(10 −21 T /Hz 1/2 ) and, throughout the operating voltage range, the dissipated power, V I qp , was well below 1 µW .…”