Abstract: NbN phonon cooled hot electron bolometer (HEB) mixers are currently the most sensitive heterodyne detectors at frequencies above 1.2 THz. They combine a good sensitivity (8-15 times the quantum limit), an IF bandwidth of the order of 4-6 GHz and a wide RF bandwidth from 0.7-5.2 THz. However, for use in a space based observatory, such as Herschel, it is of vital importance that the Local Oscillator (LO) power requirement of the mixer is compatible with the low output power of present day THz LO sources. This can be achieved by reducing the mixer volume and critical current. However, the large RF bandwidth and low LO power requirement of such a mixer result in a direct detection effect, characterized by a change in the bias current of the HEB when changing the RF signal from a black body load at 300 K to one at 77 K. As a result the measured sensitivity using a 300 K and 77 K calibration load differs significantly from the small signal sensitivity relevant for astronomical observations. In this article we describe a set of dedicated experiments to characterize the direct detection effect for a small volume quasi-optical NbN phonon cooled HEB mixer. We measure the direct detection effect in a small volume (0.15 μm · 1 μm · 3.5 nm) quasi- optical NbN phonon cooled HEB mixer at 1.6 THz. We found that the small signal sensitivity of the receiver is underestimated by approximately 35% due to the direct detection effect and that the optimal operating point is shifted to higher bias voltages when using calibration loads of 300 K and 77 K. Using a 200 GHz wide band-pass filter at the 4.2 K the direct detection effect virtually disappears. Heterodyne response measurements using water vapor absorption line in a gas cell confirms the existence and a magnitude of a direct detection effect. We also propose a theoretical explanation using uniform electron heating model. This direct detection effect has important implications for the calibration procedure of these receivers in real telescope systems. We are developing Nb HEBs for a large-format, diffusion-cooled hot electron bolometer (HEB) array submillimeter camera. The goal is to produce a 64 pixel array together with the University of Arizona to be used on the HHT on Mt Graham. It is designed to detect in the 850 GHz atmospheric window. We have fabricated Nb HEBs using a new angle- deposition process, which had previously produced high quality Nb-Au bilayer HEB devices at Yale. [1] We have characterized these devices using heterodyne mixing at ~30 GHz to compare to 345 GHz tests at the University of Arizona. We can also directly compare our Nb HEB mixers to SIS mixers in this same 345 GHz system. This allows us to rigorously calibrate the system’s losses and extract the mixer noise temperature in a well characterized mixer block, before undertaking the 850 GHz system. Here we give a report on the initial devices we have fabricated and characterized. * Department of Applied Physics, Yale University ** Department of Astronomy, University of Arizona [1] Applied Physics Letters 84, Number 8; p.1404-7, Feb 23 (2004)

Abstract: We investigate both antenna coupled and directly coupled HEB mixers at several LO frequencies within the range of 2.5 THz to 70 THz. H20 (2.5+10.7 THz), and CO2 (30 THz) gas discharge lasers are used as the local oscillators. The noise temperature of antenna coupled mixers is measured at LO frequencies of 2.5 THz, 3.8 THz, and 30 THz. The results for both antenna coupled and directly coupled mixer types are compared. The devices with in—plane dimensions of 5x5 ,um 2 are pumped by LO radiation at 10.7 THz. The directly coupled HEB demonstrates nearly flat dependence of responsivity on frequency in the range of 25+64 THz.

Abstract: We realized ultra thin NbN films sputtered grown on a 3C-SiC/Si substrate. The film with a thickness of 3.5-4.5 nm shows a 1', of 11.8 K, which is the highest I`, observed among ultra thin NbN films on different substrates. The high-resolution transmission electron microscopy (HRTEM) studies show that the film has a monocrystalline structure, confirming the epitaxial growth on the 3C-SiC. Based on a two-temperature model and input parameters from standard NbN films on Si, simulations predict that the new film can increase the IF bandwidth of a HEB mixer by about a factor of 2 in comparison to the standard films. In addition, we find standard NbN films on Si with a T c of 9.4 K have a thickness of around 5.5 nm, being thicker than expected (3.5 nm).

Abstract: Direct measurements of the thermal boundary resistance were performed by means of the stationary method. In this approach the temperature of an electrically heated film is controlled by its dc resistance while an additional film on the same substrate is used as a thermometer monitoring substrate temperature. The temperature field in the substrate is then calculated to deduce the Kapitza temperature step at the interface between the heated strip and the substrate. The main statement of all afore-said papers is that experimental values of the thermal boundary resistance are too large to be explained by the acoustic mismatch model. In this paper we investigate transparency of YBaCuO film/substrate interfaces for thermal phonons by means of photoresponse measurements. We show that our data are in reasonable agreement with the acoustic mismatch theory.

Abstract: We have performed a study of how direct detection affects the linearity and hence the calibration of an HEB mixer. Two types of waveguide HEB devices have been used: a 0.8 THz HEB mixer and a 1.0 THz HEB mixer which is ~5 times smaller than the former. Two independent experimental approaches were used. In the ΔG/G method, the conversion gain of the HEB mixer is first measured as a function of the bias current for a number of bias voltages. At each bias setting, we carefully measure the change in the operating current when the input loads are switched. From the measured data, we can derive the expected difference in gain between the hot and cold loads. In the second method (injection method [1]), the linearity of the HEB mixer is independently measured by injecting a modulated signal for different input load temperatures. The results of both approaches confirm that there is gain compression in the operation of HEB mixers. Based on the results of our measurements, we discuss the impact of direct detection effects on the operation of HEB mixers.