Our programme

We are engaged in a programme to develop HEB devices, and measure and optimise their performance. Development work has proceeded extremely well, and we now seek funding for a further 2 years to extend their operation to the higher frequencies (~ 2 THz) where they should be the most sensitive detectors available. During the first phase of our development work we have;

•   developed the capability to reliably and reproducibly fabricate HEB’s with sizes less than 0.1 mm,  using a combination of direct-write electron beam lithography, reactive ion etching and uv lithographic processing techniques. The devices are manufactured in the QMW clean rooms, which house a direct-write Electron Beam Lithography (EBL) facility, UV-lithography, mask aligners, an UHV thin film deposition system, an RIE plasma etcher and wire-bonders, and other test and fabrication equipment for sub-micron patterning of thin films.

We have constructed a superconducting film stress measurement system - using atomic force microscopy, and an electrochemical anodisation technique. Films can now be grown with Tc ‘tuneable’ over the range ~ 2 - 9.2 K, allowing efficient operation at 4 K.

•   developed of a fast (< 1ms) temperature stabilisation system for the detector, using a magnetic field to control the detector’s operating point.

We have measured performances, speeds and spectral characteristics, of hot-electron bolometric HEB's, achieving measured NEP's of < 5 10-13 W Hz-0.5, responsivities > 15 kV W-1, and speeds > 2 kHz with our first very poorly optimised detectors (measured without any corrections for cryostat, filtering etc.). These devices have excellent performance out to 2000 - 3000 GHz, well above the frequencies where SIS sensitivities fall off (above ~ 700 GHz).

•   constructed a wide-band, 0.5-18 GHz IF amplifier was purchased and installed in the 530 GHz receiver on the 4 K stage, with a special, wide-band cryogenic bias-tee, with a resultant useful bandwidth of 0.5-5 GHz, sufficient to measure the IF roll-off frequency of the HEB.

•  built a broadband, blackbody rf source and a spectrum analyser system as a tunable, narrow tuneable filter in the IF system to accurately measure the IF roll-off of the bolometer mixer. The HEB is close to the expected value for reasonable self-heating in the device, and is the fastest bolometer mixer ever developed.

•   developed a new theoretical analysis of the bolometer mixer operation based on a finite-difference approach, which is less model-dependent that current theories being published and has  allowed us to explore the high frequency performance of this unique mixer.

We are currently working on several kinds of detector:

The detector consists of some tens of long (~10 mm) and narrow (~1 mm) strips folded into meanders connected in parallel between two contact pads, on a thin crystalline quartz substrate and placed in good thermal contact with a copper block. The objective is to develop a low-noise (NEP ~ 10-11 W Hz0.5 at 4.2 K) and wideband (~ 100 MHz) bolometer with a simple geometry that allows us to measure some properties of our Nb films. First tests in thin and long strips have demonstrated the behaviour expected from electron heating by radiation and cooling by electron-phonon interaction. An ultra-fast IR Light-Emitting Diode has been installed inside the cryostat, and preliminary tests suggest bandwidths up to 100 MHz.

•   Antenna-Coupled HEB

This bolometer is optimised for the 50 GHz - 5 THz region, and consists of a few strips (~0.5 x 6 mm2) placed in parallel across the terminals of a log-periodic planar antenna, mounted on the back of a crystalline quartz hemispherical lens. Wideband operation (~100 MHz) and low noise (NEP ~ 10-14 W Hz0.5 at 4.2 K) has been obtained. The ultimate performance of this bolometer will be determined by the available amplifiers, and a collaboration to implement a dc SQUID pre-amplification stage is being pursued.

•   Diffusion-Cooled HEB mixer

    This type of mixer is based on a superconducting element of submicron dimensions, where hot-electrons can diffuse out of the microbolometer to normal metal contact leads before thermalising with phonons (Skalare et al 1995, 1996), resulting in a much faster cooling mechanism, which allows instantaneous bandwidths of several GHz and low noise. The main advantage relatively to SIS mixers is that these HEB’s operate above ~ 700 GHz, the frequency at which the performance of SIS mixers starts to degrade, with very large bandwidths (many 100’s of MHz).

(left) Photolithographically formed Nb HEB mixers on printed circuit (right) Finite element modelling of HEB mixers showing the time evolution of the electron temperature distribution in a 0.2 micron long microbridge, biased by equal amounts of LO and dc bias power, responding to an IF of 1 GHz.

•   Modelling

We have also developed a computer model to examine and optimise the behaviour of HEB mixers. This model predicts the performance of diffusion-cooled mixers by dealing with the microbridge as a discrete element, i.e. assuming a constant electron temperature throughout. We are pursuing a more rigorous analysis based on the numerical solution of the  energy-balance equations to optimise the device operation and assess the validity of the discrete element approximation. Some results from our model are shown below (left), where it can be seen that very fast time constants (and hence wide instantaneous bandwidths) are achievable with HEB mixers and predictable with our model. The present state of the art performances of submm mixers are shown below (right).

Length Scaling of the thermal time constant of HEB mixers    The present state of the art for SIS and HEB mixers

Other Work

To search for alternative detectors capable of operating at higher frequencies, studies have been made of SIN tunnel junctions (Howman, White et al 1990), Josephson mixers (Blaney 1980), hot-electron semiconductor devices such as InSb (Phillips & Jefferts 1971, White & Padman 1991, Padman, White et al 1992), and hot-electron superconductor devices (Gershenzon et al 1991, Kollberg 1992, Prober 1993, Nahum, 1995). The sensitivity of Nb SIS mixers decreases rapidly above ~ 700 GHz, of InSb mixers falls µ n 2 above ~ 500 GHz (although can be enhanced using a cyclotron resonance effect), and Josephson mixers µ n.  Despite numerous attempts to develop NbN technology as a higher frequency alternative to Nb devices, no useful devices were been made that proved useful for mainstream astronomical applications at high frequencies. This lack of NbN devices resulted from a fundamental materials technology problem, since it proved impossible to make films of sufficient quality. Recent work over the last 6 months at JPL has developed a new alloy, NbTiN, which has apparently extended the upper operating frequency of SIS devices to 1.2 THz. Thus, a materials technology now exists which has been proven to get above 800 GHz - although it is only available at JPL. However, the upper limit problem is still there, and the performance will drop off above 1200 GHz. Our results to date suggest that HEB devices offer significant advantages in increased speed, wide instantaneous bandwidth and broad spectral coverage, as well as being considerably easier to manufacture than sub-micron SIS / SIN devices, and have operating temperatures ~ 4K, simplifying cryogenic requirements. We believe the HEB will be the heterodyne detector of choice for frequencies > 1 THz, or l < 300 mm.

b) The Programme

i)          We areextending testing and optimisation of the heterodyne performance of the HEB mixers to the range 1000 - 3000 GHz. Testing using standard Y-factor techniques with blackbody loads, and local oscillator power obtained from a far-infrared laser system.

ii)         We will construct a 5-element HEB array for incoherent detection, and test its noise performance, frequency response and speed, over the range 500 GHz - 16 THz (600 - 20 mm wavelength), using our existing test equipment to explore the operational characteristics at the higher frequencies. An optimised array will be tested with a rapid-scan Fourier transform spectrometer, with the aim of taking a high spectral resolution (better than 0.03 cm-1) spectrum of the atmospheric emission from 0.5 - ~ 2 THz with an imaging array in the focal plane. Such an array will have significant advantages over a single detector - a) it will be a dedicated array covering the whole frequency range above 500 GHz, b) the optics of the hot-electron array will be optimised for Fourier transform spectrometry giving a high throughput, c) the array needs simple cryogenic cooling at ~ 4K and d) the fast response time of the HEB will allow extremely high spectral resolution (probably to ~ 0.003 wave numbers) to be obtained, or equally importantly for the study of fast time-varying phenomena, lower resolution time-resolved spectroscopy (such as is required to study the spectral evolution of a plasma in a Tokamak).

iii)        Device fabrication will be a continuing feature of the programme, both to improve the performance, and to fabricate HEB's with different types on antennae and matching networks. The frequency response, NEP's, noise temperatures, IF bandwidth, response speeds and saturation levels at the higher frequency ranges will be tested in open structure mounts for the direct and heterodyne modes of operation will be measured using existing test equipment at QMW - requiring ~ 9 months of time. The final 12 months of work will be used to optimise the design of the microbolometric detectors and mixers, to continue measurements of performance, saturation and out of band response, and to look into the problems of operating an array in front of a Fourier transform spectrometer.

The objective of the our work is to develop the critical enabling technologies necessary to construct sensitive high resolution spectroscopic detectors for the far-infrared spectral region. We consider the development programme will place us in a strong position to develop a next-generation ultra-high spectral resolution far-IR spectrometer. We are exploring three possibilities, (i) to provide an instrument for SOFIA - a 3 m class telescope mounted in the fuselage of a Boeing 747 Jumbo Jet, (ii) to construct a ground-based instrument for use at the South Pole or

(iii) at the high site in Chile.

c) Other applications

There are many other applications for this detector technology; for example the use of bolometric detectors in plasma diagnostics experiments at the Joint European Torus (JET), where the main requirements are sensitivity, wide band spectral response and speed, to allow high time resolution observations of the thermal evolution of the plasma. Similarly, there are many problems requiring fast-scan Fourier transform spectroscopy - such as the detection of trace gases via limb sounding measurements of the atmospheric emission spectrum, which can help to understand the chemical evolution of the atmosphere (such as the fluorocarbons, O3, ClO, HO2, HOCl, HNO3 and BrO etc.). In physics the study of short lived surface states on metals and the lifetimes of short-lived chemical species in the gas and liquid phase are areas of active research, and the fast response time of superconducting microbolometer permits its use with a fast-scan Fourier spectrometer, offering a technique with which to study short life-time phenomena, which can now be studied using mm-synchrotron and free-electron maser sources. We would expect the successful conclusion of our work on HEB mixers to lead to participation in a future experiments with a plasma diagnostic facility, and with the 2D fast-scan Fourier transform spectrometer we hope to be able to propose a future instrument for high spectral resolution atmospheric sensing.