German Receiver for Astronomy at Terahertz Frequencies(GREAT)

Introduction

A consortium between the Max-Planck-Institut für Radioastronomie (MPIfR), the I. Physikalisches Institut der Universität zu Köln (KOSMA), the DLR-Institut für Weltraumsensorik und Planetenerkundung and the MPI für Aeronomie (MPAe) has been established for the development of GREAT (German Receiver for Astronomy at Terahertz Frequencies). GREAT will be a first-generation dual-channel heterodyne instrument for high resolution spectroscopy for SOFIA (Stratospheric Observatory For Infrared Astronomy). Aboard the stratospheric observatory the instrument will measure the submillimeter and far-infrared spectral range.

Due to water vapor absorption of the atmosphere it is difficult and in some cases not possible to observe the submillimeter and far-infrared radiation with ground-based observatories. As can be seen in Fig. 1, the opacity of the atmosphere is very large in the wavelength range of 20 - 1000 micrometers. SOFIA will be employed in an altitude of 12.5-13.7 km. In this flight altitude the influence of water vapor absorption is small or negligible. Therefore, the air-based observatory for astronomers will be an excellent tool in order to observe the terahertz frequency range.
atmospheric transmission.

Scientific Motivation

Main goal of terahertz astronomy is to observe the warm rather than hot (such as stars) and cold (such as the cosmic background radiation) sources, e.g. the interstellar medium (ISM) which is made of gas and dust, star formation regions etc. The preceding wide range of topics of modern astrophysics require a high-spectral resolution of order 106-107 (0.3-0.03 km*s-1), because often high resolution is crucial to discriminate a given interstellar line against narrow atmospheric absorption features (e.g. the ground-state transition of HD at 2.6 THz).

The first flight version of the dual-channel receiver GREAT will have to focus on three scientifically-selected frequency windows: a low-frequency band (1.6-1.9 THz), a mid-frequency detector (2.6 THz) and a high-frequency channel (4.7 THz). We are responsible for the low-frequency band. The frequency range 1.6-1.9 THz (187 to 158 microns) covers - among other lines, e.g. high-J rotational lines of CO - the important atomic fine-structure transition of ionized carbon CII (J= 3/2 -> 1/2). This transition line is the most important cooling line of the ISM, therefore crucial for its energy balance. CII is abundant and easily excited in the photodissociated surface layers of molecular cloud clumps in consequence of the strong emission of far-ultra-violett photons from massive star forming regions. These photons dissociate molecules such as CO, and ionize carbon, why CII is intrinsic to regions of high UV-radiation. There is a good evidence that a substantial fraction of the CII flux of a galaxy originates presumably from less dense gas, which is illuminated from the average interstellar radiation field. The existence of CII can be proven with the fine-structure transition which lies in the far-infrared at 1.9 THz.

Detailed assignment of various emission components in the CII line requires full velocity resolution in order to disentangle different emission components overlapping along the line-of-sight. The RF-tuning range of the low-frequency band will allow the study of redshifted objects up to 50.000 Km*s-1 with an angular resolution of 12´´. In practice some velocity intervals will be blocked by the atmospheric absorption even at the flight altitude of SOFIA and will be accessible only with space-based heterodyne instruments like HIFI on the HERSCHEL satellite.


GREAT receiver concept

GREAT will be developed as a modular and flexible single-pixel dual-channel heterodyne receiver. The modularity of the receiver design allows easy changeover between different mixers and LOs covering different frequency bands and permits later upgrades and - with time - an increasingly more complete coverage of the FIR-spectral range. Furthermore modularity is necessary, because between flights the change of cryostats and the replacement of optical components must be easy and straightforward.

The instrument will consist of 2 independent cryogenic dewars (containing the mixer devices). They are mounted to the telescope flange via a common "optics box" (Fig. 2) which contains the polarization splitter, (optionally) a singel-sideband filter (e.g. a Martin Puplett Interferometer), a backward wave oscillator (BWO) (Fig. 3) as local oscillator unit (LO) and the calibration unit. The waveguide mixer will be implemented as diffusion-cooled Nb or lattice-cooled NbN hot electron bolometer (depending on performance).

Fig. 2: Structural layout of the GREAT frontend, displaying the science instrument flange, the optics box hosting the beamsplitter and LO couplers, racks with control electronics and the local oscillator
We decided to design the GREAT receiver with warm optics, because the increase of system noise due to the warm mirrors will not be critical at an expected HEB receiver noise temperature of approx. 2000 K.

Because a heterodyne mixer is sensitive for one polarization only, a dual-channel receiver can be realized by dividing the incident beam into two perpendicular polarizations. A wire-grid will split the incident beam. Thereafter, one polarization can be used for each frequency range separately.

After beam splitting, the RF and the LO signal have to be coupled. Two design options are available for diplexing: in the first case beam coupling will be done with a Martin-Puplett Interferometer. In the second case (if sufficient LO power is available) beam coupling will be simply done with e.g. a Mylar beam splitter (just a foil) with typically 98% transmission and 2% reflection. Due to the absence of any degree of freedom in the coupling optics the latter option would be very convenient. In both cases the superposed signals occur after the coupling process the dewar, where they illuminates the mixer device.

The incident beam operate a heterodyne mixer element (heterodyne principle), which produces the intermediate frequencies (IF). These frequencies will be amplified with cooled low noise HEMT amplifiers. Then the prepared IF signal will be analyzed with backend devices like an Acousto-Optical Spectrometer (AOS). To shift the IF-signal to the required input bandpass of the spectrometers and to provide calibration signals for the backends, IF processors need to be included. The needed IF device will be developed from a previously constructed miniaturized plug-in single channel IF-proccessor version. This module consists of two frequency mixing stages, a step attenuator to set the output power level, a detector for power monitoring, a zero switch to measure the AOS dark current and a directional coupler to inject a comb signal for the frequency calibration of the AOS. It also includes all the filtering and amplification required, including a 25dBm power amplifier at the output.

Fig. 3: Front-end cryostat


Fig. 4: A water cooled backward wave oscillator integrated into a permanent magnet. For GREAT we will purchase a BWO for 633 GHz.

Cryostats, Calibration Unit and Local Oscillator

The cryostats (Fig. 3) are designed for a minimum hold time of 24 hours and will contain about 6 ltr. LHe and 8 ltr. LN. The MPIfR is responsible for manufacturing and airworthiness. The calibration unit (CAL) is needed in order to calibrate the receiver. The CAL consists of a small vacuum-vessel (<2 l) containing a temperature controlled absorber assembly. It will be an interchangeable part of the optics unit. The following design options are possible: a hot load (absober assembly heated to 100-150°C) or a cold-load (absorber cooled by a small amount of LN). The latter option is preferred.

As the baseline option we plan to use a frequency-tripled phase locked 633 GHz BWO as local oscillator (Fig. 4), manufactured by ISTOK Company. A tunability of approx. 10-20% and a power of more than 5 mW of the BWO is expected. The frequency multiplier is a Schottky Diode Multiplier. The output beam power of the fundamental frequency will be approximately 1mW (at 1.9 THz), and should be high enough to fulfill the requirements of the second-generation array receiver STAR, too. Because BWOs are of operational complexity (high voltage supply: 7 kV DC; high static magnetic field: 1 Tesla; water cooling) local oscillator sources for these high frequencies still require development (FIR-Laser, solid-state LO, photonic LO). This is necessary for following space-based heterodyne receivers like HIFI on the HERSCHEL-satellite, because for such an instrument a compact and easy design of LO is indispensably.

Each polarization operates a detector element separately. At the low-frequencies band we will probably use waveguide diffusion cooled Niobium hot electron bolometers (HEBs) because current investigations have been showed that HEB mixer work up to frequencies of 2.5 THz. In addition to their low-noise performance at terahertz frequencies, this type of HEB mixers has the potential of an IF bandwidth beyond 8 GHz. The HEB devices for the low-frequency band are currently under development at KOSMA. To have the opportunity to choose the best performance mixer device other technical concepts are under development, too (niobiumnitrid phonon-cooled bolometers, NbTiN SIS-junctions). For diffusion cooled HEBs we intend to use the waveguide mixer block design that has been very successful over the last years in receivers at the KOSMA telescope.

In a bolometric detector, the mixing relies on the strong temperature dependence of the resistance of the superconductingmicrobridge (bias point in the superconducting-normal state transition layer). The fast cooling required for high IF frequencies can be achieved with Nb microbridges (10 nm) that are small enough that hot electrons diffuse out of the bridge on a very short time scale and deposit their energy in a normal conducting heat sink.

Back-End Spectrometer

The wide range of astronomical topics, and hence requirements on spectral bandwidth and resolution, to be adressed with SOFIA is difficult to cover by one class backend. Only high-resolution spectroscopy (v/dv=106-107) can provide the critical information about line shape and velocity structure that is needed for understanding of the underlying physics. Therefore, for SOFIA we will have a choice of backends, with some flexibility in observing modes.
Acousto-Optical Spectrometer (AOS)

Because of the relatively simple design of an AOS it is a very suitable instrument for airborne applications. The maximum requested instantaneous frequency coverage in the GREAT instrument is 4 GHz in total, the frequency resolution should be about 1 MHz. But the maximum bandwidth of acousto-optical deflectors is limited due to the rather strong acoustic attenuation in the crystal materials at higher frequencies. This presently limits the maximum usable frequency, which can be efficiently processed in a Bragg-cell to about 3 GHz, if a resolution of 1 MHz is assumed. Therefor, 1.5 GHz is approximately the maximum bandwidth of an AOS at higher frequency resolution. The GREAT instrument is planned for up to 4 GHz IF bandwidth, therefore 4 times 1 GHz can be used for full frequency coverage.
High-Resolution Chrip Transform Spectrometer (CTS), Wideband Analog Autocorrelator (AACS)

The principal of a chrip transform spectrometer is based on the Chrip Transform Algorithm, which is equivalent to the Fourier Transform. For SOFIAs first flights, MPIAe will deliver one high-resolution module with 180 MHz bandwidth and 4096 channels (45 kHz spectral resolution).

An analog autocorrelator spectrometer for wideband, but low resolution observations will be developed. The design will extend on the developments for WASP, a prototype that has been used successfully for astonomical observations at the CSO. The project goal is to provide 4 modules, 4 GHz wide with 250 channels each (spectral resolution: 16 MHz).