Introduction

The next generation continuum camera for the JCMT ("SCUBA-2") is now well into the development phase. Following a successful Conceptual Design Review (in October 1999) and Pixel Architecture Downselect (May 2000), work on the instrument is now gaining momentum rapidly. During the past year the project has been enthusiastically received by both the JCMT Board and Advisory Panel, whilst PPARC approved a "proof-of-concept" funding that will most likely last about 2 years. The JCMT Board will review the project again at the November 2000 meeting. This is a short article which is meant to give a brief overview of the proposed instrument.

Scientific goals

The scientific goals of SCUBA-2 evolved primarily from consultation with the JCMT community during 1999. The key requirements are: the ability to make very deep images - reaching the background confusion limit in only about an hour of integration time; to generate high fidelity images at two wavebands simultaneously; to map large areas of sky (several square degrees) to a reasonable depth in only a few hours; carry out photometry of known-position point-sources to a high accuracy.

These goals dictate that the per-pixel sensitivity should be at least as good as the (recently) upgraded SCUBA, and that the field-of-view should be maximised (at least 8 arcmin in diameter is possible with JCMT). In terms of producing high fidelity images near or at the confusion limit, both the techniques of jiggling and sky chopping (currently employed with SCUBA) are major obstacles. Work is currently underway on defining the SCUBA-2 observing modes, but a major departure from the existing SCUBA methodology is that the arrays will fully sample the image plane (jiggling not necessary)and will most likely be D.C. coupled (sky chopping is no longer required). In principle, this will make observing with SCUBA-2 considerably more straightforward than with SCUBA!

SCUBA-2 will therefore instantaneously sample the sky - in a "point-and-shoot" mode similar to a CCD camera. This will produce fully-sampled images in much less time than required for SCUBA leading to improved image fidelity. Having D.C. coupled arrays will potentially allow more large-scale source structure to be visible, easier and more accurate flat-fielding (using the sky), and as the subtracted D.C. level is proportional to the sky brightness, it should allow the sky transmission to be continuously corrected for.

The Scientific Case for the instrument is well established and can be viewed on the SCUBA-2 homepage . As an example of the anticipated capabilities consider the case of large-scale extragalactic surveys designed to study source counts. Figure 1 is taken from the scientific case (but is updated following a revision of the instrument projected performance) and shows the expected detection rate (sources/hour) as a function of 5-s depth. These simulations, carried out by Andrew Blain, show that a SCUBA-2 (see next section for baseline design) will be a very powerful survey instrument and will be able to detect ~ 20 galaxies per hour at the optimum depth. As the figure shows SCUBA-2 will be very competitive with any instrument currently being proposed (and is likely to be available ahead of facilities like FIRST-SPIRE and even ALMA . It should also be noted that SCUBA-2 will be complementary to many facilities - particularly the airborne and space observatories and ALMA. Furthermore, although ALMA will be a very powerful survey instrument operating in a compact mode, it seems clear that that this might not be the best use of such a huge interferometer. Wide-field bolometer arrays (such as SCUBA-2) will become vital to carry out large-scale surveys, areas of which are then subsequently followed-up by high-resolution ALMA studies.
 
 

Figure 1 : The detection rates as a function of 5-s depth for a variety of instrument surveys including an 8-arcmin diameter f-o-v for SCUBA-2. Lines stop at the confusion limit on the left and where there is only 1 source on the sky on the right. The underlying model is the "Modified Gaussian" model described by Blain et al. (MNRAS 302, 1999). Many thanks to Andrew Blain for providing this figure.

Table 1 gives several more examples of the types of observations that could be carried out with SCUBA-2 together with an estimate of the decrease in integration time required for SCUBA-2 over that of SCUBA.
 
 

¹Note that a much bigger area will be imaged with SCUBA-2!

Instrument specification and design

Following the recent review meetings, the baseline specification is now well-established. SCUBA-2 will consist of two arrays operating simultaneously at both 450 and 850mm. The field-of-view on the sky will be a minimum of 8 ´ 8 arcminutes for each array. This means that to fully-sample the sky, with a pixel spacing of 0.5Fl (required for full Nyquist sampling), 25,600 and 6,400 pixels will be needed at 450 and 850mm respectively (c.f. 131 in total for SCUBA!). Each pixel will have diffraction limited resolution and a sensitivity dominated by the sky background photon noise.

The SCUBA-2 arrays will utilise new technology Transition Edge Sensors (TES) as the detecting element. TES devices have a number of very attractive advantages over more conventional bolometers: they are very sensitive to small amounts of incoming power since the biasing arrangement (in the transition region between the normal and superconducting states) make their resistance is a very steep function of temperature; the extremely sharp n-s transition region gives rise to strong electro-thermal feedback which can significantly speed up the device (allowing potential novel fast-scanning observing modes); under a voltage bias the pixel becomes "self-biasing" - meaning that only a single bias line is needed (thereby simplifying the electronics and the number of wires). The TES devices will be bonded to a low heat-capacity silicon membrane which will contain a radiation absorbing metal film.

Multiplexing the signal readout is a crucial area of the instrument design. This will be achieved by using SQUID amplifiers which are well-matched to the low-impedance TES devices. The construction of a single pixel is illustrated in Figure 3, together with what a 16 ´ 16 array might look like. The first-stage SQUID multiplexer will be buried in the silicon substrate interconnect chip. Ribbon cables will take the detector signals to the outside world. The TES arrays, together with the SQUID readouts, will be developed under a contract to the National Institute of Standards and Technology (NIST) who are world-leaders in these technology areas.

For a number of reasons the detector operating temperature still has to be below 200 mK, and so a dilution or adiabatic demagnetisation refrigerator will be required. Although the SCUBA fridge has had a number of problems over the years, new designs of dilution system are likely to be considerably more reliable, as well as less bulky. We are currently looking at options for fridges and also the prospects of constructing a liquid cryo-free system (i.e. not using liquid helium and saving on running costs).

.

Figure 2 : On the left is a basic schematic diagram of a single pixel design. The TES device is represented as the square in the centre of the absorber. On the right is what a 16 ´ 16 array might look like. Each pixel will be about 1mm on a side for the 850mm array and 0.5mm for the 450 array (with a small wall in between).
 
 

Technical challenges

Without a doubt SCUBA-2 is a very ambitious project, and as such there are several major technical challenges to be overcome. Primary amongst these are the manufacture of the TES detector arrays and SQUID amplifier readouts. A number of issues associated with the mechanical and electromagnetic design of the pixels have yet to be addressed. The adoption of a filled array design of 0.5 Fl spaced pixels also has potential problems. Since the field-of-view of the array will be defined by a cold stop, and not by a conventional feedhorn, the control of stray light (which could degrade detector sensitivity) presents a major challenge.

Another of the main challenges we face with SCUBA-2 is how to re-image the large telescope field-of-view out onto the Nasmyth platform, in such a way that the resulting image plane is flat and (relatively) aberration-free. A complex optical design is currently being investigated and in figure 3 we show what the optical path may actually look like. As mentioned above, key aspects of the internal optical system will be the need for high quality baffling and control of potential stray light.
 
 

.

Figure 3 : A preliminary outline of a possible optical design (CAD drawing by Ian Pain). The field-of-view is some 600mm in diameter at the tertiary mirror of the telescope, and has to be re-imaged to an array of about 125mm in diameter with minimum distortion and aberrations.

 
 

Delivery schedule

SCUBA-2 is being built as a collaboration between primarily the UKATC (current Project Manager : William Duncan), NIST (PM : Kent Irwin), QMW (PM: Steven Rinehart), University of Edinburgh (PM : Alan Gundlach) and the JAC (current PM : Wayne Holland). Within 2 years of the start of the NIST contract we hope to produce a small prototype array which will act as a proof-of-concept device. If all goes well, we anticipate delivery of the final instrument before the end of 2005.

Summary

SCUBA-2 will allow the JCMT to fulfill its ultimate potential. The improved sensitivity and large field-of-view will allow the productivity of the JCMT to increase many fold - especially in an era when submillimetre interferometry with the SMA and the first heterodyne cameras will mean less time available for continuum astronomy. All areas of astronomy are expected to benefit, but perhaps the most exciting prospect that SCUBA-2 will offer is in the statistical significance of wide-field surveys. Only a tiny fraction of the "submillimetre sky" has so far been surveyed, and the construction a SCUBA-2 camera, for a very modest investment, will have a major impact on all areas of astronomy.
 
 

Wayne Holland, Ian Robson & William Duncan
 
 

More information

A preliminary homepage is now available for SCUBA-2. This page is still under construction but more information can be found on the scientific goals, current status and available documentation. Please contact the SCUBA-2 Project Scientist (Wayne Holland) for more details.