I was the world’s first sub-millimetre astronomer.

This statement tends to irritate some of my colleagues, particularly those who had been lugging around sub-millimetre detectors to put on borrowed optical and infrared telescopes for a decade before I even became involved in the field. But it is true in the sense that I was the very first person to make observations with the James Clerk Maxwell Telescope (JCMT) on the day it opened for business on the summit of Mauna Kea in Hawaii 25 years ago. The JCMT was the world’s first dedicated sub-millimetre telescope, although it was not behaving itself on that first day, tending to drift away from wherever I pointed it.

There was also a much more fundamental problem faced by all sub-millimetre astronomers back then – there was simply no such thing as a sub-millimetre camera. All I had at my disposal was a single detector that could measure the strength of the sub-millimetre radiation only in the precise direction the JCMT was pointing, which on that day was not very precise at all. So to produce a sub-millimetre picture of the sky, I had to point the telescope in one direction and measure the strength of the radiation in that direction, before turning the telescope slightly and measuring the strength of the radiation in the new direction and so on and so forth. Painting by numbers, if you like.

About 10 years later, we did finally get a sub-millimetre camera, but it had only 37 pixels. That still left us well behind astronomers who work with visible light, for whom – even then – one million pixels were routine. (Of course, technology has now come so far that most mobile phones today have a camera that can record several megapixels.) It was only on 14 May 2009 that sub-millimetre astronomers finally caught up with their optical colleagues when the European Space Agency’s Herschel Space Observatory blasted off from French Guiana in South America. The observatory was named after the German-born astronomer Sir William Herschel, who discovered infrared radiation and the planet Uranus, with help from his sister Caroline.

Cool prospects

The sub-millimetre waveband, which consists of electromagnetic radiation at wavelengths between 100 μm and 1 mm, is the last waveband to be opened up for astronomy. It is the final electromagnetic frontier, if you will. One reason why sub-millimetre astronomers have lagged so far behind optical astronomers, who were after all taking pictures of the sky in Victorian times, is an embarrassment of riches. Virtually everything emits sub-millimetre radiation, whereas only very hot objects give off optical light. Optical astronomers therefore have it easy: to avoid the one object that gets in their way – the Sun – they just have to work at night.

Sub-millimetre astronomers, however, have to cope with the fact that everything around us emits floods of sub-millimetre photons. So the only way to detect faint signals from the sky without them being swamped by sub-millimetre radiation from the camera itself is to cool the camera down to less than one degree above absolute zero, which reduces the radiation from the camera to a minuscule level. The cameras on Herschel, for example, are cooled to 0.3 K using a large bath of liquid helium.

Another problem for sub-millimetre astronomy is that water vapour in the atmosphere absorbs this kind of radiation, which means that the technique can only be carried out from the tops of high mountains across just a handful of wavelengths that are free from water absorption. That was why, in the pre-JCMT period, astronomers had to perform heroic feats, transporting cumbersome cryogenic equipment to some very remote and inhospitable places, such as the summit of Mauna Kea, which lies more than 4000 m above sea level, where conditions are suitably dry. And when they were there, the sub-millimetre astronomers only had a short time to get their bolted-on kit working before the regular astronomers took over.

Herschel gets around the problem of moisture entirely by being located in space. In fact, it has the biggest mirror ever sent into space, some 3.5 m in diameter – larger even than that on the Hubble Space Telescope. But while Hubble circles the Earth every 90 minutes only 560 km above the surface, Herschel is at the second Lagrangian point, L2, some 1.6 million kilometres from us. It is one of five points at which a small object (Herschel) can remain stationary relative to any two much larger objects, in this case the Earth and the Sun. L2 is ideal for Herschel as it is so far from our planet, which is a beacon of sub-millimetre radiation.

One of my favourite Herschel pictures reveals just why astronomers bother working with electromagnetic frequencies outside the traditional optical band (figure 1). It shows the Andromeda galaxy, which is one of the two big galaxies in the Local Group of galaxies – ours, of course, being the other. Andromeda is just visible with the naked eye, but if you view it with even a small optical telescope, you can see a bright, central bulge of old, red stars surrounded by a disc of much younger blue stars. In the Herschel image, in contrast, the bulge has vanished because stars emit hardly any sub-millimetre radiation. The radiation that Herschel is detecting in this image comes instead from tiny interstellar dust grains that are located in the outer disc, but not in the bulge. These grains emit sub-millimetre waves because they absorb starlight and get heated to temperatures of about 30 K.

These grains are a huge problem for optical astronomers because they act like smoke – scattering and absorbing optical light. This interstellar smoke is thickest in the big gas clouds in which new stars are being born – so the bright ring in the Herschel image of Andromeda confirms the notion that stars in this galaxy are being born in the disc but are no longer being created in the inner bulge. Although the optical image shows some of the light from these newly formed stars, there are many more that are hidden by dust when viewed with an optical telescope.

Dust grains are basically not that interesting, except to the small group of astronomers who are absolutely passionate about the physics and chemistry of these tiny particles. (I’m not joking – some of my best friends are mad about dust.) Most astronomers, instead, are more interested in dust grains for what they might conceal. The classic example is the Horsehead nebula (figure 2), which is typical of the kind of place that William Herschel himself thought might be “holes in the heavens” with genuinely no stars. We now know, however, that these “holes” are just places where dust is in fact hiding the stars beyond.

Interestingly, the sub-millimetre picture of the Horsehead taken with a camera on the JCMT by my Cardiff University colleague Derek Ward-Thompson shows that the horse appears to have “swallowed” an object that now lies stuck in its “throat”. The object is a place in the nebula where the dust is particularly dense and new stars are being born. Light from the young stars is heating the dust, which increases the amount of sub-millimetre radiation it produces.

Seeing the birth of stars

This ability to find and study newly formed stars, which are born in dense clouds of dust and gas, is one of the reasons why Herschel was launched. Quite simply, it would be impossible to study the first moments in the life of a star with a conventional optical telescope like Hubble because all that dust gets in the way. In peering into the big clouds of gas and dust that are the “maternity wards” of stars and then detecting the sub-millimetre light emitted from the dust around the newly formed stars, Herschel is therefore doing much to study star formation, which is one of astronomy’s “big questions”.

Indeed, Herschel has already taken a major step forward in showing how low-mass stars like the Sun were formed. Figure 3 shows pictures taken by Herschel of two of these stellar maternity wards, which exist as big clouds of gas and dust in a ring called the Gould Belt roughly centred on the Sun. Taken by a large international team led by Philippe André in Paris and including my colleagues in Cardiff, these pictures are actually pseudo-colour images made by combining three Herschel images taken at different wavelengths. The colours reveal the temperature of the dust, with red indicating cold dust and blue showing warm dust.

The picture of one of the clouds, the Aquila Rift, which is about 750 light-years from Earth, is a coruscating colour-drenched image that looks remarkably like a painting of a sunset on a stormy evening by William Turner in one of his more exuberant moods. The picture of the other cloud, the Polaris Flare, which is about 500 light-years away, is a monotonous brown. The explanation for the differences is that stars are being born in the Aquila Rift but not in the Polaris Flare. The full palette of colours for the Aquila Rift shows that the dust has a range of temperature, with the blue and yellow “paint drops” revealing warm spots heated by newly formed stars. In contrast with this Turneresque image, the monochrome image of the Polaris Flare shows that all the dust is at exactly the same temperature, and thus that there are no newly formed stars.

But why are stars being formed in one cloud but not the other? The answer appears to lie in the properties of the streaks of dust and gas, known as filaments, that snake across all the Herschel images of clouds in our galaxy, which can probably best be seen in the picture of the Polaris Flare. There appears to be a critical mass density of about five solar masses per light-year of filament length, above which – as in the Aquila Rift – gravity causes the filaments to collapse to form protostars that appear as beads on the filaments. Below the critical value, as in the Polaris Flare, the filaments never collapse and no stars are born.

But what causes the filaments to form in the first place? A clue appears to lie in the recent discovery by the Gould Belt team that while the density of the filaments can vary wildly, their width – no matter where they are seen in the galaxy – is always very similar, being about one third of a light-year. Remarkably, the team thinks it can explain this by turning to some simple physics of the turbulent interstellar gas. According to its model, the gas is usually flowing faster than the speed of sound in the gas, but when it slams into a big cloud of stationary gas, it slows down to below the speed of sound to form a filament. Indeed, the model says the gas piles up in exactly the way seen in the filaments.

This explanation is what simple physics suggests, but remember that astronomers are not working in a laboratory so it is rarely possible for us to “prove” anything. The best we can usually do is to find a model that fits our observations.

Another of astronomy’s big questions concerns how galaxies are formed. One of the big discoveries made 15 years ago with the first sub-millimetre camera on the JCMT was that there are some galaxies in the early universe that are so shrouded in dust that they are emitting 1000 times more radiation in the sub-millimetre waveband than at optical frequencies. Indeed, they are such luminous sub-millimetre sources that the dust must be hiding a very large number of newly formed stars. Calculations suggest that the stars are being created so quickly that an entire galaxy could be made in barely 100 million years, or about 1% of the age of the universe. These luminous sources therefore almost certainly hold a clue to how galaxies were formed.

The future is bright

Sub-millimetre astronomy is advancing at a quite astonishing pace. Just 15 years ago when astronomers started using the first sub-millimetre camera on the JCMT, it would take a whole night to find a single one of these shrouded galaxies. The top image above, which was the first taken by my team shortly after the Herschel Space Observatory was launched, took only 16 hours to make and reveals 7000 dusty galaxies – from those that are nearby to others that are so far out in space that we are looking 10 billion years back in time.

Unfortunately, the 2160 litres of liquid helium that Herschel originally contained will finally run out in March this year. When that happens, the cameras will warm up, and the telescope will become just another piece of space junk among millions of other bits of rubbish now floating in space. However, the treasure trove of Herschel data will be picked through by astronomers for years to come.

Meanwhile, a new sub-millimetre telescope – the Atacama Large Millimeter Array (ALMA) – has started operation 5000 m above sea level in the remote and inhospitable Atacama Desert in central Chile. Although ALMA can only operate in atmospheric windows at a few wavelengths, its advantage over Herschel is that it has much higher angular resolution. ALMA will therefore be able to produce detailed pictures of the galaxies that only appear as little blobs to Herschel. It is likely that ALMA’s observations of the sources detected in the Herschel surveys will be the key to providing answers to the origin of both stars and galaxies. William and Caroline Herschel would have been proud.