Dissecting 3000 galaxies with the Anglo-Australian Telescope
Recently, the third and final data release from the SAMI survey was released, providing scientists from around the world with a vital resource that will help analyse how galaxies formed and their evolution. Dr. Ángel López-Sánchez, who forms part of the SAMI global collaboration, walks us through a very personal account of the journey to this release.
One of the reasons why, when I was still a kid, I was obsessed with Astronomy, was the amazing colourful pictures of nebulae, star clusters, galaxies and exotic space objects that the famous astrophysicist David Malin was publishing.
Indeed, David Malin was producing the very first astronomical images in colour using the 3.9m Anglo-Australian Telescope (AAT), located at the Siding Spring Observatory near Coonabarabran (NSW). Who would have known at the time that a few decades later, he will be using the same cabin for conducting pioneer astronomical observations aiming to dissect 3000 galaxies!
This has been the goal pursued by the SAMI Galaxy Survey: to dissect a large number of relatively nearby galaxies to unveil the details of how stars are born in galaxies, what mechanisms make galaxies stop forming stars, how galaxies grow with time, what is the effect of the environment, and, in essence, what drives galaxy evolution.
Conducting the “SAMI Galaxy Survey” has been one of the key tasks of the “ARC Centre of Excellence for All-Sky Astrophysics in 3 Dimensions” (ASTRO 3D). Around a hundred international astrophysicists make up the team of the “SAMI Galaxy Survey”. The majority of them are based in Australia. And I’ve been a member of the SAMI team since the beginning.
How do we get the data for the SAMI Galaxy Survey? We use a special and pioneer instrument called “Sydney-AAO Multi-Object Integral-Field” (SAMI), that feeds the AAOmega spectrograph at the AAT. The SAMI instrument was first developed in 2011-2012 and it was unique in the world because, for the first time, it was able to dissect not only 1 galaxy at a time, but 13 of them!
How do we dissect galaxies?
Let’s recap for a moment. Galaxies are extended objects in the sky. It’s true, the majority of them are small, but they do have a size and we can resolve them. It is not the same thing observing the centre of the galaxy that its outskirts, particularly when many star-forming regions are scattered throughout the spiral disk. If we really want to understand how galaxies work, we need to get their local and global properties.
In the old times, we could get images with different filters (blue, green, red, infrared, even special filters like Hydrogen-alpha that allows us to see the light emitted by the nebulae and the hot gas in the core of galaxies). But this is not enough: we need the spectroscopic data, that is, a detailed decomposition of the light in all their colours.
The key information of the Universe is encoded in the light. Studying the spectra of galaxies we can determine their distances, their movement and rotation, the chemical and physical properties, the dominant stellar populations, the position and properties of their star-forming regions, their amount of dark matter, and much more.
Of course, the problem here is the drastic increase of the complexity: from only a few data points using filter imaging to several thousand using spectroscopy. For many years we could do only one thing or the other: taking spectra or images. But no both of them.
For the last couple of decades, however, astronomers have been developing a new kind of detectors that allow us to obtain, at the same time, images and spectra. These devices are called “Integral Field Units” (IFUs), the “galaxy dissectors”, and are truly changing the way we understand galaxy evolution. Feeding the IFU to a spectrograph we get the technique of the Integral Field Spectroscopy (IFS).
The first large galaxy survey to use the IFU technology was the CALIFA Galaxy Survey. CALIFA stands for 'Calar Alto Legacy Integral Field Area'. Led by Spanish astrophysicists and completed in 2015, the CALIFA Galaxy Survey used a particularly large IFU mounted at the 3.5m telescope of the Calar Alto Observatory in Almería (Spain) to dissect around 600 nearby galaxies. I am also a member of the CALIFA team. CALIFA opened the path for the new large galaxy surveys we are pursuing today. But observationally it was very inefficient: we could only observe one galaxy at a time, and we needed to observe each galaxy twice to get the full wavelength coverage. Indeed, it took us almost 10 years to get all the data!
Developing New Technologies
Because of the pioneering work of the CALIFA Galaxy Survey, experts in astrophotonics and instrument scientists at both the University of Sydney and the Australian Astronomical Observatory (now rebranded Australian Astronomical Optics and managed by Macquarie University, the Australian National University, and the University of Sydney), had the idea of building an instrument for the AAT that were able to use around ten small IFUs.
In less than a year they built the first SAMI prototype, which already included 13 hexabundles (the small IFUs) that can be deployed anywhere within a 1-degree field of view using the AAT's Prime Focus Camera top end.
Each hexabundle is composed of 61 ‘cores' or individual optical fibres packed together, creating a circular field of view of around 15 arcsec in the sky. Additionally, they added 26 individual fibres for getting the background sky (its subtraction is critical for getting good quality data) and 3 special bundles to observe stars that can be used for accurately guiding the AAT.
However, there is a big problem here. We all were very used to observe with the great “2dF” instrument, the famous AAT robot that can allocate 400 optical fibres anywhere within a field with the size of two degrees (four full moons in the sky) in diameter (hence, “2dF” stands for “Two Degrees Field”). We didn’t have to do much, just preparing the catalogue with our targets, running the software, and the robot builds the plate with the desired fibre allocation.
But we couldn’t do this with SAMI! Every individual bundle and fibre in SAMI must be plugged manually into a pre-designed plate. On top of that, the SAMI plate had to be manufactured well in advance of the scheduled observing run. We cannot improvise: the details of the observations must have been ready weeks in advance.
Observing with SAMI
Conducting astronomical observations with the SAMI instrument at the AAT was far from trivial. Indeed, we needed a couple of years to actually get a good plan for it. The first part was mounting the Prime Focus Camera top-end hosting SAMI in the telescope and attaching the 819 optical fibres (13 x 61 hexabundles + 26 sky fibres) to the AAOmega spectrograph.
This was carefully performed by the excellent specialist staff at the Anglo-Australian Telescope, under the close supervision of A/Prof. Julia Bryant (University of Sydney), who is the leader of the SAMI instrument, and world-renowned for her astrophotonic expertise and leadership.
Once SAMI was successfully mounted at the AAT and connected to AAOmega it was time to configure the plate for the night. To make our lives a bit easier, each SAMI plate could be used for configuring two different field positions, which would be observed at different times of the night. In the afternoon one of us had to jump into the cabin and, with the help of another observer, plug each hexabundle and sky fibre on the plate, following our scheduled plan.
It was critical to properly associate the hexabundle number to the galaxy position, to ensure that we are correctly matching the data to each galaxy. Also, hexabundles should not be bent or twisted, otherwise, they could lose their efficiency, and hence extreme care with them had to be considered. This process usually took 45 minutes, although by the end of the survey some experienced ‘SAMI pluggers’ were able to do it in 20-25 minutes.
After plugging the SAMI plate it was time to start gathering the calibration data we needed for the proper data reduction. This included important observations of the twilight sky which are needed to properly adjust the flux of all fibres at the same level.
During the twilight, we also got some data of bright calibration stars in a few of the hexabundles. This is needed for obtaining the absolute flux calibration (that is, converting the information captured in the detectors into physical units).
Once astronomical twilight ended, the telescope was ready and placed in the field position. We could be there, observing with SAMI 13 galaxies at once, for 3 - 4 hours. The individual frames were all half an hour exposures, and we obtained at least 7 of them. However, every frame was observed with a small telescope offset. This process was important to recover spatial information of the galaxy, as the gaps within the fibre cores account for around 1/3 of the area of the hexabundle.
I recall, one of the most special moments of observing with SAMI - at least once every night we’d have to jump into the cabin and re-plug SAMI, running through the process all over again.
Sometimes we used the same plate but moved the hexabundles and the guiding bundles to the other position of the plate (sky fibres remained untouched), but at other times changing the plate completely. We had to do this in a dark, inside a cold dome, perhaps at 1 or 2 am.
I always felt quite excited when replugging SAMI in the middle of the night, as it reminded me of what David Malin and other astrophysicists had to experience decades ago when obtaining photographic plates and they needed to stay alone in the dark for hours.
Dealing with SAMI data
If at this point you thought that all the work was done and the data collected at the telescope was ready for science analysis, think again. What comes out of the telescope is the ‘raw’ data, which are useless for science unless they are properly processed and calibrated. Ah, oh boy, this is not an easy task.
In the same way that observing with SAMI has provided me with some of my most memorable moments working in a professional telescope, dealing with the SAMI data processing has been one of my most frustrating moments of all my career. Why, you ask?
I have extensive experience dealing with spectroscopic data obtained with large telescopes, indeed I started doing this in the early stages of my PhD Thesis at the “Instituto de Astrofísica de Canarias” (Canary Islands Institute for Astronomy) in Spain in 2001. By 2013 I had already created many data processing scripts and documentation about this.
But, all of this was or in IRAF (“Image Reduction and Analysis Facility”. a famous astronomy data reduction package developed by the US National Optical Astronomy Observatory) or in IDL (“Interface Description Language”, a successful programming language that was widely used by the astronomical community in the 2000s). I started using both of them while doing my PhD Thesis, and when I had the extra time for learning these software tools.
However, the SAMI team, that had a large fraction of brilliant young astronomers, decided to develop a specific package for processing the SAMI data using Python. And it so happened that my knowledge of Python was zero.
I will never forget a busy SAMI workshop in 2013 where I wanted to help on the software development and, even though I knew what I wanted to do, I couldn’t follow anything. I must confess here that I cried at the end of the day, full of frustration, and feeling a complete incompetent. This also meant I couldn’t help as much as I wanted in all the SAMI data processing process.
Side note - Handy hint from me to young astronomers: if you are a student that wants to be involved in science, please learn Python as soon as possible!
Side note 2: And now that I have learnt Python, I’ve been able to develop my very own data reduction package for the KOALA (Kilofibre Optical AAT Lenslet Array) instrument, another IFU using the AAOmega spectrograph at the AAT.
So after years of refining the SAMI data reduction Python package, the team got an excellent product that is able to process almost automatically the data of all galaxies, if needed. And, on top of that, the processed data and some important derived products have been made available publicly in the Australian Astronomical Optics (AAO) Data Central.
And this brings us to the reason why I’m writing this article today. On February 1st the SAMI team published a scientific paper compiling the final release of all the SAMI data (the “Data Release 3”, DR3), which comprises the full sample of the 3068 unique galaxies observed in the survey. That same day all the SAMI data have also been publicly released in AAO Data Central.
The SAMI DR3 paper has been led by the principal investigator of the “SAMI Galaxy Survey”: Prof. Scott Croom (University of Sydney). This research paper has 41 authors (me included), drawn from Australia, Belgium, the US, Germany, Britain, Spain and The Netherlands.
A significant part of the paper was prepared by Dr. Matt Owers (Macquarie University), who worked hard to improve the data processing, and who leads the research of the SAMI galaxies within clusters (888 in total). The SAMI data of these cluster galaxies have been published for the very first time here.
The SAMI science discoveries
To date, the SAMI Galaxy Survey has published 65 peer-review papers, with around an extra dozen science papers that are still in revision or closed to be submitted. It is impossible to summarize here all the amazing discoveries that SAMI has produced so far.
The very first paper, led by Dr. Lisa Fogarty, introduced the survey and the instrument and presented the serendipitous discovery of a galactic wind in a peculiar SAMI galaxy that was observed in the pilot observations.
After that, we explored gas and star kinematics in galaxies, discovered extraplanar gas and more galactic winds associated with star-formation, found a link between the angular momentum of the galaxies and their optical morphology, and spatially resolved star-forming regions and areas that recently stopped forming stars.
We also explored the connection of the mass of the galaxies with their star-formation activity, their chemical properties, the kinematics, their bulges and their environment, constrained galaxy formation and evolution cosmological models, studied the relationship between misalignments of how stars and gas rotate to trace accretion of cold gas from the intergalactic medium, explored the stellar populations in both spiral and elliptical galaxies, decomposed the kinematics of disks and bulges within the same galaxy, and much more.
The SAMI DR3 science paper presents some general but quite interesting plots. For example, its Figure 20 (shown above) shows the map of the ionised gas (that is, the nebulae and the galactic nuclei) using the emission of the famous Hydrogen-alpha line at 6563 Angstrom.
These maps are organised by stellar mass (that is, the number of stars in each galaxy) on its horizontal axis (galaxies located at the left part of the diagram are less massive than galaxies located at the right part), and the local density of galaxies (which measures how populated is the nearby environment with other galaxies) on the vertical axis (galaxies that are almost isolated are at the bottom, galaxies that are located in the centre of galaxy clusters are located at the top).
Similarly, Figure 21 (shown above) in the SAMI DR3 science paper shows the ratio between the emission of the single-ionized nitrogen, [N II], and the H-alpha emission, that can be used as a first tracer of the number of metals (i.e. any element that is not hydrogen or helium) in a galaxy when applied to its nebulae. Purple colours indicate low amounts of metals (low “metallicity”), which are found in the low-mass galaxies, while red colours codify metal-rich galaxies. These are typically found in the high-mass regime, as massive galaxies have experienced larger star-formation activity than the low-mass galaxies and therefore their gas have been processed further.
This three-and-a-half-minute video shows highlights of researchers adjusting and deploying the SAMI instrument at the AAT. My colleagues Luca Cortese (University of Western Australia) and Jesse van de Sande (University of Sydney) help me to plug the SAMI plate before starting the observations. Credit: Ángel R. López-Sánchez (AAO-MQ).
The next step: Hector
But this story does not end here.
For the last few years, we have been developing the “successor” of SAMI, called Hector. Hector uses the very same 13 SAMI hexabundles, but adds 8 new larger hexagonally-packed bundles with up to 217 fibre cores each, making it possible to observe larger galaxies.
Hector will also be a pioneer in positioning the bundles not manually but using a robot. Furthermore, a new high-resolution spectrograph will be added for the new hexabundles (the existing 13 bundles will still use the AAOmega spectrograph) to increase the survey speed and galaxy coverage.
We expect to observe ~15000 galaxies in the next few years. This big number of galaxies is needed to perform a detailed statistical analysis of well-resolved galaxies in the nearby universe.
Only with these data using integral field spectroscopy in large galaxy samples, it is possible to connect galaxy evolution and kinematics to the large-scale structure of the Universe to explain the evolutionary history leading to the individuality of galaxies, constraining the complex cosmological models' theoretical physicists are now developing.
As part of the Hector science team, I’m already involved in the preparations of the first Hector observations, already scheduled by mid-2020 at the AAT. These are exciting times for being an astrophysicist working in galaxy formation and evolution in Australia!
Dr Ángel López-Sánchez
Dr Ángel López-Sánchez is an astronomer and science communicator at the Australian Astronomical Optics (AAO) and the Department of Physics and Astronomy at Macquarie University (MQ) based in Sydney, Australia. He is a recognised expert in the study of how the gas is converted into stars in nearby galaxies and how this affects galaxy evolution.
He leads the "HI KOALA IFS Dwarf galaxy Survey" (Hi-KIDS) program, that uses the instrument KOALA at the 3.9m Anglo-Australian Telescope (AAT) to dissect 80 gas-rich nearby dwarf galaxies to understand their assembly story. He also provides support for visiting astronomers to the AAT. He is an active member in large spectroscopic galaxy surveys (such us GAMA, CALIFA, SAMI and Hector) and upcoming optical and radio galaxy surveys.
He is a globally-recognised science communicator, with visibility in Spanish and Australian printed, broadcast, and social media (one of the Top-100 most-followed astrophysicists in Twitter). He is also a passionate amateur astronomer that uses his own equipment for capturing the beauty of the Cosmos. His Twitter feed is @El_Lobo_Rayado.
Read the paper now on arXiv.org