SARAO News

David DeBoer, project manager and designer of the antennas for the Hydrogen Epoch of Reionization Array (HERA) built in the Karoo, was in an elevator when he first learnt that low-frequency radio interferometry could be used to deliver high measurements of redshifted 21 centimetre hydrogen emissions from the early Universe.

The mid-floor discussion introduced him to the possibility of detecting the ignition of the first stars and galaxies in the universe some 500 million years after the Big Bang – and that it could be done at a very reasonable cost compared to some other astronomy projects.

DeBoer was sharing the elevator at the University of California Berkeley over a decade ago  ago with an excited Prof Don Backer, then director of Berkeley’s Radio Astronomy Laboratory. The first papers that predicted this effect and signal were just being published.

A few early researchers – among which was Backer – began realising that low-frequency arrays could be used to measure the weak signature of the time when the Universe first started to light up and change the nature of the cosmos. This period is known to scientists as the  Epoch of Reionisation. They worked out that it could be done by measuring the power spectrum of fluctuations in the intergalactic medium introduced by the first luminous sources in the universe.

DeBoer remembers fondly how Backer explained that it would not only make a great experiment, but could be done reasonably cheaply using basic building materials such as wood, mesh and PVC water pipes.  “All you need is a paperclip!” was his excited refrain.

Soon after, Backer and one of his graduate students, Aaron Parsons, partnered with the US National Radio Astronomy Observatory (NRAO) to develop the first equipment, and the University of Virginia, the University of Pennsylvania and other partners also quickly came on board. After Backer’s passing in 2010, Parsons was appointed principal investigator of what was to become the Donald C. Backer Precision Array for Probing the Epoch of Reionisation (PAPER). Along with the Murchison Widefield Array, it would become a frontier project in the effort to detect this cosmological signal..

In turn, DeBoer moved to the Southern Hemisphere as director of the Australian Square Kilometre Array Pathfinder (ASKAP) and help found the Murchison Radio Astronomy Observatory (MRO). He also oversaw the installation of the first four PAPER elements on Australian soil in 2007. Another 32 dishes were deployed at NRAO, while the center of PAPER activities moved to South Africa. It culminating in the eventual 128-dish array moving to South Africa when SKA SA became a key partner in 2009. This was in part done because the country was fast becoming a leader in developing the digital processing that is so vital to large-scale data-driven astronomy projects.

DeBoer says the partnership with SKA South Africa has ensured a significant expansion in the project: “It has taken it from something very small; it has grown into a big thing.”

What goes where?

“When you want to get a global signal in the sky, you want a small element,” explains DeBoer. “Small elements see more of the sky, which is a funny kind of inverse relationship. It makes it very powerful because you can see a large part of the sky.”

In PAPER’s case, the small size of the elements made it possible to move them around easily in search of the perfect placement pattern to pick up the best possible signal.

DeBoer jokes that in radio astronomy, scientists squabble most about where precisely to place dishes to maximise returns.

By the time PAPER was already partially deployed, Parsons had a great insight, which would give rise to the whole new design and layout seen in the HERA array. He realised that detecting the signal the team wanted could be done best by repeating the same spacing between elements over and over again.

This gave rise to a new design of 14 meter diameter non-tracking dishes that will deliver on the sensitivity and the precision needed to reach HERA’s primary science objective. It uses the latest proven bright foreground filtering techniques, while new subtraction techniques are also being developed to unlock new capabilities.

The redevelopment of PAPER into HERA is very much about providing more raw sensitivity, and about collecting as much radio wave power as possible to increase the chances to detect the exceedingly weak signal aimed at.

Measurements and modelling showed that the best way to get more raw sensitivity would be to simply build more dipoles. To do so would however be quite costly, as between 10 000 and 20 000 would be needed to do the trick.

“Quite frankly, a correlator that can handle so many feeds is currently not buildable for any reasonable financial budget or even power budget,” explains DeBoer. “Therefore we decided to take another path, and simply throw more metal at it.”

He’s referring to the new antenna design, which has seen a substantial increase in the size of the collection area per dish.

Gone are the so-called chicken coop designs of the PAPER dishes randomly placed across the landscape. If all goes according to plan, by 2020 there will be an array of 350 dishes, each resembling the insides of an overturned umbrella. These will be packed into a hexagonal grid 300 meters across, leaving 14,6 meters between each.

“There are downsides, but in our analysis this is by far the most cost-effective to make a detection of the signal we are targeting,” adds DeBoer.

Nineteen dishes are already working, while construction of another eighteen is underway thanks to funding from Cambridge University. In September 2016 the programme received 9.5 million dollars from the American National Science Foundation to construct a further 240 dishes, and for the science, development and analysis work that are part and parcel of such a process.  With support from other partners, these funds allow the work towards full construction to commence.

The first 37 elements are predicted to provide a factor of about 5 more sensitivity than PAPER-128, and to provide a significant chance of detecting the EOR power spectrum signal.

HERA-128 should provide a robust detection of the EOR signal and allow some characterisation. If and when HERA reaches the 350 element mark, it will be able to do EOR science as a function of redshift and spatial scale, and potentially produce the first images of the EOR.

DeBoer guesses that as technology shifts and budgets improve in the next ten years, the HERA designs will again change back to “many, many, many” smaller elements, the likes of which are already being seen with the design of SKA Low in Australia.

“We are charging ahead, but even if there are still a lot of details and design aspects to work through, we are confident that the element aspects will do the job,” says DeBoer. “We are eager to make that detection, and to do good science.”

The first 19 operational dishes are gathering data and taking images. DeBoer however says these are essentially still “in a tuning phase”.

“We are giving ourselves a year to get a deep understanding of the instrument, so that we can ultimately get the signal we want,” he explains. “The images are not yet very impressive because there are still only 19 antennas in place, but researchers from Cambridge and the NRAO are already looking at the data, or looking to see if there are any artefacts to be detected.”

HERA timeline

• March 2015: Work starts on the first 19 HERA dishes
• February 2016: Work completed  
• July 2016: Construction starts in earnest on the next 18 antennas, to be completed by early 2017
• March 2019 observing season: If no further funding is received, the project ends.
• 2020: If full funding is received, the 350th HERA antenna will be built and observations will be taken until March 2022.

Key partners

The HERA collaboration consists of the following partner institutions:

•         Arizona State University
•         University of California Berkeley
•         University of California Los Angeles
•         University of Cambridge
•         Massachusetts Institute of Technology
•         National Radio Astronomy Observatory
•         University of Pennsylvania
•         Scuola Normale Superiore di Pisa
•         SKA South Africa
•         University of Washington

Additional collaborators are at Harvard University, the University of KwaZulu Natal, the University of Western Cape, Imperial College London, California State Polytechnic University.

Paper at hand
For more information about the latest science and developments behind HERA, read DeBoer, D. et al (2016). Hydrogen Epoch of Reionization Array (HERA) www.arxiv.org

Taking HERA forward

The 128 dual-polarization analogue and digital signal paths in the PAPER antennas at the SKA P2 site are slowly but surely being dismantled, and are being reused in the HERA dishes being built close by. A few feeds have also been sent off to partner institutions for research and development purposes. More will of course have to be built once all the available ones have been reinstalled.

The antennas are the most costly part of constructing HERA, which is being done based on specifications provided by the development team at Berkeley. For the rest, it’s really a question of using building material that can be sourced via most hardware shops: cement, underground PVC water pipes, 8 meter long tar poles and 1220mm wide birdcage mesh.

According to HERA project engineer Kathryn Rosie in Cape Town, low-cost construction material can be used because the project is so focused; the instrument only has to satisfy the requirements and tolerances associated with one particular experiment, resulting in a design life expectancy of approximately seven years from today.

The use of such materials has opened up opportunities to local business people from Carnarvon and the greater Karoo region to provide building material or to provide specific services. One such service provider, for instance, has at his home assembled more than 1980 of the mesh panels that make up the reflective part of the dish.

Most members of the construction team currently being employed had previously picked up skills through the Extended Public Works programme. Others joined SKA SA via an optic fibre training programme, and have received training on how to install for instance the optic fibre that runs between the dishes and the digitiser.

The recent funding received from the American National Science Foundation for 240 dishes means construction work will pick up pace over the next few months. For Rosie and MeerKAT lead technician Matthys Maree, who looks after matters relating to HERA’s on site construction, it means that even more service providers have to be sourced, more construction team members have to be employed and more tenders have to be put out.

Maree has great respect for the construction teams who often work under extreme conditions. Three tar poles per dish need to be planted in holes dug 1,5 meter deep into hard soil, for instance. The site provides very little by way of shade from the hot Karoo sun. “And in winter, it can be quite cold,” he admits.