17 mins read 29 Jun 2023

Australian Scientists Help Uncover Cosmic Gravitational Rumblings

Australian astronomers, using CSIRO’s Parkes radio telescope, have today announced the best evidence yet of the stochastic gravitational wave background, opening up a new chapter into gravitational wave astronomy. As the fourth and final instalment of this series, we spoke with the two scientists leading the Australian papers in this exciting discovery.

A Galactic-scale detector, using pulsars to measure the distortions of space-time generated by distant supermassive black hole binaries. Credit: Shanika Galaudage.

As part of a global collaboration, Australian astronomers have today announced the strongest evidence yet of the stochastic gravitational wave background (GWB) - a cosmological signal that is expected to be generated by the cacophony of supermassive black hole binaries across the Universe (amongst other sources). These behemoth objects (where masses are billions of times that of our Sun) reside in the hearts of galaxies that orbit each other in a slow,  inspiral decay towards their ultimate fate of merging. In the process, they radiate low-frequency gravitational waves into the cosmos. 

These new results, presented jointly by Pulsar Timing Array (PTA) teams from around the world (as part of a global collaboration known as the International Pulsar Timing Array project), are also the first time scientists have confirmed a special spatial correlation that supports their findings. This opens up the opportunity to now extend our understanding of the gravitational wave (GW) spectrum, in particular in the nanohertz frequency regime, complementing the existing detections of higher frequency gravitational waves (which are generated by the in-spiral and merger of stellar-mass compact remnants), that have been confirmed by terrestrial-based instruments like LIGO, VIRGO and KAGRA (LVK). 

As part of these studies, scientists have been monitoring the zombie cores of former massive stars, long after they have exploded in violent supernova events, leaving behind a rapidly rotating, and highly magnetised compact object known as a pulsar. In particular, a special kind of pulsar, known as a Millisecond Pulsar (MSP) which are very stable over the long term, and behave like cosmic clocks littered across our Milky Way Galaxy. 

Astronomers from the PPTA have been timing MSPs for nearly two decades now, keeping a watchful eye on the ‘timing residuals’ - an outcome of the predicted model of when the pulsar's signal will arrive relative to the actual observation. Unlike the confirmed detections made by LVK of stellar mass compact remnants (which have wave periods that last in the milliseconds to seconds range), nanohertz-frequency GWs generated by supermassive black hole binaries have wave periods that last years and decades, hence the data stretching over such a long time span. This is also why astronomers could never build an instrument here on Earth to detect these kinds of GWs - and instead, cleverly developed a Galactic-scale detector using MSPs that are spread across the sky, known as a Pulsar Timing Array.

The smoking gun of the gravitational wave background - first signs of detection are presented in this Hellings and Downs correlation plot, which shows the measured spatial correlations based on angular sky separation. The number of pulsars in each of the angular bins are the grey-shaded histogram in the background. Credit: Reardon et al. 2023.

The exciting announcement confirms that the smoking gun of a Gravitational Wave Background (GWB) detection, known as the Hellings and Downs (HD) correlation - has been made with all PTAs involved starting to find evidence for it. The HD correlation is the quadrupolar spatially correlated signal observed across all the timing residuals that the MSPs present. 

This is predicted as part of Einstein’s General Relativity, in which passing GWs create tiny variations in the distance between the Earth and each of the MSPs (by distorting space-time) and thus induce a small variance in the MSP's timing residuals. A shift of this nature observed from a single MSP provides interesting science, but the correlated signal that shows all MSPs separated across the sky undergoing this shift is the tell-tale sign of the GWB. 

“We can detect gravitational waves by searching for pulses that arrive earlier or later than we expect. Previous studies have shown an intriguing signal in pulsar timing array observations, but its origin was unknown,” said OzGrav and Swinburne University of Technology researcher Dr Daniel Reardon, who led the searches. “Our latest research has found a similar signal among the set of pulsars we’ve been studying, and we now see a hint of the fingerprint that identifies this signal as gravitational waves.”

Hints of a common noise process (i.e., a common signal, such as that induced by the GWB) across all of the MSPs were first reported in publications in 2021 - which had pulsar astronomers fairly excited around the world, as this is expected to be the first indication of the GWB. However, without the HD correlation, this could not be confirmed with certainty, and further research indicated it could potentially be a false positive masquerading as the GWB signal. 

CSIRO astronomer Dr Andrew Zic, who co-led the analysis, said that while it is exciting all the major collaborations are seeing hints of the waves the true test will come in the near future when all of the data is combined into a global dataset. 

“This signal could still be caused by things like variations in a pulsar’s rotation over a long period of time, or may simply be a statistical fluke,” Dr Zic said. “Our Parkes radio telescope, Murriyang, has an advanced receiver and an excellent view of the best pulsars in the southern sky, which are essential for this work.”

A Global Array of Astronomers

IPTA telescopes from around the world with IPTA members being represented from North America, Australia, Europe and India. Emerging teams in China, South Africa and Argentina (not shown) have also started to time millisecond pulsars. Credit: Dr Thankful Cromartie.

Observatories and teams of astronomers in different parts of the world that track pulsars that are local to their own skies form the basis of the International Pulsar Timing Array (IPTA) project. This collaboration, which features several hundred scientists, is made up of four smaller PTA teams. These are the Parkes Pulsar Timing Array (PPTA) project, located in Australia, the European Pulsar Timing Array (EPTA), the North American Nanohertz Observatory for Gravitational Waves (NANOGrav), and the Indian Pulsar Timing Array (InPTA). Additionally, there are several emerging PTA teams in China, South Africa and Argentina. 

The benefits of having PTA teams around the world includes boosting the sensitivity of detection of the GWB signal by combining data sets (as a higher number of MSPs improves the sensitivity of an array), providing all-sky coverage, and importantly - so that the teams can also cross-verify each other's work independently, ensuring that any interesting findings are indeed reproducible and hold up against rigorous scientific analysis.  

Here in Australia, the PPTA observations have been running at regular cadence since 2004 using CSIRO’s Parkes radio telescope (also known as Murriyang, its Wiradjuri name, which means ‘Skyworld’). The iconic, 64-metre dish located in central NSW is also responsible for the discovery of nearly half of all known pulsars to date. The Dish, as it's colloquially called, is also the only telescope in the world equipped with the Ultra Wideband Low-frequency (UWL) receiver - an innovation by Australia’s science agency, CSIRO, that allows radio astronomy data to be collected across a massive 3.3 GHz of bandwidth.

Looking up at the focus cabin of the Parkes radio telescope, in which the UWL receiver is hosted. Credit: Rami Mandow.

“The ultra-wideband receiver has been a boon for the PPTA. We can now cover frequencies from 704 to 4032 MHz in one single shot, whereas previously we’d have to cover different parts of the band in separate observations with multiple receivers” said Dr. Zic. “The wideband feed also increases our instantaneous sensitivity and allows us to capture the way that pulsar signals evolve across this huge 3.3 GHz bandwidth.”

“Observing at a vast range of radio frequencies allows us to separate signals like gravitational waves, from noise caused by the ionised interstellar medium – plasma between us and the pulsars,” added Dr Reardon, about the UWL. 

The PPTA monitors a suite of about 30 MSPs, which are mostly located in the global south - some of which are inaccessible to other IPTA teams because their observatories are too far north. However, these southern MSPs are also visible to emerging PTA teams, in particular the MeerKAT telescope which is located in South Africa. 

Parkes Pulsar Timing Array Results

The CSIRO Parkes radio telescope (‘Murriyang’). Credit: Alex Cherney / CSIRO.

The team from Australia have used the third data release of the PPTA (observations made by the Parkes radio telescope over the last 19 years) to search for and characterise both the smoking gun of the GWB detection - the HD correlation - as well any common-spectrum noise processes. 

What the team found was evidence for the expected amplitude and spectral index of the isotropic stochastic GWB signal that is generated from a background of GWs radiated by inspiralling supermassive black hole binaries.

“The spectral index describes the relationship between the strength of the signal and the length of the wave – gravitational waves should get much stronger as we look at longer waves,” said Dr Reardon.

“The signal we observe is consistent with this, and the amplitude is close to that expected of a population of binary supermassive black holes distributed throughout the whole Universe. What we could be seeing is the sum of all of the gravitational waves emitted by every one of these binary supermassive black holes in the Universe. These add to a very slow rumble of waves, like a random rolling ocean, that take years or decades to oscillate,” he said.

A number of interesting findings in the PPTA GWB results have also been presented. For example, the amplitude signal strength is time-dependent, a finding that was motivated when the team noticed that previous upper limits of the GWB set by all PTAs were lower than the common noise being detected in recent and current data sets. These results indicate that the common noise (i.e. the GWB signal presented across all MPSs) is increasing in amplitude with time - which in itself would be strange, and so the team is further investigating and not ruling out any other possibilities (such as data processing misspecification). 

“If the gravitational waves are truly distributed equally on the sky, the gravitational waves should look the same in every direction and the signal should be mostly stable with time,” said Dr Reardon.

“We seem to be seeing something get stronger with time in our data, and that is unexpected. We really don’t know why this is and it could just be something in the pulsars themselves changing with time in a way we don’t account for completely.” 

“But an exciting alternative is that this is an unexpected property of the signal. One possibility is that the gravitational waves could be stronger in some parts of the sky than others,” he said.

An interesting finding in the PPTA paper - when taking both 6 and 9 year slices of the data, the team finds that the common noise amplitude increases as a function of time. This could potentially be indicative of the GWB signal being stronger in some parts of the sky relative to others. Credit: Reardon et al. 2023.

To find these results, PTAs (like the PPTA) must model a number of noise properties that can make the detection of the GWB signal challenging. This can include low-frequency noise intrinsic to the MSP (because they do not have perfectly stable rotations), noise induced by variations in the interstellar medium through which the MSP signal must travel, as well as any errors in Solar System ephemeris models, clock errors, unmodelled variations in the solar wind, or instrumentation offsets. 

Accompanying the GWB paper, the PPTA have also released a noise analysis paper, which describes the complex noise models used for each of the MSPs. This is used to characterise the different noise sources as accurately as possible to search for and extrapolate the GWB signal (which itself is a correlated noise source seen across all MSPs). 

“Many physical phenomena imprint signals on pulsar times of arrival, including the gravitational wave background,” said Dr Reardon. “To search for gravitational waves, we need to make sure we are characterising all the signals that are present in the pulse delays so that we don’t mistake other effects for the thing we’re looking for”

In this analysis, the team looked at all the different noise sources for each of the MSPs (such as white noise, achromatic red noise, and dispersion measure variations that originate both in the interstellar medium as well as the solar wind to model the noise parameters accurately. 

From it, they were able to demonstrate that these noise processes are more complex than previously assumed, impacted by sudden changes in MSP magnetospheres or scattering in the interstellar medium. As the authors discuss in the noise analysis paper, the PPTA team find that choosing different noise models also significantly affects the common noise processes (such as that of the GWB signal), and the for this release, the PPTA selected advanced noise modelling that included additional terms of band and system noise, chromatic noise, high fluctuation frequency and variable solar wind terms. 

The Galactic distribution of classical pulsars (grey) and millisecond pulsars (lilac). As can be seen, most lie along the central Galactic plane. The 26 original millisecond pulsars used in the PPTA’s second data release are marked with yellow stars, where the additional new millisecond pulsars are marked in red stars. Credit: Rami Mandow.

To come to these results the PPTA released a third paper - the data release. Here, all 32 of the MSPs used in the PPTA are described in terms of their properties - such as their period, dispersion measure, orbital periods, the timespan of observations and any interesting features they each have.

A major change since the last data release for the PPTA was the introduction of the UWL (installed on Parkes at the end of 2018) - an advantage that the PPTA team has relative to other IPTA teams. This has resulted in an increase in the sensitivity to each pulsar, an improved quantification of the dispersion measure for each MSP as each of their signals propagates through the interstellar medium, and better measures of the variability of pulse profiles across the wider bandwidth.  

Additionally, since the prior data release, the PPTA has increased the number of MSPs in the team’s observing program, including an additional seven MSPs after reviewing three years of UWL data collected for them and determining that they would be suitable for inclusion in the GWB analysis. 

“There are lots of ingredients that contribute to the sensitivity of a pulsar timing array,” said Dr Reardon. “The number of pulsars is one of them because the significance of the fingerprint of gravitational waves depends on the number of unique pairs of pulsars that can be made.”

“The real power of the Parkes Pulsar Timing Array data set will show through when we combine our data with the collaborations in the northern hemisphere. This way, we form many pairs of pulsars that cover both the northern and southern skies.” 

Results from Global Pulsar Timing Arrays

Artist impression of supermassive black hole binaries distorting space-time to create ripples that are affecting pulsar signals as seen from Earth. Credit: M. Kramer / MPIfR / EPTA.

The PPTA is not alone in making this announcement, with global PTA teams in North America as well as the combined Europe and India team each publishing three papers (GWB analysis, noise analysis and data release) unveiling a total of nine papers across the IPTA.  

In NANOGrav’s data set, 67 MSPs were used across a total of 15 years of timespan, with the North American team confirming multiple lines of evidence for the stochastic GWB signal following the HD correlation. The EPTA (which uses telescopes located around Europe) and InPTA (using the uGMRT - a telescope in India) combined their data sets (with a maximum) timespan of 24 years reported in their papers, including a total of 35 MSPs in their analysis. They also found marginal evidence for the GWB signal.

Commenting on some of the key differences between the data sets from the IPTA teams vs. that of the PPTA, Dr Reardon said that the Australian team has the longest time span of data that covers a wide range of frequencies. 

“This allowed us to accurately explore how the signal changes with time. We found that it appears to be getting stronger with time, and we don’t yet know why.”

Whilst not part of the IPTA, scientists from the pulsar timing team in China - which uses the Five-hundred-metre Aperture Spherical Telescope (FAST - the largest telescope in the world) have also confirmed the evidence of the HD correlation. 

An New Era of Gravitational Wave Astronomy

Gravitational waves detected by Pulsar Timing Arrays are lower frequency than those detected by LIGO/VIRGO/KAGRA. Credit: Shanika Galaudage.

Now that each of the teams from the IPTA has found the first evidence of the HD correlation and verified each other's work, the global collaboration is working to bring their data together, increasing sky coverage and the amount of MSPs into a new combined, IPTA-level data release.

“The next step is to combine pulsar data sets collected by telescopes in both the northern and southern hemispheres to improve the sensitivity of our observations,” said Dr Zic. “[This] will combine the full power of the global array, and rule out any anomalies.” 

By combining these data, better characterisation of the signal can be made, and a more appropriate determination of its origin can be considered - confirming if it is being (mostly) generated by supermassive black hole binaries, or potentially, other cosmological sources. These can include exotic phenomena, such as GWs generated by inflation and phase transitions that occurred in some of the earliest epochs of the Universe’s history, or cosmic strings as they radiate GWs. All of these are expected to produce nanohertz-regime GWs, which will need to be decoupled from each other to make an accurate determination of the source that is generating them. 

For the PPTA, the combination of data can also help provide answers to some of the open questions raised as part of this recent analysis, including the non-stationarity of the amplitude of GWB signal detected.

“The IPTA is many times more sensitive than the PPTA alone, thanks to its collection of 115 pulsars spread across the whole sky,” said Dr Reardon. “If the signal we observe is genuinely gravitational waves, the IPTA will confirm a detection. Its sensitivity also means that it should easily be able to confirm or refute our observation that the signal appears to be getting stronger with time.”

Obviously, whilst this new evidence is encouraging, there are still many open questions that will require longer timespans of data, the introduction of new MSPs into observing programs and rigorous combination and analysis across the IPTA. Similarities can be drawn between this new evidence of the GWB, with the discovery of the Cosmic Microwave Background (CMB) - the echo of the last scattering from the primordial fires of the Big Bang. When the CMB was first announced, the exciting results spread quickly across the science and general communities, but as time went on and further analysis was conducted (with new instruments) - and significant findings, such as anisotropies were found - giving rise to new understandings of how the first stars and galaxies formed and evolved through to today. 

Similar anisotropies could be expected in future analysis of longer timespan data sets of the GWB signals, created by individual loud supermassive black hole binaries or from a highly eccentric ‘nearby’ supermassive black hole binary source. Furthermore, there could be a range of exciting new unknown science that we are yet to explore, as we open up this new window into the Universe. 

“A new era of nanohertz gravitational-wave astronomy could be just around the corner,” said Dr Reardon. “We expect to first resolve the background hum of every binary supermassive black hole in the Universe. As the signal strength improves, we should learn more about the population of these behemoths and their environments at the centers of galaxies.”

“An exciting future prospect is being able to tune in to the low-frequency rumbles of individual pairs of supermassive black holes, pinpointed to individual galaxies. These black holes are the engines of galaxies and are important for regulating their star formation – they’re an essential part of our understanding of the formation of our Universe.”