The direct detection of gravitational waves by the LIGO Scientific Collaboration has opened a new window of observation on the universe. We can now explore the universe with light across the entire electromagnetic spectrum; with cosmic rays and neutrinos; and, as of last September, with gravitational waves as well.

Like light, however, gravitational waves span a broad range of frequencies, and just as you cannot use an infrared telescope to observe ultraviolet light or X-rays, you cannot use LIGO to look for every kind of gravitational wave. LIGO’s detectors, with arms four km. long, saw (or rather, heard) waves generated by two merging black holes, each approximately 30 times as massive as the Sun. But for the very low frequency gravitational waves created by the merger of the million- to billion-solar-mass black holes at the centers of most galaxies, you need a detector as big as the Milky Way.

Fortunately, we have just such a detector, known as a pulsar-timing array (PTA). As the name implies, PTA’s use signals from the fast-spinning neutron stars known as millisecond pulsars, which emit precisely timed radio pulses thousands of times every second.

PTA’s take advantage of the regular arrival times of radio pulses from millisecond pulsars to search for gravitational waves.  When two supermassive black holes coalesce into one, the mergers bathe the universe in low-frequency waves, stretching and squashing the fabric of spacetime. The pulsars and the Earth behave like buoys on the surface of a choppy spacetime sea, bobbing up and down as the waves pass by. This causes changes in the timing of pulsar pulses that can be detected in carefully designed experiments here on Earth.

The major goal of PTA's is to measure the amplitude of waves that bathe the universe in a constant background of gravitational radiation from the entire cosmic history of supermassive black hole mergers. The amplitude can in turn inform us of galactic demographics, and of the history of how galaxies were assembled over cosmic time. There are three PTA’s on different continents: NANOGrav in North America, the European PTA and the Parkes PTA in Australia. Together they form the International PTA (IPTA), which allows the individual PTA’s to come together and share their data.

Even when we don’t make a detection, we set a bound on the amplitude of the background, called an upper limit. We know that the true amplitude of the gravitational wave background must be less than this upper limit, or we would have detected it.

In September 2015, the Parkes PTA reported a new stringent upper limit, or non-detection, of the gravitational wave background. This result was widely publicized as a case of missing gravitational waves as predicted by the simplest models for the gravitational wave background. Nature is likely much more complicated than our simplest models, however, with gas and stars interacting with the supermassive black hole binaries, diminishing the signal we are looking for.

It is also possible that some of the supermassive black holes binaries form a stable binary and never merge. This is commonly referred to as the “final parsec problem,” which refers to our lack of understanding of what happens fluidly, from beginning to end, when supermassive black holes come very close together. Indeed, these effects, and more, have been an active area of research for several years and thus the current non-detections, while disappointing, are not entirely surprising1.

Moreover, we must draw a distinction between analyses carried out to set upper limits, and analyses that aim to make a direct detection. While we may use a few very good pulsars to set upper limits, a new study shows that it is very unlikely that these same pulsars will yield a detection2. In fact, in this study we find that the path to detection lies in timing a large array of average pulsars, rather than a small number of exquisitely timed pulsars. Hence the IPTA collaboration is crucial for detecting low-frequency gravitational waves, and is likely to do so within the next 10 years.

Even in the absence of a direct detection of the gravitational wave background, we can do interesting and meaningful science with upper limits. For example, the gas and stars interacting with supermassive black holes affects the shape of the gravitational wave spectrum.  By placing constraints on the shape of this spectrum, we can thereby gain insights into supermassive black hole binary environments.

One can also look for gravitational wave hotspots using methods borrowed from cosmic microwave background analyses, as well as individual supermassive black hole binary systems that are loud enough to rise above the background radiation. PTA’s are also sensitive to the theory of gravity that is used to calculate the distinctive signature in the pulsar arrival times, thus providing a novel way of testing Einstein’s theory of General Relativity.

While supermassive black hole binaries are the primary source of very low frequency gravitational waves, other cosmological sources could generate these, such as cosmic strings. These hypothetical constructs would be extremely thin and dense, having likely formed in the early universe. When the strings interact, they can form loops that decay by radiating gravitational waves. The new NANOGrav limit on cosmic string tension is already a factor of four better than other cosmological experiments, such as the Planck satellite, and will continue to increase in sensitivity with time3.

With the IPTA, a direct detection of the low frequency gravitational wave background is likely within 10 years. With this detection, we will probe supermassive black hole environments, gain insights into the cosmic merger history of galaxies, probe exotic physics, and even test General Relativity. Indeed, low frequency gravitational waves will offer a new avenue to explore the universe, not accessible by any other means.


1. The NANOGrav Collaboration, arXiv:1602.06301 

2. Taylor, Vallisneri, Ellis, Mingarelli, Lazio, van Haasteren (2016), ApJL, 819, L6

3. The NANOGrav Collaboration, arXiv:1508.03024