LÉO RAMOSOn the morning of February 11, 2016, Argentine physicist Gabriela González — a friendly scientist from Córdoba living in the United States for almost three decades — had an important meeting at the National Press Club in Washington. The physics and astronomy professor from Louisiana State University and spokesperson for the scientific team at the Laser Interferometer Gravitational-Wave Observatory (LIGO) was chosen, together with four other well-known researchers, to make a highly anticipated announcement: for the first time ever, the passage of gravitational waves through Earth had been detected. It was González’s task to describe the discovery, made 100 years after Albert Einstein predicted the existence of this space-deforming phenomenon in his theory of general relativity.
Separated by 3,000 kilometers (km), the twin detectors are part of LIGO’s second generation. Located in Hanford, Washington State, and Livingston, Louisiana, both consist of L-shaped arms 4 km in length containing mirrors and lasers. They recorded — almost simultaneously — the waves generated by the collision and fusion of two black holes located 1.3 billion light-years from Earth. The detection, which occurred on September 14, 2015, is further strong evidence of black holes, from which nothing — not even light — escapes (see Pesquisa FAPESP Issue nº 241).
González was in São Paulo in early November 2016, where she gave a talk on the recording of gravitational waves produced by the fusion of black holes during the symposium held to celebrate the five-year anniversary of the International Center for Theoretical Physics/South-American Institute for Fundamental Research (ICTP-SAIFR). The center is a collaboration between the Institute of Theoretical Physics at São Paulo State University (IFT-Unesp), where it is based, the Abdus Salam International Center for Theoretical Physics (ICTP) in Trieste, Italy, and FAPESP.
“We are facing a new astronomy, that of gravitational waves,” says the 51-year-old researcher, married to fellow Argentine Jorge Pullin, also a professor of theoretical physics at Louisiana State University. “We are discovering the dark side of the Universe.” In this interview, given while visiting the city of São Paulo, González talks about the importance of detecting gravitational waves, comments on her administrative work at LIGO, a project that has cost $1.1 billion over three decades, and envisions the future of this new way of observing the Universe.
|Detection of gravitational waves|
|Undergraduate degree in physics from the National University of Córdoba (Argentina) in 1988. Doctorate in physics from Syracuse University (United States) from 1989 to 1995|
|Louisiana State University|
|114 articles. Supervised one master’s student, six doctoral students and 10 postdoctoral researchers|
Gravitational waves were detected shortly after LIGO’s second generation of detectors came on line. Did you suspect that it might be a false positive result?
In fact, we had been recording data for several weeks, but in a diagnostic mode. We were calibrating the detectors to make them more sensitive. You have to push the mirrors, simulating signals at different frequencies. We were trying to understand how the measurements were affected by external, seismic or acoustic noise [interference]. A few hours before we detected the waves, there were people in Livingston’s lab checking to see if a car braking near the building could affect the measurement.
And can it?
Yes, if the car brakes strongly, but the interference does not occur in the frequency band we measure, between 20 hertz and 5 kilohertz. Airplanes also affect measurements, but to a lesser extent and also at other frequencies. Seismic noise at low frequencies is quite large and we look for ways to neutralize it. The way in which we push the mirrors, especially to align them, can introduce noise. Quantum noise from light is another problem to overcome. The laser is made of photons. We count photons in the photocell and there is quantum uncertainty about how many photons we detect. This noise is well known, and is called shot noise. We counteract shot noise by increasing the laser power, or in other words by using many photons. Another noise that limits us is Brownian motion, which is random and produced by atoms that move and vibrate when the temperature is not absolute zero.
Did you soon convince yourself that the signal detected in September 2016 represented gravitational waves?
Not immediately. First, we knew that it was not a false signal, created by the team performing the experiment. We had not yet learned how to create a false signal with the second generation LIGO. Still, I spent hours calling my colleagues to make sure it was not a test. They might have forgotten to note down some test. The signal was too obvious, too good to be true. We had to make sure to eliminate all other alternatives. It took months to be absolutely sure. We had to fine-tune and review all the code and software to record data, something we had not yet had time to do. It took a long time to analyze all the magnetometer, seismograph and cosmic ray recordings. We had to make sure that nothing had happened that could have caused the signal. We found several occurrences that had to be investigated thoroughly.
What, for example?
We discovered information about an electrical storm in Africa that produced one of the most powerful lightning flashes ever registered with modern instruments. This occurred seconds before the detection of gravitational waves. But, through our analyses, we saw that the lightning was not the cause of the signal we had detected. In parallel, it also took time to interpret the signal of the gravitational waves, to determine the mass of the black holes and their energy. We didn’t want to have any doubts. We had decided, before even starting to collect data with the second generation of detectors, that we would write an article and send it to the reviewers when we had a possible recording of the waves. We needed feedback from people outside the project. All of us, more than 1,000 people, all wanted the experiment to work.
But, before you announced the discovery, there was a second and third detection of gravitational waves. Did this happen before LIGO’s paper was accepted?
In fact, from the beginning, we knew that we didn’t have enough data to exclude the possibility that the September signal was a random statistical fluctuation. We needed to record at least one month of LIGO data, 15 days for each detector. But the signal was so strong that we all thought it was difficult to extrapolate and know how many more might occur. Then, on October 12, there was a new, much weaker signal. It was not confirmation, and it did not reassure us. On December 26 there was another signal. Not as strong as the first, but it was statistically very significant. That was the signal that convinced us. It was from another system of black holes, different from what we had seen previously. This signal made us less nervous, although it took a few more months to confirm it.
But when you made the announcement of the discovery of gravitational waves in February 2016, you only spoke of the first signal, right?
Yes. Although we had already recorded the third signal, we were still analyzing the data. We were unsure as to its nature.
If, after the first detection, you had not registered any more, would you have announced the discovery anyway?
Yes. Even if we had not registered the occurrence in October or in December, we would have had no reason not to announce the discovery. All the evidence we had was that the signal was caused by gravitational waves produced by black holes. It would have been a little uncomfortable if we had only registered one signal, but we would have announced it anyway.
In the three cases, was the source of gravitational waves located in different places in the Cosmos?
Yes, but the three signals are consistent with the hypothesis that they were produced by a system with two black holes.
How do you know this?
The first signal, in September, was the clearest. Any binary system with a pair of rotating masses — neutron stars, white dwarfs, black holes or the Earth and the Sun — produces oscillatory gravitational waves. As the stars draw closer to each other, they spin faster. The period decreases, but the frequency and amplitude increase. When they are close enough, they merge into one object. In this case, we would expect to see an oscillation that grows and then rapidly decreases. We have a database with hundreds of thousands of signal models. For each pair with a certain mass, there is a distinct frequency and a distinct signal. We correlate the model with the data from each detector as a function of time. If we record a signal with one detector, we see if there is a signal consistent with the same model, with a difference of no more than 10 milliseconds (which is the speed of light between the two detectors), recorded by the other detector. If so, they coincide. We found about 1,000 coincidences during the first months in which we recorded data with the second generation LIGO detectors. Most of them are not astrophysics coincidences. They are random fluctuations produced by a detector. The two strongest detections, the ones from September and December, are very significant. The possibility that they were produced by the fluctuation of a detector is almost zero. We estimated that a random coincidence of this magnitude could happen once every 200,000 years. The October signal was much less significant. In that case, we think that a coincidence of that type could occur every few years without being of astrophysical origin.
How do you differentiate a signal from a pair of black holes from another from a system of neutron stars?
By the mass. Neutron stars are smaller, and usually have a mass of about one solar mass. At most, we estimate that they have less than three solar masses. All of the systems associated with the signals we have seen are much smaller. The smallest of them has eight solar masses. Thus we concluded that they must be gravitational waves created by the fusion of black holes.
How do you avoid information leaks in an experiment involving so many researchers? In January 2016, a researcher outside LIGO tweeted that you had recorded gravitational waves.
We all have to follow a confidentiality agreement until an announcement is made. All members of the experiment team were informed of the progress of the analyses. Only on January 21 did we feel confident enough to send the article for publication. In December, we were still investigating the nature of the lightning strike in Africa. But people talk to colleagues. Many people knew that we had found something, but everyone was careful, because it needed to be confirmed. It would be shameful to tell a journalist that we had discovered something and later have to tell the public we were wrong. That tweet was very unfortunate. We told reporters that there was nothing to report and that we were still analyzing data — and we were.
What happens when LIGO’s detectors record a signal that might be gravitational waves? Does a red light appear, or a computer message?
We have computer programs that do fast, superficial analyses of the data to detect signals that could come from astrophysical events. This is all electronic. Once we have a candidate signal, an alert appears on an Internet page within minutes. But we were not ready for the detection on September 14th.
Didn’t you have an internal procedure to inject false data, in order to see if LIGO researchers could differentiate a planted signal from a real one?
Yes, we had had a team doing that since the first generation of LIGO. Some projects use this type of method to prevent the team from having an attitude unconsciously committed to the desired result. In our case, we felt we were too accustomed to null results and we were not prepared for a positive result. Thus, to support this psychological and scientific process — after all, science is done by humans — we sent false signals in 2007 and again in 2010, when we worked with the first generation of detectors. In 2007 we sent two signals: one marginal signal which was discussed thoroughly by our team, and a second that we simply didn’t see, that was not noticed.
Was the low sensitivity of the first LIGO phase the reason why it was not detected?
No, because we had adopted a procedure in which we first looked at the higher quality data and, later, examined the rest. But, in 2007, we forgot to look at the rest of the data. That was proof that we were not ready. Then, in 2010 there was another injection of false data, which became known as Big Dog. We later discovered that it was a false signal, but we were not prepared to interpret it.
You even did a paper on Big Dog, right?
Yes, but this procedure was foreseen in the plan. We had agreed to not ask whether the signal was true or false and write an article analyzing the data. This process took up much more of our time than expected. The second time, we were prepared to discover the signal, but not to interpret it, to measure its parameters and say how large the black holes or stars responsible for the signal were.
After recording three gravitational waves in 2016, have there been new detections?
LIGO has not been performing detections since January 2016. We should start collecting data again in November 2016, probably with a sensitivity similar to what we had before, or perhaps about 10% better. We have not made progress in this respect as quickly as we would have liked. The second generation Virgo detectors [a gravitational wave observatory near Pisa, Italy] are set to begin operating in the first half of 2017. So, sometime in 2017 we will have three detectors operating.
With the discovery of gravitational waves, does knowledge of the Cosmos change?
I have the impression that this event was like when Galileo observed the sky with a telescope and discovered that Saturn had rings. It’s something like that. We now have an instrument that can view the sky in a completely different way. We can see black holes that do not emit light. We are discovering the dark side of the Universe. I can also say that we are listening to the Universe. The frequency band we work with is audible. We see the Universe with electromagnetic waves. Now we can hear it, too, with gravitational waves. It is a revolution in astronomy. The black holes of stars that we find are very large, about 30 solar masses. There are huge black holes, but in the center of the galaxies. We had believed that stellar black holes would be a few solar masses in size. There may be unknown phenomena that emit no light, but emit gravitational waves.
Recently, an agreement was signed for the construction of a third LIGO observatory in India. When will it begin operating?
Initially, the United States had approved building two detectors in Hanford, and one in Livingston. Then came the idea of having a third observatory in a different location, first in Australia and then in India. This detector already exists and will be installed in the observatory that will be built in India. We believe that it will only start recording data in 2024. Before LIGO India, an observatory in Japan called Kagra should begin operating in 2019 and will cooperate with us. We have been cooperating with Virgo, in Italy, since the first generation of detectors. They also wrote an article on the discovery of gravitational waves with us in 2016. So, at some point, we will have five gravitational wave observatories.
How did you start working on LIGO?
In 1989 I finished my undergraduate degree in physics in Argentina and went to the United States to do my doctorate. I was studying the theory of relativity. A professor who had just gone to Syracuse University, Peter Saulson, talked about how one could accurately measure the fluctuations of space-time. I was not an experimental physicist, but it enchanted me. At that time, LIGO was a project for the future. We knew it would take years to get it running. I began to work on LIGO in 1992, when the proposal was approved by the United States., Construction began in 1995. In 1999, the detectors were ready. In 2001, we began to record the first data. During this time, I worked on different aspects of LIGO.
When you were still a theoretical physicist, what aspect of Einstein’s theory did you work with?
I looked for solutions to the equations. When I began working on LIGO, my dissertation was on the Brownian motion of pendulums and how it effects different frequencies. The important thing was to know how Brownian motion, among all the noise, affected the band of frequencies in which LIGO collects data. What LIGO measures is the distances between the mirrors of the detectors. But the mirrors are made of atoms and are connected by fibers, which are made of atoms. They all vibrate. The difficult problem in my dissertation was to measure how much of this noise appeared in LIGO’s frequency band. After finishing my doctorate in 1995, I went to MIT [Massachusetts Institute of Technology] to work with Professor Rainer Weiss in the LIGO group. There, I worked on designing the pendulums for the first generation of detectors.
Did you expect to win the Nobel Prize for Physics for the discovery of gravitational waves this year?
No. In fact, we do not know exactly what the selection process is like, what the criteria are. The prize for the Higgs boson went to the authors of the theoretical paper, not those who discovered the particle. So, if Einstein were alive, he could have received the Nobel prize again for having predicted gravitational waves. We were not disappointed that we did not win. But, of course, I would be delighted if some of the people working in this field for 40 years were awarded the prize.
This is the third time you have been spokesperson for LIGO. Why did you apply for the position?
It is an elected position. There has always been more than one candidate. I have been leading the scientific collaboration since 2011, when we had finished recording data using the first generation of detectors, but still hadn’t analyzed it fully. It seemed important to me that the collaboration did not dissolve. During this last period in which I have been spokesperson, my priority was to prepare ourselves to record data in September 2016, which we did, and reorganize ourselves to work more efficiently. We are too big. Reorganization, however, was not possible. We spent a lot of time preparing to collect data and, when we began, we detected waves right away, which took even more time. My term ends in March 2017.
Can Brazil participate more in LIGO?
Yes. In addition to taking part in current experiments and data analysis, I think that Brazil could contribute to the development of a third generation of detectors, which could register much more distant black holes, that is, much older ones that appeared at the beginning of the Universe. Thus, we could learn what black holes have been like over time. This project would be very expensive and would probably require even more international collaboration. This seems to me to be an opportunity for Latin America. Why couldn’t we install a detector here? We could install one in the United States and another here. I like to dream about this.
When would this third generation of detectors come online?
These projects take decades. It takes about 10 years to put together the partnerships and the financial part, and another 10 to build the devices. LIGO was first proposed in the 1970s. I think that it would not take 40 years today, but it would take at least 20. A project like this involves not only astrophysics researchers, but also engineers to develop technologies for a new generation of detectors. It’s something that generates a lot of human resources for science and technology. For this third generation we would need new detectors. We cannot update the current ones. These detectors would probably be larger, which would make them more sensitive. Instead of 3 km or 4 km long, they would measure 10 km or 40 km.
What kind of structure could house a detector like that?
It depends on whether it is on the surface or underground. In the latter case, it would not need to be too big, but it would probably be more expensive. For a surface detector, the ideal is to find a flat or easily leveled location that has low seismicity. We are not worried about earthquakes because they affect observatory measurements regardless of their location. We record earthquakes in China or in the Indian Ocean in observatories in the United States. Now that we have built two observatories, we have found that it is better to build them on firm, rocky ground that transmits the movements of the Earth to a lesser extent. The zone near an ocean has waves striking the coast and produces microseismic noise. When you are near the coast and the soil is not rocky, the amplitude of this noise is greater. It’s also important to be close to a center with infrastructure, Internet and human resources.
Which Latin American countries are collaborating with LIGO?
At the moment, only Brazil. There are two groups collaborating with LIGO. An experimental group from INPE, in São Paulo, headed by Odylio Aguiar, is looking for other ways to detect gravitational waves, using a spherical detector. They have a lot of experience with how to suspend objects with low noise and how to work at low temperatures. This is the kind of experience they bring to LIGO. The other group, led by Riccardo Sturani, who was in São Paulo [at IFT-Unesp] and moved to the International Institute of Physics in Natal, does data analysis searching for binary systems. In Latin America, there are a lot of people researching the theory of relativity and gravitational waves, but not doing data analysis. I hope this changes.