Is Einstein Still Right?

by Clifford M. Will

  Thinking that this analysis was not worth publishing as a scientific paper, he wrote it up in a twenty-three page report printed in one of MIT’s quarterly newsletters in 1972. The concepts laid out in his report would become the foundation of LIGO’s design.

  While preparing his MIT report, he requested and obtained funding from MIT to build a small prototype with arms 1.5 meters long, and around 1975 he wrote a proposal to the NSF to continue this work. Despite positive reviews, the proposal was turned down. The proposal came to the attention of Heinz Billing at the Max Planck Institute in Munich. One of the pioneers of computer science, Billing had recently returned to physics, and his group was engaged in using resonant bars to check Weber’s claims. Turned on by Rai’s description of the potential of interferometry for detecting waves, he and his colleagues started to build a prototype. Soon his laboratory was paid a visit by Ron Drever, who also had been working on bar detectors, and now became intrigued by this new approach. Drever (1931–2017) was a brilliant and inventive physicist at the University of Glasgow, who at the age of twenty-nine performed an exquisite experiment using nuclear magnetic resonance techniques to show that the mass of an atom does not depend on its orientation relative to the Galaxy or relative to the Earth’s velocity through the universe. Drever’s group began to build a prototype interferometric detector.

  Kip Thorne at Caltech also began to think about interferometers, spurred by a late-into-the-night discussion with Rai Weiss at a hotel in Washington during a NASA committee meeting. His group had been at the forefront of the theory of gravitational wave sources, but he felt that Caltech should also have a presence in the experimental side, and so he recruited Drever to move to Caltech in 1979, where he began to build a 40 meter prototype.

  It had become clear that tabletop or laboratory-scale prototype interferometers were fine for technology development, but that they would not be sufficiently sensitive to ever detect the kinds of gravitational waves that might reasonably be expected from astrophysical sources. Instead, devices with arms as long as many kilometers would be needed. The reason is that the change in distance between two objects that a gravitational wave induces is proportional to the distance. If you double the distance between the beam splitter and the mirror, you double the displacement, and therefore you double the difference between the light beams when they recombine. Go from a 40 meter prototype to a 4 kilometer system and you increase the effect by 100. Unfortunately, you also increase the cost by a similar factor. One reason is that the light must propagate through an ultra-high vacuum, otherwise the fluctuations in the light speed caused by its interactions with the atoms in the residual gas or air would have a larger effect than the displacements of the mirrors caused by a gravitational wave. There was also general agreement that, just as with Weber’s bars, two widely separated interferometers would be needed in order to claim a credible detection. It was thus also becoming clear that gravitational wave detection would be very costly.

  As a result, at the urging of the NSF, Caltech and MIT agreed in 1984 to cooperate on the design and construction of LIGO, with joint leadership by Weiss, Thorne and Drever. This leadership arrangement proved unworkable, however, and in 1987, astrophysicist and former Caltech Provost Rochus E. Vogt was appointed LIGO director. By 1992, initial funding for construction had been provided by the US Congress, and sites in Hanford, Washington and Livingston, Louisiana had been selected. Barry Barish, a high-energy particle physicist, replaced Vogt as LIGO director in 1994 and oversaw the construction and commissioning of the detectors and the initial gravitational wave searches. The plan for LIGO involved two stages: building and operating the interferometers with proven technology, with a sensitivity where gravitational waves might be detected, and then to upgrade them with advanced technology to a level where waves would be detected, if general relativity and our understanding of astrophysics were correct. Searches for gravitational waves with the initial LIGO were carried out between 2002 and 2010. To nobody’s surprise, no waves were detected. On the other hand, the interferometers reached the sensitivities they were designed to achieve and much experience was gained in operations and data analysis. Between 2010 and 2014, the interferometers were shut down in order to install advanced technology that had been under development, such as more powerful lasers, improved mirrors and better isolation from seismic disturbances. By September 2015, the instruments were as much as ten times more sensitive than in the initial LIGO.

  But the Americans were not the only ones who wanted to detect gravitational waves. Alain Brillet was a French physicist at the National Research Center in Orsay near Paris who had worked with Jon Hall at the University of Colorado in 1979 to do a twentieth-century version of the famous Michelson–Morley experiment, but using a laser as the light source in the interferometer. Adalberto Giazotto was an elementary particle physicist at the National Institute of Nuclear Physics in Pisa, Italy who had taken an interest in gravitational wave detection, and particularly in the problem of seismic isolation of the mirrors. Together they proposed a large European interferometer, which was ultimately built near the town of Cascina, about 15 kilometers south-east of Pisa, and named Virgo after the Virgo cluster of galaxies. Drever’s group in Glasgow and Billing’s group in Munich combined forces to propose a large interferometer. Because of funding limitations, caused in part by the cost of the reunification of Germany, they had to settle for 600 meter arms, compared to the 4 kilometers of LIGO and 3 kilometers for Virgo. That instrument, called GEO-600, was built near Hannover, Germany. Researchers in Australia initially built advanced resonant bar detectors, and then moved into interferometers, but could never convince their government to go beyond an 80 meter prototype called AIGO, sited near Perth in Western Australia. Japanese teams also became very active, and have recently completed an ambitious interferometer, the Kamioka Gravitational Wave Detector (KAGRA), a 3 kilometer instrument built deep inside Mount Ikeno near Hida, Japan, where numerous underground physics experiments studying neutrinos, dark matter and proton decay have been run using inactive shafts and tunnels from the Kamioka mine (see Chapter 9).

  You might be picturing an intense international race and competition to be the first to detect gravitational waves, but in fact quite the opposite happened. Recall Weber’s dictum that you had to have more than one detector to be sure that you have detected gravitational waves. In addition, since the mirrors in the interferometers respond immediately and freely to a passing gravitational wave, by recording the difference in time of arrival of the same signal in two widely separated interferometers, you can get some idea of the direction of the source. The principle is the same as the one we discussed in Chapter 3 (see Figure 3.5), whereby the difference in arrival time of a radio wave at two separated radio telescopes can be used to determine the source direction. Two interferometers give only limited information about the location of the source on the sky; the more interferometers you have, the more accurately you can pinpoint the source. For this to work, different teams have to cooperate, regardless of the desire of any individual or national agency to garner the glory of being “first.” So while LIGO and Virgo were still under construction, the leadership of the two projects began the delicate negotiations that would ultimately lead in 2007 to the LIGO–Virgo Collaboration, a rather remarkable structure that views the two LIGO instruments and Virgo as a single network of three interferometers, with full data sharing and transparency, coordination of schedules, and so on. (For colleagues of ours who are members of the collaboration, it also means an unbelievable number of teleconferences across many continents, resulting often in highly inconvenient work hours!) The GEO-600 and AIGO teams joined the collaboration to work on technology development. Despite its lower sensitivity, GEO-600 also made observations when the LIGO and Virgo instruments were offline, just in case a strong event, such as a nearby supernova, might occur. Virgo also adopted a two-stage development strategy, similar to the initial-to-advanced LIGO track. When the first detection was made on 14 Septem
ber 2015, advanced-Virgo was still about a year away from being up and running, so the signal was only seen by the LIGO interferometers, yet the discovery paper published in 2016 included all the members of Virgo as co-authors. The paper had over a thousand authors. In the next chapter we will see how important this cooperation proved to be, when we discuss what signals were actually detected and what they implied.

  By early 2015 both LIGO interferometers were working, and entered what is called an “engineering run,” during which the operators of the instruments poke and prod them, tweak the dials and alter various settings, all in an effort to get the maximum performance. That run was scheduled to end on 18 September, when all tweaking would cease and an “observing run” would begin. But the engineers completed their work about a week ahead of schedule, and both interferometers were performing quietly, awaiting official kick-off of the observing run. On 14 September at 5:51 a.m., Eastern Daylight Time, the Livingston instrument recorded a signal, and 7 milliseconds later the Hanford detector recorded the same signal. The signal, known thereafter by the name GW150914 (GW for gravitational wave, followed by the date in yymmdd format), arrived close to a hundred years after Einstein published his theory. In the next chapter we will describe the detection and what we learned from it.

  Before we do that, there is one final aspect of gravitational waves that we need to discuss. In various newspaper reports about gravitational waves, you may have encountered the phrase “listening to gravitational waves.” Many popular books on this subject use musical motifs, such as Marcia Bartusiak’s Einstein’s Unfinished Symphony or Janna Levin’s Black Hole Blues; Chapter 9 of this book talks about a “loud” future instead of a bright future. What is that all about? Normally you think about astronomers “gazing” at the heavens, “seeing” a supernova explosion, or “watching” a planet transit in front of the Sun. Why do we “listen” to the universe with gravitational waves?

  The reason has to do with the fundamental difference between electromagnetic waves and gravitational waves. When electromagnetic waves, a.k.a. light, impinge upon some material, such as the retina of your eye, the electric and magnetic fields of the light waves push on the charged electrons in the material and generate an electrical current. If you prefer the quantum mechanical picture in which light consists of “photons,” then the photons knock the electrons loose from their host atoms. The current is then sent from the retina of your eye to the optic nerve and then to your brain. The current could be produced in the CCD device in your camera or smartphone. Or it could be in the conducting antenna of a radio telescope. The act of “seeing” in all its manifestations basically amounts to using light to move electrons and thereby to produce electrical currents.

  Gravitational waves act very differently, causing bits of mass (not charge) to move back and forth relative to each other via the stretching and compressing of spacetime (recall the hockey pucks in Figure 7.3). So when a gravitational wave passes through your head, it causes the eardrum and bones in one ear to move relative to those in the other ear. It also tries to stretch and compress your skull in the same manner, but since your skull is rather rigid, it can resist that effect to a large degree. The elements of your inner ear, on the other hand, are more free to move, and in doing so, they strike the membrane of the cochlea, forcing the fluid inside to move back and forth, triggering a set of hairs that convert the oscillations into electrical impulses that travel through nerves to your brain. The only difference between sound waves and gravitational waves in this regard is that sound waves use the expansion and compression of air to move the eardrum, while gravitational waves use the expansion and contraction of spacetime itself to produce the same effect.

  But we cannot sense all possible vibrations of our eardrums. This is because the conversion of the oscillations into electrical impulses is not efficient enough for frequencies below about 20 hertz (for the lowest-pitched sounds) or above about 20,000 hertz (for the highest-pitched sounds). Other mammals, such as dogs, can hear up to 45,000 hertz, while the greater wax moth can hear sounds up to 300,000 hertz! Dog whistles, in fact, leverage this physics principle: they produce vibrations in the air that are too high a frequency for the human ear to pick up, but that can be easily heard by dogs.

  As we will learn in the next chapter, the waves from GW150914 had a frequency in the region between 40 and 300 hertz, well within the human audible band. So why didn’t people around the world hear the signal? The answer is the incredible weakness of gravitational waves. The minimum eardrum motion that we can detect is about a nanometer, or a millionth of a millimeter. The gravitational waves detected by LIGO moved our eardrums by about a trillionth of a nanometer! Alternatively, we could have heard those waves if the source had been extremely close, about twice as far from the Sun as Neptune, instead of over a billion light years away. We should be happy that this was not the case. Since the source was two black holes each about thirty times the mass of the Sun in a close orbit around each other, their normal gravitational tug on the Sun and planets would have disrupted the Solar System (or prevented its formation in the first place) long before we came along.

  Instead of eardrums, we have the mirrors of LIGO and Virgo. Instead of the tiny bones of the inner ear and the cochlea, we have laser beams bouncing off the mirrors, capable of detecting their motions to better than a thousandth of the diameter of a proton. The laser interferometers of LIGO and Virgo are our tools for listening to gravitational waves. And what have we learned from these sounds?

  CHAPTER 8

  What Do Gravitational Waves Tell Us?

  T minus 2 hours: At the LIGO Hanford observatory in Washington State it is 12:50 a.m., and the last two scientists depart to get a well-deserved night’s sleep after a long day of engineering tests, leaving only the night-duty operator on site. The detectors had been turned on a few months before to calibrate the instruments in preparation for the first science run of the “Advanced LIGO” detectors, which is scheduled to begin in four days.

  Two billion kilometers away, about one and a half times the distance to Saturn, a packet of spacetime ripples approaches the solar system at the speed of light (Figure 8.1).

  Figure 8.1 A packet of gravitational waves approaches the solar system on 14 September 2015. The waves come from below, about 15° from the “south pole” of the solar system. Shown are the positions of the Earth, Mars, Jupiter and Saturn at that time. The waves from this source may have been passing through the solar system for millions of years; we only show the final burst of strong waves, the ones detected by the LIGO instruments.

  T minus 15 minutes: In Livingston, Louisiana it is 4:35 a.m. and two technicians at the LIGO Livingston observatory are about to run one last test before finishing their shift. The “bump test” consists of driving a car at 5 or 10 miles per hour over speed bumps right outside the LIGO building, GPS unit in hand, to check whether the motion of the car over the speed bumps creates ground vibrations strong enough to be detected by the interferometer. Right before starting the test, they realize their GPS unit is malfunctioning and needs to be recalibrated, so instead of running the test they call it a night and go home.

  The packet of spacetime ripples is now 270 million kilometers, or 1.5 astronomical units, away.

  T minus 0 minutes: The packet of spacetime ripples passes through the LIGO Livingston detector, and 7 milliseconds later it passes through the LIGO Hanford detector. For about a fifth of a second the mirrors at opposite ends of the interferometers move back and forth a few times by a tiny amount and the interference patterns of the laser beams inside the instruments change a bit, generating electrical signals that record the passage of the spacetime ripples. The signals are recorded automatically, but nobody notices. Not just yet.

  T plus 10 minutes: It is noon in Hannover, Germany, close to lunch time. Postdoctoral researcher Marco Drago at the Albert Einstein Institute notices that an automated computer program has pinged, announcing the presence of something strange in the LIGO data. He is curi
ous and so he checks the logs to see if there was a scheduled injection, a test to make the mirrors in the instrument vibrate artificially just as if a wave had gone through. But the logs show no scheduled injection.

  Marco goes next door to the office of his friend, Andrew Lundgren, also a postdoctoral researcher. He tells Andrew about the ping and together they investigate it further. Could it have been an injection that was not logged? No, there was truly nothing scheduled. Could it have been a bump test? No, there was nobody testing the interferometers. Could it have been a micro-earthquake or some atmospheric effect? No, the data quality monitors were all showing perfect conditions.

  T plus 54 minutes: Marco sends an email to the entire LIGO scientific collaboration, over 1,000 researchers spread around the world. He describes the event that was recorded and he ends the email asking for confirmation that this was not an artificial injection. A flurry of emails ensues. Within hours, there is confirmation from the LIGO leadership that this event was not an injection or test of any kind.

  T plus 10 hours: The LIGO executive committee gathers via telephone conference call. They discuss the event and make a decision: maintain the instruments in their current state and continue taking data. The software is locked. The hardware is locked. Cabinets housing the electronics are physically locked. For the next two weeks, nobody is to touch either of the two instruments or alter a setting, so that more data can continue to accumulate to be able to compare the event to data that presumably contains only random noise.