Testing Einstein (Again)

Arthur Fisher. Natural History. Volume 114, Issue 2. March 2005.

Ptolemy made a universe, which lasted fourteen hundred years. Newton also made a universe, which has lasted three hundred years. Einstein has made a universe, and I can’t tell you how long that will last.

~ George Bernard Shaw (1930)

On April 20, 2004, at 9:57:24 A.M. local time-meeting a one-second launch window-a Delta II rocket rose from Vandenberg Air Force Base in California, bearing a payload with the less-than-evocative name Gravity Probe B, or GP-B. Launched into a 97.5-minute, pole-crossing orbit 401 miles above the Earth was a three-ton satellite designed to accomplish one of the most technologically challenging experiments in the history of physics. The launch, which proceeded flawlessly, was the culmination of forty-five years of effort by hundreds of physicists and engineers from Stanford University, Lockheed Martin in Sunnyvale, California, and NASA’s Marshall Space Flight Center in Huntsville, Alabama. Its purpose: to perform two new tests of Einstein’s general theory of relativity.

Publicly presented in 1915, the general theory of relativity was a spectacular intellectual feat. It interprets gravity not as a force but as a field distorting space and time. It holds that massive objects such as planets follow geodesies-paths that act as straight lines, or shortest “distances”-through a curved, four-dimensional generalization of geometry called space-time, which encompasses both space and time. As the physicist John Archibald Wheeler of Princeton University once wrote: “Spacetime grips mass, telling it how to move … Mass grips spacetime, telling it how to curve.”

“I do not consider the main significance of the general theory of relativity to be the prediction of some tiny observable effect,” Einstein remarked in 1930, “but rather the simplicity of its foundations and its consistency.” One reason the theory gained rapid acceptance, however, is that it passed a couple of experimental tests with flying colors, proving itself superior to existing theories of the universe. No one doubts, by now, that it is a powerful theory. But undertaking new tests of its predictions is no idle exercise. Even if all they do is confirm Einstein’s views, they could save physicists from exploring some theoretical blind alleys.

Long before Gravity Probe B bounced into orbit, the idea behind it was bounced around in a conversation among three naked professors sunbathing at a males-only Stanford University swimming pool. Leonard I. Schiff, executive head of the physics department, had been working out a way to use gyroscopes to test two obscure, minute effects predicted by the general theory of relativity. The experiment he had in mind would have to be performed in space because there the gyroscopes are weightless and better isolated from extraneous disturbances. He and his companions, William M. Fairbank, an authority on superconductivity, and Robert H. Cannon, an expert in gyroscopes, met at the pool in December 1959, to talk over the idea and its engineering difficulties. The three later learned that George E. Pugh, a Department of Defense scientist, had articulated the same idea independently in a November 1959 memo.

Nursed along for four decades, the project has cost nearly $700 million and has followed a turbulent path, suffering technical setbacks, suspensions, threatened cancellations, and intense political and academic criticism. In the intervening years, both Schiff and Fairbank have died, and Cannon is only indirectly involved. The principal investigator is C.W. Francis Everitt, who signed on in 1962, when he was just a twenty-eight-year-old postdoc from England. Now seventy, with a cascade of long gray locks, Everitt says, “I would never have joined GP-B if I’d had any idea of how long it would take.”

The eminent English physicist Patrick M. S. Blackett, with whom Everitt had studied in London, once told him, “If you don’t know what kind of physics you want to do, invent some new technology. It will always lead to new physics.” Inventing the technology required for Gravity Probe B has turned out to be a supreme test for the ingenuity of the project’s scientists and engineers. Happily, the technology is now doing its job, collecting data methodically and with exquisite sensitivity. Will it lead to new physics? If the experiment reveals that Einstein’s calculations were even slightly off, the quest for a more exact theory could transform human understanding of the physical universe.

Einstein formulated two theories that interweave space and time. His special theory of relativity, published a hundred years ago this year, describes the behavior of objects moving close to the velocity of light; it also predicts the equivalence of mass and energy, according to the famous equation E=mc2. The special theory has been confirmed repeatedly, in the innards of particle accelerators, in nuclear power plants, and, of course, in atomic weapons. But it was Einstein’s second theory, the 1915 general theory of relativity, that was first put to a public test, making the name “Einstein” a household word.

In spite of Einstein’s insistence that the importance of his theory was its conceptual simplicity, he was keenly aware that every scientific theory must make observable predictions. In 1915, however, most conceivable tests of general relativity were too subtle to be practical. One experiment was technically feasible. The theory predicted that starlight would be deflected when it passed close to the Sun on its way to Earth. Isaac Newton’s classical theory of gravity, together with his particle theory of light, also predicted a deflection, but the magnitude of the effect Einstein calculated was readily distinguishable from Newton’s.

How could star positions be measured accurately, however, when the stars appeared close to the Sun? Answer: observe them during a total solar eclipse. Astronomers knew an eclipse would take place in 1919, presenting a golden opportunity for a test. Photographs of stars near the edge of the Sun, made during the eclipse, could be compared with photographs of the same stars made when the Sun was absent from that region of the sky.

The eminent English astronomer Arthur Eddington led an expedition to an island off the west coast of equatorial Africa, where the eclipse viewing was predicted to be excellent. On the day of the eclipse, May 29, there was a tremendous rainstorm, and Eddington was in despair. But the storm lifted before totality, and the pinpoints of five stars were photographed. When the eclipse photographs were compared with a photographic plate taken at the University of Oxford before the expedition, Einstein’s ideas were confirmed, at least as well as they could be at the time. On November 7 the London Times headlined: “Revolution in Science/New Theory of the Universe/Newtonian Ideas Overthrown.” And on November 10 The New York Times proclaimed: “Lights All Askew in the Heavens/Men of Science More or Less Agog over Results of Eclipse Observations/Einstein Theory Triumphs.”

Even before Eddington s dramatic announcement, Einstein himself had pointed out that his theory could solve a longstanding problem in astronomy. Like the other planets in the solar system, Mercury moves in an elliptical orbit around the Sun. Along with that primary motion, Mercury’s perihelion, its closest approach to the Sun along the orbit, gradually migrates as well, drifting around the Sun in the same direction as its primary motion. This wobble, comparable to that of a toy top, is called precession, and is caused by external forces-in the case of Mercury’s orbit, by the gravity of the other planets orbiting the Sun.

In 1859 Urbain-Jean-Joseph Le Verrier, the director of the Paris Observatory, published his observations of an anomaly in Mercury’s orbit. Le Verrier had found that Mercury’s precession is 574 arc seconds per century (an arc second is 1/3,600th of a degree). In other words, the precession would complete a full 360-degree cycle around the Sun in some 2,260 centuries. When Le Verrier applied Newton’s theory to calculate the gravitational effects of the other planets on Mercury, however, he arrived at a precession rate of only 531 arc seconds per century. The discrepancy perplexed astronomers for more than fifty years.

Relief came in 1915, when Einstein was working out ways to test his own theory of gravity. The theory maintained that space-time near the massive Sun would be distorted in previously unforeseen ways. Taking that distortion into account in calculating Mercury’s precession, lo and behold, the results matched the observed amount. “I was beside myself with joyous excitement,” Einstein wrote later.

There things stood, notwithstanding numerous technical advances, until June 18, 1976. On that day, NASA and the Smithsonian Astrophysical Observatory in Cambridge, Massachusetts, confirmed, with high accuracy, Einstein’s equivalence principle-one of the fundamental assumptions of the general theory of relativity. The experiment, known as Gravity Probe A, confirmed that clocks in gravitational fields of differing strengths do not keep the same time. The investigators compared the frequency of an atomic clock that stayed on the ground with that of an identical clock that was lofted 6,200 miles above the Earth, atop a rocket. Consistent with Einstein’s theory, the clock aboard the spacecraft, in the weaker gravitational field, ran slightly faster.

Later in 1976, another space-age test became feasible, when two Viking landers arrived on Mars. As Earth and Mars traced their orbits around the Sun, signals from Earth were bounced off the landers and back to Earth, and each signal’s round trip was precisely timed. When Earth and Mars were on nearly opposite sides of the Sun, the signals were slightly delayed, by an amount predicted by Einstein’s theory. Known as the Shapiro time delay (after the physicist Irwin I. Shapiro of the Harvard-Smithsonian Center for Astrophysics in Cambridge, Massachusetts), the effect is related to the bending of light by the Sun.

Gravity Probe B seeks to test, with very high accuracy, two further, very small effects predicted by the general theory of relativity. The first, known as the geodetic effect, is a consequence of the calculated curvature of space-time near a massive object such as Earth. The Dutch astronomer Willem de Sitter first derived the effect mathematically from Einstein’s theory in 1916, but it was Schiff who extended the theory to the case of a spinning gyroscope. Translated into the behavior of a gyroscope aboard Gravity Probe B, the geodetic effect should be detectable as a rotation of a gyroscope in the plane of its orbit-or to use a nautical term, a deviation in its pitch. Given Earth’s mass and the height of the satellite’s orbit, the gyroscope should rotate slightly more than 6.6 arc seconds per year.

The second effect GP-B is looking for, known as frame-dragging, is an even smaller one. According to Einstein’s theory, a massive rotating body such as the Earth should drag nearby space-time around itself. Imagine a soccer ball spinning in a viscous fluid such as honey. As it spins, the ball pulls the honey around as well, creating a vortex. Similarly, the Earth’s rotation exerts a pull on space-time, which in turn affects any orbiting satellites.

Frame-dragging was derived in one form from the general theory of relativity in 1918 by two Austrian physicists, Josef Lense and Hans Thirring, and thus is also known as the Lense-Thirring effect. Their calculations concerned the effect of frame-dragging on a moon’s orbit. Schiff, instead, calculated the effect of frame-dragging on the axis of a gyroscope. To return to the analogy of the soccer ball spinning in honey, the effect would show up as a gradual change in direction of a pointer immersed in the honey.

In both cases the effect of frame-dragging is very small. For a gyroscope in polar orbit, it works out to be about 0.041 arc second per year. If you walked “up” a slope at that angle, you would have to walk nearly eighty miles to climb an inch. But in 1959 Schiff calculated that an ideally constructed gyroscope would be able to detect not only the geodetic effect but also frame-dragging. Aboard Gravity Probe B frame-dragging should cause a gyroscope to turn, or yaw, in the same direction as Earth’s rotation.

Conducting the experiment, however, requires much more than a near-perfect, drift-free gyroscope. Every force that might affect the gyroscope, except gravity, must be understood and excluded. Minute changes in the gyroscope’s spin angle have to be measured without disturbing the gyroscope. A point of reference is needed, in the form of a bright, nearby star whose motion is known, and an onboard telescope is needed to keep track of the star. Each of those challenges at first seemed insurmountable.

The heart of GP-B is a twenty-one-inch-long block of fused quartz bonded to a quartz telescope [see illustration on pages 52 and 53]. For redundancy, the block of quartz houses four gyroscopes, each a gemlike sphere the size of a Ping-Pong ball, also made of fused quartz. The four gyroscopes are the most perfectly round objects ever manufactured, honed to within forty atomic layers, or 0.6 millionth of an inch from the highest “peak” to the deepest “valley.” Each sphere is coated with a uniform layer of niobium, a metallic element that becomes a superconductor at 9.3 degrees Kelvin (-442.9 degrees Fahrenheit).

To reduce friction on the spheres, they are levitated by voltages applied to saucer-shaped electrodes. Jets of helium gas spin the gyroscopes up to a speed of 5,000 revolutions per minute. Then, to further minimize friction, the gas is evacuated from around the spheres. The gyroscopes are isolated from external magnetic fields by a succession of magnetic shields.

The telescope itself is kept rigorously aimed at the guide star, IM Pegasi, in the constellation Pegasus. Because the satellite loses sight of the star when it passes behind the Earth, the telescope must be able to reliably reacquire, or locate, the star after every orbit.

The entire gyroscope-telescope assembly is maintained at high vacuum within a nine-foot-long chamber, known as the “probe.” The probe in turn is enclosed within a kind of large thermos bottle known as a Dewar. Filled with 645 gallons of liquid helium, the Dewar maintains the probe at a temperature of only 1.8 degrees Kelvin (-456.4 degrees Fahrenheit). The temperature is kept from rising by gradually venting helium from the Dewar.

The problem of measuring changes in the gyroscope’s spin axis without disturbing the gyroscope has a particularly clever solution. Attaching a pointer to it is not feasible. Enter the spherical coating of niobium: when niobium becomes superconducting, and the spherical superconductor spins, it does something very handy. It generates a small magnetic field that is proportional to the rate of spin and precisely aligned with the spin axis. Hence every tiny shift in the gyroscope’s spin axis can be measured by an exquisitely sensitive magnetometer.

Originally, Everitt explains, the gyroscopes were to be arranged so that one pair would detect frame-dragging, and a second pair, whose spin axes would be perpendicular to those of the first pair, would measure the geodetic effect. For technical reasons, that approach was abandoned, and all four gyroscopes were set spinning along the telescope’s line of sight to the reference star. Each gyroscope senses both axis-shifting effects. Clever mathematics then enables the Gravity Probe B team to sift them apart.

For all the advanced technology aboard Gravity Probe B, frame-dragging may have already been observed. Two NASA satellites, LAGEOS (Laser Geodynamics Satellite) I and LAGEOSII, each studded with 426 minute reflectors, have been tracked for years with laser range-finding equipment. In 1998 Ignazio Ciufolini of the University of Lecce, Italy, and Erricos Pavlis of the University of Maryland, in Baltimore, reported an analysis of the satellite orbits over a period of eleven years. The data suggested the orbits have shifted as a result of framedragging, though the measurements had a large margin of error. More recently, Ciufolini and Pavlis have refined their calculations, based on new data on Earth’s gravitational field. They maintain that framedragging has wrenched the LAGEOS satellites out of their expected orbital positions by two meters a year. There have also been hints of frame-dragging in observations of neutron stars and black holes.

Does that finding mean NASA wasted $700 million on Gravity Probe B? Hardly. The mission is expected to measure frame-dragging directly with an accuracy of 1 percent-good enough, for many experts, to prove the physical reality of Einstein’s prediction. For LAGEOS the displacement of two meters per year, expressed as an angle, is just 0.03 arc second (1/120,000th of a degree). It seems at least questionable whether such a vanishingly small signal can be isolated from the overwhelming noise of the panoply of ordinary Newtonian effects, which amount to 120 degrees a year. In addition to measuring frame-dragging, of course, Gravity Probe B is also investigating the geodetic effect.

What about the possibility that Gravity Probe B will show the way to new physics? According to one Nobel laureate, the physicist Chen Ning Yang of Stony Brook University in New York, Einstein’s general theory of relativity is likely to be amended in a way that somehow entangles spin and rotation. “The Stanford experiment is especially interesting in that it focuses on the spin,” he writes. “I would not be surprised at all if it gives a result in disagreement with Einstein’s theory.”

Given such high stakes, Everitt is on guard against any premature release of his own data. The Gravity Probe B team expects the helium on board to be exhausted by the end of July. “Then,” he says, “it will be four to six months after that before we publish any results.” After forty-five years, that might almost seem like rushing into print.