News FocusPhysics

Gravity Measurements Close in on Big G

Science  18 Dec 1998:
Vol. 282, Issue 5397, pp. 2180-2181
DOI: 10.1126/science.282.5397.2180

The precise strength of this pervasive force has proved surprisingly elusive, but conflicting results are finally giving way to a single answer

Two hundred years ago, in a stone house on the outskirts of London, Henry Cavendish weighed the world. By the light of a candle, he watched as a small lead barbell suspended by a fiber twisted minutely under the gravitational tug of two bowling ball-sized weights. The size of the twist revealed the strength of gravity between the two known masses, the barbell and the balls. Because Cavendish knew how strong Earth's tug was, he could then precisely pin down its mass.

Pulling together.

New readings of G (red) are in rough agreement, hinting at an end to a quest that goes back to Henry Cavendish.

The mass of the Earth is no longer a burning question, but the Cavendish experiment is legendary. The torsion balance is still one of the best ways to measure gravity's strength, a number called the gravitational constant, or Big G. G is perhaps the most elusive of all the fundamental quantities. While the charge of the electron is known to seven decimal places, physicists lose track of G after only the third. For some, that's an embarrassment. “It grates on me like a burr in the saddle,” says Alvin Sanders, a physicist at the University of Virginia in Charlottesville.

Over the past few decades, he and a handful of other physicists have dedicated themselves to measuring G more accurately. To their dismay, they've come up with wildly different values. “You might say we've had negative progress,” says Barry Taylor, a physicist at the National Institute of Standards and Technology (NIST) in Gaithersburg, Maryland. But when 45 members of the Big G community met last month* to honor the Cavendish anniversary and discuss the remarkable lack of progress over the last two centuries, they had a pleasant surprise. Six groups using a variety of techniques weighed in with new values of G, and they were all in rough agreement.

Because the results—one of which is reported on page 2230 of this issue—are preliminary and disagree with some older measurements, the G-men are cautious about declaring the case closed. But many suspect experimental finesse may finally be settling the long debate over the value of G. “The numbers are sort of converging,” says Tim Armstrong, a physicist at New Zealand's Measurement Standards Laboratory (MSL). Finally “a consensus is emerging.”

Heavy lifting. Measuring gravity is harder than it might seem. Its relentless pull is apparent to anyone who has tried to shift a refrigerator or watched a rocket gun for the heavens, but when compared to other forces, gravity is unimaginably feeble. If gravity were as strong as the electric force that pulls oppositely charged things together, a bathroom scale would read out your weight in a number some 40 digits long. To accurately gauge gravity's tiny tug, most experiments have to be carefully isolated from electrical and seismic disturbances and performed in a vacuum to minimize the push of atoms in the air as they bounce off the test objects.

The going is also tough for another reason: No container can shield out the gravitational attraction of other objects. Riley Newman, a physicist at the University of California, Irvine, likes to tell the story of how his group once discovered a peculiar early morning wiggle in gravity's strength. After months of head-scratching, a graduate student leaving the lab at 3 a.m. was literally doused by the answer: The sprinkler system on the surrounding lawn, set on a regular clock, was soaking the ground with enough water to create an additional nearby mass, which skewed the morning readings. Newman says the group's new experiments will be done in an abandoned Nike antiaircraft missile bunker in Washington state, far from the lawns of suburban America.

What's the payoff for all the trouble? Very little. “Nobody gives a damn about Big G,” says Clive Speake, a physicist at the University of Birmingham in England who organized the conference. One day, physicists hope to merge quantum mechanics and gravity, which might allow them to calculate what Big G should be. They could then test the theory by comparing the predicted G to the measured value. But for now, measuring it is a kind of sport—a Mount Everest of precision measurement. “You have to be an oddball to do this,” Speake admits. Cavendish was no exception, he points out. A colleague once described the reclusive physicist as speaking fewer words than a Trappist monk.

The official value for G—chosen by an international panel in 1986—comes from a 1982 measurement by Gabe Luther, now at Los Alamos National Laboratory in New Mexico, and William Towler of the University of Virginia. Their setup was similar to Cavendish's: a tiny barbell hung from a long fiber made of quartz or tungsten. When disturbed, the barbell would rotate lazily back and forth about once every 6 minutes, driven by the fiber's resistance to twisting. When two huge tungsten balls were brought near, their gravitational tug on the barbell slowed the swing time by a split second. By measuring that difference, Luther and Towler pegged Big G with an estimated accuracy of better than a hundredth of a percent.

All was well until 1994, when heavyweights at the German standards lab, the PTB in Braunschweig, announced a value of G that was supposed to be just as accurate. It was, by all reckoning, a tour de force measurement. Instead of suspending test weights from a delicate fiber, they floated them on a layer of mercury. That allowed the researchers to use larger masses and generate a stronger pull. To everyone's shock, their value came out way above the Luther-Towler number—a full half-percent larger. Things got worse in 1995 when New Zealand's MSL came out with a number that significantly undercut the accepted value (see chart).

Some physicists in the standards community took the conflict as a challenge. After all, they'd measured time accurately enough to build clocks that slip by only a second every 100 million years. Not to be bested by gravity, many joined the effort to hunt down Big G. Worried that perhaps the traditional torsion balance might not be up to the task, many tried new techniques.

Weighing in. One of the more outlandish approaches reported at the conference came from a group including James Faller, a physicist at NIST in Boulder and the University of Colorado. Loosely speaking, the team dropped a weight through the hole of a large tungsten donut, then raised the donut above the release point and dropped the weight again. With the donut below, its gravitational tug made the weight fall faster. When the donut was overhead, it slowed the object's descent by a hair. From the tiny difference between the two drop times, the team teased out the value for G they report in this issue.

A group at the University of Zurich also scrapped the torsion idea and suspended small masses, a kilogram each, from a precision scale. By moving other large masses above and below the kilograms, the researchers could change their effective weight and measure G. In work published in a recent Physical Review Letters, the large masses were vats of water. Now they've gotten their hands on a few bathtubs worth of mercury, which pulls on the test masses with a force equivalent to that of 2000 tons of water, making G easier to measure.

Researchers at the University of Wuppertal in Germany are lugging around huge masses as part of a different scheme, which relies on two pendulums hung side by side. When the team wheels in a half-ton mass on either side, the tug of the masses pulls the pendulums apart slightly–“by about the width of an atom,” says team member Hinrich Meyer. By bouncing microwaves between the bobs, he and his colleagues can size up the gap and sift out G.

The torsion balance hasn't disappeared from the scene, however. Clive Speake, working with Terry Quinn, a physicist at the international standards lab, the BIPM near Paris, and others announced a new measurement made with a torsion balance in which the thin fiber was replaced by a broad strip. The strip could support heavier masses, and it eliminated problems that can plague a fiber. When a fiber twists quickly, many of its atoms become temporarily dislocated. This effect changes how strongly the fiber resists twisting, which can throw off the measurement. But a strip resists twisting by wrapping and shortening, which is entirely reversible. A Russian group, led by Oleg Karagioz, also submitted new results using a more traditional torsion balance.

The new results all point to a value of G hovering just above the Luther-Towler number. And one of the old outlying measurements was also brought within the fold at the meeting. The New Zealand group revealed that they had uncovered a simple but grave error in their old measurement. They'd forgotten to include the thickness of the walls of the cylinder they'd hung from their torsion fiber when they calculated the cylinder's moment of inertia, which affects how fast it reacts to the pull of a nearby mass. The slip was “embarrassing,” Armstrong says. But the group has now remeasured G, and “we believe our new numbers are right.”

That leaves one dangling thread—the far-out results of the PTB. “Nobody understands it,” says Meyer. “They must have made an unbelievable mistake, but we cannot find it.” But if the tentative consensus on big G strengthens, says Terry Quinn, “we may just have to throw the PTB result out.” Already physicists are hatching plans for new measurements. “I now own a 1050pound tungsten [donut],” says NIST's Faller. “What else am I supposed to do with it?”

  • * The Gravitational Constant: Theory and Experiment 200 Years After Cavendish,” 23 to 24 November, Institute of Physics, London.


Navigate This Article