Chicago
ONLY the taxman thinks Gabe Luther is retired. As he bounds down a hallway at
Los Alamos National Laboratory in New Mexico, carefully planting a foot in the
middle of each square of the patterned carpet, he seems a tenth his real age.
鈥淕ravity is an attractive force,鈥� he mutters, 鈥渁nd the thing it attracts most is
办辞辞办蝉.鈥�
Men of Luther鈥檚 66 years usually play golf, and his shorts, T-shirt and
cartoonish railroad engineer鈥檚 cap would seem in line with that. But Luther
plays a different sport. His game involves grappling with gravity, or, more
specifically, trying to pin down the value of the gravitational constant G
that describes its strength. 鈥淏ig G鈥�, as physicists call it, is
the number that keeps the Earth wheeling about the Sun, compresses hydrogen to
fusion temperatures in stars, and pulls basketball ace Michael Jordan back to
Earth.
Advertisement
But while scientists know many fundamental numbers to eight decimal places,
they disagree on G after only three. On the scientific landscape, G
is an eyesore, an embarrassment in an age of precision.
It may seem odd that of the four known fundamental forces, the one that was
discovered first is the one we know least about. Gravity is familiar to any
child who trips and falls. But measuring it is not for the faint of heart. The
gravitational force is so feeble that the attraction between two protons in a
helium atom is some 1040 times weaker than their electrical repulsion. And
gravity is peculiar in other ways as well. Whereas the electromagnetic force
makes two objects attract or repel one another depending on their charges,
gravity makes any two objects attract one another with a force proportional to
G and their masses.
Einstein鈥檚 theory of general relativity explains this relentless tugging
geometrically鈥攅very mass distorts the fabric of space-time around it into
a gradually sloping well. Other masses slide down into this well. It鈥檚 a bit
like a tired couple lying down side by side on a soft mattress, which sags and
pulls them into the centre. For measurement, this means trouble. Walk into a
room, and you subtly distort the curvature of space-time, pulling everything
gently toward you. There鈥檚 no way around it. Paul Heyl, who measured G
in the 1930s at the National Bureau of Standards (NBS) in Gaithersburg,
Maryland, painted two yellow footprints on the ground and planted his feet there
throughout his measurement.
Fifteen years ago, Luther joined the chase for the elusive G and
teamed up at the NBS with physicist William Towler of the University of Virginia
in a determined attempt to improve on the accuracy of Heyl鈥檚 measurement. Theirs
has been the accepted value ever since. Indeed, around Los Alamos Gabe has
become known as The Big G. And Legend of Luther draws letters from researchers
the world over, not to mention a few nutcases who, inspired perhaps by
Einstein鈥檚 ability to divine fundamental physics while toiling in a patent
office, have formulated their very own theories of gravity. Luther is one of a
small and eclectic pack of scientists who continue to pursue a better value of
G.
It is a race tinged with absurdity. For, as Luther himself admits with a
grin, 鈥淢easuring G is of no practical importance whatsoever.鈥� There is
no reason to run. Physicists have so far been unable to relate G to
anything else, so there are no predictions for what it should be. Measuring Big
G is like shooting at a target without ever being able to check if
you鈥檝e hit it. Big G is just one of those curious numbers, such as the
charge of the electron, that the Universe seems to depend on. So there is
something noble about measuring it, like devoting a lifetime to scaling the
tallest mountain without expecting to find anything at the top.
Some theorists, however, speculate that gravity may be the loose thread that
could start all of physics unravelling. If experiments find that G is
changing slowly over time, for example, physicists would have to rethink how
space and time are stitched together in a single fabric. Einstein would groan in
his grave.
What鈥檚 more, his theory of general relativity does not appear to mesh with
the uncertainties of quantum mechanics. A quantum particle can be in several
places at once. Ask a physicist what the gravitational field of a smeared out
particle looks like, and they鈥檒l hand you the chalk鈥� no one knows.
Theorists hope one day soon to come up with a quantum theory of gravity and
sweep all of the known world, all four forces, into a single equation. Then, out
of the electric charge of the electron, say, and the mass of something else,
they will concoct a formula for G. 鈥淚t will be a sorry state of affairs
if there鈥檚 no good measurement to compare it with,鈥� says George Gillies, a
physicist at the University of Virginia, Charlottesville, who serves as a
historian for the field.
So in addition to the kooks, G has also attracted a gaggle of
serious scientists鈥攖he kind who revel in measuring the speed of light to
nine decimal places. But instead of clearing things up, they鈥檝e plunged the
field into confusion.
In 1994, Germany鈥檚 standards lab, the renowned PTB in Braunschweig, weighed
in with a value of G that was higher than the Luther-Towler number. And
it wasn鈥檛 just a bit higher. Scientifically speaking, it was seriously adrift.
But things really came unstuck when the Measurement Standards Laboratory in New
Zealand announced a value that was far below the Luther-Towler figure, and the
University of Wuppertal in Germany reported a value slightly higher than that,
but still off-target by enough to make physicists go pale.
Today, scientists don鈥檛 so much agree on Luther鈥檚 value as cling to it as a
fixture of stability, like a ship鈥檚 mast in a storm. In 1986, when an
international panel of physicists tried to sift through all the experimental
results to decide the 鈥渂est鈥� value for use around the globe, they couldn鈥檛 do
it. 鈥淲e just threw up our hands,鈥� recalls E Richard Cohen, a member of the
panel, who was then at the Rockwell International Science Center in Thousand
Oaks, California.
So measuring G clearly has a few pitfalls. And Luther, like all the
researchers, has a cautionary tale or two to tell. One weekend at the NBS, he
went to measure G but found it had shifted drastically. 鈥淭hat鈥檚 not
really possible,鈥� says Luther, 鈥渟o I looked around and found that a guy two
storeys above had moved a ton of books into his office.鈥� It helps to be either a
hermit or nocturnal鈥攖he best times to work are at night, over the weekend,
or away from civilisation.
Most days, Luther leaves the lab at Los Alamos and drives to a remote bunker
located 20 kilometres away in a patch of desert near a canyon that hides the Rio
Grande River. Over the years, he says, he has 鈥渂egged, borrowed and stolen鈥�
enough equipment to measure G independently. Inside a small garage-like
building, he and a graduate student tinker with and tune a device remarkably
similar to the one pioneered by the British physicist and chemist Henry
Cavendish in 1798. With the help of an ultra-sensitive torsion pendulum, they
measure the minute gravitational attraction between a couple of tungsten spheres
and a tiny barbell.
A torsion pendulum is essentially a thin fibre, often made of quartz or
tungsten, hung vertically with its free end attached to the middle of a barbell
(see Diagram).
With the slightest provocation, the barbell will rotate,
twisting back and forth, about once every six minutes in Luther鈥檚 apparatus. But
when the tungsten spheres move in close, their gravitational attraction reduces
the swing time by a fraction of a second. Luther infers G by measuring
that tiny shift.
Of course, there are subtleties鈥攍ots of them. The torsion pendulum has
to be in a vacuum, as otherwise even a small difference in temperature across
the room would create forces that would alter the pendulum鈥檚 motion. And the
swing time, measured by bouncing a beam of light off a mirror attached to the
fibre and tracking its position, can be disrupted if the mirror isn鈥檛
symmetrical and bounces more photons off one side than the other.
鈥淭he history of this measurement,鈥� Luther says, 鈥渋s that everyone thinks they
know everything they have to worry about, but nobody does.鈥�
Two years ago, Kazuaki Kuroda at the University of Tokyo wrote a paper that
ripped the rug out from under 200 years of measurements. The torsion pendulum
method assumes that the fibre resists twisting with the same strength when the
masses are near, as when they are far away. But Kuroda pointed out that if the
torsion pendulum is swinging faster than its normal frequency (as happens when
the masses are brought near) then it will resist a little more. It鈥檚 like a
stopwatch that runs slightly fast, but only when you鈥檙e trying to time a race.
Without taking this into account, torsion experiments would slightly
overestimate G. This year Luther and a graduate student, Charles
Bagley, did a few tests, found that Kuroda seemed to be right, and had to revise
their measurement.
Over the years, scientists have grown wise to nature鈥檚 mischievous ways,
although that hasn鈥檛 seemed to help much. Tim Armstrong and Mark Fitzgerald,
physicists at the Measurement Standards Laboratory in New Zealand, neatly avoid
the Kuroda problem by using an electrostatic field to hold the torsion pendulum
still, and measure how much it would like to twist, if it were allowed to. The
verdict from the non-swinging pendulum? A value of G lower than the
Luther-Towler value鈥攂y a lot.
Swinging faster
The physicists at PTB in Braunschweig avoided the Kuroda problem by ditching
the fibre, and floating their barbell in mercury instead. This also allowed them
to use a heavier barbell that would feel a stronger gravitational tug and
produce a more easily discernible signal. And after 15 years of work, the group
announced a very precise value indeed. There was just one problem鈥攊t was
vastly larger than the Luther-Towler value. 鈥淲e thought it was wrong,鈥� says
PTB鈥檚 Winfried Michaelis. But after two years of arduous tests without finding a
single error, they published their paper and quit the game to move on to more
reliable quantities.
Breaking tradition, researchers at the University of Wuppertal have done away
with the torsion pendulum altogether. Instead, they dangle two ordinary
pendulums side by side. Then they place four large masses nearby, each weighing
about a ton, which pull the pendulums apart by as much as 20 billionths of a
metre. By bouncing electromagnetic waves between the hanging spheres, they can
measure the distance between them and infer G. They can also detect
earthquakes on the other side of the planet, and the surf on the French coast.
鈥淭here鈥檚 this stream of seismic waves running across the Earth and they come
here and shake our apparatus,鈥� says Hinrich Meyer, who leads the group. 鈥淚t鈥檚 a
big effect in the autumn and in the spring.鈥� The Wuppertal value fell somewhere
between the other measurements, but didn鈥檛 coincide with any of them.
How can so many competent scientists disagree so completely? 鈥淨uite frankly,
no one knows,鈥� confesses Gillies, 鈥淚 doubt it鈥檚 new physics, but it might be new
instrument science.鈥� Perhaps there is some unforeseen factor that, if taken into
account, could bring all the experiments in line. But no one has been able to
dream up such a cure.
Of course, the matter could also be put to rest if the various experiments
each uncovered different twists that made their results less precise than they
had thought. After all, every measurement comes with 鈥渆rror bars鈥濃�攍imits
that are supposed to bracket how far off the experimenters think they can be.
But these are guys who live and die by their error estimates, and they鈥檝e racked
their brains and come up empty. 鈥淢ark and I argue a lot over what to put in,鈥�
says Tim Armstrong. 鈥淲e argue our lives away.鈥�
Most believe that the only hope now lies in new experiments. Fortunately, a
handful are planned or are under way, and some may even have results in time for
the Big G meeting in London next year, commemorating the 200th
anniversary of Cavendish鈥檚 measurement.
Terry Quinn, director of the International Office of Weights and Measures
(BIPM) in Paris, and Clive Speake at the University of Birmingham have measured
G in the past, and they鈥檙e gearing up to do it again, this time
replacing the fibre with a wide strip that can hang a heavier load. Jens
Gundlach, a physicist at the University of Washington, thinks he can eliminate a
number of uncertainties by replacing the barbell with a solid, rectangular
plate. He believes this change makes the experiment immune to many of the issues
that plague other experiments, such as knowing the exact size and density
distribution of the masses. The G controversy has also ensnared
scientists in Russia, Taiwan and Switzerland.
Still further experiments plan to scrap all that has gone before. James
Faller at the University of Colorado has gone back to what could be called the
Galileo method鈥攄ropping an object and seeing how fast it falls. Since the
Earth鈥檚 mass isn鈥檛 known very accurately, Faller takes a half-tonne tungsten
doughnut and compares how fast an object falls above it and below it
(see Diagram).
Above, the gravitational attraction of the doughnut helps pull
the weight toward the Earth, but below it acts to slow it down. Can you really
measure such a tiny difference? 鈥淚t sounds ludicrous,鈥� Faller admits. But he
insists that the technology is strictly off-the-shelf. The machine that drops
the ball, called a gravitometer, has been used for years by companies searching
for oil or mineral deposits to measure the local gravitational attraction of the
Earth.
But while Faller can measure G without the annoyance of an extremely
sensitive pendulum, he won鈥檛 be able to pin down G accurately enough to
make them obsolete. That may have to wait for an experiment led by Alvin
Sanders, a physicist at the University of Tennessee. In perhaps the most
ambitious proposal to date, Sanders鈥檚 group plans to take the whole enterprise
into space. To get away from the nagging noise of the Earth, Sanders and
colleagues are proposing to measure G by watching the paths of two
spheres in orbit about the planet. The spheres, floating serenely in a
temperature-controlled satellite, will allow G to be determined more
accurately than any of the planned pendulum experiments. But moving Big
G from the lab bench into space will also break with the tradition of
tabletop budgets. Sanders says the cost of a launch is currently $100
million.
But for Luther, who has been seen balancing sacks of flour on his head in the
supermarket, gravity is something to be pursued on Earth. Like the others, he鈥檚
devising a new, better way to squeeze out the next decimal place of Big
G. It鈥檚 his way of repaying society and the government for covering his
education all those years ago.
鈥淢ark Twain said you get information by carrying a live cat home by the tail
that you can鈥檛 get any other way,鈥� he says, locking the door to his bunker. 鈥淚
feel like I鈥檝e carried the cat home and I鈥檝e got enough information that it
behoves me to keep going.鈥� Going, yes, but Luther doesn鈥檛 really expect to get
anywhere. The Big G just likes carrying cats.