On the plains of San Agustin, a vast new mexican valley filled with sagebrush and inhabited mostly by cattle, MIT radio astronomer Jacqueline Hewitt sits in the control room of the Very Large Array, a collection of 27 rail-mounted radio telescopes, each with a dish antenna more than 80 feet across. The antennas are positioned as far apart as possible, forming a Y with the control room at the center and the farthest telescope in each arm some 15 miles away. The dishes are all focused on a single point in the sky. Hewitt is tuning in on a quasar--a pointlike object billions of light-years away that radiates the energy of a million suns, mostly as radio waves. She is waiting for its signal to flicker.
At the four-meter optical telescope atop Cerro Tololo, in the dry mountains of northern Chile, another MIT astronomer, John Tonry, is looking at galaxies, trying to measure the subtle differences in brightness across their surfaces. And at Cerro Tololo’s sister telescope, the four-meter on Kitt Peak in southern Arizona, a team of observers from several institutions is scanning galaxies for planetary nebulas, the bright blobs of gas blown off by giant stars in their death throes.
The observations could hardly be more different, but each project is an attempt to answer two of the most fundamental questions about the universe: How big is it, and how old? The questions are equivalent. In asking how big the universe is, we are talking about the visible universe: the sphere, centered on Earth, whose radius is the distance light could have traveled since the Big Bang, when the universe came into being as a single point and began expanding rapidly outward. The distance to the edge of the visible universe, in light-years, is the same as the number of years since the whole universe began. The universe as a whole may well be bigger than the visible universe; in the first nanosecond of its existence it may have undergone a burst of faster-than-light inflation. But astronomers cannot hope to measure what they cannot see.
And they’ve been remarkably unsuccessful at measuring the size of what they can see. They have uncovered some of the secrets of black holes, of pulsars, and of colliding galaxies; they can follow the birth, life, and death of stars in great detail; they can even tell you what happened in the very first seconds after the Big Bang. But they don’t know when that Bang occurred, at least not very accurately. The universe is either 10 billion years old or 20 . . . or somewhere in between. The visible universe is either 20 billion light-years across or 40 billion . . . or somewhere in between.
Yet the answer is vital to any deep understanding of cosmology. Determining whether the universe will expand forever or slow and then collapse in a Big Crunch, for example, depends on how dense it is--and you can’t calculate the density of anything without knowing its size and volume. Deducing how galaxies formed out of the smooth soup of matter created by the Big Bang, one of the major conundrums of theoretical astrophysics, depends on how much time gravity has had to pull the galaxies together. The question of the size and age of the universe is nothing less, according to Dick Bond, an eminent theorist at the Canadian Institute for Theoretical Astrophysics, than the most important problem in the field today.
In theory, nature has provided astronomers with a measuring stick they can use to gauge cosmic distances to high precision. The stick is called the Hubble constant, named for Edwin Hubble, who discovered in the 1920s that the universe is expanding at a uniform rate. Like polka dots on the surface of an inflating balloon, galaxies are constantly spreading apart as new space is created between them. And the farther apart two galaxies are, the faster they’re flying apart. If galaxy A is receding from Earth twice as fast as galaxy B, then it’s precisely twice as far away.
But how far away is that? That’s where the Hubble constant--or H0, as it’s abbreviated--comes in. In the formula that says velocity is proportional to distance, H0 is the constant of proportionality. It tells astronomers how much faster the galaxies are separating with increasing distance; or it would if they knew what its correct value is. Its units are kilometers per second per megaparsec. (A megaparsec is about three and a quarter million light-years.) The bigger the Hubble constant, the more quickly speed increases with distance.
A big Hubble constant also means a young universe: if the cosmos is expanding quickly, it has gotten to its present size quickly, and the Big Bang is relatively recent. A Hubble constant of 100 kilometers per second per megaparsec--the high end of current estimates--means the universe is on the order of 10 billion years old. But if H0 is as low as 50, or even a little lower, the age is closer to 20 billion.
The trouble with using the Hubble constant as a measuring stick, then, is that no one knows how long it is. That’s why when astronomers are writing their technical papers they almost never say how far away some distant galaxy is in light-years; they give the distance in terms of the Hubble constant, essentially telling their readers to insert whatever value you think is right for H0, and this will give you the distance. And that’s why alert readers of magazines such as this one will on occasion be surprised to find that the edge of the universe seems to have moved from being, say, 15 billion light-years away one month to being 10 billion light-years away a few months later, when a different astronomer is being quoted. As long as astronomers can’t agree among themselves how big the universe is, journalists are loath to adopt an official policy on the question. Nor do they want to bore their readers with a discussion of the complexities of cosmic distance measurement every time they mention a distance.
To get a firm value for the Hubble constant, and thus the key to all cosmic distances, one really has only to measure the distance to one far-off object conclusively and then determine how fast it’s moving away. But doing that is not so easy. For one thing, the object has to lie well outside the neighborhood of our Milky Way galaxy, where the outward flow of galaxies is hidden by local motions that have nothing to do with the expansion of space. The nearby galaxy M31, for example, also known as Andromeda, is not receding from the Milky Way at all. The two massive spiral star systems are actually falling together under their mutual gravity, and both, along with the entire Local Group of about two dozen galaxies, are falling toward--not expanding away from--the great cluster of galaxies in the constellation Virgo.
The Hubble constant has to be measured in deep space, where the expansion speeds are so great they dwarf any local effects. Unfortunately, you have to go so far to get there--hundreds of millions of light-years-- that the galaxies are too far away for their distances to be measured with precision. Astronomers can tell their relative distances precisely, based on their recession velocities, but their actual distances only sloppily.
None of this would be a problem if astronomers could find a standard candle, some class of objects that all had the same intrinsic brightness--which could be determined from nearby specimens--and that were bright enough to be seen halfway across the universe. Then the object’s apparent brightness would be a measure of its distance; the dimmer it looked, the farther away it would be. So far, though, no one has found a standard candle that everyone agrees works for the whole universe. Instead astronomers have always had to apply this basic approach--judging distance by how much it dims an object’s intrinsic brightness--in a stepwise fashion, with a series of standard candles, each one serving as a yardstick for measuring the distance to the next. Each step is only a little inaccurate by itself; it’s when they’re used all together that they get out of hand.
The first (and only precise) step is trigonometric parallax: a nearby star will appear to change position against the background of faraway stars as the Earth moves from one side of the sun to the other. The same thing happens to a finger held in front of your nose--as you look at it first with one eye, then with the other, it appears to jump across the field of view. The position change plus some high school trigonometry gives the star’s distance with great accuracy.
These nearby stars--the greatest distance measurable by parallax is 4,000 light-years--can then be used to calibrate the distance to stars of the same type that are far away. From the distance of the nearby star and its apparent brightness you can calculate its intrinsic brightness, which will be the same for all stars of that type. When you then see a faraway star of the same type, you can determine its distance by comparing how bright it looks to how bright it really is. These faraway stars are not all that far away: they are still within our galaxy, at most a few tens of thousands of light-years from Earth. But sometimes, fortunately, they are found in double-star systems along with stars called Cepheid variables.
As their name implies, Cepheids, which are ordinary stars in a late stage of life, vary greatly in brightness over periods of several days. The longer a Cepheid’s cycle, the greater its average luminosity--and the relation is well enough understood that all an astronomer has to do is measure a given Cepheid’s period to find its intrinsic average brightness and thus its distance. Cepheids take the distance ladder outside the Milky Way: they’ve been seen in the Andromeda galaxy since the 1920s, and thus astronomers know that Andromeda is about 2 million light-years away. Assuming it is a spiral galaxy of typical brightness, one can then estimate the distance to more distant spirals, and to elliptical galaxies that sometimes cluster with them. Now we are in the range of tens of millions of light-years from Earth.
The final step on the distance ladder is redshift, the reddening of starlight in galaxies that are moving away from the Milky Way as the universe expands. The degree of redshifting is precisely determined by a galaxy’s speed of recession. So astronomers, perched on their shaky ladder of assumptions, calculate the brightness-based distance of moderately faraway spirals and use these distances and the galaxies’ recession velocities to calculate the Hubble constant. Then they go on from there to gauge the distance to yet more distant galaxies--and even to the quasars, enormously bright, starlike beacons of light at the very edge of the visible universe. Now we are 10 billion light-years from Earth--or maybe 20.
One obvious way to sharpen the distance scale is to sharpen the precision of one of these steps. Recently, for example, Wendy Freedman of the Carnegie Observatories and Barry Madore of Caltech took another look at Cepheids. One problem with that rung on the ladder is that distance isn’t the only thing that dims a Cepheid’s light; so does any dust cloud that happens to be floating between the star and us. Dust-dimming can make a star seem farther off than it is. To get around this, Freedman and Madore observed the Cepheids’ near-infrared light, which passes through dust more readily than ordinary visible light does. After recalculating the distances to the Cepheids, they came up with a Hubble constant of 85, which translates into a visible universe that is around 22 billion light-years across and 11 billion years old.
An even better way to improve the distance ladder, in theory at least, is to replace one of the shaky steps with a more secure one. That’s what the planetary nebula group, led by George Jacoby of Kitt Peak and Robin Ciardullo of Penn State, is trying to do. Planetary nebulas are big, bright blobs of gas visible in other galaxies. They are all created in the same way--when an elderly star blows off its atmosphere--so their range of brightnesses, if you look at a bunch of them, is always the same. Instead of using all of Andromeda as a standard candle--the most uncertain step on the distance ladder, and the best one to cut out--Jacoby and Ciardullo use planetary nebulas. The Hubble constant, by this technique, is about 80, and the universe is about 12 billion years old.
John Tonry of MIT is doing something similar, though more complicated. His standard candles are stars in faraway galaxies, stars that are too small to see individually. To measure them, he relies on agonizingly precise measurements of the brightness of patches of galaxy containing thousands of stars and on the statistics that govern random natural processes. Suppose you have a bunch of open paint cans outside, and there’s a hailstorm, he says, by way of explaining his method. Then the hail melts, and there’s water in the cans. Now imagine that you’re looking at the differences in water levels from one can to the next. If the hailstones were big, you’d have a lot of variation in the numbers. One can might have three stones and the next can just one--and the water levels after melting would vary a lot. But if the stones are tiny, you could have a hundred more in one can than in the next, yet water levels that are not much different. The idea is that by comparing levels you can get a notion of what the individual units are like, even if you can’t actually see them.
It turns out to be the same with stars in galaxies. By measuring the variations in brightness from one patch of a galaxy to another, Tonry can calculate the average brightness of individual stars in the galaxy-- that is, the size of the hailstones. He can also see the average color of the stars in a patch. Since the color and size of a star determine its intrinsic brightness, he can deduce the intrinsic brightness of the distant stars by looking at similar stars in our own galaxy, and of the patch as a whole by assuming it contains the same number of stars as a similar-size patch in the Andromeda galaxy. The apparent brightness of the distant patch then tells him how far away the faraway galaxy is. Tonry’s Hubble constant, like the one from the planetary nebula technique, is around 80.
In recent years, then, three different techniques--near-infrared observations of Cepheids, planetary nebulas, and Tonry’s patches--have all converged on the idea that the universe is young, around 11 or 12 billion years old. Is there a trend in the making? Are there the beginnings of a consensus? Hardly: at least three other groups of astronomers have reported results suggesting a much older and larger universe. Two of those groups are attempting the most ambitious improvement yet to the cosmic distance ladder. They are throwing out the ladder altogether and trying to vault to the far reaches of the universe in one jump.
One of the groups is led by Robert Kirshner, head of the astronomy department at Harvard and an expert on supernovas. These exploding stars can be seen halfway across the cosmos, and they’ve long formed a part of distance measurement schemes. Kirshner’s innovation (see the July issue of DISCOVER) is a clever technique for measuring both the apparent size and the actual size of the expanding shell of debris from the explosion, and then using the two to deduce the supernova’s distance. Although he has yet to gather enough data to produce what he considers a definitive result, his Hubble constant is currently hovering around 65, which implies a 15-billion-year-old universe.
The other ladder-busting technique is the one that Jacqueline Hewitt and her graduate student Grace Chen, along with Ed Turner of Princeton, are pursuing at the Very Large Array. The technique depends on a cosmic mirage. When a galaxy happens to lie in front of a more distant quasar, the gravity of the former can act as a lens, distorting the image of the latter by bending its light (or its radio waves). Usually the quasar is split into two or more separate images as its light divides, traveling over at least two distinct paths to Earth.
Unless the quasar, the galaxy, and Earth are perfectly aligned, one of these paths will be slightly longer than the other. That being the case, any flicker in brightness in the quasar will be seen first in one image, then in the other. From the speed of light and the time delay between the flickers, one can calculate the difference in path lengths. From that, together with the angle between the two apparent images of the quasar and the angle by which the quasar’s light has been bent, one can calculate the distance to the lensing galaxy.
Hewitt and her graduate student Joseph Lehár have made the calculation so far for one gravitational lens system--a double-image quasar they saw flicker first in one image and then, about 540 days later, in the second. From the time delay they pegged the distance to the lensing object at 5.7 billion light-years--and the Hubble constant at about 40. That implies a universe that’s more than 24 billion years old, twice as old as some of the other estimates.
Hewitt’s estimate, however, is still very uncertain, as she would be the first to note. To calculate how much the quasar’s light is bent she has to make reasonable assumptions about how much mass is in the lensing object and how the mass is distributed (just as you would need to know both the thickness and the curvature of a glass lens to predict how it would bend light). The problem is that the lensing object in Hewitt’s first case is not just a galaxy but a cluster of galaxies. The mass distribution is very messy, and the Hubble constant derived from it correspondingly uncertain.
A cleaner system would obviously be better, and Hewitt has a good candidate. In 1986 she discovered an object called the Einstein Ring. It is a quasar that, like many of its kind, emits two jets of hot gas from its core. Unlensed, the quasar would look something like a dot bracketed by two dashes. But the lensing has left one of the jets almost too dim to be seen, and smeared the other into the shape of a perfect ring. The quasar itself has been doubled, its two images appearing on opposite sides of the ring. The lens in this case, says Hewitt, appears to be just a single galaxy, which will make the system a lot easier to analyze, if the quasar ever flickers. After a lot of work with the VLA, she is still waiting for the flicker.
Meanwhile, her initial estimate of a low Hubble constant has been buttressed by work done with the Hubble Space Telescope. Allan Sandage of Carnegie Observatories, who has long argued for an old universe, used the Space Telescope to observe so-called Type Ia supernovas--a subset of stellar explosions that, Sandage and his colleagues believe, always shine with the same intrinsic brightness and are thus well suited to be standard candles. The researchers determined that intrinsic brightness by observing one such supernova in a galaxy that was close enough for its distance to be measured by the Cepheid method. Then they measured the distance to farther galaxies by observing supernovas in them. Sandage’s group arrived at a Hubble constant of 45.
Most astronomers hope that Sandage and Hewitt are right, and that all the high estimates of the Hubble constant turn out to be wrong. The reason, says Ed Turner, is that a young universe causes problems. There’s some disagreement about how severe they are, but they’re there. According to accepted theories of stellar evolution, there are stars in the Milky Way that are at least 15 billion years old. If the universe is only 10 billion years old, something is seriously wrong somewhere. With some tinkering, you can get the age of the stars down, says Turner, but not enough.
Astronomers have faced this sort of dilemma before. In the 1940s, Hubble himself came up with a constant of 500, which implied a universe that was only 2 billion years old. The trouble was, geologists already knew that the Earth had rock formations that were nearly twice that old. That crisis was averted by a careful reanalysis of the distance ladder, which increased the size and age of the universe until it could comfortably allow our planet to exist.
Because it measures the distance to faraway galaxies directly, without relying on intermediate steps, the gravitational lens technique probably has the greatest potential for pinning down the Hubble constant once and for all. But until Hewitt or some other observer makes measurements everyone can accept, astrophysics will be left dangling-- poised between a comfortable solution in which the universe is old enough to contain its stars, and a real, inescapable crisis.