The Zircon Fingerprint

Although the list of features of Chicxulub is long and closely matches those to be expected of the K-T impact crater, each individual piece of evidence is circumstantial. To be precise, what has been demonstrated is that on the Yucatan Peninsula, buried under a half-mile of sedimentary rock, lies a crater of K-T age that is somehow linked to the unique Haitian tektites. What has not been demonstrated is that this crater is the parent of the K-T boundary clay around the world. It is highly likely, to be sure, but not proven. But surely circumstantial evidence is the best we can hope to find for an event buried so deep and so far in the past.

Not so. A new kind of evidence, based on the fractured zircon that caught my eye on the cover of Nature, offers proof of a genetic link between the crater and the boundary clay and completely rules out volcanism. To understand this evidence, we need to go a bit deeper into the way radioactive parent-daughter pairs are used to measure rock ages. The principle is simple: Atoms of some elements spontaneously emit subatomic particles, such as neutrons and protons, and in so doing change into atoms of other elements. Each decay occurs at a known rate, called the half-life, which is the time it takes for one-half of any original number of parent atoms to convert to daughter atoms. If we know how much of the parent and how much of the daughter are present in a sample today, and we know how rapidly the parent changes into the daughter (the halflife), we can calculate how long it took for the original amount of parent to decay to that amount of daughter. As an analogy, imagine that you enter a room at exactly 9:00 A.M. and find an hourglass standing on a table, with three-quarters of its sand in the bottom cone. You would quickly conclude that the sand has been flowing for 45 minutes and therefore that the hourglass had been turned over at 8:15 A.M. Radiometric dating is similar but adds two wrinkles. First, imagine that as the grains pass through the constriction between the upper and lower cones of the hourglass, they change color, analogous to one element turning into another. Second, imagine that the constriction is adjustable and is tightened a little more as each grain of sand falls through. The speed with which the grains fall from top to bottom is then related to how many grains remain in the upper half: The fewer that are left, the more slowly they fall through. You could no longer figure out in your head when the hourglass had been turned over, but if the tightening followed certain rules, a simple mathematical formula would do the trick.

Of course, in practice radiometric dating is not so simple. Sometimes atoms of the daughter element, inherited from some ancestral rock, were already present when the decay clock started to run, thus causing the rock being dated to appear older than it is. In other cases, parent or daughter atoms are gained or lost after the decay clock has started to run, throwing off the calculation. Daughter loss is commonly caused by heat, which expands and opens crystal structures and allows the loosely bonded daughter atoms to escape (this is a particular problem with argon atoms when using the older potassium-argon method). If all the daughter atoms are lost, the radiometric clock is set back to zero and the time subsequently measured is not the true, original age of the specimen, but rather the time that has elapsed since it was heated. However, geochemists know how to tell when each of these problems has arisen and usually can correct for them.

The decay of uranium into lead is unique among the geologically useful parent-daughter pairs because two isotopes of uranium, each with its own half-life, decay into two isotopes of lead: U 238 decays to Pb 206 and U 235 decays to Pb 207. Thus two uranium-lead clocks keep time simultaneously. If one plots the uranium-lead isotopic ratios of samples that have suffered no lead loss, they lie along a special curve called Concordia (after the goddess of agreement), shown in Figure 15. If several samples of the same rock or mineral plot at the same point on Concordia, we know that the material has not lost uranium or lead and that the date obtained is its true original age. Because uranium is present at measurable levels in a variety of rocks and minerals, the method has wide applicability.

FIGURE I 5 Composite diagram showing a section of Concordia (the smooth curve) and the position of K-T zircons from Chicxulub, Haiti, Colorado, and Saskatchewan. The fit of the lead-loss line is nearly perfect. A few points that appear to have a different history are also shown. [Data from Krogh and colleagues; recalculated by the author.]

FIGURE I 5 Composite diagram showing a section of Concordia (the smooth curve) and the position of K-T zircons from Chicxulub, Haiti, Colorado, and Saskatchewan. The fit of the lead-loss line is nearly perfect. A few points that appear to have a different history are also shown. [Data from Krogh and colleagues; recalculated by the author.]

Sometimes, however, the uranium-lead ratios of a suite of related specimens plot not on Concordia, but on a straight line that intersects it twice, like a chord to an arc. The mathematics of uranium-lead decay reveal why: When a rock or set of minerals has been altered and lead has escaped, samples with different degrees of lead loss plot along a line that intersects Concordia at two points. The upper, older intercept of the straight line and Concordia gives the original age of the rock; the lower, younger intercept gives the time at which the rock lost lead. When measured ratios plot neither on Concordia nor along a straight line, we know that the geologic history of the samples is more complicated—they may have passed through more than one heating event or they may have lost variable amounts of both uranium and lead. Such results give little or no useful information, and the geochronologist tries again with different samples. But when several samples do plot on a straight line that intersects Concordia twice, we know that the material has passed through only a single lead loss episode and that we have measured both its original age and the time at which it lost lead. Uranium-lead dating thus provides two essential pieces of information and is one of the most powerful tools in the geologist's kit.

The ideal substance for uranium-lead dating would be one that contained no original lead but enough uranium to have produced measurable amounts of radiogenic lead (lead derived from uranium decay) over periods of geologic time. It should occur in a variety of rock types. While we are at it, why not ask for a mineral that is so hard and chemically inert that it survives weathering, erosion, and the heat and pressure of metamorphism? Believe it or not, exactly such a mineral exists: zircon. When zircon crystallizes, it contains uranium but no lead, thus eliminating the problem of original daughter atoms. A minor by-product in granitic rocks, zircon sometimes grows large enough to form a gemstone (a fact well known to viewers of home-shopping networks). Because it survives erosion and every geologic process known except complete remelting, zircon winds up in a wide variety of rocks and looms much larger in understanding earth history than its infrequent occurrence would suggest.

The application of zircon dating to the K-T boundary problem now begins to become clear. If the rocks that existed at ground zero contained zircons, which many continental rocks do at least in small amounts, these zircons might have been shocked, heated, and had their clocks at least partially reset. The fireball cloud might have lofted them high and distributed them over thousands of miles. If suites of such zircons show up in the K-T boundary clay, their uranium and lead isotopes might have retained both the original age of the target and the time of the impact—65 million years. They would then fall along a straight line that intersected Concordia at 65 million years and at some older age defined by the true age of the target rocks. But surely this is too much too expect.

Bruce Bohor of the U.S. Geological Survey and Tom Krogh, who now operates one of the world's most sophisticated lead-dating laboratories, at the Royal Ontario Museum in Toronto, examined samples from the upper K-T layer at the Raton Basin in Colorado and discovered zircons with the same multiple shock deformation lamellae that characterize impacted quartz.47 Shocked zircon had never before been seen. Krogh and his Royal Ontario Museum colleague Sandra Kamo set out to measure the uranium and lead isotopic ratios of these zircons, but ran into two difficulties. First, although the analysis had to be done grain by grain, the individual zircons weighed only from 1 millionth to 3 millionths of a gram and could not even be seen with the naked eye, making them hard to handle, to say the least. Second, the zircons contained only between 5 picograms and 200 picograms (a picogram is a trillionth of a gram, or ' 0"'2 gram) of radiogenic lead, almost too little to be measured. This made analysis especially difficult because, as with iridium, it is almost impossible to rid a laboratory of the effects of contamination from environmental lead. To overcome these two problems, Krogh and Kamo invented new methods of lead analysis, in the process reducing their laboratory lead background (the amount of environmental lead contamination that cannot be eliminated) to the lowest of any lab: 2 picograms of lead per experiment.

Using an electron microscope, they found that they could arrange the tiny Raton Basin zircons visually in a series with unshocked specimens at one end and increasingly shocked and finally granular zircons (whose crystal structures had been completely destroyed) at the other. On the Concordia diagram, the uranium-lead isotopic ratios of these zircons plot along a nearly perfect straight line, showing that they came from a single target and experienced a single episode of lead loss. The line intersects Concordia once at 550 ± '0 million years, which must be the original age of the zircons, and again at 65.5 ± 3 million years, which must be the time of lead loss. The more shocked the zircons, the farther down the line they plot, closer and closer to the lower, 65-million-year intercept. This point is absolutely critical: Zircons that originated in a volcanic eruption 65 million years ago would have crystallized at that time. They would possess, and would display, an original age of 65 million years—not 550 million years. Subsequent lead loss would make them appear younger than 65 million years, not older. The least altered and unaltered zircons would give the true age of 65 million years. But just the opposite is the case for the Raton Basin zircons: The unshocked and least shocked zircons give the oldest ages, while the most shocked, including those that are completely shocked, give the youngest, approaching and in extreme cases reaching 65 million years. Because 65-million-year-old zircons could not produce this result, neither the zircons nor the clay itself could have come from a 65-million-year-old volcanic eruption.

Krogh and his group went on to take the next logical step: They analyzed zircons from both the Chicxulub breccia and from the Haitian Beloc Formation.48 Plotting the Chicxulub, Haitian, and Raton Basin zircons on the same Concordia diagram, they found that '8 of 36 fell on a straight line that intersected Concordia at 545 ± 5 million and 65 ± 3 million years, "as though they had come from a single sample."49 Finally, the group studied zircons from the K-T site most distant from Chicxulub, in South Central Saskatchewan, 3,500 km away.50 They found an upper intercept age of 548 ± 6 million years and a lower of 59 ± '0 million years, the same within the analytical precision as the results from the other three sites.

In all, Krogh and his colleagues studied 43 K-T zircons. A few seemed to point to an age of about 418 million years for the parent rock; several others scatter randomly when plotted on the Concordia diagram, indicating they have had a more complex history, perhaps having lost lead twice. But as shown in Figure 15, of the 43 zircons from all four sites, 30 fall exactly on a straight line (the 418-million-year-old zircons are omitted from the diagram). These 30 zircons, found at four sites separated by thousands of kilometers and representing three completely different geologic settings—a Chicxulub breccia, Haitian tektites, and K-T boundary clays from two locations, one 3,500 km from the Yucatan—plot precisely on a single straight line with a coefficient of correlation, the statistician's test of "goodness of fit," of 0.998. (When all 43 zircons are included, even those that obviously have a more complex history, the correlation coefficient is still a remarkably high 0.985.) If one were to collect and analyze 30 zircons from a single rock unit, their fit could be no more perfect. Even though scattered over 3,500 km, these are the same zircons.

One of the most surprising results of Krogh's work, after one gets used to the near perfection of the fit, is that so many of the zircons have the same original age. Since we know that the impact that formed Chicxulub excavated a crater some 20 km deep, a huge slice of crustal rocks, with diverse ages and compositions, should have been caught up in the ejecta. Yet most of the zircons give the same 550-million-year upper age. Krogh and his co-workers speculate that the upper few kilometers at the Yucatan ground zero may have been made up of zircon-free limestone, so that most of the zircons came from a single underlying, zircon-bearing layer.

The remarkable sleuthing of Krogh and his colleagues has to rank as one of the great analytical triumphs of modern geochemistry (though Officer and Page do not cite them in their 1996 book or once mention zircon). Here is what the zircons tell us:

1 . The K-T boundary clay was not formed by volcanism. Had it been, none of the K-T zircon ages would exceed 65 million years. Furthermore, volcanic zircons are angular and unshocked, not the opposite.

2. At least in the Western Hemisphere, the clay had a single source crater rather than having been derived from multiple impacts, as some had suggested in the late 1980s.

3. The K-T ejecta deposits from Haiti, Colorado, and

Saskatchewan each came from the same target rock, of age 545 million to 550 million years, and each was shocked at exactly the same time: 65 million years ago.

A. Since rocks 545 million to 550 million years in age are rare in North America, and since the Chicxulub zircons themselves give both that upper age and the 6 5-million-year lower age, the K-T ejecta in Haiti, Colorado, and Saskatchewan almost certainly came from the Chicxulub structure.

Thus the K-T zircons provide direct, noncircumstantial evidence that Chicxulub is the K-T impact crater. This is no longer a fascinating speculation, but closely approaches the status of observational fact. Today, those who doubt that Chicxulub is the long-sought crater can be counted on one's fingers. Yes, a giant meteorite did strike the earth at the end of the Cretaceous. But did it cause the K-T mass extinction and the death of the dinosaurs?

Was this article helpful?

0 0

Post a comment