The past decade has witnessed a revolution in geochronology, the science
of determining the age of rocks. The newest and most powerful tool available
to geochronologists, single crystal laser fusion, has brought an unprecedented
degree of precision to the business of dating rocks from the argon they
contain. More than this, it has opened doors to a new and yet to be explored
dimension of rock history. And, because the technique requires only tiny
samples – often single crystals – rather than the large chunks of rock demanded
by other approaches, researchers also benefit in more mundane ways. ‘I recently
came back from Olduvai Gorge with 75 samples,’ explains Bob Walter of the
Geochronology Center, Berkeley. ‘They weighed just a few grams, a small
bag full. With conventional techniques I would have needed hundreds of pounds;
my back would have ached with lugging it all around; and I would have had
to ship it back to the States at a cost of thousands of dollars.’
The immediate precursor of today’s sophisticated system produced its
first results in November 1981, in a paper in Geophysical Research Letters
by Derek York and his colleagues at the University of Toronto. This paper
began a series of ‘firsts’ from the Toronto group, which single-handedly
wrested the new technique from obscurity and refined it into the powerful
precision tool it now is. ‘A lot of people said it wouldn’t work,’ remembers
York, a swashbuckling, ‘can-do’ researcher, a physicist turned geochronologist.
‘We had to push this thing really hard before the bandwagon started rolling.’
It is rolling now and has produced what one leading researcher describes
as ‘an order of magnitude leap’ in the sophistication of dating techniques.
Within the past few years virtually every leading geochronology laboratory
around the world has put together some version of the single crystal laser
fusion system.
The modern approach to dating rocks goes back to the beginning of this
century, when Ernest Rutherford suggested that the natural radioactivity
they contain might be exploited as a built-in clock. With information about
the rate of decay of certain isotopes, a way of measuring the products of
this decay and some kind of calibration, rocks may be dated. The advantage
of this method over the traditional approach is that it produces an absolute
age, as a number of years ago, rather that a stratigraphic date, which slots
the rock into a particular position on a scale of older and younger rocks.
There are different kinds of radiogenic clocks, of course, based on different
families of isotopes. Since the 1950s three principal clocks have emerged:
uranium-lead, rubidium-strontium, and potassium-argon. This last system
is used in the single crystal laser fusion technique.
Advertisement
Many rocks contain traces of potassium, particularly volcanic rocks
that contain alkali feldspar. Naturally occurring potassium contains a small
quantity of potassium-40, an isotope that decays slowly to produce argon-40,
the most common isotope of this noble gas. The temperature at which minerals
such as feldspar crystallise following a volcanic eruption is high enough
to eliminate all the argon from the lattice, effectively setting the argon
clock to zero. As time passes after an eruption, a rock containing potassium
will accumulate argon-40. This is the basis of the potassium/argon clock,
devised in 1948 by Tom Aldrich and Alfred Nier of the University of Minnesota
and developed by Berkeley geophysicist John Reynolds in the late 1950s.
The principle is simple: the more argon-40 a rock contains, the older
it is. The amount of potassium in the sample has to be known, of course,
because the clock is based on estimates of the quantity of argon-40 that
would be expected over particular periods of time from a certain amount
of rock. Then, with appropriate technical adjustments, the potassium/argon
clock can give reliable dates, particularly for relatively young rocks –
samples that are a few million years rather than hundreds of millions of
years old.
To date a rock using this conventional potassium/argon technique takes
two separate measurements: the level of potassium in the minerals and the
amount of argon-40 that has accumulated in the mineral grain. The two measurements
required two separate experiments on two separate samples of the rock, a
cumbersome, if feasible, approach that has been a workhorse of geochronology
for many years. It has its disadvantages; each sample should weigh as much
as a gram, making it about the size of a hazelnut or smaller. This necessarily
limits the precision of dates in, for example, sedimentary rocks, for each
sample may include grains of different ages. In addition, under certain
conditions argon may leak from a sample, giving an erroneously low argon
level and a false (too young) date. Nevertheless, a great deal of important
work was achieved with this technique: in the 1960s, Brent Dalrymple and
Ian MacDougall documented reversals of the Earth’s magnetic field, critical
to the understanding of sea floor spreading. This was a crucial step towards
the theory of plate tectonics. The same dating method was used in the early
days at hominid fossil sites in East Africa.
Meanwhile, in 1965, Craig Merrihue, at the University of California,
Berkeley, hit upon an idea that would allow the two measurements – of potassium
and argon – to be made simultaneously on one sample. Natural potassium consists
mostly of the isotope potassium-39, which is relatively stable under normal
circumstances. Blast it with neutrons, however, and it becomes argon-39,
a rare and unstable cousin of argon-40. Merrihue realised that argon-39
could serve as a vicarious measure of the amount of potassium in a rock.
When the irradiated rock sample is crushed and melted in the normal way,
both forms of argon are released and can be determined simultaneously in
a mass spectrometer: the argon-39 provides a measure of the potassium contained
in the sample, and the argon-40 represents the cumulative ticks of the radiogenic
clock. All the information required for dating the sample comes from one
measurement step on one sample.
Merrihue developed this technique, which is known as the argon-40/39
or argon/argon technique, with Grenville Turner, who was until recently
at the University of Sheffield but has now moved to the University of Manchester.
Merrihue never saw their technique applied, because he died in a climbing
accident in 1966. Various labs were involved in refining the new technique
over the next decade, including those of Turner in Sheffield, Jack Miller
in Cambridge, and Derek York in Toronto.
For the argon/argon technique, each rock sample has to be melted at
a temperature of about 1600 °C, achieved through radio frequency induction,
the principle on which a microwave oven operates. The sample is enclosed
in a high vacuum crucible, heated for 45 minutes, and the gases it releases
passed into a mass spectrometer to measure the proportions of the two isotopes.
It is time-consuming; one sample might occupy a machine for an entire day.
The recent revolution focused on the means of heating the sample, but it
required many steps to achieve and involved more than simply replacing one
method of fusion for another.
In the early 1970s, George Megrue of the Smithsonian Institution in
Washington DC, was experimenting with ways of analysing gases held in minerals.
He decided to use a pulsed laser to heat small areas of his samples, driving
off the noble gases so he could analyse them in a mass spectrometer. The
advantage of this step was that Megrue could measure gases from particular
areas of the samples, rather than having to melt the whole rock. Later he
heard about Merrihue’s argon-40/39 dating method, and combined it with his
system, substituting the pulsed laser heat source for the radio frequency
heating. By irradiating his rock samples with neutrons before analysing
them, Megrue could not only investigate the distribution of noble gases
in them, but also obtain an age. Oliver Schaeffer, of the State University
of New York at Stony Brook, then applied the pulsed laser dating technique
to lunar rock samples from the early Apollo missions.
A crucial turning point in the revolution, however, depended on one
of those accidents of history that so often nudge science in unexpected
directions. ‘I was about to embark on a sabbatical,’ remembers Derek York,
‘and I was looking around for something new to get involved in.’ In his
previous sabbatical, seven years earlier, York had decided to find out about
the newly emerging argon/argon method, and subsequently played an important
role in exploiting it. ‘This time I thought lasers were exciting, so I decided
to take a look.’ York spent his sabbatical at the University of Tokyo, and
later collaborated with Yotaro Yanase, a postdoctoral researcher. York and
his colleague faced a problem, however: the pulsed laser that Schaeffer
had used so successfully in dating Moon rocks was obsolete and no longer
available.
Megrue and Schaeffer had used pulsed lasers because they could deliver
enough energy to fuse small sections of the rock sample. No one had ever
considered using continuous lasers to do the job at that time. Unable to
obtain a pulsed laser of the sort Schaeffer had used, York, Yanase and Chris
Hall, then a graduate student, toyed with various alternatives, but nothing
worked. ‘Then we thought, why not try a continuous laser?’ recalls York.
‘Maybe they were powerful enough to fuse the material we wanted?’ A more
fortuitous question could not be imagined at the time, for it opened the
way to exploiting one of the great benefits of the argon/argon technique,
an application that had not been possible with the pulsed laser.
The conventional argon/argon method – using radio frequency induction
heating – can be applied in two ways. The most straightforward is to fuse
the rock in one blast, giving a single age estimate. A second and more time-consuming
method developed by Merrihue and Turner, is to apply the heat in steps,
increasing the power with each increment. With each step the evolved gases
are drawn off and analysed in the usual way, producing an age for the sample
at each step. At its simplest, this ‘step heating’ method apparently produces
a series of ages for the rock, progressing step by step from the outer surface
of the sample towards its core. ‘Step heating is tremendously important,’
says York. ‘It is the single greatest advantage of the 40/39 technique.’
By taking, say, a dozen heating steps, the experimenter is doing far more
than obtaining a dozen independent dates for a single sample. Something
of the history of the rock may be revealed too.
It is possible that over geological time a rock may experience no disturbance,
in particular no thermal disturbance. If so, none of the radiogenic argon-40
accumulated over time would have been driven out (giving an erroneously
young age), and none would have been taken up from other sources (giving
an erroneously old age). The ages produced at each stage of heating would
be the same: plotted against the temperature steps, the series of ages would
produce a flat, straight line, a plateau with no valleys or peaks to disturb
the uniformity. This is known as the ‘age spectrum’. For rocks with such
simple argon signatures, the flat line gives their age.
The secret life of rocks
But most rocks lead more interesting lives; they are subject to the
vagaries of rifting, mountain building, intrusion by igneous rocks such
as granites, and many other sources of thermal disturbance. These upsets
imprint themselves on the rock by altering the amount of argon trapped in
the matrix, mostly driving off gas. Unless the heating is extreme, the outer
layers of the rock will be more affected than the inner layers, producing
a gradient of increasing argon concentration from the outside to the inside.
Step heating will reveal this pattern, as the ages produced early in the
series (from the outer layers) will be less than those produced later (from
the inner layers). The age spectrum then begins with a steeply rising slope,
which in some cases flattens out to a plateau which marks the age of the
rock. ‘Step heating allows you to get a glimpse of a rock’s history,’ says
York, ‘and this is so much more informative than simply a figure for its
²¹²µ±ð.’
The use of a pulsed laser for fusing minerals precludes step heating,
because all the energy comes in a single package. ‘It was for this reason
that laser fusion initially didn’t catch on for 40/39 work,’ observes Brent
Dalrymple, a geochronologist at the US Geological Survey at Menlo Park in
California. ‘It was just too limited. Then Derek (York) had the bright idea
of using a continuous laser; that allowed incremental heating to be done.’
The results of this ‘bright idea’ appeared in November 1981 in Geophysical
Research Letters, where York, Yanase and Hall described how they used a
15-watt argon-ion laser to deliver a series of increasingly powerful, 30-second
blasts as a novel form of step heating. ‘This was the first publication
of an age spectrum produced by a continuous laser,’ says York, with some
pride. Here was the true beginning of the dating revolution.
Two further stages were yet to be achieved: the ability to work with
single crystals, rather than numerous small pieces of mineral, and automation
of the whole system. ‘Going from a chunk of rock to a single crystal was
simply a matter of sensitivity,’ says York. ‘If you have one crystal rather
than a thousand, then you have only one thousandth of the amount of gas
coming off, so you need a sensitive mass spectrometer to analyse it.’
Britain has a long tradition of producing exquisite mass spectrometers,
and so it was natural that York should turn to a British company to replace
his sturdy, reliable, but now inadequate instrument. ‘A beautiful little
machine,’ says York. ‘I’d had it since 1964.’ It was duly replaced by a
VG1200, made by Michael Lynch of Vacuum Generators, at a cost of $220 000.
Soon, York and Hall had their single crystal system up and running. In the
mid-1980s, Garniss Curtis visited the lab. Curtis, a renowned geochronologist
at the University of California, Berkeley, had stuck very successfully with
conventional potassium/argon dating during his long career, and had not
ventured into argon/argon territory. ‘It was very impressive to see them
measure individual components of gas from a single crystal,’ remembers Curtis.
‘I was converted.’
At the time Curtis was setting up the Geochronology Center, which is
associated with Donald Johanson’s Institute of Human Origins, in Berkeley.
‘It was clear that single crystal dating was going to be important,’ says
Curtis. ‘We decided to get involved.’ With his colleagues Robert Walter,
Alan Deino, and Paul Renne, Curtis decided to automate the technique, and
within a few years had a working system that could produce single-step ages
on 140 crystals in one fully-programmed run. Meanwhile, York and Hall were
also automating their system to handle 14 multi-step analyses of crystals
at one time. There is some good-natured rivalry over which team – Toronto
or Berkeley – actually produced the first automated system.
Dalrymple, an objective observer in all this, is clear about who won:
‘I don’t know of anyone who has done anything unique in a technical sense
with this system that wasn’t done first at Toronto, and that includes the
Berkeley people.’
The immediate benefit of using single crystals is increased precision.
The cleaner a system is, the less blurring there will be of the age produced,
and you can’t get cleaner than a single crystal. But greater precision is
only a minor benefit compared with the other advantages of the system. Perhaps
the most powerful lever it offers geochronologists is the ability to overcome
problems with contamination, long a troublesome aspect of dating.
When a volcano erupts to form an ash layer most of the rock comes from
the same source, but the eruption invariably sweeps up with it minerals
from other sources and other eruptions. Even lava exploding out of a volcanic
vent may pick up older material from earlier eruptions from the walls of
the vent. The newly forming ash layer – known in the business as a tuff
– may also become contaminated with older material lying on the land surface,
deposited there by earlier eruptions. A piece of tuff may contain crystals
and rock particles of several different ages, even though they may all look
much the same. The usual result of contamination is an artificially old
age, as the older rock contributes disproportionably more argon-40 to the
count when the rock is crushed and melted. ‘Contamination has been an ever-present
worry, especially with young rocks,’ says York. ‘Even minor contamination
can be a severe problem.’
Foreign bodies
York emphasised this last point in a paper he and Hall published with
a group of French researchers: part of the title read ‘The defeat of xenocryst
contamination’. Xenocryst is the geologist’s term for foreign, or contaminating,
crystal. The paper described the dating of pumice – foamy volcanic glass
– from the Massif Central, in France. Earlier workers had produced an age
of 850 000 years for one volcanic layer. When the French team crushed the
pumice and retrieved feldspar crystals from it, they found two types: clear
crystals with angular faces; and rounded cloudy crystals. Clearly, contamination
was a real possibility here.
Gilbert Ferand, of the University of Nice, and York and Hall determined
that the clear crystals, which outnumbered the cloudy ones, by a thousand
to one, were 580 000 years old. Because the clear crystals were the most
common in the sample, this was the probable date of the eruption that produced
the pumice. The cloudy crystals were much older, at some 330 million years.
In this case contamination at a vanishingly low level can produce an increase
in apparent age of more than 50 per cent, a significant error by any measure.
‘This study shows that the ability to measure single grains is critical
in unravelling the ages of young volcanic material when country-rock contamination
is likely,’ comment York, Ferand and their colleagues. ‘The single crystal
dating effectively eliminates the contamination problem,’ says York.
The ability to see through contamination allows new types of problem
to be addressed, explains Dalrymple. ‘We’re interested in the rate of basin
formation on the lunar surface, but when you get a big impact a lot of the
target rock melts and then crystallises to form meltrock,’ he says by way
of example. ‘This meltrock contains a lot of older material, and if you
have to use samples of a few milligrams, as you did with conventional 40/39
dating, inevitably you will get contamination. As a result the data we’d
been getting in the past was sometimes mystifying. Now, with single crystal
dating, we are getting beautiful age results.’
In addition to circumventing contamination, the use of single crystal
samples means that sparse horizons can be dated. This has been important
in work on thin ash layers in coal deposits, yielding information on the
rate at which coal formed. ‘From thin ash layers you can pluck a few crystals
and date them,’ says Dalrymple. ‘Previously, that would have been impossible.’
Substantial though these immediate practical benefits are, York has
his eye on wider goals. ‘Conceptual pioneering,’ he calls it. ‘Because the
system is automated you can leave it on over night, over the weekend, whatever
you want, working away at what might be regarded as risky projects, things
you wouldn’t normally get funded for,’ he explains. ‘You can leave the robot
working on risky projects, while you get on with more reliable work, to
keep the grants going.’
One project the Toronto team is pursuing in this way is the dating of
sedimentary rocks, one of the most difficult problems in geochronology.
Sediments are mixtures of all kinds of material, including rock fragments
and mineral grains of all ages, so that contamination is certain to impede
conventional radiogenic approaches. ‘We can do step heating on tiny fragments
of sediment, see if it will work,’ explains York. ‘Most people would say,
‘Only those lunatics in Toronto would do that.’ Well, we’re having some
success, with Huronian sediments, famous sediments north of Lake Huron.’
Although the particles in the sandstone and clay of the Huronian sediments
are so small that it is almost impossible to say which mineral grains are
being dated, York and his colleagues have extracted some information about
the rock’s history. ‘We can’t say when the sediments were first laid down,’
explains York, ‘but we can pinpoint times at which folding of the sedimentary
layers occurred. This represents the first ever direct dating of folding
events in sediments.’
Even more risky is York’s dream of finding sedimentary rocks which have
held tiny packets of the ancient atmosphere in structures which may be fluid
inclusions, tiny bubbles up to 100 micrometres in size, held within mineral
grains. Bubbles in ice cores have already revealed the variations of carbon
dioxide and methane hundreds of thousands of years into the past, but York
has his eye on greater things. He wants a picture of how the atmosphere
has changed through time since Earth formed, 4.5 billion years ago. ‘The
signature is in the ratio of argon-40 to argon-36,’ explains York. ‘When
Earth formed the ratio was about 10 -4; now it is about 300.
I’d like to track the rise of argon-40 through time.’ First, he has to find
rocks of the right kind, rocks which contain fluid inclusions with those
packets of palaeo-atmosphere. So far no one has found any. York thinks it
might be done by ‘scanning the rocks like astronomers scan the skies’. By
letting the robot system work day in and day out, sniffing the gas content
of many different kinds of rocks from many different locations, York hopes
to strike lucky. ‘We’ve been trying it for a while,’ he says, ‘but we’ve
failed so far. We need to step up the effort.’
Who knows, those lunatics in Toronto might just stumble on something
that has fascinated theoreticians for years but which, so far, has defeated
the down-to-earth practitioners.
* * *
Haitian tektites date the demise of the dinosaurs
In June, geologists in the US used the single crystal fusion technique
to date glass beads, lending support to the hypothesis that an extraterrestrial
impact brought about the end of the dinosaurs. One of the team was Glen
Izett, a geologist at the US Geological Survey in Denver, Colorado, who
was involved with various geochemical investigations of material from the
Cretaceous/Tertiary (K/T) boundary. This time, 65 million years ago, was
marked by mass extinctions and the end of the dinosaurs.
The material Izett studied, which came from Beloc, Haiti, included the
now famous iridium signature and so-called shocked quartz. Both are indications
of impact by asteroid or comet, and some researchers have cited them as
evidence that the mass extinction associated with the K/T boundary was indeed
caused by such an extraterrestrial event. This hypothesis is still controversial.
As well as the iridium and shocked quartz there were tektites – spherules
of glass that can also be generated by extraterrestrial impact. Their size
varies from less than a millimetre to 10 millimetres across. ‘Because the
tektite material could be dated by the argon/argon technique, this offered
us an opportunity to put a precise age on an iridium-containing K/T layer,’
explains Izett. Brent Dalrymple, also of the US Geological Survey in Menlo
Park, California, joined the project: Dalrymple had set up an automated
single crystal laser fusion system.
The tektites most suited to this system were between 0.7 and 1.2 millimetres
in size. Step heating showed that geological heating had not affected these
tektites since they were deposited, and the researchers obtained an accurate
age for the samples of 64.5 +- 0.1 million years. ‘The Haitian tektites
are the first datable impact products found in association with the worldwide
(iridium) anomaly,’ reported Izett and his colleagues, ‘and thus their age
directly dates the time of the K/T impact or impacts.’
The team says that its new data ‘confirm the impact hypothesis’ and
‘deliver a devastating blow’ to the opposing idea, that the mass extinctions
were caused by catastrophic volcanic activity. ‘We could not have been so
precise or been so sure of our results, without the single crystal laser
fusion dating system,’ says Izett. ‘It helped us nail down one of the most
contentious problems in geology and palaeontology today.’