Sunteți pe pagina 1din 9

NEW VIEW of the young earth covered in oceans

of liquid water as early as 4.4 billion years ago

contrasts sharply with the hot, hostile world
typically depicted in textbooks.

A Cool
Early Earth?
The textbook view that the earth
spent its first half a billion years
drenched in magma could be wrong.
The surface may have cooled
quicklywith oceans, nascent
continents and the opportunity for
life to form much earlier
By John W. Valley

In its infancy, beginning about 4.5 billion years ago,
the earth glowed like a faint star. Incan- ed for as long as 500 million years, an hardened over, the thickening layer of
descent yellow-orange oceans of magma era thus named the Hadean. Major sup- consolidated rock would have insulated
roiled the surface following repeated col- port for this view comes from the appar- the exterior from the high temperatures
lisions with immense boulders, some the ent absence of any intact rocks older deep within the interior. If there were
size of small planets, orbiting the new- than four billion years and from the suitably quiescent periods between ma-
ly formed sun. Averaging 75 times the fi rst fossilized signs of life, which are jor meteorite impacts, if the crust was
speed of sound, each impactor scorched much younger still. stable, and if the early hothouse atmo-
the surface shattering, melting and In the past five years, however, ge- sphere did not trap too much heat, sur-
even vaporizing on contact. ologists including my group at the face temperatures could have quickly
Early on, dense iron sank out of the University of WisconsinMadison fallen below the boiling point of water.
magma oceans to form the metallic have discovered dozens of ancient crys- Furthermore, the primitive sun was
core, liberating enough tals of the mineral zircon with chemical fainter and contributed less energy.
gravitational energy compositions that are changing our Still, for most geologists, an undis-
to melt the entire plan- thinking about the earths beginnings. puted fiery birth and scant clues in the
et. Massive meteorite The unusual properties of these durable geologic record seemed to point instead
strikes continued for minerals each the size of the period in to a prolonged ultrahot climate. The old-
hundreds of millions of this sentence enable the crystals to est known intact rock is the four-billion-
years, some blasting preserve surprisingly robust clues about year-old Acasta gneiss in Canadas
craters more than 1,000 what the environment was like when Northwest Territories. This rock formed
kilometers in diameter. they formed. These tiny time capsules deep underground and bears no infor-
At the same time, deep bear evidence that oceans habitable to mation about conditions on the surface.
OLD VIEW of underground, the de- primitive life and perhaps continents Most investigators assumed hellish con-
a hot young cay of radioactive ele- could have appeared 400 million years ditions at the planets surface must have
earth: Life ments produced heat at earlier than generally thought. obliterated any rocks that formed earli-
rates more than six er. The oldest rocks known to have orig-
December 8,
1952. times greater than they Cooling Down inated underwater (and thus in relatively
do today. si n c e t h e 19 t h c e n t u ry, scien- cool environs) did not form until 3.8 bil-
These fiery conditions had to subside tists have attempted to calculate how lion years ago. Those sediments, which
before molten rock could harden into a quickly the earth cooled, but few expect- are exposed at Isua in southwestern
crust, before continents could form, be- ed to find solid evidence. Although mag- Greenland, also contain the earliest evi-
fore the dense, steamy atmosphere could ma oceans initially glowed at tempera- dence of life [see Questioning the Old-

D O N D I X O N ( p r e c e d i n g p a g e s) ; T I M E L I F E P I C T U R E S / G E T T Y I M A G E S ( t h i s p a g e)
pool as liquid water, and before the tures exceeding 1,000 degrees Celsius, a est Signs of Life, by Sarah Simpson;
earths fi rst primitive life could evolve tantalizing suggestion of a more temper- Scientific American, April 2003].
and survive. But just how quickly did ate early earth came from thermody- Single crystals of zircon began to add
the surface of the earth cool after its lu- namic calculations showing that crust new information about the early earth in
minous birth? Most scientists have as- could have solidified on the surface the 1980s, when a few rare grains from
sumed that the hellish environment last- within 10 million years. As the planet the Jack Hills and Mount Narryer re-
gions of Western Australia became the
Overview/Zircon Time Capsules most ancient terrestrial material known
at that time the oldest dating back al-
!"" Geologists have long thought that the fiery conditions of our planets most 4.3 billion years. But the informa-
birth 4.5 billion years ago gave way to a more hospitable climate by about tion these zircons carried seemed am-
3.8 billion years ago. biguous, in part because geologists were
! " "Now tiny crystals of the mineral zircon, which retain clear evidence of when unsure of the identity of their parent
and how they formed, suggest that the earth cooled far sooner perhaps as rock. Once formed, zircon crystals are
early as 4.4 billion years ago. so durable that they can persist even if
! " "S ome ancient zircons even bear chemical compositions inherited from the their parent rock is exposed at the sur-
cooler, wet surroundings necessary for life to evolve. face and destroyed by weathering and
erosion. Wind or water can then trans-



Isua sediments
(oldest evidence
for life, 3.8 billion
years old)

Acasta Gneiss Fossilized gravel bed in the Jack Hills

(oldest intact rock, (above) contained the worlds oldest zir-
4 billion years old)
cons yet discovered. Geologists crushed
and sorted hundreds of kilograms of this
Rocks older than
rock (below) to find the 20 crystals that
2.5 billion years bear signs of cool conditions more than
four billion years ago.
Inferred Jack Hills zircons
Exposed (oldest earth material,
4.4 billion years old)

Ancient rocks older than 2.5 billion years crop out or lie just underneath the soil in many spots
around the globe (red) and are probably hidden below younger rocks across even broader regions
(pink). Zircon crystals as old as those discovered in the Jack Hills of Western Australia may
eventually be discovered at another of these locations.

port the surviving grains great distances The Jack Hills conglomerate was depos- to lead. When a zircon forms from a so-
before they become incorporated into ited three billion years ago and marks lidifying magma, atoms of zirconium,
L U C Y R E A D I N G - I K K A N D A ; S O U R C E : W I L L I A M H . P E C K C o l g a t e U n i v e r s i t y ( m a p) ; J O H N W. V A L L E Y ( p h o t o g r a p h s)

deposits of sand and gravel that may the northwestern edge of a widespread silicon and oxygen combine in exact
later solidify into sedimentary rock. In- assembly of rock formations that are all proportions (ZrSiO4) to create a crystal
deed, the Jack Hills zircons separated older than 2.6 billion years. To recover structure unique to zircon; uranium oc-
by perhaps thousands of kilometers from less than a thimbleful of zircons, my col- casionally substitutes as a trace impu-
their source were found embedded in a leagues and I collected hundreds of kilo- rity. Atoms of lead, on the other hand,
fossilized gravel bar called the Jack Hills grams of rock from these remote out- are too large to comfortably replace any
conglomerate. crops and hauled them back to the labo- of the elements in the lattice, so zircons
So, despite the excitement of finding ratory for crushing and sorting, similar start out virtually lead-free. The urani-
such primeval pieces of the earth, most to searching for a few special grains of um-lead clock starts ticking as soon as
scientists, including me, continued to sand on a beach. the zircon crystallizes. Thus, the ratio of
accept the view that the climate of our Once extracted from their source lead to uranium increases with the age
young planet was Hadean. It was not rock, individual crystals could be dated of the crystal. Scientists can reliably de-
until 1999 that technological advances because zircons make ideal timekeep- termine the age of an undamaged zircon
allowed further study of the ancient zir- ers. In addition to their longevity, they within 1 percent accuracy, which for the
con crystals from Western Australia contain trace amounts of radioactive early earth is about plus or minus 40
and challenged conventional wisdom uranium, which decays at a known rate million years.
about the earths earliest history.
JOHN W. VALLEY received his Ph.D. in 1980 from the University of Michigan at Ann Arbor,

Digging Deep where he first became interested in the early earth. He and his students have since ex-
t h e aus t r a l i a n z i rc ons did not plored the ancient rock record throughout North America and in Western Australia,
give up their secrets easily. For one thing, Greenland and Scotland. Currently Valley is president of the Mineralogical Society of
the Jack Hills and their surroundings are America and Charles R. Van Hise Professor of Geology at the University of Wisconsin
dusty barrens at the edge of vast sheep Madison, where he founded a multimillion-dollar laboratory called WiscSIMS. The cut-
stations, called Berringarra and Mileu- ting-edge capabilities of the labs new CAMECA IMS 1280 ion microprobe will enable a
ra, situated some 800 kilometers north diverse range of research; besides zircons, Valley and his colleagues will probe many
of Perth, Australias most isolated city. rare or extremely small materials ranging from stardust to cancer cells.

w w w. s c ia m . c o m SCIENTIFIC A MERIC A N 61
Scientists extract multiple clues about the earths ancient cons atomic structure. Simply put, the constant radioac-
environment from a single crystal of the mineral zircon (main tive decay of uranium to lead means that the more lead
cutaway below). They first embed the zircon in epoxy, then present relative to uranium, the older the crystal.
grind and polish the crystal to expose a pristine surface. Investigators then grind down the surface to expose a
A scanning electron microscope identifies the zircons deeper layer of the crystal and make a second microprobe
growth patterns and any smaller fragments of minerals pit in precisely the same location as the first one to
enclosed as it grew. Inclusions of quartz, for instance, occur measure atoms of oxygen, one of three elements that make
most commonly in zircons that came from granite, a type of up a zircon. The ratio of certain oxygen isotopes atoms of
rock characteristic of continents. oxygen with different masses reveals whether the crystal
An ion microprobe creates a small pit by sputtering records hot or cool conditions.
atoms off this polished area using a narrow beam of ions Researchers create a third pit to measure the abundance
and identifies those atoms by comparing their masses. To of certain trace elements impurities known to make up less
determine the age of the crystal, scientists measure atoms than 1 percent of the crystals structure. Some of these ele-
of uranium and lead, two impurities trapped within the zir- ments are more common in continental crust.
Polished Quartz inclusions


microprobe pits
Oxygen-isotope pit

A L F R E D T. K A M A J I A N ; S O U R C E : A A R O N J . C A V O S I E U n i v e r s i t y o f W i s c o n s i n M a d i s o n ( z i r c o n s c h e m a t i c a n d r e d z i r c o n s)
Location of oldest Second
uranium-lead date polished

Trace-element pit
Outer surface rounded

J O H N W. V A L L E Y ( l a b o r a t o r y) ; S I M O N A . W I L D E C u r t i n U n i v e r s i t y o f Te c h n o l o g y ( o l d e s t z i r c o n) ;
from abrasion

Original rectilinear
shape of crystal
The earths oldest zircon, viewed as a
cathodoluminescent image, dates back 4.4 billion years.

Red zircons (right),
photographed near
Roosevelts nose on
a U.S. dime for scale,
come from the same
rock sample that
yielded the earths
oldest crystal. An
ion microprobe,
such as this one in
the authors
laboratory at the
University of
WisconsinMadison ( far right), can analyze isotope ratios or
trace elements from spots about 1/15 the diameter of the
crystals themselves.


Dating specific parts of a single crys- We knew that a zircon could retain evi- one millionth the size of those then pos-
tal first became possible in the early dence not only of when its host rock sible in my laboratory in Wisconsin. Af-
1980s, when William Compston and formed but also of how. In particular, ter 11 days of round-the-clock analysis
his colleagues at the Australian Nation- we were using the ratios of different iso- and little sleep (typical conditions for
al University in Canberra invented a topes of oxygen to estimate the temper- this difficult procedure), we completed
special kind of ion microprobe, a very atures of processes leading to the forma- the measurements and found that our
large instrument they playfully named tion of magmas and rocks. predictions were wrong. The zircons
SHRIMP, short for Sensitive High-Res- Geochemists measure the ratio of !18O values ranged up to 7.4.
olution Ion Micro Probe. Although oxygen 18 (18 O, a rare isotope with We were stunned. What could these
most zircons are nearly invisible to the eight protons and 10 neutrons, which high oxygen isotope ratios mean? In
naked eye, the ion microprobe fi res a represents about 0.2 percent of all oxy- younger rocks the answer would be ob-
beam of ions so narrowly focused that it gen on the earth) to oxygen 16 (16O, the vious, because such samples are com-
can blast a small number of atoms off common oxygen isotope with eight pro- mon. A typical scenario is that rocks at
any targeted part of a zircons surface. tons and eight neutrons, which compris- low temperature on the earths surface
A mass spectrometer then measures the es about 99.8 percent of all oxygen). can acquire a high oxygen isotope ratio
composition of those atoms by compar- These atoms are called stable isotopes if they chemically interact with rain or
ing their masses. It was Compstons because they do not undergo radioactive ocean water. Those high-!18O rocks, if
group working with Robert T. Pid- decay and thus do not spontaneously buried and melted, form magma that re-

The tiny zircons from Western Australia

did not GIVE UP THEIR SECRETS easily.
geon, Simon A. Wilde and John Baxter, change with time; however, the propor- tains the high value, which is then
all then at Curtin University of Technol- tions of 18O and 16O incorporated into passed on to zircons during crystalliza-
ogy, also in Australia that fi rst dated a crystal as it forms differ depending on tion. Thus, liquid water and low tem-
the Jack Hills zircons in 1986. the ambient temperature at the time the peratures are required on the surface of
Knowing this history, I approached crystal formed. the earth to form zircons and magmas
Wilde. He agreed to reinvestigate the The 18O/16O ratio is well known for with high !18 O; no other process is
uranium-lead dates of Jack Hills zircons the earths mantle (the 2,800-kilometer- known to do so.
as part of the doctoral thesis of my stu- thick layer immediately below the thin, Finding high oxygen isotope ratios in
dent William H. Peck, who is now an five- to 40-kilometer-thick veneer of the Jack Hills zircons implied that liquid
assistant professor at Colgate Univer- continents and ocean crust). Magmas water must have existed on the surface
sity. In May 1999 Wilde analyzed 56 that form in the mantle always have of the earth at least 400 million years
undated crystals using an improved about the same oxygen isotope ratio. For earlier than the oldest known sedimen-
SHRIMP at Curtin and found five that simplicity, geochemists calibrate these tary rocks, those at Isua, Greenland. If
exceeded four billion years. To our great ratios relative to that of seawater and ex- correct, entire oceans probably existed,
surprise, the oldest dated back to 4.4 press them in what is called delta (!) no- making the earths early climate more
billion years ago. Some samples from tation. The !18O of the ocean is 0 by like a sauna than a Hadean fireball.
the moon and Mars have similar ages, definition, and the !18O of zircon from
and meteorites are generally older, but the mantle is 5.3, meaning that it has a Continental Clues
nothing of this vintage had been found greater 18O/16O ratio than seawater. c ou l d w e r e a l ly b a s e such far-
A A RON J. C AV O S IE Univer sity of Wisconsin Madison

(or expected) from our planet. Almost That is why Peck and I expected to reaching conclusions about the history
everyone assumed that if such ancient find a primitive mantle value of around of the earth on a few tiny crystals? We
zircons had ever existed, the dynamic 5.3 when we took Wildes Jack Hills zir- delayed publishing our findings for more
Hadean conditions destroyed them. Lit- cons, including the five oldest, to than a year so we could double-check
tle did we know that the most exciting the University of Edinburgh in Scotland our analyses. Meanwhile other groups
discovery was yet to come. that same summer. There John Craven were conducting their own research in
and Colin Graham helped us use a dif- the Jack Hills. Stephen J. Mojzsis of the
Evidence of Ancient Oceans ferent kind of ion microprobe specially University of Colorado and his col-
p e c k a n d i soug h t Wildes zircons suited to measure oxygen isotope ratios. leagues at the University of California at
from Western Australia because we We had worked together many times Los Angeles confirmed our results, and
were looking for a well-preserved sam- over the preceding decade to perfect the we published back-to-back technical ar-
ple of the oldest oxygen from the earth. technique and could analyze samples ticles describing our findings in 2001.

w w w. s c ia m . c o m SCIENTIFIC A MERIC A N 63
Zircons from the Jack Hills of Western Australia have changed four billion years ago. Many of these tiny timekeepers also
the way scientists think about the early history of the earth. bear clear chemical signs that oceans of liquid water and
These crystals are the oldest terrestrial materials yet possibly even continents existed on the earths surface at a
discovered hundreds of those identified formed more than time once thought to be molten and fiery.


The oldest age for a Jack Hills zircon 4.4 billion years (red) is an Oxygen isotope ratios in Jack Hills zircon samples (blue),
exact match of two geologic clocks. Two pairs of isotopes which range up to 7.5, are possible only if their source rock
uranium 235 lead 207 (vertical axis) and uranium 238lead 206 formed in a relatively cool, water-rich environment near the
(horizontal axis) form two radioactive timekeepers that start earths surface. Had magma oceans covered the planet when
ticking when a zircon forms. If they are well preserved, their final these zircons formed, their values would have clustered near
ratios plot along a single line (yellow). Dates from other parts of 5.3, as do those of all crystals from hot rock that originates in
the zircon (pink) fall off this line because some lead was lost from the planets deep interior (red).
these areas, but scientists can correct for this damage.
Expected value for Expected value for cool,
the earths hot mantle wet environments
1.20 20 -
Ratio of Lead 207 to Uranium 235

Dates in billions 4.4

of years ago 4.2

Number of Samples
15 -
0.80 Typical
mantle zircons
Oldest zircon Jack Hills zircons
date from earth 10 - (includes all

L U C Y R E A D I N G - I K K A N D A ; S O U R C E : J O H N W. V A L L E Y ( o x y g e n i s o t o p e r a t i o s) ; A A R O N J . C A V O S I E U n i v e r s i t y o f W i s c o n s i n M a d i s o n ( z i r c o n g r a i n s)
>4 billion-year-old
0.40 zircons analyzed
All dates in this region so far)
fall between 4.2 billion 5-
and 4.4 billion years ago

0.00 0-
10 20 30 40 50 60 70 80 4.5 5.0 5.5 6.0 6.5 7.0 7.5
Ratio of Lead 206 to Uranium 238 Oxygen Isotope Ratio

Rounded surfaces of some Jack Hills zircons under
a scanning electron microscope show that wind and

L U C Y R E A D I N G - I K K A N D A ; S O U R C E : S I M O N A . W I L D E C u r t i n U n i v e r s i t y o f Te c h n o l o g y ( u r a n i u m - l e a d ) ;
possibly running water buffeted these crystals over
long distances possibly across a large continental
landmass before they were finally laid to rest (right).
Zircons found near their place of origin retain their
original sharp edges ( far right). The large number of
ancient, rounded Jack Hills zircons suggests their
original source rocks were widespread.

As the possible implications of the other materials that were encapsulated as granitic rock, that evidence would sup-
zircon discoveries spread through the sci- the zircon grew around them. Such zir- port the hypothesis that they are sam-
entific community, the excitement was con inclusions can reveal much about ples of the worlds fi rst continent. But
palpable. In the superheated violence of where the crystal came from, as can the caution is warranted. Quartz can form
a Hadean world, no samples would have crystals growth patterns and the compo- in the last stages of magma crystalliza-
survived for geologists to study. But these sition of trace elements. When Peck and tion even if the parent rock is not gra-
zircons pointed to a more clement and I studied the 4.4-billion-year-old zircon, nitic, although such quartz is much less
familiar world and provided a means to for instance, we found that it contained abundant. For instance, zircons and a
unravel its secrets. If the earths climate pieces of other minerals, including few grains of quartz have been found on
was cool enough for oceans of water ear- quartz. That was surprising because the moon, which never developed a gra-
ly on, then maybe zircons could tell us if quartz is rare in primitive rocks and was nitic, continental-style crust. Some sci-
continents and other features of modern probably absent from the very first crust entists have also wondered if the earths
earth also existed. To find out, we had to on the earth. Most quartz comes from earliest zircons formed an environment
look more closely into the interiors of granitic rocks, which are common in more like the early moon or by some
single crystals. more evolved continental crust. other means that is no longer common,
Even the smallest zircon contains If the Jack Hills zircons came from a perhaps related to giant meteorite im-


pacts or deep-sourced volcanism, but no vosie, now an assistant professor at the five Jack Hills zircons in 1999, the data
one has found convincing evidence. University of Puerto Rico, found such supporting our conclusions have grown
Meanwhile other clues for continen- evidence even within single zoned zir- rapidly. Investigators in Perth, Canber-
tal crust came from trace elements (those cons where the core formed early, say, ra, Beijing, Los Angeles, Edinburgh,
elements substituting in zircon at levels 4.3 billion years ago, with a surround- Stockholm and Nancy, France, have
below 1 percent). Jack Hills zircons have ing overgrowth that formed later, be- now put tens of thousands of Jack Hills
elevated concentrations of these ele- tween 3.7 billion and 3.3 billion years zircons into ion microprobes, searching
ments as well as patterns of europium ago. That the zircons get younger from for the relatively few that are older than
and cerium that are most commonly core to rim is expected because zircon four billion years, and other dating
created during the crystallization of crystals grow concentrically by adding techniques have been applied as well.
crust, which means the zircons formed material to their grain boundary. But Hundreds of newly discovered zir-
near the earths surface rather than in the great age difference with time gaps cons have been reported from several
the mantle. Furthermore, the ratios of between the cores and rims of these par- localities with ages from 4.4 billion to
radioactive isotopes of neodymium and ticular zircons indicates that two dis- four billion years old. David R. Nelson
hafnium two elements used to deter- tinct events took place, separated by a and his colleagues at the Geological Sur-
mine the timing of continental-crust- major hiatus. In more commonly avail- vey of Western Australia have found
forming events suggest that significant able, younger zircons, this kind of core- similarly ancient zircons as far as 300
amounts of continental crust formed as to-rim age relation results from tectonic kilometers south of the Jack Hills. Geo-
early as 4.4 billion years ago. processes that melt continental crust chemists are scrutinizing other ancient

The Jack Hills zircons may be samples

of the worlds FIRST CONTINENT.
The distribution of ancient zircons and recycle the zircons within it. Many regions of the earth, hoping to find the
provided additional evidence. The pro- scientists are trying to test whether sim- first pre-4.1-billion-year-old zircons
portion of zircons older than four billion ilar conditions produced the ancient outside Australia.
years exceeds 10 percent in some samples Jack Hills zircons. And the intensifying search is spur-
from Jack Hills. Also, the zircon surfaces Most recently, E. Bruce Watson of ring improved technology. Cavosie has
are highly abraded and the originally an- the Rensselaer Polytechnic Institute and demonstrated much better accuracy of
gular crystal faces are rounded, suggest- T. Mark Harrison of the Australian Na- analysis and reported more than 20 Jack
ing the crystals were blown long distanc- tional University have reported lower- Hills zircons with the high oxygen iso-
es from their source rock. How could than-expected levels of titanium in these tope ratios that fingerprint cool surface
these zircons travel hundreds or thou- ancient zircons, suggesting that the tem- temperatures and ancient oceans as ear-
sands of kilometers as windblown dust peratures of their parent magmas must ly as 4.2 billion years ago. My colleagues
and still be concentrated together unless have been between 650 and 800 degrees and I are continuing the search using the
there had once been a lot of them? And C. Such low temperatures would be pos- fi rst model of the newest generation ion
how could these zircons escape burial sible only if the parent rocks were gra- microprobe, called the CAMECA IMS
and melting in the hot mantle unless nitic; most nongranitic rocks melt at 1280, which was installed in my labora-
thick continental-type crust was stable higher temperatures, and so their zir- tory this past March.
so as to preserve them? cons should contain more titanium. Many questions will be answered if
These findings imply that the zircons pieces of the original zircon-forming rock
were once plentiful and came from a Zircons Are Forever can be identified. But even if we never
S I M O N A . W I L D E C u r t i n U n i v e r s i t y o f Te c h n o l o g y

widespread source region, possibly a si n c e m y c ol l e ag u e s and I ana- find that rock, there is still much to learn
continental landmass. If so, it is quite lyzed the oxygen isotope ratios in those from the tiny zircon time capsules.
likely that rocks from this earliest time
still exist, an exciting prospect because MORE TO EXPLORE
much could be learned from an intact A Cool Early Earth. John W. Valley, William H. Peck, Elizabeth M. King and Simon A. Wilde in
rock of this age. Geology, Vol. 30, No. 4, pages 351354; April 2002.
Furthermore, the age distribution of Magmatic !18 O in 44003900 Ma Detrital Zircons: A Record of the Alteration and Recycling
of Crust in the Early Archean. Aaron J. Cavosie, John W. Valley, Simon A. Wilde and the Edinburgh
ancient zircons is uneven. Ages cluster
Ion Microprobe Facility in Earth and Planetary Science Letters, Vol. 235, No. 3, pages 663681;
in certain time periods, and no zircons July 15, 2005.
have been found from other eras. My The authors Zircons Are Forever Web site is at
former graduate student Aaron J. Ca- home.html

w w w. s c ia m . c o m SCIENTIFIC A MERIC A N 65