The oldest known bit of Earth

February 25, 2014 • 2:09 pm

The Earth is 4.54 billion years old. We know that not from radiometric dating of rocks on our planet, as the oldest rocks haven’t yet been found, but from dating meteorites that fall on earth from the solar system, which formed around the time Earth did.

But of course that’s not fodder for creationist, for we also have old homegrown rocks, clearly showing that the earth is far older than, say, 10,000 years.

And now we’ve found the oldest bit of Earth yet. As ZME Science and a new paper reported yesterday, it’s a zircon crystal from Australia dated at 4.37 billion years.  The paper with the original report is in Nature Geoscience (reference and link below; free download [I think]).

Here it is.

Photo: University of Wisconsin

And it’s not big:

The geological relic indicates, for one, that Earth’s crust formed shortly after the planet stabilized and formed. John Valley, a University of Wisconsin geoscience professor who led the research, said the findings suggest that the early Earth was not as harsh a place as many scientists have thought.

No doubt, this is an extraordinary find, however, the untrained eye would have surely missed it. Measuring about  200 by 400 microns, or roughly two times the width of a human hair, the tiny gem was luckily retrieved by geologists in 2001 from a rock outcrop in Australia’s Jack Hills region.

The researchers used two ways to check the date: radiometric uranium-lead dating and atom-probe tomography, which uses the actual position of individual atoms in the crystal to check the accuracy of the U/Pb method. I’ll let the readers enlighten us about how this method works, as I don’t fully understand it myself; but the upshot is that the APT dates comport with the uranium lead dates, making hash of the creationist objection that dating methods are unreliable, even in the hands of experts.

The APT was used because of worries that U-Pb dating might be off because lead might have moved within the crystal.  This shows that scientists do know the ways that dating could be off, and have checks for them. And APT showed no evidence of such movement. I’ll show that by simply posting part of the paper’s abstract:

Here we use atom-probe tomography to identify and map individual atoms in the oldest concordant grain from Earth, a 4.4-Gyr-old Hadean zircon with a high-temperature overgrowth that formed about 1 Gyr after the mineral’s core. Isolated nanoclusters, measuring about 10 nm and spaced 10–50 nm apart, are enriched in incompatible elements including radiogenic Pb with unusually high 207Pb/206Pb ratios. We demonstrate that the length scales of these clusters make U–Pb age biasing impossible, and that they formed during the later reheating event. Our tomography data thereby confirm that any mixing event of the silicate Earth must have occurred before 4.4 Gyr ago, consistent with magma ocean formation by an early moon-forming impact about 4.5 Gyr ago

h/t: Ant


Valley, J. W. et al. 2014. Hadean age for a post-magma-ocean confirmed by atom-probe tomography. Nature Geoscience. doi:10.1038/ngeo2075

65 thoughts on “The oldest known bit of Earth

  1. Take that Canadians! We’ve got the record for the Earth’s oldest rocks back again. The site is just 800km north of where I live.

    Does this mean that Western Australia now has the oldest rocks and the oldest fossilised stromatolites? Do we get a prize for it?

      1. The Aussies would get Reg Sprigg to wave Spriggina back in your face – one of the better known members of the Ediacara Fauna. (OK, maybe they’d have to get a relative to wave, or get a relative’s permission to dig him up. And I love the optimism of the Aussies in calling a salt pan a “Lake”, Torrens in this particular case.)
        And before anyone plays the “Burgess are prettier” card, I’ve been reminded in this morning’s reading of the beauties of spinifex textured komatiites. Totally inorganic (~1500K), but so pretty!

      1. He’ll need careful packaging for shipment. Spinifex textured komatiite would be excellent material!
        … and it’s re-boot time, whether I like it or not!

    1. I think technically a “rock” is a conglomerate of minerals or non-minerals. The rock that included that stone is gone long ago.

      You don’t have the oldest trace fossils, if any of the two recent Isua finds are accepted. (Which may be a long shot, since the Australia stromatolites and MISS finds accumulate so rapidly.)

      But yes, you can get a “Noble Prize” for oldest (so noblest) estate. All 200 by 400 microns of it…

      1. While the Isua material is spread over whole square metres. Not very square – decidedly elongate – but well over a square metre in total.

    1. I think that’s the electron back-scatter image. Or cathodoluminescence. Pretty. But the reality is a mutli-zoned translucent brown. Think of something like a large grain of demerara sugar.

  2. [A good article on the specifics of the radiodating, for those of us that hasn’t momentarily access to the original work and like pictures, may be the ars technica article: ]

    I am happy now. This seems to make a nice, predictive timeline.

    – If we don’t look at these results, the same precision models that gives the best date for Tellus (protoEarth) formation can’t constrain the Earth-Moon formation timing much. It is leaving a window of 50 – 200 million years after Tellus formed. [Chronometry of Meteorites and the Formation of the Earth and Moon, Kleine & Rudge, Elements, 2011]

    At the same time there were results that could push the Earth-Moon formation at least 150 million years after Tellus formation. [ ]. And the Late bombardment has always been rather unconstrained, with problems of absolute crater dating and Apollo rock dating (if it was mostly Imbrium ejecta or not).

    – But accepting these results, the observations come out of the same predictive model.

    The iffy part is that a cool early Earth as suggested by the stable oxygen isotopes requires that the Late bombardment wasn’t too energetic. [Link to reference is in the ars technica article above.]

    But timely there is a new paper out that claims that at most 10 % of the crust surface was molten at the same time, and that the crust was not molten through.

    “Today, it is difficult to determine the details of how this ‘impact epoch’ affected the young planets. The new study estimates the thermal effects of a period known as the Late Heavy Bombardment (LHB) on Earth. … Heat generated by the impacts left up to 10 percent of the planet’s surface covered with melt sheets more than a kilometer thick.”

    [ ]

    The LHC was mild, or at least survivable, and potentially reconcilable with the isotope data. Indeed, last year saw two independent results on putative trace fossils from prokaryotes at Isua before 3.8 billion years ago. (Isua BIFs and organic carbon, respectively.)

    Abiogenesis seems to be no severe constraint, as it most often comes out as rapid, perhaps 0.1 Ma in some dirty RNA theories. The nice part here is that the deep phylogeny of genes that may have started sometime later takes time. The earliest gene root likes to sit at least 4 billion years back.

    Now such results won’t stress geological uncertainties appended to Late bombardment and putatives of late ocean formation, late crust formation and late Moon formation. And vice versa.

    “These rocks don’t lose their ‘topes, zircons are a man’s best friend”.

  3. Carl Zimmer has a new article in Scientific American about another possible site in Canada that may be of similar age. This one is more controversial. You can read about it here.
    On the tomography data, I am no expert here but Zimmer explains that some have worried that uranium and lead atoms might drift apart in these ancient crystals if they were repeatedly softened by heating over time. This would create erroneous results in the small samples taken to determine their age. The tomography procedure maps the positions of atoms to check if they had drifted apart. The results seem to show they do not move much. Yea, Zircons!

    1. It sounds as though the tomography is not an independent dating method, but an extension of the U/Pb ratio method. Doesn’t the tomography verify that the U and Pb atoms are part of the same decay process by being in the right spacial relationship? Or am I off in lala land?

      1. Precisely. The tomography assures us that the lead sampled really is derived from the uranium sampled, with no migration, so that their relative abundances can be used with confidence in the standard equations. (The uranium is the ultimate parent as far as the dating is concerned, any heavier elements in the relevant decay series having far shorter lifetimes and having decayed long before the Earth itself was formed)

    2. On the tomography data, I am no expert here but Zimmer explains that some have worried that uranium and lead atoms might drift apart in these ancient crystals if they were repeatedly softened by heating over time.

      This is precisely the point that the APT (Atom Probe Tomography) was intended to address.
      The concern, not ill-founded, was that there might be small areas of daughter-lead concentration compared to the bulk of the crystal, and that by coincidence, these sites were preferentially sampled for age dating by microprobe (SIMS – Secondary Ion Mass Spectrometry, IIRC ; my copy is in the cabin, along with my night-shift minion) because they looked undisturbed.
      Part of that concern was correct : it does look as if the recoiling lead daughter nuclei (recoiling because they’ve just spat an energetic alpha particle out on the opposite vector) have been concentrated into clusters. BUT, the clusters are on the order of 10nm (nanometres, 10^-9m) across, spaced at around 22nm (though they’re not isotopic ; the actual average volumetrically is a spacing of 50nm ; in Fig 3a there is a distinct visible alignment of clusters transverse to long axis of the “needle”). Since the size of a SIMS analysis pit is around 10000 to 20000nm, then each pit would have sampled around a million concentration clusters and the depleted areas between them. Or, in short, the SIMS ages are accurate.
      There are some more arguments that the ages presented a decade ago are accurate, but that’s the main one.

  4. > free download [I think]

    I don’t think so: “Purchase article full text and pdf: €30”

    Seriously, who’s gonna pay that, except the universities themselves, who have to (after they funded the research in the first place?)

    1. I have no idea who they think is actually going to pay these ludicrous prices and I would love to see some statistics on it.

    2. It greatly annoys me too. However while downloading this article a couple of days ago, I’ve discovered that the network I’m accessing it through has a site licence.
      So, when I’m not busy with “work” work, it’s “slurp the archive” time.
      (Drop me an email. I take the “fair use” criterion as being “would I walk down the corridor to the photocopier?” ; you pass.)

    3. Well you kinda answered your own question. Their customers are seen primarily as institutions etc., who will buy a whole years’ run of journals. Then, second, scientists who will buy individual articles using grant money or institutional money. Their business model isn’t set up around laymen wanting to read the occasional article.

  5. No doubt, this is an extraordinary find, however, the untrained eye would have surely missed it. Measuring about 200 by 400 microns, or roughly two times the width of a human hair, the tiny gem was luckily retrieved by geologists in 2001 from a rock outcrop in Australia’s Jack Hills region.

    Geology ignoramuses like me would love to know more about this step…they knew to ‘recover’ this particular bit because…??

    1. What you do is gather up as many kilograms of rock as you can haul away, grind it down to a certain sized powder, and then feed the powder into a vibrating water-washed table that functions like an old fashioned gold pan. Here is a picture of one:

      This table separates the denser minerals from the ordinary quartz and feldspar that dominates ordinary rocks. These ‘heavy minerals’ can include magnetite and a host of others in addition to the zircon crystals that were dated in this study. There are a couple of ways you can separate the various heavy minerals from each other, but the end result always requires someone weeding out individual grains to date under a microscope.

      In the original 2001 study, they zapped roughly 20 (I think) individual grains that were selected as looking the oldest (older grains are more radiation damaged by alpha recoil from U decay and therefore darker). They got a range of ages from about 3.3 to 4.4 billion years. This age range reflects the ages of the rocks that eroded for form the Jack Valley Conglomerate in western Australia.

      So, they picked this sedimentary deposit because it was so darn old knowing that the sedimentary grains in it originally crystallized in rocks that were still older. After dating a bunch of zircon grains, they found the one that is about 4.4 billion years old. Ta da!

      1. Impressive. I had read somewhere that zircons grow new layers (like the one shown in this post) when the rocks they are in are melted. So scientists can measure the age of different layers and get a different date for each. At least that has been my impression. Here is an article that seems to be saying that.

        1. So scientists can measure the age of different layers and get a different date for each. At least that has been my impression. Here is an article that seems to be saying that.

          Short version : Yes.
          The zircon crystal in question, “Sample 01JH36-69” (2001 season, Jack Hills, and some geographical or acquisition sequence number), has a “rind” or overgrowth some 10 to 20µm thick, which dates to “3400 Ma” (average 3391, 3379, and 3430 ; there may be asymmetric error bars on those, but their mean is 3400).
          You’ll note that the thickness of the “rind” is similar to the size of the SIMS measurement spot.

      2. Thanks for that wonderful explanation!

        You’d be a good person to ask: how common are ~4 GA rocks, and how readily might some random schmuck acquire one? Are they like trilobites that any gift shop attached to a science museum is likely to have, or are they the sort of thing that only the people doing the research are going to have?


        1. I could not resist looking. You can get a purported piece of banded iron from the Jack Hills formation on e-bay. It is pricey. But what is similarly profound are fossils of stromatolites. Those can be purchased on e-bay and other places for pretty cheap. I think the age for most of the common ones are 1-2 Gya.

          1. I frequently bring home little bits of ironstone (from banded iron formations, whether or not actually banded in hand specimens) in various stages of weathering and recrystallization from the Hamersleys (a bit further north and slightly younger), because they’re pretty.

            There’ve been a number of sites I’ve wanted to go back to with a truck and a hoist, I can see it being spectacular for bathrooms, benchtops and feature walls. If I ever build a house…

            1. There’ve been a number of sites I’ve wanted to go back to with a truck and a hoist,

              When I went for interview at my university, I described myself as having an “insatiable desire for pretty rocks.”
              They should have taken the warning.
              I know (well, knew) people who’d had the attic of their house collapse into the living areas because of the pile of rocks in there. Silly people : that’s why you get a house with a garage.

          2. I was considering relating the position of the UK’s only easily accessible outcrop of billion-year old biogenic stromatolites. But on second thoughts, I’d pass you the directions in a rain-swept car park in mid-summer in a village on the far north west coast. You’d need to demonstrate some commitment first, because otherwise the deposit would be worked out in months.

  6. Two non-meteoritic dating methods for the whole earth; both behind paywalls but the abstracts, free, will serve most of us: whole earth lead (see and helioseismology (

    None of which detracts from this work. Locating the atoms so we know they haven’t moved; and the implication that the earth not only existed but had liquid water at that time.

    I think you find the zircons by grinding up the present-day rock and flotation. I don’t know how you decide which ones to look at; information,anyone?

    1. Locating the atoms so we know they haven’t moved;

      … more than 30-50 nanometres.
      Short version : “much”.

      I don’t know how you decide which ones to look at; information,anyone?

      You look at each one under the microscope : “fractured, has inclusions : reject” ; “good shape, no fractures, internal zoning ; accept”
      They examined around 100,000 grains.
      Believe me, that is an Everest of sampling. Herds of blinded postgraduate students are probably roaming feedlots i nthe Outback, used up and spat out by the herculean efforts involved. Bald camels roam the landscape, denuded for the camel-hair brushes to shift each tiny grain into the “accept” or “reject” pile.
      Doing point-counting on a couple of thousand points in a pseudotachylite is enough to simply send one screaming for the hills at the idea of doing this!

      1. Oh, I dunno. After feeding the crusher, running the screeching grinding disks, minding the shaker and fussing with heavy liquids, It’s actually sort of relaxing.

        1. Hmmm, I can see that. My specimens, I just had to sandpaper the tips off my fingers preparing the slides.

            1. I’ve never really believed that bovine excrement about suffering and “the great outdoors” being “character building”. “Character discovering” or “revaling” I’d be perfectly happy with. But “building”? Pass me the waders and the river – I’ve got some stables to clean.

  7. A far as I am concerned it is the detection technique that is the real story here. The age of the zircon is only of note until the next oldest zircon is found.

    I have spent a few hours reading what I can about APT and I am truly awestruck. I did not realise that we were able to build up a 3D isotope-by-isotope model of a material sample. Admittedly it is a small material sample, no more than 100nm or so in radius and about 1um in length, consisting of perhaps a few 100 million atoms.

    Unfortunately the Springer “Local Electrode Atom Probe Tomography: A Users Guide” ebook costs £72.00, so unless someone decides to write a Dummies Guide I will continue to remain largely ignorant about what appears to be a truly remarkable analytical technique.

    1. Unfortunately the Springer “Local Electrode Atom Probe Tomography: A Users Guide” ebook costs £72.00,

      Do you not have an “Inter-Library Loan” system at your public library? Should cost about one peppercorn.
      But you might have to read it in a week.

  8. To me (a geologist), the coolest thing about this very ancient sample is that the oxygen isotopes in the zircon crystal suggest that surface water helped trigger the melting that produced the igneous rock within which the crystal crystalized.

    The only way to get surface water deep enough to cause partial melting of rocks is along subduction zones (where ocean slabs sink beneath, for example, South America.

    Thus, it seems that there were oceans deep enough to fill at least the trenches along subduction zones. Since the Earth finished forming roughly 4.53 Byr ago, it must have rained quite a lot over the first ~150 million years to have a least partially filled the oceans!

    1. Thinking it through…assuming the water was already here, with all that heat there’d be an awful lot of steam, and I’d guess near 100% humidity everywhere. That would seem, in turn, to suggest that dew would condense on any surface cool enough to not re-vaporise the water. Even between Biblical-style torrential downpours, I’d expect the oceans to fill quite rapidly just from a perpetual mist / drizzle.

      And, when one thinks of it in those terms, it would seem like the miracle would be for life to not arise….


    2. The only way to get surface water deep enough to cause partial melting of rocks is along subduction zones

      Well …
      In our current plate tectonic regime, yes. However it is commonly accepted that the higher heat flow of the early Earth (consequent on a lot of the medium-half-life nuclei being more common then) lead to a distinctly different style of tectonics. What that style was is a much more difficult question. And very unsettled. When the styles changed is also an open question.

  9. U-Pb dating using zircons has been done for years and is a standard geological technique, though the methods and supertoys change. That might not be obvious from the article. The big news is the extreme age of the zircon.

    Note that the age is that of the zoned part of the crystal, not necessarily the overgrowth (gray rim in the photo) or the rock that contained the zircon! Zircons can weather out of the original igneous host rock, such as a granite, and be “recycled” into sedimentary rocks which can be metamorphosed and still sometimes preserve the original U-Pb values in the crystal.

    Mineralogy data for zircon (and 1661 photos!):

  10. “a 4.4-Gyr-old Hadean zircon”

    I had not seen years reported in Gigs before this. Nice high-tech touch. Amazing technology, too.

  11. I was surprised (and very happy) to find that the paper was available for download. My undergrad degree is in “Interdisciplinary Sciences” (I started as an EE major but all that rapid fire math was just too much for my ADD-addled brain) and I was required to take upper-level courses in a number of disciplines. I was continually frustrated by how difficult it was to obtain full-text, peer-reviewed papers in the earth sciences. I eventually concluded that the reason it was rather easy to find papers in high-energy physics (e.g. was easy was because nobody is likely to get rich after confirming the existence of the Higgs mechanism, but there are fortunes to be made out of digging stuff up out of the ground-hence the frequency of earth science research being behind pay-walls.

  12. Nice-looking bit of rock. That’s a bit better picture than the one our paper (the New Zealand Herald) chose to accompany its account – true to its tradition of bizarrely incongruous library pics, the Herald had a gem-cut white stone (diamond? zircon? who knows?) heading its account. Wish I could poke the editor with a stick every time they do that.

  13. Is there a term for the kind of deep time ‘vertigo’ one can experience when contemplating these vast stretches of the planet’s history?

    1. I don’t know, but it may help to think of the time dimension as being horizontal rather than vertical. Imagine driving along it (e.g. a continent-spanning 4400 km to see where this specimen formed, at 1 megayear per km = 1 year per mm), not climbing up and down a cliff.

    2. We call it “getting Deep Time” ; I’ve seen it knock people over.
      If you ever read the Hitchhiker’s Guide books, the “Total Perspective Vortex” has a similar effect : you get hit by how utterly minusculy insignificant you, your works, and you species’ works are. (Unless you’re Zaphod Beeblebrox, of course.)
      I imagine that some of the god-squad get the same feeling of grovelling insignificance, and invent an invisible pink unicorn big brother to assuage their fear of the indifferent outside world.

      1. the Hitchhiker’s Guide books,

        “To The Galaxy” ; not the ones for Ulan Bataar and such like parochial neighbourhoods.

        1. Well of course.

          I must admit I ignored HHGG for years after it came out, under the impression that anything said to be a publishing smash hit with that title could only be a pop-science coffee-table book for the chattering classes, that had stolen the ‘Hitch-Hikers Guide’ part of the title from the popular travel series…

          1. I watched H2G2 when it was first broadcast on the radio : I still remember coming back from school to find it on as we were making a brew. In a class of it’s own.

              1. DNA always reckoned that the sound quality in the books was much better than that on the radio series ; while the staging and lighting on the radio was better than on the TV.
                He was making a point, with humour.

Leave a Reply