Radiocarbon dating only takes us back some 50,000 years. This makes it a much smaller threat to Young Earth creationists than, say, lead-uranium dating, which takes us back billions of years. So why do creationists single it out for attack? Because there are indeed problems with the most simple-minded application of the method, and it does not matter to the creationist that these problems have long since been solved. Creationists think, and argue, more like lawyers than like scientists. In the courtroom, changing your story under cross-examination will destroy your credibility, and yet this is what scientists do all the time. Scientists accept that even the most well-established findings are subject to revision and refinement; lawyers, like theologians, seek certainty whether the data justify it or not.
This post is derived from a talk given by my friend Leonard Sym to Glasgow Skeptics in the Pub, and like Leonard I will follow Rapoport’s rules for debate, which specify that one should first summarise one’s opponents’ position in terms they would accept, next, list points of agreement, then point out what one has learnt from one’sopponents, and only at that stage embark on criticism.
I start with a simplified summary of the principles behind radiocarbon dating, without which the discussion would be meaningless. Most radiometric methods depend on measuring the amount of a parent radioactive isotope present in the sample, and the amount of the daughter into which it decays. Add up the amount of parent still present and the amount of daughter, and that gives you the amount of parent present initially. If you know the rate constant for the decay, you now have enough information to work out how long has passed since the parent material was put in place. You can determine the rate constant by taking a known amount of parent, and counting the number of decays per second, as measured with a Geiger counter or a more reliable and up-to-date instrument such as fluorescence counter.
In the case of radiocarbon dating, the parent is carbon-14 and the daughter is nitrogen-14, which is lost from the sample. So we can’t find the initial amount of parent in the way described above, because we don’t know the amount of daughter. This seems like a dead end, until we remember where carbon-14 comes from. Carbon-14 is formed in the upper atmosphere by the effects of cosmic ray bombardment on nitrogen, is rapidly converted to carbon-14 dioxide, and then mingles with the rest of the CO2 in the atmosphere (see Figure). If we assume a steady rate of bombardment, that means we will have a steady rate of production of carbon-14, and a steady state abundance of carbon-14 in the atmosphere, where the amount decaying each year is equal to the amount being formed.
Now consider what happens during the life of an organism, and after its death. As long as it is alive and metabolising, it will exchange carbon with its environment, taking it in directly as carbon dioxide by photosynthesis (for a plant) or indirectly as food (for an animal). At this stage, the proportion of carbon present as carbon-14 will be directly dependent on that in the atmosphere. But as soon as it stops metabolising, it stops exchanging, and the proportion present starts decaying according to the radioactive decay law, with a halflife of 5730 years. So it looks as if we can just use the proportion present in the atmosphere right now as a measure of the initial proportion, and compare it with the proportion remaining.
So far, so good. Now let me list the creationists’ objections:
1) As in all radiometric dating, the decay rate is assumed to be constant. What if this isn’t true?
2) The production rate is assumed constant. But this is unrealistic, since the intensity of cosmic ray bombardment is known to change over time
3) For 150 years, and especially in the last 50 years, we have been adding carbon dioxide from fossil fuels to the atmosphere, diluting the radiocarbon since all the radiocarbon in the fossil fuels will have long since decayed
4) Considerable amounts of carbon-14 were added to the atmosphere by nuclear testing in the 1950s, further undermining the assumptions
5) What if carbon-14 is less readily taken up than carbon-12 by plants? Won’t this undermine the reasoning?
6) We can check the method by applying it to materials whose age we know, but this will only serve where we have a good historical record, and this record only goes back, at best, some 5000 years
7) The Genesis flood, which in Young Earth accounts is responsible for the formation of our fossil fuel deposits, would have further upset the clock by burying huge amounts of carbon-12. Moreover, it could have been associated with an increase in the rate of carbon-14 production, making pre-flood specimens look much older than they really are.
With the exception of the first and last, all these objections have some degree of plausibility, but unfortunately for the creationists they have all long since been answered, many of the answers being set out by Willard F. Libby, inventor of the method, in his 1955 book on the subject.
1) Radiometric decay constants are just not the kind of thing that could change, unless everything else changes at the same time. We have known since the work of George Gamow in 1928 that radioactive decay is what is known as a quantum mechanical tunnelling effect, and that its rates depends on such things as the strength of nuclear and electrical forces, the mass of fundamental particles, and Planck’s Constant h, which gives the scale for all quantum mechanical phenomena. If any of these had been different, we would not have had the same kind of physics and chemistry that we have today. But we know from their structure that ancient rocks were formed under the same rules as we have today, because they contain the same kinds of elements combined to make the same kinds of minerals. The creationists have published theoretical curves for changes in decay constants, but these have no basis in science, and are generated merely to make observations fit the biblical timeline.
2) From the outset, radiocarbon dating has relied on calibration, using objects for which dates were known from historical records, then tree ring counting extending back 10,000 years. This method works because all but the outermost layers of a tree are metabolically inert, and out of circulation. The most recent calibration comes from organic debris in varves (annual layers of sediment) deposited in a lake (Lake Suigetsu, in Japan) that happens to be free of turbulent inflows. This has made possible the establishment of a calibration curve going back 52,800 years.
3) and (4) There have indeed been major disruptions since 1950, but no one uses radiocarbon dating for such recent material. The situation in 1950 is regarded as a baseline, material from that year is the standard for comparison, and 1950 is the “present” in conventional dating of “years before present”
5) It will surprise many people to learn that plants really do take up carbon-14 less readily than carbon-12. One of the lies you were probably told at school is that all isotopes of the same element have precisely the same chemical properties. This is not true, and generally speaking, heavier isotopes are slightly more sluggish in their chemical reactions. this gives rise to the process known as isotopic fractionation.
This effect has been measured for photosynthesis. In addition to very small amounts of radiocarbon, atmospheric carbon dioxide contains roughly 1% of the stable isotope carbon-13, the remainder being carbon-12. Carbon dioxide in plants is, as expected, slightly depleted in carbon-13 relative to carbon-12, and the effect is far from trivial; around 27 thousands of carbon-13 abundance for most kinds of plant. We expect the effect to be twice as large for carbon-14, which, using the known 5730 year halflife of carbon-14, corresponds to 435 years; not trivial when dating historical artefacts. However, exactly the same effect will apply to the material used to set up the calibration, and the errors will systematically cancel out. Ideally, the fractional abundance of carbon-13 should be measured, as well as that of carbon-12, to calibrate out any minor fractionation effects, and this is less arduous than it sounds because nowadays carbon-14 abundance is measured by direct counting in a mass spectrometer rather than, as in the original studies, indirectly inferred from sample radioactivity.
For plants, it is straightforward to match like with like. Not so for material derived from animals, where the total amount of isotopic fractionation will depend on their diets, and also on what they have been eating.
6) This objection would have had some force in 1946, when the method was newly developed. However, as already explained, we now have direct calibration back to 52,800 years before present, beyond which the amount of remaining carbon-14 is so small that using current techniques the method becomes useless.
7) This is pure special pleading. If carbon-12 had been buried in the flood, the appropriate amount of carbon-14 would have been buried with it. And the ideas of a changed rate of production or decay have been dealt with under (1) and (2) above.
There are other “objections” based on the obvious fact that organisms like cave water snails, alive right now but deriving their carbon from limestone, will have radiocarbon apparent ages measured in thousands of years. I have discussed this before. And marine specimens will always contain less carbon-14 than terrestrial specimens of the same age, because of the time it takes for mixing between the atmosphere and surface waters, and again between surface waters and the depths.
So how should we respond to the self-styled “creation scientist”? The first, and most difficult, thing is to realise that he is been perfectly sincere. He is certain that his reading of the Bible is true; but the facts of geology are also true; and it is therefore his mission to create an account that reconciles the two. If this means the mountains must have skipped like rams, then that is how they must have skipped. He will feel no more absurd at this point, than the cosmologist feels in invoking a time when our Universe was smaller than a tennis ball and its temperature was trillions of degrees. Counter-arguments will be dismissed as so many minor anomalies that will no doubt be explained away in due course. If the creationist repeats long-refuted claims, that is because he believes that there are refutations of the refutation, even if he cannot immediately call them to mind, or does not have time to explain them properly. He will remember the weaknesses of his opponents’ arguments, and attack them, while suppressing the recollection of their strengths, and in the process he will create, and then triumphantly destroy, a series of straw men. You and I of course would never do such things, but your friends might when it comes to defending emotionally precious but logically fragile beliefs; consider, for example, what passes for political discussion in your favourite pub or chatroom.
And what does this mean for debating with creationists? Simply this: don’t do it. Such a debate, unlike a discussion between people willing to learn from each other, is a zero-sum game. He will project simplicity, sincerity, and certainty, and when you come to reply, you will sound as if you are making excuses. He will present anomalies (did I mention those 2000-year-old water snails?), and when you explain the special circumstances, you will be the one who seems guilty of special pleading. His followers will end up confirmed in their convictions, as will yours, and those in the middle will come away confirmed in their own initial conviction that there are two sides to the story, both worth hearing. Which there aren’t.
But does that mean that we can learn nothing from the creationists? Not at all. In terms of Rappaport’s rules of debate, the scientific community had already come up with arguments (1) through (6), and taken the necessary countermeasures, and so cannot be said to have learnt from the creationists. But both Leonard and I have learnt a great deal from examining the creationist claims. Be smart, and learn from everyone.*
1] It is of course necessary to eliminate errors caused by the movement of material, or the presence of daughter in the initial material. There are standard techniques for doing this, for instance by measuring non-radiogenic isotopes of the daughter material, and, these days, by microsampling of single crystalline grains
2] Even if it is not lost as N2 gas, it will be undetectable against the background of organic nitrogen compounds already present
*Ben Zomah, Mishnah Pirkei Avot 4a
** extra credit for spotting (a) the misleading labelling in the diagram (h/t John Gribbin), (b) the reference in the text to Psalm 114
There are other creationist objections to radiocarbon dating, based on sample contamination or simple misinterpretation of data, but these have been discussed elsewhere and need not detain us.
Lecture scene from Glasgow Skeptics in the Pub Facebook page. Atmospheric carbon-14 diagram public domain, by Hokanomono via Wikipedia. North Ronaldsay sheep by Liz Burke, CC BY-SA 2.0, https://commons.wikimedia.org/w/index.php?curid=3499298. Radiocarbon cycle schematic from ANU Radiocarbon Dating Lab materials
It looks as if life on Earth just got older, and probably easier. Tiny scraps of carbon have been found inside 4.1 billion year old zircons, and examination shows that this carbon is most probably the result of biological activity. This beats the previous age record by 300 million years, and brings the known age of life on Earth that much closer to the age of Earth itself. The implication is that life can originate fairly quickly (on the geological timescale) when the conditions are right, increasing the probability that it will have originated many times at different places in our Universe.
The Solar System, it is now thought, formed when the shockwave from a nearby supernova explosion triggered a local increase in density in the interstellar gas cloud. This cloud was roughly three quarters hydrogen and one quarter helium, all left over from the Big Bang some 9 billion years earlier. It had already been seeded with heavier elements produced by red giant stars, to which was now added debris from the supernova, including both long-lived and short-lived radioactive elements. Once the cloud had achieved a high enough local density, it was bound to fall inwards under its own gravity, heating up as it did so. The central region of the cloud would eventually become hot enough and dense enough to allow the fusion of hydrogen to helium. A star was born.
The heavy elements (and in this context “heavy” means anything heavier than hydrogen and helium) in the dust cloud surrounding the nascent Sun gave rise to the rocky cores hidden within the outer giants Jupiter, Saturn, Neptune and Uranus, of the outer reaches of the Solar System, and to the rocky inner planets, Mercury, Venus, Mars, and, of course, to Earth and everything upon it. We are stardust.
The asteroids are made out of material that was never able to come together to form a planet, because of the competing gravitational pull of Jupiter. Asteroids are continually bumping into each other, scattering fragments, and some of these fragments fall to earth as meteorites. The Hubble Telescope has given us images of star and planet formation in progress. Such is our modern creation myth, magnificent in scale, and rooted in reality.
The oldest solid objects in the Solar System are calcium-aluminium rich grains, the most refractory of all the materials to condense out of the gas cloud. These are now known from a refined form of uranium-lead dating  to have formed as much as 4,568.2 million years ago, give or take a very few hundred thousand years either way, and that is now the accepted best estimate for the Solar System’s age. A remarkable feat, to fix this to within around 1% of 1%. As time went by, and the outer regions of the gas cloud radiated away their energy, more materials condensed out, and the grains grew and stuck together by contact and eventually by their own gravity. Thus we went from grains to pebbles to larger objects to planetesimals and eventually to the planets as we know them. The final stages were marked by increasingly violent collisions, culminating in the collision between the proto-Earth and a Mars-size object that gave rise to the present Earth-Moon system, and rounded off by what has been called the Late Heavy
The energy of the collisions will have caused melting, even before the formation of full-scale planetesimals, and the separation of the molten bodies into a metal-rich (mainly iron) core, and a less dense, oxygen-rich outer mantle. It is Earth’s iron core that is responsible for its magnetic field, and this field in turn shields us from the constant bombardment of charged particles emanating from the Sun, which would otherwise have stripped away our atmosphere. Elements like platinum and gold (so-called siderophiles, or iron-lovers) concentrated in the core, which is one reason why they are so rare at the surface, while elements such as oxygen, calcium, magnesium, aluminium and silicon are lithophiles, or rock-lovers, and concentrated in the mantle. Fortunately, the highest melting point rocks, which are thus the first to solidify, are less dense than average, which is why Earth has a solid crust floating on the surface of the mantle. The precious metals are all much stronger siderophiles than iron itself, which forms a strong bond with oxygen and is one of the most common elements in the crust and mantle, as well as being the main constituent of the core. The Late Heavy Bombardment explains the craters on Mercury, the Moon, and Mars. No such craters survive on Earth, but that is because weathering and plate tectonics have completely reworked the surface.
We can learn a lot about the history of these processes from the distribution of the different elements, and even of individual isotopes, especially radioactive isotopes and their decay products. For example, hafnium-182 is radioactive, with a half life of slightly under 9 million years, decaying to tungsten-182. Hafnium is a lithophile, and tungsten a siderophile. So if core formation is slow on this timescale, most of the hafnium-182 from the supernova debris will have had time to decay to tungsten, and will vanish into the core. But if core formation is relatively fast, the hafnium-182 will remain in the rocky phase, where the tungsten-182 derived from it will end up stranded.
We can also sometimes learn about how a material was formed by looking at the ratio of different non-radioactive isotopes. Almost all elements occur as more than one isotope, with the same number of protons and electrons, but different numbers of neutrons. You may well have been told at school that isotopes, despite their have different masses, have identical chemistry, but this is not quite true. Generally speaking, because of quantum mechanical effects , different isotopes have very slightly different chemistries, and small deviations in their relative abundance provides clues to a sample’s history.
Using many detailed arguments of this kind, we come up with the following sequence:
- Beginning of solar system, 4,568 million years ago (see above)
- Collisions between planetary embryos, and partial melting of resulting meteorites, within a very few million years of that beginning
- Accretion of Earth under way within 10 million years of beginning
- Earth-Moon system formed, between 30 and 100 million years from the beginning. Formation of the Earth’s liquid core would be complete at this stage, although the formation of the solid inner core is remarkably recent by comparison (around 1,000 to 1,500 million years ago)
- Oldest rocks on moon 4,460 million years old, (dating Moon’s oldest crust to within a very few tens of millions of years after its formation)
- Oldest rocks on Earth, 3960 million years old, with evidence for an older (4000 to 4200 year old) component
- Late Heavy Bombardment, around 3,900 million years ago, as estimated by dating craters on the Moon.
It was at one time assumed that the Late Heavy Bombardment would have heated the Earth’s surface sufficiently to destroy any life forms in existence at that time. But careful estimates of the total heating effect show that this is not the case, even at the surface, while bacteria obtaining their energy from reactions involving minerals have been found 2.8 kilometers below the surface.
The Jack Hills of Western Australia are of enormous interest to geologists. The rocks that they are made of are thought to have been originally laid down some 3,600 million years ago, as deposits from river deltas, although they have undergone many episodes of transformation since then. They are of special interest because the delta deposits contained zircons that were already, at that time, hundreds of millions of years old; tough grains of impure zirconium silicate from the already ancient mountains, eroded out by the streams that fed the deltas, and transported and buried there unchanged. These have inspired a truly heroic effort from geologists; one paper, in its title, refers to “The first 100,000 grains.” Two separate research groups have reported that the oldest zircons found there, dating back to 4,400 million and 4,300 million years ago, show evidence for the presence on the planet of liquid water , which is generally regarded as a necessary condition for the emergence of life.
Necessary, but not sufficient.
We turn now to the oldest evidence for life on earth.
Hard fossils of complex organisms appear in abundance around 545 million years ago, at the base of the Cambrian, although the record actually dates back to at least 575 million years, and we can stretch this back to 610 million years if we include fossilised traces, such as burrows (here, Ch. 7, updated here and here), and much further if we regard some of the mixed bag collectively termed “acritarchs” as complex. If we want to go back much further, we will be relying on evidence from single-celled organisms, which is always less clear-cut and more open to alternative explanations. However, such organisms can form mats, with a characteristic texture that develops from horizontal layers of dead organisms, with trapped soil particles between them. This leads to the development of what are known as stromatolites, domed multi-layered structures that persist to the present day. Modern stromatolites, at least, are quite complex communities of cyanobacteria, single celled organisms capable of photosynthesis, with different kinds of bacteria, using different wavelengths of light, found at successive levels. Stromatolites are found throughout the fossil record; they were at their most abundant some 1,500 million years ago, but are now found mainly in highly saline lagoons, where grazing creatures, which disturb their formation, cannot survive. The oldest fossil stromatolites are found embedded in 3,430 million year old chert (silica rock), and if we make the reasonable assumption that continuity of form represents continuity of kind of organism, it follows that diverse communities of photosynthesising bacteria were already in existence at that time.
There are claims of microfossils of chains of bacteria, going back to 3,600 million years ago, but these are little more than dark smudges embedded in chert, and their interpretation remains controversial. Moreover, rocks of this age or older have all undergone considerable change, having been subjected at various times to great pressure and high temperatures. To go back further, we have to resort to more indirect kinds of evidence.
Carbon occurs on Earth as a mixture of two main isotopes, carbon-12 (99%) and carbon-13 (1%). There are also traces of carbon-14, used in radiocarbon dating, but this has a half life of only some 5700 years and apart from contamination is effectively absent from materials over a million years old . It has been known since 1939 that the isotopic composition of carbon in plants is different from that found in the carbon dioxide from which it is derived; plant carbon, and materials derived from it, are “light”, meaning that they have a measurably smaller proportion of carbon-13. This is as expected  from quantum mechanics, which predicts that common dioxide containing carbon-12 will be slightly more chemically reactive than that containing carbon-13. The excess of carbon-12 is, of course, inherited by all materials derived from plants, such as animals (which eat them), and fossil fuels. Indeed, one of the many ways in which we know that the recent unprecedented rapid increase in atmospheric carbon dioxide is the result of our burning fossil fuels, is the increasing proportion of carbon-12 in atmospheric carbon dioxide over time.
In 1995, I had the privilege of visiting the laboratories of Gustaf Arrhenius at the Scripps Institution of Oceanography, La Jolla. There I met a Ph.D. student, Steve Mojzsis, who has gone on to pursue a distinguished career in isotope geochemistry. Steve is now Professor at the University of Colorado at Boulder, and his research group was responsible for several of the findings described above. As his Ph.D. problem, Steve was examining 3,800 million year old sediments from Greenland, which were known to contain carbon slightly, but perhaps not conclusively, lighter than expected. Within these rocks, he found grains of hydroxyapatite, which is a very tough form of calcium phosphate, essentially the same as the material your teeth are made of. And within these grains were tiny granules of carbon.
What happened next was made possible by advances in scientific instrumentation, and specifically in the development of what is known as iron microprobe mass spectrometry (more fully, Sensitive High Resolution Ion Microprobe or SHRIMP). This is just what the name implies. A beam of charged particles (ions) is accelerated and focused, and used to drill away at an area of the sample a hair’s breadth across. The fragments blasted out by this process are then fed into a mass spectrometer, which sorts out the different isotopes. When the carbon granules were examined in this way, they were found to be within the range expected for organic material arising by photosynthesis. So these granules were biological in origin, and the earlier inconclusive results were the result of averaging out organic and inorganic material.
Science does not provide proofs, at least not in the sense that mathematics provides proofs, and there are alternative non-biological routes to light carbon. But these involve reactive metals that would not have been present in the crust after core formation, and in any case such processes would not account for the segregation of the light carbon within granules. And so, while scientific conclusions are always in principle subject to being overturned by new evidence, my own view is that it would be unreasonable to deny this evidence for life 3,800 million years before the present.
Steve’s record stood for 20 years, but has just been spectacularly broken, as a result of the zircon screening that I mentioned earlier. Some of the oldest zircons contain flecks of carbon, visible under the microscope. One of these was selected for special examination, cut open, and the carbon examined. Radiometric dating of the freshly cut zircon surface gave a date of 4,100 million years old, while the carbon itself turned out to be light, in the range expected for what had once been living material, with the carbon having been derived from carbon dioxide by photosynthesis. Thus we can now say, with a surprising degree of confidence, that there was life on Earth, and indeed life capable of carrying out the complicated sequence of reactions necessary for photosynthesis, 4,100 million years ago.
So what does this tell us? Are we all descended from the life forms in existence at that time? Almost certainly yes. The alternative would be a far more complicated story, with life having arisen more than once. It follows that the life from which we are all descended was present on Earth within 350 million years of the formation of the Earth-Moon system, and, within an even shorter time after Earth had developed a solid crust, cool enough for liquid water (a prerequisite of our form of life).
In 1981, Francis Crick wrote that “we can only say that we cannot decide whether the origin of life on earth was an extremely unlikely event or almost a certainty – or any possibility in between these two extremes.” Now, at last, we can go beyond this. If the origin of life was unlikely, then life originating so early would be even more unlikely. So while it may be putting it too strongly to say that its emergence was “almost a certainty”, we can say that it was certainly a reasonable possibility. And if it was a reasonable possibility here on Earth, then it must equally be a reasonable possibility on all the Earth-like planets we have discovered, whose number grows almost daily.
To quote Steve’s comment on these discoveries, “This is what transformative science is all about. If life is responsible for these signatures, it arrives fast and early.”
1] Technically speaking, lead-lead dating. This depends on the ratio of lead-206 (formed by decay of uranium-238) to lead-207 (formed from uranium-235), with non-radiogenic lead-204 as a measure of lead from other sources. The calculation depends on the known difference in half life between the parent uranium isotopes. We know that these half lives must have been constant, since they are not free variables but consequences of the more fundamental constants of nature, and had these been different then the meteorites would not have formed as they did in the first place.
2] One problem with this scenario is the extreme similarity in composition between Earth and Moon rocks, difficult to explain if they are derived from two separate parent bodies. See, however, here.
3] As a consequence of the uncertainty principle, all materials store an unremovable amount of what is called “zero point vibrational energy”, and the amount of this energy is proportional to vibrational frequency. Lighter isotopes are therefore associated with higher zero point energies, leading in general to slightly higher chemical reactivity.
4] The amount of the minor isotope oxygen-18 present in these samples is different from the bulk of the mantle from which they crystallised, and indicative of mantle formed from the remelting of crust that had exchanged oxygen-18 with liquid water.
5] There is a steady trickle of claims from Young Earth creationists to have detected carbon-14 in dinosaur bones, diamonds, and coal. The first two of these are explained by contamination, while the more interesting case of coal is associated with nuclear reactions involving other radioactive atoms trapped within the material.
General references hyperlinked in the usual way. Selected more technical references (some behind paywall but all with open abstracts): Potentially biogenic carbon preserved in a 4.1 Byo Zircon here; Solar system age here; Earth’s accretion here, here, and here; Moon formation here; late origin of earth’s inner core here; zircon mass screening here, Earth’s oldest surviving crust here, here and here; 4.4 Byo zircon here, and the existence of water on earth when oldest zircons formed, here and here; habitability of Hadean Earth here; 3.43 Byo stromatolites here; Biological carbon isotope effect here; previous oldest evidence for life on Earth here. Image of zircon with granules via ibtimes.
An earlier version of this post appeared in 3 Quarks Daily. I thank James Downard for alerting me to the diversity of the Precambrian biota.