Radiocarbon dating, right? How does that work, and how can we ensure its accuracy?
In: Planetary Science
When an organism is alive, it’s constantly got carbon moving in and out of its body. Plants take in CO2 from the environment and photosynthesize it, animals eat plants, and cellular respiration turns that food back into CO2 that releases into the air. So the carbon in the bodies of living things and the carbon that’s outside our bodies are all interchanging.
When an organism dies, it stops exchanging carbon with the environment. There’s a particular form of carbon, C-14, that is radioactive – it decays over time into the more stable form of C-12. And because those C-14 atoms are just randomly dispersed around, everything living has about the same proportion of C-14 – just a tiny percentage of their total carbon. Which means that anything that’s been dead for a significant amount of time is going to have less C-14 than average, and more C-12 than average, because some of that C-14 will have decayed and not been replaced.
We know that the half-life of C-14 (the time it takes for half of the amount in a sample to decay) is around 5,700 years. So if a dead plant has half the amount of C-14 we’d expect, then it’s been dead around 5,700 years. If it has a quarter the amount, it’s been dead around 11,400 years. And so on. Once you get past around 50,000 years, there’s not enough C-14 left to measure it well, so that’s the cap for archaeologists to study. But other radioactive materials can be used in some cases to date older materials, through a similar idea.
Accuracy is an expectations management exercise.
It’s one thing to say “That wooden artifact is 4000 years old”, and a completely different thing to say “That bowl was made on September 27 in 1976 BCE at three in the afternoon”. The second clearly has more precision, absolutely impossible precision. The first implies a ±50 year precision, which is more plausible. Once you have the right expectation, other techniques like examining soil layers can give corroborating evidence that your radiocarbon dating is correct.
Everything living absorbs carbon, including a radioactive version called Carbon-14. When something dies, that C-14 starts to decay at a predictable rate. So, by measuring how much C-14 is left in something like a bone or a piece of wood, we can figure out roughly how long ago it died. It’s only good for organic materials, though, and only up to about 50,000 years old. Beyond that, the C-14 levels are too low to measure accurately. We cross-reference with other dating methods like thermoluminescence, which measures trapped electrons in things like pottery, or potassium-argon dating for really old volcanic rock. And we look at the surrounding context, too – if we find something in a layer of dirt we know is from a certain period, that helps narrow things down. It’s like a detective piecing together clues, not just one magic bullet.
There’s a couple of different ways.
Radiocarbon dating is probably the best.
When an organism is out in the environment, it’s constantly ingesting carbon. Some of thay carbon is carbon-14, a radio isotope. When the organism dies, it stops taking in new carbon and there is no way for new carbon-14 to be introduced. We know how much carbon-14 is naturally in the environment (which is constantly being made by cosmic rays hitting out atmosphere) so the now dead organism will slowly lose carbon-14 as it decays into nitrogen-14. By measuring how much carbon-14 is left relative to carbon-12, we can tell how long the organism has been dead (up to around 50,000 years, then there is too little carbon-14 left to measure.
We can apply this to ceramics by measuring the carbon thay was in the bacteria in the clay the ceramics were made from, and this tells us when the ceramics were fired.
There’s also other hints, like what we know about the area, dates of volcanic eruptions or drought years that match up with layers of soil we find these artifacts in, but radiocarbon dating works regardless of those factors
Radiocarbon dating is calibrated using tree rings, oddly enough. Dendrochronology uses tree rings to count backwards from when the tree was cut, to its first ring. Trees outside tropical regions grow in the summer and stop in the winter, leaving an annual record. The growth rings vary in thickness depending on how good the growing season was (temperature, rainfall, etc).
Fortunately for dendrochronologists, trees can live a long time. So if you find a recently felled 100 year old tree, you have a pattern of tree rings unique to that time period and region. Now, suppose you have a 50 year old structure that has a large beam that itself was made from a 100 old tree. You can match up the overlapping years with the first tree and know not only what year the tree was felled, but now you have a record going back 150 years. Repeat that a number of times and you can date any piece of the same type wood in the region.
Now, radiocarbon date all those pieces of wood. Since you know very accurately the real date, you can build a chart of radiocarbon content to actual years and correct the radiocarbon date The C14 content of the atmosphere varies based on solar activity, so this is a really important technique. Also, C14 in the atmosphere is homogeneous, so if you’re dendrochronology is based on European oak and pine (which is the most complete dataset), the correction still works for finds in, say, Australia.
It’s not the only technique. For example (and I apologise if I get the fine details slightly wrong), there’s a technique called luminescence dating that measures the energy of photons being released from crystals in soil samples, that can tell you when they were last exposed to sunlight, and give a time context to buried objects and artifacts. Tree rings, if you’re lucky enough to find wood, carry a record of good and bad growth years that can be correlated to other dated sites, say. Objects themselves may give information (if you find a roman coin with a known emperor’s head on it, it’s not been there longer than that emperor’s reign – and it’s often reasonable to deduce that things on top of it aren’t earlier than that emperor’s reign either, for example
And conversely, if you find something underneath something you can date, it’s often reasonable to assume that it’s no later than whatever is above it). I’m sure there are other things I’ve forgotten or plain don’t know about.
Archeology is about human history not the world in general so we so we can use multiple methods that relate to human behaviors. First and the oldest is to look for inscriptions and texts get a dates off those, build a time time line off them all and then start looking for dateable events to tie those to our calendar system. Famous documents like the [Abydos list of pharoahs](https://en.wikipedia.org/wiki/Abydos_King_List) the [Turin king list](https://en.wikipedia.org/wiki/Turin_King_List) as well as others can provide dates across hundreds and thousands of years for at least a rulers reign. So we know that if we find an inscription dated to the reign of Rameses II is comes before X pharaoh and after Y pharaoh. Then we just need a length of reign for each and an entire chronology can be built. Even without dates we have some information. We know no ruler is legitimately going to live more than 100 years, and more likely only going to reign for 20-30.
Next we take events like solar eclipses, the appearance of comets, known battles or rulers and chronologies of other civilizations and compare them to what we generated. If two civilizations saw the same supernova we can accurately tie the reign of a king in civilization A to the same time as that of another king in civilization B. Once we can build a timeline for one civilization we can often tie it to anther through things like diplomatic correspondence, or trade in certain goods such as pottery.
Once we have a rough chronology we can start to look at art and architecture. Styles change over time and humans produce [a tremendous volume of pottery](https://en.wikipedia.org/wiki/Monte_Testaccio). Shards from that pottery last a long time. If we can date one shard of one style to reign X then everything in the same style will be roughly the same time period. So we someone goes out and digs up a new site and finds pottery from that style they know the site is from that period. This use of pottery goes back before we have written documentation. We can look at a pot shard from ancient Crete and say it from the Black cord period (made up I’m not familiar with Cretan pottery) but we know the black cord period came before the black sphere period and after the solid black period giving us at least an order for items found around the shards. Over the past several hundred years we have built up tremendous databases of pottery styles. Now we just need other evidence to apply dates, that evidence can be the written record or more scientific tools.
Yes in recent times we can use radiocarbon dating. We can also us dendrochronology. We have built extensive databases of the war/cold periods going back thousands of years in some locations and can look at the tree ring patterns of wooden objects to see where the match giving us a year.
All these things go hand in hand and refer back to each other and need to match. A radiocarbon date of 4000 B.C., a dendro date of 3900 B.C., a pottery date of 3950B.C. and an inscription we calculate to 2800B.C. tells us something is wrong somewhere, likely with the inscription evidence since its the one that is the outlier. This is were the accuracy comes in. All of the methods have a margin of error but if all the dates fall in the same range we can say with confidence that its an accurate date.
But once we have good records in multiple areas we can with confidence find a single piece of evidence, say a pot shard and date a site to a specific time even if we dont find other evidence a the site because we have tied that style at other sites to other evidence.
EDIT: oh and we can do the same with some geology. Certain fossils and shells are only found in rocks of certain time periods so finding the same shell somewhere else tells us that rock is from the same time (geologically) as the first. Then tie in other isotope dating, erosion and weathering dating, deposition times to create new rock etc and a timeline can be built on the order of billions of years.
Radiocarbon dating is an early form of what’s now known as radiometric dating. As noted in other replies, carbon dating is only good for a relatively short span of time. But once scientists figured out the basic idea, they discovered other isotopes with different, much longer half-lives. As noted the core principal is the same: some natural process causes the accumulation of a radioactive isotope at a known rate that stops when whatever it is dies or gets buried. The isotopes then decay, also at a known rate. Measuring the ratio gives you a date that has (ideally) a precision that makes it useful for the time span you’re working with. For example, a +/- 5000 year margin of error would make dating Egyptian pottery pretty useless, but a +/- year margin of 50,000 years would be quite excellent for dating things from the Mesozoic era or earlier.
But there are other ways, most worked out before radiometrics were developed. The most important is stratigraphy. It works on the principal that depth = age. If you have a date for one layer of your dig (due to, say, pottery that was described by a 16th century explorer) and a different date for a much lower layer (due to an example being carved on the base of a sacrificial altar made with wood beams you can date by counting the tree rings), then anything you find between those two layers will be older than the former, and younger than the latter.
You probably already see how problematic this method is. You only get relative dates, not absolute ones. This is why it took the development of radiometrics before it was possible to *prove* things like the age of the Earth, or how long ago the dinosaurs were extinct. But it was, and is, still quite useful. It’s certainly simpler and cheaper.
A combined approach is the most common way: Radiometrics provide exact dates of specific creatures or artifacts. If you stumble across a site that includes these items, it allows you to quickly narrow down how old anything unknown at the site is.
None of this is consistent, simple, or foolproof. Contamination can throw off a dating sample. Plowing, digging, or other sorts of excavations can push newer artifacts into older deposits. The Earth’s crust has been caught in the taffee-pulling machine that drives continental drift for its entire history, making some stratigraphies quite confusing. The machine testing the sample can break, or be set up with the wrong set of assumptions. We’re constantly refining the techniques, which has led to significant revisions of all sort of dates in ways both large and small.
But, overall, the combination of the methods described here and in other answers has provided us with dating tools with a precision undreamt of as recently as a few decades ago. It’s really neat stuff!
Others have explained radiocarbon dating, but there are many things to which it is not applicable. Generally, it can tell you roughly when an organism died. This can tell you when a person died from their remains (if they’re in good enough condition) or when, for example, the organic material used to make a piece of wood, fabric, or paper died. It generally isn’t much good for dating a stone building or a bronze arrowhead, though. So, a combination of other methods are used, including written records and stratigraphy (the last one is referring to the idea that older artefacts tend to be buried deeper than more recent ones, often in clearly defined layers). Sometimes you can cross-reference things with known natural events like volcanic eruptions (which can leave layers of ash) or appearances of comets (which people often wrote about).
Latest Answers