In this article we shall discuss the principles behind amino acid dating (also known as racemization dating); we shall discuss how it ought to work, and why it often doesn't.
An object is said to have chirality if it is not possible to make it into a mirror-image of itself by turning it round. For example, a shoe is chiral: you cannot turn a left-foot shoe into a right-foot shoe by turning it round or flipping it over. On the other hand, an object such as a table-knife is not chiral: if you have it lying on the table so that the blunt edge is on the right and the serrated edge is on the left, then you can produce the mirror-image of this situation by rotating the knife around its long axis.
Some molecules are chiral. For example, consider the two molecules in the picture to the right. They both have exactly the same chemical formula, but one is left-handed, and the other is right-handed. They are said to be enantiomers of one another.
When we make chiral molecules using ordinary chemical processes, we usually produce equal quantities of both enantiomers. Such a mixture is said to be racemic.
However, biological processes produce molecules with a distinct chirality: all the amino acids are "left-handed" (with the exception of glycine, which is not chiral) and all the sugars are "right-handed".
So when an organism dies, its amino acids are left-handed. But after its death, the amino acids can spontaneously change their chirality, flipping from being left-handed to right-handed, and indeed back again.
The result of this process is that eventually the amino acids will collectively become racemic: each particular amino acid will have one chirality or another, but after a sufficient amount of time, collectively the amino acids won't favor one enantiomer over another. This process is known as racemization.
We should note that although the underlying basis for this process is random, and that in principle the amino acids could by some statistical fluctuation become less racemic and more chiral, the laws of statistics ensure that in practice if we are looking at a large enough sample of amino acids, the chances are astronomically remote that such a thing will occur.
So the process of racemization looks like a good candidate for one of nature's clocks. We know that when an organism dies, its amino acids will all be left-handed; and we know that as time progresses the amino acids will become continually more and more racemic.
So it would seem that if we want to know how long it was since an organism died, all we have to do is see how racemic its amino acids are. And this would work, on one proviso. The process of racemization would have to go at a constant rate, and we'd have to know what it was.
And this is where the whole idea breaks down.
How do we know it works?Edit
The problem with racemization is that it depends on chemical processes that are affected by temperature, humidity, and the nature of the original material undergoing racemization. As a result, it isn't possible to say that racemization happens at such-and-such a rate.
However, it does have some applications. Suppose we examine a particular material (let us say tests of the foraminiferan Neogloboquadrina pachyderma) in a particular environment (let us say in mud in Arctic waters) and by comparing it with a dating method we know we can rely on, we establish that under these conditions racemization does happen at a reasonably steady rate.
In that case we could use the foraminiferans to date sediment in places where we aren't able to use radiometric dating. (For it would be strange and anti-scientific to conjecture that the rate of racemization of the shells in the Arctic mud is constant whenever we can check it, but variable when we can't.)
Just this was established by Kaufman et. al. (2008) in their paper Dating late Quaternary planktonic foraminifer Neogloboquadrina pachyderma from the Arctic Ocean by using amino acid racemization, Paleoceanography, 23(3). It gives the reader some idea of the difficulties of the method that they were obliged to use the single common foram species N. pachyderma, having found that racemization rates differed even between different species of forams.
So dating by racemization can have a few applications, but the conditions under which it can confidently be applied are rather rare. What's more, racemization happens quite fast by geological standards, so, like the other methods of absolute dating we have discussed so far, dating by racemization cannot take us far back in geological time.
All this is not to say that the reader should dismiss out of hand results obtained by amino acid dating; but it can be trusted only when the people applying it have taken care to ensure that they are using it in a context in which it is known to work. In early papers, before geologists and archaeologists had learned the pitfalls associated with amino acid dating, inaccurate dates were presented with much more confidence than they deserved, and such papers should not be relied on.