Getting an absolute date on the age of a rock or a fossil is actually pretty simple, despite that date being in the million- or billion-year range. Before getting to the actual process, it is necessary to discuss the phenomenon that allows us to date in geological time: radioactive decay.
Radioactivity is natural and occurs all around us, all the time. It’s caused by elements falling apart and decaying, giving away protons and neutrons, or gaining or giving away electrons. The original element is referred to as the parent, and the element after decay is finished is the daughter. The most used pairs in geology are:
- Carbon-14 → Nitrogen-14
- Uranium-235 → Lead-207
- Uranium-238 → Lead-206
- Potassium-40 → Argon-40
- Rubidium-87 → Strontium-87
The decay rate for each element is constant as it’s unaffected by external factors like pressure or temperature, and it follows a geometric progression. The decay constant tells us how long it takes for half the atoms to decay, in other words the half life.
So, if you can somehow measure the decay constant and the half life of one of these parent-daughter combinations, and then also measure the parent:daughter ratio in a sample, you can then do basic arithmetic to figure out how old your sample is. If the element decays completely in 2 million years, and your sample has a 50:50 ratio (a perfect first half life), then your sample is 1 million years old.
This process is exactly what we do in geology. It’s called radiometric dating. The decay rate can be measured by a scintillometer, a device that detects gamma rays, and the elemental composition of a sample is determined like any inorganic chemical sample, with a mass spectrometer.
That’s really all there is to it. Of course, in practice, things get a lot murkier, because you have to find appropriate samples that contain the correct radioactive elements. Rubidium-87 needs 48.8 billion years to get to a 50:50 ratio with its daughter, uranium-238 takes 4.47 billion years, uranium-235 needs 704 million years.
In other words, each element series has its own effective range. For example, carbon dating cannot be used for anything beyond recent archaeological time: the quick decay rate means that the ratio becomes 0:100 in no time, and a 0:100 ratio is useless doesn’t allow us to calculate a date (0.1:99.9 is the minimum required).
The necessary elements need to be locked in minerals at the locality investigated. If you have a deep rock or volcanic layers, you have a very good chance of getting a usable sample that can be dated. In these cases, what is actually being dated is the time of crystallisation – when the melt hardened to form the minerals and rock.
However, dating a fossil is much more difficult. Since fossils are mostly found in sedimentary locations with no diagnostic minerals that formed in situ to use for dating, we’re often stuck with having to extrapolate or interpolate from geological layers older and younger than the fossil-bearing locality. For example, you have a fossil locality sandwiched between two volcanic layers. The top of the one below it is 80 million years old, the bottom of the one above it is 70 million years old. Therefore, the fossil locality must be 80-70 million years old. Detailed stratigraphic work incorporating regional or even global geology can then reduce the range of this date.