Ion beam

Accelerator Mass Spectrometry

Accelerator mass spectrometry (AMS) is an ultrasensitive technique for measuring the concentration of a single isotope. Usually the isotope in question is 14C. By counting 14C atoms, 14C/12C ratios can be measured to the 1:1016 level, resulting in attomole sensitivity for 14C for milligram samples.

Our AMS system is the first one in Finland. At present, our accuracy is better than 0.5% (for samples with one milligram of carbon). The measurement duration is roughly one hour. The machine background is 80 ka BP, which is among the lowest reported internationally. The spectrometer in Helsinki is hence especially suited for measuring old samples.

The principle of accelerator mass spectrometry (AMS) is simple, namely, to separate one isotope from a huge amount of atoms and molecules. The separation is achieved with an electrostatic accelerator, several electrostatic and magnetic analysers, and collisions with other atoms. Advantage is also taken of the fact that not all atoms form negative ions. Because high energies are used, the background from atomic or molecular ions with the same mass can be removed. This makes it possible to determine concentrations orders of magnitude smaller than conventional mass spectrometers can. AMS can measure orders of magnitude smaller samples, important e.g. In archaeology. Measurements are also quicker since there is no need to wait for the nuclei to decay.

The figure below illustrates the difference between conventional mass spectrometry (MS) and AMS.

Data-analysis: improved detection of instrumental error and stabile uncertainties

We have developed a Bayesian data-analysis method for AMS, called the CAR model, features of which include:

  • Gives a continuous parameter for the magnitude and type of machine errors.
  • The known measurement uncertainty is taken explicitly into account. This stabilizes the scatter in the uncertainty compared to the standard error of the mean (see the figure below).
  • Reliable results for fewer measurements, smaller samples

Due to the random scatter of the standard error of the mean, there will be cases where the true error is five to seven times bigger than the standard error of the mean. For example, an archaeologist might get a radiocarbon date of 1000+-50 B.C., while the true age might be 700 B.C., affecting historical interpretation significantly. From the figure below, it can be seen that the uncertainties from CAR (dark dots) have much less random scatter than the mean and standard error of the mean (MB), and hence the uncertainties from CAR are seen to be more reliable.

AMS data analysis

The car4ams program package can be downloaded from

Pictorial overview of the setup

Ion source

The main parts of the AMS system are shown below.

The ion source

Ion source

The accelerator

Accelerator tank

High-energy beamline