The AMS radiocarbon dating facility of the CEDAD  is  based on a high current HVEE 4130HC TandetronTM accelerator, with a terminal voltage of 3 MV, equipped with a dedicated 14C high throughput AMS system.  

The AMS beam line consists of a low energy and a high energy mass spectrometers connected to the Tandetron accelerator.

The AMS system uses a bouncer magnet, in which the carbon isotopes are injected sequentially into the accelerator with cycle frequencies up to 1 kHz.

The injector has been installed in order to allow  future expansion of the AMS facility to other radioisotopes.

In the Low Energy AMS  spectrometer the intense negative carbon beam, produced in the ion source from the solid graphite targets, undergoes the first momentum and energy selections  and is sequentially injected in the accelerator. It consists of:

 The 846 HVEE sputtering ion source;

.The 54°electrostatic analyser;

The 90° injection magnet;

All the optical elements for beam transport (einzel lenses and steerers).


In the 846A HVEE ion source a high temperature spherical ionizer is used to produce, by sputtering of  a 7 KeV  Cs+ beam, an intense negative 35 KeV carbon beam. The source has demonstrated ion beams with 12C-currents up to 50 mA, although during normal operation the beam current is limited to about 20 mA to reduce ion source maintenance and for considerations regarding space charge induced mass fractionation phenomena.

To increase the total samples throughput, up to 58 samples can be placed at a time  in the ion source carousel wheel from which they are transferred by a computer controlled retractable arm to the focal point of the Cesium beam of the ion source. The total loading and unloading time is about  90 s per sample (being about 20 % of the total  sample measurement time).

 A diameter of 0.5 – 0.6 mm of the Cesium spot allows to  keep it within the graphite surface (2 mm wide) also when, during routine AMS measurements, the position of the cathode is changed by two micrometers step motors  in order to reduce the target surface cratering and to allow a homogeneous consumption of the target surface. In particular the target is moved such that the Cesium beam is focused in nine different positions. The following figures show the spherical ionizer in the source..

ESA 54°

The first analysis element, placed just after the ion source, is a 54° electrostatic analyzer (ESA) consisting of two spherical aluminium electrodes with a reference curvature radius of 469 mm. The radius (r) of the trajectory  of the incoming carbon negative particle extracted from the ion source with an energy E0 and charge state q is given by the relationship:

where E is the electric field in the ESA and d and ΔV are the distance and the voltage difference between its  electrodes. The quantity  is called electrostatic rigidity.  Thereforethe ESA is used to analyse the charged particles on respect to their E/q ratio.


The ions energy selected in the 54°ESA are, then,  analysed respect to their momentum by  the following 90° double focusing magnet. In a homogeneous magnetic field region the ions, initially moving perpendicularly to B,  are deflected on circular trajectories  with a radius (R) given by the well known relationship:

where B is the magnetic field, E, M and q are respectively the ion kinetic energy, mass and charge and the quantity  is called magnetic rigidity. Thus,  by using a magnetic analyzer a momentum analysis of the incoming particles can be carried out.

Figure shows a momentum spectrum of the negative beam produced in the ion source carried out by the first 90° injection magnet. A clear separation between different masses is achieved with peak intensities depending on the selected mass. The most intense peak refers to mass 12 (12C-)  while the less intense refers to mass 14 amu. The peak at 13 amu corresponds either to the 13C- particles  and the  12CH- molecular isobars that cannot be distinguished from the  13C- peak. Different species have mass 14 amu: 14C-, 13CH- and 12CH2- all convoluted into the same peak. No evidence of 14NH- molecule exists in the spectrum probably due to low sensitivity of the used measurements device. Mass 16 and 17 amu correspond to 16O and 16OH- atomic and molecular species, respectively.

The mass resolution of the LE spectrometer is  not sufficient to distinguish the contribution of the monatomic carbon ions from the relative molecular isobars. For instance the relative mass difference between 14C- and 12CH2- can be estimated to be ΔM/M=1134 which is significantly higher than the mass resolution of the 90° magnet. It has also to be noted that such a kind of very high mass resolutions could also be achieved, in principle,  by expedients like narrowing slits but this would lead to an unacceptable reduction of the counting efficiency and, consequently, to an increase of the measurement time. In AMS radiocarbon dating the separation of the molecular and of the monatomic ions in achieved by breaking the molecular isobars during high-energy collision in the accelerator stripper canal.

The fast bouncing sequential injection

The low energy 90° injection magnet is used to sequentially inject the three carbon isotopes beam in the accelerator. This is achieved by keeping the magnetic field of the magnet constant, and by applying a pulsed voltage to the magnet vacuum chamber and to two symmetrically placed accelerating-decelerating gaps by mean of a variable DC (± 5 kV) power supply. In this way the energy of the ions can be  changed while they pass through the magnetic field region such that, according to Eq. (1.2), the radius of their trajectories is constant and correspondent to the optimal value for the injection in the accelerator (R0):


where M and E indicate the masses and the energies for the three carbon isotopes, respectively. From Eq. (1.3), and by expressing the ion energy as a function of the extraction voltage (Vex),  the following relationship can be obtained to calculate the voltage (bouncing voltage) to be applied to the magnet chamber in order to inject the three beams in the accelerator:


where is the bouncing voltage to inject the mass m and mis the reference mass (the mass of the isotope which is injected when no voltage is applied to the magnet chamber). A comparison between the theoretical values calculated according to (1.4) and the experimental ones is given in the following table.

Ion Mass

Theoretical Voltage

Experimental Value

12 amu

2916.7 V

2909.6 V

13 amu

0 V

-1.2 V

14 amu

-2500.0 V

-2489.6 V

The following figures show he transmission of different masses versus bouncer voltage. The firstshows a momentum scan at the exit of the injection  magnet when no voltage is applied:  the 13 amu beam is along  the accelerator axis and it is  transmitted.

When a positive voltage of 2909.6 V is applied to  the magnet chamber the ions are accelerated and their energy is increased. I n this case the mass 12 beam is injected.

On the contrary, when a voltage of -2489.6 V is applied the negative particles are slowed down such that  the mass 14 amu beam is transmitted).

The bouncer changes the energies of the ions only when they fly through the magnet chamber, since the accelerating effect that the ions see while entering the accelerating gap  is compensated by the decelerating one they see while leaving the magnet through the decelerating gap of the magnet chamber. This means that after passing the magnet the ions have the same energy they had before entering it.

The fast bouncing sequential injection: definition of the injection cycle.

Figure shows the injection cycle scheme (not in scale) with the indications of  all the parameters defining  it.

The quantities indicated in Figure aredefined as follows:

              I.      t12, t13, t14 are the injection times for the three isotopes: the time during which each isotope is injected in the accelerator;

II.      td is the delay time, the time (ion flight time) which the ions need to reach the measurement devices in the high energy spectrometer (which can be easily estimated by studying the motion of the three carbon isotopes along the system to be  5- 6 μs). The time td is also the time when the data acquisition starts after  the ion injection. td is set to 10 μs.

         III.      tw is the wait time, the settling time which the magnet chamber needs to stabilise at the set voltage level. In order to estimate tw the system formed by the high voltage pulsed power supply and the magnet chamber have to be studied. It can be schematised as indicated in Figure 11 where Rps and Cps indicate the internal output resistance and capacitance respectively of the power supply and Cm the capacitance associated with the magnet chamber.

The Tandetron accelerator

The Tandetron accelerator is a linear, electrostatic accelerator of the tandem type in which the negative ions are accelerated in a double step. In a first step the ions, injected in  a negative charge (q0=-1) are attracted by the positive terminal voltage (Vt=2400 KV) acquiring a kinetic  energy (E1= 2.4 MeV):


In the accelerator terminal at the energy E1 the ions  enter a  13 mm wide stainless steel cylinder (stripper canal) in which low density ( 1-5 μg cm-2 ) Argon is fluxed. During the collision with the Argon atoms the negative ions lose electrons leaving the accelerator with a net positive charge (charge stripping process). Now the positive ions are repulsed by the positive terminal and are further accelerated towards the accelerator exit which is kept at ground potential.

The energy acquired along the high energy accelerator column (E2) for an ion leaving the stripped in the charge state +qHE can be expressed as:


Thus the total final energy of a particle at the exit of the accelerator (E) is:


In this chapter it will be described how in AMS radiocarbon dating the stripping process plays a crucial role in the definition of the system sensitivity.

In order to reach the high energies needed for the stripping process voltages as high as some MV, must be obtained. In  a Tandetron accelerator these high voltage levels are  obtained by a  Cockroft-Walton type power supply.

The high voltage generator

The general scheme of the high voltage generator of a Tandetron system can be described as formed by a radio frequency driver which supplies power to an RLC oscillator circuit consisting of a transformer coil and two dynodes (Figure 12). All the systems (except the RF driver) are under an insulating atmosphere of SF6 in order to avoid arcing (high voltage sparks).

The RF driver supplies a voltage to the two dynodes whose amplitude (Vdynode ) varies according to the set terminal voltage being 4 KV for the maximum terminal voltage of 3 MV and with  a frequency ( fRF )  of 34,4 KHz corresponding to resonant conditions for the RLC circuit.  The RF voltage present on the dynodes is capacitively  coupled via corona rings to a series of stacked diode circuits which are used to rectify the RF voltage and multiply the signal to a DC voltage at the accelerator terminal. 

The portion of the dynode voltage which is present on each corona ring (Vcorona) can be calculated by the formula [11] :


where Fcoupl express the voltage capacitive coupling between each  corona ring and the dynodes and is given by:


where CDC denotes the capacitance between the dynode and the corona ring and CCC the capacitance between two opposite corona rings (see Figure 12).

Figure 12 . Schematic diagram of a Tandetron accelerator [10].

The diode stacks are mounted between two opposite corona rings and rectify the RF  dynode voltage to generate the DC terminal voltage (VT) given by:


where n is the number of corona rings along the multiplying stack.

The main advantages of a system like that if compared with the Van der Graaf-type high voltage generators are:

·        The complete absence of moving parts along the multiplying stack, resulting in a correspondent reduction of maintenance and in simplified operation conditions;

·        The terminal voltage can be easily regulated with high accuracy by changing the voltage output of the RF driver;

·        The beam currents which can be supported are  significantly  higher than the mechanical charge transport systems;

·        Voltages as high as 5 MV have been, recently,  demonstrated with Tandetron-type accelerators [11].

The main disadvantage is that this kind of accelerators, because of the high capacitance between the terminal and the tank, can store much more energy than the Van der Graaf type and  the release of this energy in event  of high voltage sparks  can induce serious damages to the control electronics.

The stripping process

As already described in the previous paragraph  in the accelerator terminal  the ions, injected in a negative charge state and accelerated along the low energy accelerator column,  are converted to positive charge states,  being   then further accelerated towards the accelerator exit along the high energy column.

The charge exchange process is achieved in the accelerator stripper consisting of a narrow stainless steel canal along which Argon gas is continuously flown. The gas is  recirculated by a turbo molecular pump in order to reduce beam losses due to charge exchange processes along the low and high energy columns. The interaction at high energy (2.4 MeV for the carbon ions) with the Argon atoms leads to the stripping  of electrons from the negative ions which leave the stripper canal with a net positive charge.

The charge state distribution at the exit of the accelerator is a complex function of different parameters like: the ion energy and mass and  the argon stripper density [12-16].

Figure 13 shows the momentum scan of the high energy positive particles obtained at the accelerator exit when a 35 KeV negative 12C- beam is injected.


Figure 13. High energy spectrum from a pulsed 12C-  beam injected: the two peaks   correspond to the two charge states 3+ and 4+ .


            At the used terminal voltage (2.4 MV) corresponding to a beam energy of 2.4  MeV in the stripper the most abundant positive charge state for the carbon isotopes produced is the 3+ corresponding to the stripping of 4 electrons during the charge exchange process.

            Dedicated experiments have been carried out in order study how the transmission of the carbon ions through the accelerator is related to the stripping parameters (namely the argon gas density).

            Let us define the overall optical transmission Topt through the accelerator as the ratio between the number of particles per time unit of a certain mass (m) emerging from the accelerator () and the number of negative particles of the same mass injected ().


where  is the sum of all the possible positive charge states () at the accelerator exit:



The high energy mass spectrometer

The high energy spectrometer consists of a 110° analysing magnet, two Farady cups for the measurement of the two stable isotopes currents, a 33° cylindrical electrostatic analyser, a 90° magnet and a gas ionisation detector for the radiocarbon counting.

The 110° magnet separates the 12C3+,  13C3+ and 14C3+ beams; the 12C3+,  13C3+  beam currents are measured  by  two Faraday cups.

In order to identify the different ionic species coming out from the accelerator, and to check the effectiveness of the molecules breakdown and of the isobars separation, the masses 12, 13, 14 amu have been separately injected by varying the voltage applied to the bouncer magnet. The high energy momentum spectrum during the injection of the 12C beam is shown in Figure 15. Only two peaks corresponding to the two charge states 3+ and 4+ are present at a magnet current of 84.4 A and 103.0 A.