PIXE

Particle induced x-ray emission (PIXE), is a powerful yet non-destructive elemental analysis technique now used routinely by geologists, archaeologists, art conservators and others to help answer questions of provenience, dating and authenticity.

Quantum theory states that orbiting electrons of an atom must occupy discrete energy levels in order to be stable. Bombardment with ions of sufficient energy (usually MeV protons) produced by an ion accelerator, will cause inner shell ionization of atoms in a specimen. Outer shell electrons drop down to replace inner shell vacancies, however only certain transitions are allowed. X-rays of a characteristic energy of the element are emitted. An energy dispersive detector is used to record and measure these x-rays and the intensities are then converted to elemental concentrations.

The Advantages of PIXE:

Compared to electron based x-ray analytical techniques such as energy dispersive spectroscopy(EDS), PIXE offers better peak to noise ratios and consequently much higher trace element sensitivities as seen in the spectra below. Absolute trace sensitivity to a given trace element is dependent upon a number of factors, such as matrix composition, detector efficiency and peak overlap. However, tests performed on the Harvard system have produced parts per million (ppm) level measurements for certain elements.

An MeV proton beam can be brought out from the high vacuum environment of the accelerator into the ambient of the laboratory. This technology makes it possible to measure a valuable artifact (see figure below) or precious material at atmospheric pressure out of the confines of an evacuated chamber without sampling. A possible disadvantage to running in this configuration is that low energy x-rays from lighter elements will be attenuated in air. However, we have the capability to purge the area between the sample and the detector with helium to minimize this effect.

A measurement in progress:

This is a gilded bronze vessel from the Han Dynasty (207 B.C.to 220 A.D.) from the Harvard University Art Museum's Asian Studies collection. A common technique to affix gold to a bronze base material incorporates a mercury amalgam to promote adhesion. Conservators were using PIXE to determine the method of gilding by seeing the presence or absence of a mercury signal.

Major elemental analysis is performed for any element from sodium to uranium in a single spectrum on our system. X-rays from elements below sodium cannot be seen because they are absorbed in either the detector window, the atmosphere between the sample and the detector, or through any filters used. For trace element analysis, we choose a filter to attenuate the x-rays at the energies of the major elements allowing the detector to measure the trace elements with greater sensitivity. Usually these filters will cause an insensitivity to lighter elements , but will allow simultaneously analysis of any element above the filter's absorption edge.

The Harvard system was designed to fulfill the needs of the archaeology and art conservation community. We use the largest possible detector solid angle so that useful PIXE data can be acquired with relatively low beam currents. Usually these currents will cause only a few milliwatts of heating from the beam. The situation is improved even more by running it with a helium purge so that heat may be convected away.

The Use of Standards

Because of a lack of precise knowledge of critical system parameters such as absolute detector efficiency, x-ray filter attenuation and integrated beam current, many PIXE systems rely on standards to determine accurate specimen compositions. Although we have had good results running in standardless mode for a variety of samples of known composition, we advise you to have a standard or set of standards in close approximation to that of the specimen. Many times it is easy to develop an internal standard to fit a particular measurement.

The Ideal Specimen

The ideal specimen should be flat and of uniform composition. Ceramic shards for example must be crushed mixed and pelletized. Since most of the characteristic x-rays emanate from the top few microns, for accurate measurements it is important that the sample be homogeneous to within the micron level.

Real life samples do not always fit this ideal. A precious object may have a hidden layered structure, or not have a flat or even surface. Pelletization is obviously not an option. Here is advisable to analyze a number of spots and take the average. Such measurements are considered to be qualitative or at best "semi-quantitative".

The Benefit of X-ray Filters

At very high count rates, Si(Li) x-ray detectors will behave in a less than ideal fashion. Energy resolution could become significantly worse causing peaks to overlap. Pileup peaks may appear which manifest themselves as multiples of principle peaks. Spectrometer dead time may cause counting errors.

X-ray peaks from major elements can occupy much of the detectors useful counting capability. Dominant low energy x-rays may be filtered out so that the detector will only see contributions due to the higher energy trace or minor elements. Beam current may now be increased with an overall effect of much greater trace element sensitivity while keeping the detector at a low count rate.

In some cases single element foils may be used as "notch" filters to attenuate only a dominant matrix element but allow relatively high transmission of all other x-rays. A good rule of thumb in this case is that the best absorbing filter is lighter than the dominant matrix element by 2 atomic numbers. That is to say that the best filter for a Fe matrix would be a Cr foil.

 

System Description

A finely collimated beam of protons is produced by a General Ionex tandem ion accelerator and brought through a graphite plate with a series of closely spaced .3 mm holes. An o-ring seals a thin polymer window over this plate providing a vacuum barrier through which the beam is brought.

The graphite block provides mechanical support to the film and helps to conduct heat away from it greatly increasing its lifetime under beam bombardment.

Relative beam current integration is measured off of the graphite block providing accurate normalization amongst samples. An internal Faraday cup provides secondary electron suppression. Because of this we can precisely measure and control the relative number of protons striking each sample.

Accurate sample placement relative to that of the beam is accomplished by the aid of a HeNe laser beam which runs collinear to the ion beam. The sample is mounted on a block tilted as to provide the sample normal tilted 45 degrees from the incident ion beam.

The sample area can be purged with He when looking for light elements with PIXE. A surface barrier detector is placed 60 degrees relative to the incident ion beam for RBS measurements at atmospheric pressure.

A Si(Li) x-ray detector is placed 90 degrees relative to the incident ion beam and is used to determine x-ray peak energies and intensities. X-ray absorber filters are used to attenuate the dominant peaks and allow greater trace element sensitivity.

GUPIX, an interactive software package is used to analyze and convert raw spectral data into elemental concentrations.

 

Main Reference:

S.A.E. Johansson and J.L. Campbell, and K.G. Malmqvist: Particle Induced X-ray Emission spectrometry (PIXE) (WILEY 1995)

Back to the MRSEC website