An Overview of XRF Basics

3. Sample Preparation Techniques
for XRF Analysis

3.2 Preparation of Solid Samples

3.2.1 Metals

Preparation of metal samples must be simple, rapid and reproducible. Usually, metal samples are prepared as solid disks by conventional methods of machining: cutting, milling and polishing. Grinding is used in the case of hard alloys and brittle materials such as ceramics.

The best polishing operations require very fine abrasives to produce the scratch-free surface necessary for most analyses, and a mirror-like surface if the sample is to be analyzed for light elements. The surface finish is of prime importance because polishing striations give rise to the so-called shielding effect, which results in a decrease in fluorescence intensities. As expected, the decrease in intensity is more important for lighter elements when the primary radiation is perpendicular to the striations and weaker when they are parallel to them. Modern spectrometers are equipped with spinning sample holders to smooth out the influence of sample orientation, resulting in observed intensities on samples and standards that are reproducible.

However, the shielding effect may still be present; sample rotation will compensate for it only if the magnitude of the effect is the same for standards and samples. Striations therefore must be of the same size and the compositions of the standards and samples must be similar (same effective wavelength).

In practice, striation depths of 100 mm are acceptable for elements with characteristic lines of short wavelengths, but striations deeper than few mm may significantly impair the accuracy of Si, Al and Mg measurements.

Very fine grits of Al₂O₃, SiC, B₆C (80 to 120 grits) are commonly used to obtain the desired surface finish for most metals (Fe, Ni and Co bases).

Mechanical polishing may be undesirable for soft, malleable, multiphase alloys because of smearing of the softer components. The intensities of the elements in softer phases increase while those of the harder phases decrease. In such cases, special precautions must be taken during milling and final polishing of metals with Pb, Cu, Al, Zn or Sn bases.

Polishing may be a source of contamination since currently used abrasives, SiC and Al₂O₃, contain two elements that are often determined in commercial alloys. Sample surface cleaning may be necessary to remove contamination as well as grease and handling residue.

3.2.2 Pressed Pellets

Since powders are not affected by particle size limitations, the quickest and simplest method of sample preparation is to press the powders directly into pellets of equal density, with or without the use of a binder. In general, provided that the powder particles are less than about 50 mm in diameter, the sample will pelletize at 10 to 30 t. Where the self-binding properties of the powder are poor, higher pressure may have to be employed or in extreme cases a binder may have to be used. It is sometimes necessary to add a binder before pelletizing and the choice of the binding agent must be made with care. The binder must be free from significant contaminant elements and must have low absorption. It must also be stable under vacuum and irradiation conditions and it must not introduce significant interelement interferences. Of the large number of binding agents that have been successfully employed, probably the most useful are wax and ethyl cellulose.

The analysis of powders is invariably more complex than that of metal samples because, in addition to interelement interferences and macroscale heterogeneity, particle size effects are also important. Although inhomogeneity and particle size can often be minimized by grinding and pelletizing at high pressure, often the effects cannot be completely removed because the harder compounds present in a particular matrix are not broken down. These effects produce systematic errors in the analysis of specific type of material, such as siliceous compounds in slags, sinters and certain minerals.

Analytical data for longer wavelengths will sometimes be improved if a finely ground powder is compacted at higher pressures (say up to 30 t). A 40-ton press should be therefore considered if light element analysis is required in pressed powder samples. A good quality die set is required to produce good quality pressed powder samples. Powders can be pressed into aluminum cups or steel rings. Alternatively boric acid backing can be used, or free pressing if a binder is used.

Sample preparation with the binding agent Moviol

Preparation of the solution
Over a mild heat mix 100 ml aqua dest. with 2 g Moviol flakes. Stir the solution for approx. 15 to 20 min and, afterwards, transfer to a container to cool. If necessary, allow the undissolved parts to settle (to reduce cloudiness) and draw off the clear liquid. Transfer the solution to a PE dropper bottle and label it. The solution is now usable and can be kept for approx. 2 years.

Use of the binding agent
A prerequisite is a finely ground sample powder. Depending on the sample type, three drops of Moviol will be used for 5 to 7 g samples, for example to be used with Polysius rings, aluminum rings or "freely" pressed 30 mm samples. For "freely" pressed 40 mm samples (9 to 10 g), start with 4 drops.

Gently grind the sample containing Moviol using a mortar and pestle, so that no clumps are visible. Afterwards, transfer the sample to a press. If the sample doesn't bind firmly, then the amount of Moviol needs to be raised (quartzite or very quartz-rich samples may need up to 50 drops / 10 g). Due to the volume of 150 ml = 1 drop, the concentration of C, H and O can be neglected.

User Tips
For TiO2 and other “pasty” substances, it is helpful to cover the surface of the sample with a thick Mylar foil before pressing. A standard 12-mm foil is successful for this purpose. The use of the foil keeps the sample from sticking to the plunger of the press and ensures an even smoother surface.

In the case of smaller amounts of a sample substance, one can prepare the sample by using the sandwich method:

	Sample material
	Boric acid

The aforementioned sample will be pressed onto approximately two spatulas full of boric acid. The use of non-backscattering boric acid is recommended. For XRF applications, it is necessary to make the sample layer on the boric acid so thick that the sample appears to be infinitely thick. As a general rule, the thickness should be at least 1.5 mm.

For samples containing higher amounts of Moviol, the sample should be dried before measurement with the XRF analyzer, otherwise chamber evacuation time takes too long.

3.2.3 Fused Beads

The dissolution or decomposition of a portion of the sample by a flux and fusion into a homogeneous glass eliminates particle size and mineralogical effects entirely. The fusion technique also has additional advantages:

Possibility of high or low specimen dilution for the purpose of decreasing matrix effects
Possibility of adding compounds such as heavy absorbers or internal standards to decrease or compensate for matrix effects
Possibility of preparing standards of desired composition

The fusion procedure consists of heating a mixture of sample and flux at high temperatures (800 to 1200°C) so that the flux melts and the sample dissolves. The overall composition and cooling conditions must be such that the end product after cooling is a one-phase glass.

Heating of the sample-flux mixture is usually done in platinum alloy crucibles, but graphite may also be used when conditions permit.

The more frequently used fluxes are borates, namely sodium tetraborate, lithium tetraborate and lithium metaborate. A mixture of these fluxes is sometimes more effective in certain cases.

← Prev Next →