Читать книгу X-Ray Fluorescence Spectroscopy for Laboratory Applications - Michael Haschke, Jörg Flock - Страница 39

Higher analytical accuracies are possible by the manufacturing of pressed pellets, since these can produce relatively smooth sample surfaces and uniform sample densities. For this purpose, the ground material is filled into a press mold having a diameter that corresponds to the respective size of the sample holder of the spectrometer. Pressures of 100–200 kN for pellet diameters of 32 mm and of 200–300 kN for pellet diameters of 40 mm have to be typically applied for up to one minute to compress the material. As a result, pellets with a relatively high density are prepared, which are stable and can be stored over a longer period. In order to avoid contamination by the mold, the pressing surfaces can be covered with a thin plastic film, which is removed after pressing the pellet. Figure 3.9 shows a press die. This die allows evacuation of the sample material during the pressing operation. As a result, gas inclusions are avoided, which can destroy the pressed pellets during the pressure changes in the course of vacuum measurements.

Figure 3.9 Press die for to produce pellets from small-sized materials.

Source: Courtesy of Specac Ltd.

Various possibilities exist for the manufacturing of a pressed pellet, which are distinguished, in particular, by the stability of the pressed pellet, but also by the effort for its production; it depends on the starting material as well. Pressed pellets can be made by

 direct pressing of the ground raw material without any additives;

 mixing the sample with a binder and pressing the mixture into a pellet. The binder increases the strength and thus, if necessary, allows multiple use of the pellet. At the same time, a dilution of the matrix is achieved. The binders used are cellulose, wax, or other soft materials;

 pressing the sample material with a binder into solid sample holders such as rings or aluminum molds. The rings are made of steel or plastic. The material of the Al molds must have high purity in order to avoid superpositions of the sample spectrum with that of high-energy fluorescence radiation of impurities. This support of the sample material with molds is particularly useful for often used samples, where particles can flake off and contaminate the spectrometer. The pressing of small-particle material produces stable pellets with very smooth surfaces, as shown in Figure 3.10a;

 simultaneous pressing of the sample material into a substrate made of binder or boric acid utilized as a sample vessel. In this case, a test sample can also be produced with a small amount of sample material. Samples with a binder carrier pellet are shown in Figure 3.10b. Calibration samples can also be produced by this preparation method in order to ensure a high analytical accuracy even in the case of small sample quantities. However, an increased spectral background has to be expected for samples prepared in this way because the high-energy incident radiation is scattered on the light matrix of the sample mold;Figure 3.10 Pressed pellet (a) and cross section of a pressed pellet in binder (b).

 pressing the sample material at elevated temperatures. Particularly in the case of organic materials, this can lead to a gluing of the individual sample particles and thus to stable pellets; see for example in Section 10.13.2.

In order to avoid the sample thickness influencing the analytical results, always the same sample quantities as well as the same mixing ratios with binders should be used for the production of pressed pellets. Depending on the size of the press die, the usual sample amount is 3–5 g. The binders are added in constant mixing ratios with proportions between 10% and 15%. As a result of the same sample masses, the sample thickness is either greater than the information depth of the analytes or the two parameters are always in the same relation to one another. The addition of binders must be considered for the quantification, both for the matrix interaction and for a possible normalization to 100%.

The durability of pressed pellets depends on the hardness of the sample particles as well as on their grain size distribution. Ultimately, the size of the surfaces of the individual particles contacting each other is important. If this area is large, then the van der Waals forces between the individual sample particles are sufficient and the pellet is mechanically stable. The various possibilities for producing stable pellets are demonstrated in Figure 3.11.

 In the case of hard particles with a narrow grain size distribution (for example, spheres of the same size as in Figure 3.11a), the closest sphere packing is obtained by pressing, but it has only relatively few and small points of contact between the individual sample particles. No stability of the pellet can thereby be expected.

 If the sample particles have a broader particle size distribution, smaller particles can be incorporated between the gaps of the larger particles. This increases the number of points of contact and thus the stability of the pellets. This case is illustrated in Figure 3.11b by the inclusion of small spheres in the interstices of the distribution of Figure 3.11a.Figure 3.11 Influence of grain size distribution and different hardness of sample particles on the durability of pressed pellets, (a) spheres of same size, (b) spheres of different sizes, (c) particles of differing hardness, and (d) incorporation of binders.

 If the hardness of the sample particles is different, the pressing can deform the softer sample parts. They then adhere to the harder parts so that the contact surfaces are significantly enlarged and thus improve the stability of the pellet (Figure 3.11c). If no soft components are present in the sample itself, a binder can be added to the sample. This has the same effect.On the other hand, these soft components tend to segregate to the edges during pressing, i.e. also on the sample surface (see Figure 3.11d). These soft components, in particular the binder then forms an additional absorption layer, which especially influences the low-energy fluorescence radiation of the sample components and thus the analytical result.

Various binders are available. It is important that the binder does not contain any analyte elements. Binders usually are mixed into the sample during the last grinding process; they are ground with the sample material and thus homogenized. These additives often also support the grinding process by preventing particle aggregation. The binders are available as pellets with a defined weight; therefore, no additional weighing is necessary during portioning binder and sample. This is particularly important for an automation of workflows. Liquid binders are also used; they are often the only way to produce a stable pellet. In this case, it is necessary to check whether these binders dissolve elements from the sample; then the evaporation of the solvent can lead to a depletion of these elements.

The selection of an optimum binder as well as the mixing ratio with the sample for a given analytical problem usually requires tests, since there is a dependence between sample material, grain size distribution, and the available preparation technique. Table 3.8 shows a summary of the most commonly used binders.

Table 3.8 Binders and additives for pressed pellets.

Binder	Function	Properties/application
Boron acid Borax	Additive and binder Also advantageous for sample stabilization as sample mold	No longer allowed, slightly toxic
Paraffin wax	Mostly binder	Slightly toxic, no influence by moisture
Cellulose	Mostly binder	Absorption of moisture
Methylmetaacrylate in a solution of acetone	Binder for materials that enlarge their volume due to water absorption	Mixing with the sample and wait for evaporation of the acetone, then pressing
Polyvinyl alcohol solution	Additive and binder, avoids aggregation of the grounded material and cools down the mill

The quality of sample preparation influences the analytical accuracy and the reproducibility of the analyses. The manufacturing of pressed pellets produces a much higher quality measurement sample compared to the simple loose powder samples of small size materials; it improves the analytical accuracy. For mineralogical material it is in the order of 0.5% for main components and <1% for secondary components.