The short answer is that fused beads provide a near perfectly homogeneous representation of the sample to the XRF and is considered by many to be the ideal sample preparation method.

Pellets however may contain different grains with respect to their grain sizes and their composition.

This may have an influence on your results.

You may also look up the link for some more information.

http://labauto.flsmidth.com/blog/the-5-most-common-ways-to-prepare-samples-for-xrf-analysis

Thank you, Fused disks it is!!!.

In pressed pellets, sheet-like and/or platy minerals (i.e., micas, amphiboles, pyroxenes) cover more areas relative to feldspar and quartz grains, and cause in homogeneous surfaces for whole rock sample. Thus, your results should be influenced incorrectly. Finally, fused beads provide homogenous nature for samples as stated by Philipp Hoenicke.  

Just to play devils advocate!

While I strongly prefer fused beads for all the above stated reasons, there are a couple of disadvantages.  The dilution factor in making fused beads is  greater, i.e. there is more sample to binder/flux in a pressed pellet so you get a better signal for trace elements.  One of the old labs I started out in (Leeds Uni., 20 years ago ... ) did both, fused disk for majors and some trace, and a pressed pellet for trace.  With modern instruments this is rarely a problem.

Fusion can also result in the loss of some volatile elements, such as Na.  As long as you use appropriate (i.e. similar to your unknowns) standards and prepare them exactly the same way as your unknowns, this is not a problem.

You have one complicating factor in that you are examining sediments that will contain a significant proportion of structural water and other volotiles.  This will of course be lost during fusion unless you drive it off first with an LOI proceedure, the results of which you will then add back in during data processing.

That being said, I would still use fused beads every time!

best regards from Sweden, George!

Thank you George, The standard procedure for the lab I will be using is Fused beads / disks for Majors and pressed pellets for Trace so it is exactly as you said.  The sediments are Oligo-Miocene rocks so the volatile element is less of a concern.  We just need to find some spare platinum crucibles (we're hoping Chemistry might have some)

Thanks again!

Dear Mrs. Blayney,

the basic arguments have already been put forward by the colleagues above. I would like to add some details concerning the application of XRF in this case.

The preparation of  beads has proved to be the common processing for major and minor elements from Si to P and Bi to Zr. Among the trace elements the pressed pellets to my knowledge are recommended   for Ag, Ba, Cd, Ce, Cs,I, In, La, Nd, Sb, Sm, Sn, Te.

I wish you much success

Harald G. Dill

Fused beads give a homogenous surface hence it is better and accurate for estimation.Na,K as mentioned by some colleagues is not a problem because alkalies are  for accuracy often estimated by other techniques. However, pressed pellets can also be used for trace elements as suggested by some colleagues and better compared with fused disc results for full confidence?

Fused disks use smaller amount of powder (~400 mg). They are prepared by melting the mixture of the sample powder and flux (e.g., lithium tetraborate) in a platinum crucible at 1200C. They are usually used to measure major and some minor element concentrations by XRF, but the precision for minor elements (e.g., Cr, Ni, Co, V), not to mention trace elements, is much worse than for analyses done in pressed powder pellets. The latter are made out of ~4 grams of sample powder and are more representative samples. Usually, you need both for a quality analysis, especially if the amount of material is not an issue.

I have already mentioned that it is better to use both i.e fused disks and pressed pellets for accuracy of analysis, if sample quantity is not a problem ,and i agree with my colleague. If sample quantity is too small then better other methods we followed for accuracy.

The pros and cons of both techniques are already mentioned. In sum you could say:

Fused beads:

• + well established
• + fast
• - only major elements with good analytical quality (excluding volatiles)
• - inhomogeneous if not steered during melting
• - loss of volatiles
• - matrix modification 
• - dilution by flux, worsening LOD
• - possible contamination by flux in the ultra trace element range

Pressed pellets:

• + highly homogeneous
• + wide element range with good analytical quality 
• + no dilution
• + no binder needed 
• + no matrix modification
• - orientation of micas

The major issue concerning pressed pellets, is the grain size. Reduction to nanometer scale leads to: (i) increase in homogeneity (spacial distribution of phases) to a range that LA-ICP-MS with high precision and accuracy is possible; (ii) increase of particle cohesion allowing for binder-free manufacturing of pellets (no dilution/contamination). To achieve grain size reduction to nanometer scale, a clean workspace, a high power planetary ball mill (agate vials), ultrapure water and a freeze drier is needed. With precaution to rocks/samples that show high mica content there is no clear advantage of fused beads over pressed pellets, considering the afore mentioned. 

If you just want to do major element XRF, fused beads might be the method of choice, but as other colleagues already mentioned if you are interested in minor, trace or ultra-trace elements and subsequent geochemical characterisation of your sample by e.g. LA-ICP-MS, I recommend using pressed pellets only.

Fused discs are used to analyse for major elements, whereas pressed powder discs are used to analyse for trace elements. The fused discs dilute the major elements to a level where they are within the upper detectability bounds of the XRF. I you want to analyse for major elements, you must use fused discs.

Fused disc gives a completely homogeneous sample for a reproducible and reliable data with matrix effect minimum as my colleague Ms Jennifer has mentioned.The grind powder pressed on a pellet can never be homogeneous to that extent due to micaceous, flaky miners etc.hence it is  unlikely to give precise reproducible data.I have already mentioned earlier that if there is no limitation of sample it is better to conduct both but with running standard  of both i.e fused disc and press pellets to compare the accuracy.

My experience holds the same opinion as expressed earlier by me.As the flux used to fuse the sample for making glass disc does not interfere with the sample, dilution of sample is unlikely. This is also evident by running standards.

Dear Tamsin,

I have the same opinion of Ashkan!

From my experience, both methodologies can be used to determine major elements in rock samples. I have tested reference samples and obtained similar results using both methods.

The main differences between the results are due to errors that occurred during sample preparation and during the calibration curve settings and equipment configuration. These problems can occur for both methods.

Indeed, fused disc has all the advantages cited by Brijraj and Tamsin. On the other hand, It has the disadvantage of diluting the sample. It may be a problem for samples with very low concentrations as trace elements.

Best Regards,

Ygor Jacques

As mentioned earlier fused disc has a great advantage of homogenous samples, there cannot be better method to obtain homogenous sample than fusion it is obvious that data will be very precise.On a pressed pellet data generation  because of micaceous minerals which  cannot be grind  beyond  a certain limit may make minor difference. Therefore, standards must be run every time for comparison.In fact both methods be used and compared. Of course instrumental error could occur because of  use of different instruments which should be taken care off. Low amount of samples may not be a problem? Even in ICP estimation of  trace & REE etc.the sample in solution is low but the precision is not disturbed?

Brijraj Krishna Das

Almost all advantages and disadvantges of using glass-disks or pellets are mentionned in the various comments. But one important observation is missing, that is specific to X rays: the matrix effect, that is the variable auto-absortion of X rays by samples irradiated for analysis, glass-disk or pellets.

Glass disks: the amount of sample in the glass is low (dilution). The matix effect is buffered by the melting reagent. It is even recommended that the dilution is high. It is even recommended to add an heavy absorber (ie: La oxyde) to reach a better buffering . In other words, the contribution of the matrix effect of the sample in the glass disk is neglible with respect to the over-all matrix effect of the glass disk. The intensity of emitted X rays is directly proportionnal to the concentration of the element in the glass disk. No correction needed. This method can be used for major and minor elements which support the lowering of emission due to the dilution and / or to the heavy absorber.

Plellets: to be used for trace elements whose intensities of emitted X rays do not support the buffering technique above. But it does need the appropriate corrections for the matrix effect of the sample. The auto absorption by a basalt is twice larger than that by a granite ! The matrix effect of a pellet can be accurately calculated from its major element composition that can be obtained through the glass disk method above.

Pellets, background correction. Correcting for the matrix effect is of course an absolute necessity. In addition to this correction, the determination of the background needs a specific care. The backgroung is an image of the background of the X ray tube irradiating the sample. This image is made through the elastic / inelastic "reflexion" of photons from the X ray tube to the detector. This "reflexion" is function itself of the matrix effect of the pellet. In addition to this observation, the X ray tube may contain a ray (weak but can be of importance for trace element determinations) whose wavelengh cannot be separated from the emitted ray by the trace element involved. This correction, as well, can accurately be accounted for.

We run a Van XRF Laboratory on board the Glomar Challenger (DSDP, Deep Sea Drilling Project) from 1974 to the mid eighties, initially for major element determinations and then for both major and trace element measurements. The stategy of Leg 82 of DSDP relied on shipboard trace element determinations (Nb, Zr, as "equivalent" of La and Sm). Details can be found in the Initial reports of DSDP (ie: Leg 37, Leg 82).

some references:

Bougault, H. and P. Cambon (1973). "Dispersive X-ray fluorescence analysis on board oceanographic vessel." Marine Geology: 37-41.

Bougault, H., P. Cambon, et al. (1977). "X-Ray spectrometric analysis of trace elements in rocks. Correction for instrumental interferences." X-Ray Spectrometry 6: 66-72.

Bougault, H. (1980). Contribution des éléments de transition à la compréhension de la génèse des basaltes océaniques. Analyse des éléments-traces dans les roches par spectrométrie de fluorescence X. Paris, Paris VII, France: 221 (in french)

Etoubleau, J., H. Bougault, et al. (1985). Analysis of Trace Elements in Basalts by Shipboard X-Ray Fluorescence Spectrometry: a Discussion of Niobium. Initial Reports of the Deep Sea Drilling Project. H. Bougault and S. C. Cande. Washington, U.S. Govt. Printing Office. 82: 35-43.

Mr. Henri has elaborated the same thing which we have been mentioning.His analysis can make issue better understandable to which I agree.

Brijraj Krishna Das

A major issue we had with pressed pellets using them for major element analysis was getting low SiO2%, hence low totals. We think this is mainly due to matrix/mineralogical effects.

Fused glass disks are used for major and minor element analysis. Pressed powder pellets are used for analysis of trace elements, as well as minor elements to a higher degree of accuracy. Pressed powder pellets use ca. 4 grams of powder. Fused glass disks - 0.4 grams. If you do not have enough material, you have only one option.

I work as an XRD and XRF technician and I surely trust both the methods based on the same theories Henri has mentioned above.@Henri Bougault

I completely agree with all the above arguments. All said and done, it now depends on the requirement. You have two options....

1. Do you just need to screen, classify the geological samples based on the elemental composition or proceed for mineral exploration or have large number of samples for screening etc.[ Need is determination of anomalous values as high as possible even with low precision]

2. Generate data for mapping, use data for elemental ratios etc.[need is determination of as low values as possible, with high precision]

For (1) you just need to validate the method first by participating in Proficiency testing or by following the protocol.

(Pl grind the samples to the same size as that of the CRMs. Calibrate the instrument with CRMs, do the alpha corrections and proceed for analysis using pressed pellets. Of course need to be careful while finalizing the calibration curve and alpha corrections). So you may manage with the equipment available with you.

For(2) you need to compulsorily adopt the procedures mentioned by Bernard, Brijraj and others.

Fusion is more suitable for major/minor element identification while pressed pellets could be considered more suitable for trace elements. The fusion method has a dilution effect of between 0-10, and this may affect the detection limit of target analytes (but pretty much ok for major/minor elements), therefore, you want to want to avoid impacting the detection limit of the trace elements, therefore preferably use the pressed pellets, but at the expense of precision and homogeneity, not so sure again.