One of the most common methods of preparing samples for XRF analysis is by making pressed pellets. This process is particularly popular as it produces high quality results, is relatively quick and is a low cost approach. The process also lends itself to simple and cost effective automation for higher throughput laboratories.
The process of making pressed pellets for XRF analysis includes grinding the sample to a fine particle size, mixing it with a binder/ grinding aid in a grinding or mixing vessel, pouring the mixture into a pressing die and pressing the sample at a pressure of between 15 and 35T. The resulting pellet or tablet is then ready for analysis. While this is a common and relatively straightforward approach to preparing samples for XRF analysis there are five important things you should consider when designing a sample preparation recipe. These include the particle size of the sample, the choice of binder,the dilution ratio, the amount of pressure applied to the sample and the thickness of the final pellet.
One of the key aspects of preparing pressed pellets is to make sure that the sample is ground to a particle size of <75um, but <50um is ideal. Depending on the sample type and volume this can usually be accomplished with many types of pulverizing mills. Small particle size is important to producing pellets that provide the best analytical results because it effects how the sample will compress and bind together when pressed. More importantly samples with larger or variable particle sizes can lead to heterogeneities in the sample. The sampling depth or critical escape depth, for elements contained in a sample is energy dependent and thus not the same for every element. Longer wavelength elements such as Na will have smaller escape depths than shorter wavelength elements such as Fe. This means that Na analysis is only sampling the first 10um of a sample and thus more susceptible to sample heterogeneities on that scale.
Larger particle sizes can result from the presence of hard to grind phases in a sample resulting in variable grinding of the sample. The presence of larger particle sizes at the analysis surface of the sample can lead to analysis errors through the "shadow effect" where larger grains shadow the X-ray signal coming from the smaller grains at the surface of the pellet.
Sample binder is literally the glue that holds the sample together. The optimal binder is usually is a cellulose/wax mixture that homogenizes with the sample during grinding and will recrystallize under pressure to bind the sample together. These binders are available under various commercial names but are generally similar. Sometimes acrylic binders are used, but these are difficult to homogenize with a sample in a mill and need to be mixed by hand. Some binder grinding agents are available as pre-mixed pellets that can be automatically added to a mill during grinding.
How much binder should I add? Since most modern X-ray instruments will provide plenty of intensity for major elements you are safe in using plenty of binder in order to insure a good strong pellet. A 20-30% binder to sample ratio will almost always produce a very strong pellet that can be dropped onto the ground from a few feet without breaking. If you want to use less binder to lower your consumable costs then you will need to experiment in order to find the optimal dilution level.
The pressure applied to a sample should be sufficient to compress the sample completely and recrystallize the binder. One of the most important factors is making sure that the sample is fully compressed so that no void spaces remain in the pellet. A pellet may look good but still contain void spaces which result in a lowered intensity for lighter elements. You should experiment with your samples using increasing pressure until the intensity for the light elements reaches a maximum and stabilizes. Most samples will reach this maximum at 25-35T of pressure applied for 1-2 minutes. It's also important to release the pressure applied to a sample slowly so as to prevent cracking of the sample surface.
The thickness of your pellet is important for achieve the best analytical results. The pellet must be infinitely thick to the X-ray beam for all of the elements you are measuring. X-rays produced in the sample must be able to escape the sample without being reabsorbed in order to make it to the detector to be measured. Higher energy elements (usually those with higher atomic numbers) will have deeper escape depths in a sample relative to lower energy elements. The signal coming from higher energy elements sample more of the pellet than those coming from lower energy elements. The samples ability to reabsorb X-rays is directly proportional to the samples average atomic mass. The absorption capability of any sample can be calculated from its average elemental composition and the mass absorption coefficient for each element present.
The problem arises when the escape depth of a particular element is larger than the thickness of the pellet. Because this is an element dependent problem, a particular pressed pellet may provide good results for some elements but not others. In general, you should make sure your pressed pellet is thicker than the escape depth of the highest energy element you want to measure.
Usually a pellet made with 8-10g of sample for a 32mm diameter pellet or 13-15g of sample for 40mm diameter pellet will be sufficiently thick enough for the elements that can be measured by even the most powerful WDXRF instruments currently available.
If you keep these 5 key areas in mind when designing your sample preparation recipe for pressed pellets then you will be able to produced robust high quality pellets that will provide you with the best analysis results your XRF is capable of producing.
Interested in Learning More about Sample Preparation for XRF and XRD ?
FlSmidth provides laboratory sampling and sample preparation solutions from our manual Jaw crushers and pulversing mills to our fully automated XRF sample preparation solutions for fused beads and pressed pellets.
Visit us at http://labauto.flsmidth.com/