Which REEs can XRF realistically measure and why some are harder

Each element has several groups of characteristic lines (K, L, M). For most practical XRF tasks involving REEs, L linesare used, because REE K lines require higher excitation energies than are typically applied in routine measurements—especially in portable instruments. This leads to a key implication: REE L lines fall in a part of the spectrum where many other elemental lines and elevated background are often present. That is why correct spectral processing and a well-designed measurement methodology are particularly important for REEs.

In general, XRF can work across the full lanthanide series La–Lu, but practical “ease” varies significantly.

Typically easier (with proper preparation and settings)

Mid and heavy REEs (roughly Sm–Lu) are often determined more consistently. This is because their working lines are, on average, less sensitive to absorption in the sample matrix and to low-energy signal losses during “in-air” measurements. In addition, in some cases their lines are easier to separate correctly during spectral processing.

Typically the most difficult

The most frequent issues involve light REEs (La, Ce, Pr, Nd). Their L lines sit in a densely populated region of the spectrum and are more likely to overlap with lines from other elements and with background structure. At trace concentrations, this can lead to unstable results: even small changes in the matrix or in spectral deconvolution settings can noticeably shift the calculated concentrations.

A special note should be made about Y (yttrium), which is almost always considered alongside REEs in geochemical applications. A common challenge for Y is overlap with Rb (rubidium): depending on the matrix and the effectiveness of spectral separation, systematic bias in either direction may occur.

Concentration and measurement time

REEs are often present at ppm to hundreds of ppm, so counting statistics matter. Increasing measurement time does reduce random uncertainty, but it is important to understand the point of diminishing returns: at some stage, the dominant contribution to error comes not from statistics, but from line overlaps and matrix effects. Therefore, “measuring longer” helps, but it does not replace a correct methodology.

Spectral overlaps (the main source of errors)

For REEs, it is critical that their L lines are often located close to one another and close to lines of common matrix elements. In real ores and concentrates, this can lead to cases where part of the intensity is incorrectly attributed to an REE during spectral deconvolution (or vice versa). The highest risk arises when two conditions coincide: low REE concentrations and the presence of strong lines from interfering elements in the same energy region.

Matrix effects (absorption and enhancement)

XRF measures line intensities rather than “concentration directly.” Intensities are strongly influenced by the matrix: major elements, density, mineral phase composition, degree of grinding, porosity, and moisture. For REEs, this effect is especially noticeable because the working part of the spectrum is already complex, and any matrix change also alters the relationship between useful signal and background. If calibration is built on materials with a different matrix, systematic bias may occur.

Measurement conditions (air vs helium/vacuum)

The low-energy part of the spectrum is more strongly attenuated in air and within the sample itself. Therefore, in laboratory configurations where vacuum or helium purge is available, measurements for certain elements can become more stable. In portable practice, this typically translates into higher requirements for sample preparation and for maintaining repeatable measurement conditions.

Detector resolution and quality of spectral processing

For REEs, the system’s ability to separate close peaks and to perform stable deconvolution is essential. In this sense, spectral processing quality and detector parameters directly affect the likelihood of misassigning the contribution of neighboring lines—especially in the region of light REEs and in the presence of interfering elements.

Sample heterogeneity and lack of homogenization

REEs are often concentrated in specific minerals and distributed unevenly. Measuring a rock fragment or coarse material can produce large scatter because, in different spots, you are effectively analyzing different phase proportions. This is especially evident at low concentrations, where hitting (or missing) an REE-bearing grain can strongly change the result.

Practical takeaway: if comparability is required, grinding, homogenization, and repeatable sample preparation conditions are necessary.

Unstable geometry and surface condition

For powders and loose materials, consistent layer thickness, packing density, and absence of moisture are important. Any change in these parameters alters absorption and scattering conditions, and therefore background and line intensities. For REEs this is critical, because measurements are often made in a region where the background contribution is significant.

Incorrect calibration and transferring a method between different matrices

Even if an instrument can “do REEs” in general, transferring a calibration between different ore types (for example, iron-rich vs phosphate vs carbonate matrices) can produce systematic errors. This most often affects light REEs and elements for which strong overlaps are typical.

In many tasks, the goal is not laboratory-grade accuracy for each element, but a fast way to identify promising zones and streams. In such cases, it is reasonable to use indirect indicators if direct determination of individual REEs is unstable. In practice, this is done in three ways:

1. Estimating REEs via associated elements and mineralogical logic, when it is known which phases and impurities host REEs in a particular deposit (e.g., phosphate association, etc.).
2. Threshold criteria and screening indices, where classification (low/medium/high) is more important than precise ppm values for each REE.
3. The “XRF → ICP confirmation” workflow, where XRF is used for high-throughput screening and ranking, and accurate quantification is performed by ICP-OES/ICP-MS on selected “target” samples. This is often the most cost-effective approach.

A more accurate way to put it is: XRF can provide useful REE results, but quality depends on the specific analytical goal.

• For screening and comparisons within the same site using a consistent methodology, XRF is often very effective.
• For trace-level concentrations of certain light REEs and complex matrices, stricter conditions are required: sample preparation, matrix-specific calibration, control of overlaps, and, when needed, confirmation by laboratory methods.