X-Ray Fluorescence

The JMU XRF Lab analyzes major and minor elemental oxide concentrations of rocks, sediments, and soils on fused beads.

Elements analyzed currently: Al2O3, CaO, Fe2O3, K2O, MgO, MnO, P2O5, Na2O, SiO2, TiO2, ZrO2

Loss on ignition: measured as total loss (950°C for 105 min), but we can do any temperature/time combo from 0-1100°C.

Contact Bill Lukens if you are interested in having samples analyzed. We occasionally analyzed samples for outside users, for analytical trade or by fee.

Methods

Sample Powder Preparation

Kaliyah Booker (Earth Science B.A., Fall '22) and Joslyn Herold (Geology B.S., '24) providing elbow grease for the sake of science!

Samples (10-50 g) are ground to a homogeneous powder in a Spex Shatterbox using a zirconia ceramic grinding vessel. Powders should be homogeneous and roughly the consistency of baking flour, with no nuggets or grit remaining.
Sample powders are dried in an oven at 105-110°C for a minimum of 2 hours.
Oven-dry powders are cooled to room temperature in a desiccator.

Oven-dry samples are then divided into two aliquots:

Split 1 is prepared as a fused bead.
Split 2 is prepared for loss-on-ignition (LOI) analysis.

Fused Beads

Fused beads are prepared with the following procedure.

6.3000 ± 0.0002 g of flux is weighted into a 95 / 5 (Pt / Au) crucible. Flux is 49.75 / Li-metaborate / 49.75 Li-tetraborate / 0.5 LiBr mix of pre-fused microbeads.
0.7000 ± 0.0001 g of oven-dry powder is weighed onto the flux and gently mixed to homogeneity.
The flux-sample mixture is placed in the Katanax 300 electric fluxing furnace. The fluxing program has a bathing temperature of 1050°C for roughly 12 minutes. The entire process takes roughly 20 minutes per sample with loading, ramping, fluxing, and cooling.

Spex Katanax 300 electric fluxing furnace. This model can hold up to three samples at a time; we a currently set up for one.

Oven-dry powder (Silurian paleosol, left) and corresponding fused bead (right). Darkening of fused bead is due to radiation damage during XRF analysis.

Standard powders, flux, and fused beads are stored in cabinet desiccators.

LOI

Katanax 300 fluxer (left) and muffle furnace (right).

Loss-on-ignition is calculated with the following procedure (Neydon et al., 2022).

A sterile porcelain crucible is weighed and recorded.
Approximately 0.5 to 5 g of oven-dry powder is weighed into the crucible, depending on crucible size. Total weight (sample + crucible) is recorded. Follow instructions on minimum sample size by crucible type due to static-induced uncertainty from the analytical balance.
The sample is placed in a preheated Ney 525 muffle furnace at 950°C for 105 min (1 hr 45 min). Bathing time is started when the furnace reaches full temperature after sample loading.
Samples are removed from the muffle furnace and cooled to room temperature in a desiccator.
The porcelain crucible and roasted sample are weighed and recorded.

LOI is calculated as:

LOI % = [(oven-dry sample wt. - roasted sample wt.) / oven-dry sample wt.] x 100

Oven-dry sample powders (left) and roasted samples (right). Sample positions do not correspond to each other between images. These samples consist of basalt, schist, and shale prepared by the Fall 2021 GEOL 390: Lab Techniques in Geology class. Oxidation of Fe-bearing minerals can result in slight gain-on-ignition (negative LOI values).

XRF Analysis

Fused beads are analyzed on a Malvern PANalytical Zetium spectrometer.

Instrument specifications

1-kW power
Wavelength-dispersive analysis
Rh (rhodium) X-ray tube
Two detectors:
- Scintillation detector for Na to Mn (with Be window)
- Gas-flow detector for Fe to Zr (uses P10 gas: 90 / 10 ratio of Ar / methane)
Five crystals:
- LiF200, LiF220, Ge, PE, PX1

Quantitative Analysis

Calibration curves for 11 elemental oxides are constructed using 23 analytical standards, mostly from the USGS and Geological Survey of Japan.
No corrections are currently used in SuperQ software.
An Ausmond drift monitor (silicates and general) is analyzed at the start of every batch.
Two standards are analyzed at the beginning and end of each batch as unknowns to assess accuracy and precision for each run (14 samples + 2 check standards per run).

Page updated

Google Sites

Report abuse