TL;DR
This paper introduces and evaluates a new, simpler shaker cup (SH) method for preparing rock samples for X-ray fluorescence (XRF) analysis, comparing it to traditional grinding (GR) and stirring rod (ST) methods. The SH method demonstrates superior accuracy and precision, making it a highly recommended alternative for routine XRF analysis of geological samples.
What Problem Does This Paper Address?
Accurate elemental analysis of geological samples is crucial for various fields, including geochemistry and mineral exploration. Wavelength Dispersive X-ray Fluorescence (WD-XRF) is a powerful technique for this, but its accuracy is heavily dependent on sample preparation. Traditional methods for preparing fused glass discs, a common sample form for WD-XRF, involve either extensive grinding (GR), which risks contamination and is time-consuming, or simple mixing with a stirring rod (ST), which can lead to uneven mixtures. This study aims to address the need for a sample preparation method that balances homogeneity, minimal contamination, and efficiency.
Methods
The researchers compared three sample preparation methods for creating fused glass discs for WD-XRF analysis:
- Grinding (GR) Method: Involves grinding the sample and flux in an agate mortar, followed by transfer to a Pt-Au crucible. This method is known for good homogeneity but risks contamination and is time-consuming.
- Stirring Rod (ST) Method: Involves directly mixing the sample and flux in a Pt-Au crucible with a glass rod. This is a simpler, faster method with lower contamination risk but can result in uneven mixing.
- Shaker Cup (SH) Method: A novel method inspired by protein shakers. The sample and flux are mixed in a disposable shaker cup, offering a potentially more homogeneous mixture than the ST method with reduced contamination risk compared to GR.
The study evaluated these methods based on:
- Homogeneity of the sample-flux mixture.
- Potential for sample contamination and loss.
- Sample preparation time.
- Accuracy and precision of the resulting WD-XRF data.
Matrix effects were corrected using a combination of theoretical alpha coefficient and experimental coefficient methods. A dimensionless criterion index (Js) was used to quantitatively compare the accuracy and precision of the methods.
Datasets
- Certified Reference Materials (CRMs): A total of 73 CRMs were used.
- 54 CRMs were used to establish calibration curves.
- 19 CRMs were used to validate the three sample preparation procedures.
- The specific details of these CRMs are provided in Table S1.
Key Insights and Ideas
- Sample preparation is the Achilles' heel of XRF: The paper strongly emphasizes that sample preparation is the most significant source of error in quantitative XRF analysis, influencing the final results more than the instrument itself.
- The Shaker Cup method offers a compelling trade-off: The SH method successfully combines the advantages of both GR (homogeneity) and ST (speed, reduced contamination) while mitigating their respective drawbacks.
- Visual homogeneity doesn't tell the whole story: While mapping images showed similar elemental distribution for all methods, visual inspection of the sample-flux mixture revealed significant differences, with the SH method showing better mixing than ST.
- Volatile element behavior differs: The GR method showed significantly lower chlorine content, attributed to the muffle furnace burning step, highlighting how different preparation steps can affect specific elements.
- Cracking can be an issue with ST: Fused glass discs from the ST method sometimes exhibited crystallization spots and cracking, requiring multiple fusions to achieve a homogeneous disc.
- The Js index is a powerful comparative tool: The dimensionless criterion index (Js) provided a robust, quantitative way to compare the overall accuracy and precision of the three methods, especially when considering elements individually or collectively.
Key Results
- Homogeneity: The SH method produced a more uniform sample-flux mixture than the ST method, and both were visually better than the GR method's initial mixture.
- Contamination and Loss: The GR method (three transfers) was deemed most prone to contamination and loss, followed by SH (two transfers), and then ST (one transfer). However, the SH method's contamination risk could be mitigated with deionized water.
- Time Efficiency: The SH method (3 min) was significantly faster than GR (6 min) and only slightly longer than ST (2 min).
- Accuracy and Precision (Js Index):
- For major oxides, the SH method had the lowest Js value (3.95), indicating the best accuracy and precision.
- For minor elements, the ST method had the lowest Js value (11.36).
- When considering all elements except Zr, the SH method showed the lowest overall Js value (15.20), suggesting it is the best overall method. The poor performance for Zr with the SH method was noted as an area for future investigation.
- LLD: The LLDs for MgO and CaO were higher with the SH method compared to GR, while Al2O3 and Na2O were lower. Other elements showed similar LLDs across methods.
Limitations and Future Work
- Zr anomaly: The study noted a poor result for Zr with the SH method and indicated an intention to investigate this further.
- Minor element data quality: The paper mentions that the data for some minor elements were poor, with concentrations below the Limit of Detection (LLD) or Limit of Quantitation (LOQ), indicating potential issues with those specific CRMs or low concentrations.
- Analytical depth of Sr: The analytical depth of Sr approached the thickness of the fused glass disc, which could affect accuracy and precision. The authors suggest weighing constant sample/flux masses, maintaining constant conditions, and using instrument rotation to mitigate this.
- Further investigation of Zr: The authors plan to investigate the reason for the poor Zr results with the SH method.
- Optimization of SH method: While the SH method is recommended, further optimization might be possible.
Summary updated