FUSIONS – HOW TO IMPROVE THROUGHPUT AND CONCENTRATION RANGE OF ANALYSIS BY ELIMINATING THE LOSS ON IGNITION PROCESS STEP, USING DIFFERENT DILUTION RATIOS AND MAINTAINING ACCURACY AND PRECISION OF RESULTS Laura Oelofse (Rigaku Americas) and Yoshijuro Yamada (Rigaku Corporation )
52
Embed
Fusion-How to Improve Throughput by Eleminating the Loss...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
FUSIONS – HOW TO IMPROVE THROUGHPUT AND
CONCENTRATION RANGE OF ANALYSIS BY ELIMINATING
THE LOSS ON IGNITION PROCESS STEP, USING DIFFERENT
DILUTION RATIOS AND MAINTAINING ACCURACY AND
PRECISION OF RESULTS
Laura Oelofse (Rigaku Americas) and Yoshijuro Yamada (Rigaku
Corporation )
Abstract
The use of fusions for XRF in industrial process monitoring is common
practice and there are several time consuming steps to complete in
order to render a sample fusion ready.
This paper details a method that would eliminate the need to carry out
the Loss on Ignition, Gain on Ignition step thus eliminating 2 hrs from
the preparation time and it also details the ability to use different
dilution ratios of sample and flux for materials on the same calibration
curve in order to increase the scope of materials that can be included in
a universal calibration curve using both naturally sourced certified
reference materials and synthetic pure chemicals as calibration
standards
Analysis Schemes in Various
Industry Sectors
Product Incoming Raw
Materials
Composition
varies in
narrow range
Widely
varying
composition
XRFs Role in High Throughput
Solutions
Results
Sample Preparation
Sample Loading
Sample Introduction
Analysis
Fusions – Flux + Sample
All compounds changed to oxide form
Eliminate Particle Size Effect
Eliminate Mineralogical Effect
6
Some Typical Fused Glass Beads
7
Why Preparation of Fused Glass Beads
• Particle size and mineralogical effects are removed or diminished by fusion of the sample with a suitable flux to form a glass bead.
• Synthetic calibration standards can be made by mixing pure oxides at concentration levels to suit the analytical range.
• Depending on sample type, fusion can even be quicker than pressed powder procedure.
• Generally accuracies and reproducibility are superior with fusion procedure.
• Only drawback is possible dilution of trace elements and therefore inferior LLD. Low dilution fusions are possible.
8
Preparation of
Fused Glass Beads
9499D00500
To standard holder
Melting method
Weigh outand mix
Flux + Specimen
Heat forMelting
Platinum crucible Remove bubbles
Glass diskspecimen
1000° -1100° C
Cast & Cool
8 - 15 mins
5 mins
3 - 7 mins
9
Preparation of Fused Glass Beads
• Preparation of powder as fused glass bead involves weighing out
sample and flux, placing in Gold/Platinum crucible, heating to 1000
– 1200 degrees and casting as a flat glass bead by pouring melt
into a heated Gold/Platinum mould and cooling under controlled
conditions.
• Newer alternative is moldable where melt remains in crucible and
bead is formed in situ.
• DEPENDING ON NATURE OF SAMPLE – LOSS ON IGNITION
MAY BE NECESSARY. ( CEMENTS, LIMESTONE, DOLEMITE)
In reality there are three types of samples Type 1 – Sample is stable, no loss or gain during fusion process
Type 2 : Sample loses CO2 or Intrinsic Waters during fusion, known as Loss on Ignition
Flux
S F
Sample
LOI
S Replaced with flux
Sample
B
Flux
Flux
LOI
S F
Sample
Type 3: Sample loses CO2 or Intrinsic Waters during fusion, known as LOI and changes oxidation state and pick up oxygen, gaining on ignition, known as GOI
GOI
Analytical Error Factors in Fusion Method The errors can be removed by Rigaku Bead Correction method
Powder Sample
Fusion Bead
Heterogeneity Effect
Grain-size Effect
Mineralogical Effect
Weighing Error
Loss on Ignition
Gain on Ignition
Evaporation of Flux
Error factors
for Bead
Error Factors in Fusion Method
Weighing Fusing
Weighing error
Bead
Sample
Flux(Li2B4O7 etc)
H2O CO2
LOI O2
FeO Fe2O3
GOI Flux evaporation
1000-1200˚ C
Pt crucible
The four error factors can be corrected.
(1) (2) (3) (4)
Strategy for the Corrections of LOI, GOI and Dilution Ratio
Model 1 : Use of Ratio of flux to sample weight ( F/S)
Definition of LOI : Imaginary component with no x-ray absorption Correction LOI(GOI) : Concentration is manually input or calculated as balance Dilution ratio : Corrected by manual input of F/S
Model 2 : Use of Ratio of bead to sample weight( B/S)
Flux
LOI
S F
Sample
Definition of LOI : Imaginary component of flux Correction : LOI(GOI) : Concentration is manually input or calculated as balance Dilution ratio : Corrected by manual input of B/S (Note) Flux evaporation can be corrected
LOI
S Replaced with flux
Sample
B
Flux
Analysis in Fused Beads
Use of fusion bead correction
Software generates theoretical alphas for LOI/GOI and dilution ratio
)RαWαWαKC)(1bI(aIWi FFLOILOIjj2
The alphas correct for
LOI/GOI
and Flux evaporation during fusing
Dilution ratio
Calibration equation
The alphas are generated by using a fundamental parameter software and it
generates variety of models.
Dilution ratio models : Flux weight to sample weight(F/S) or bead weight to sample weight(B/S)
LOI/GOI : Loss eliminated or manual input
Rigaku Fusion Bead Correction Software
• Allows for varying flux : sample ratios and the use of
catch weights
• Allows calculation of the Loss on Ignition /Gain on
Ignition and flux loss component by balance or ratio input
Oxides BCS376 LOI 0 LOI 10
SiO2 67.1 67.42 60.68
TiO2 0.02 0.02 0.02
Al2O3 17.7 17.79 16.01
Fe2O3 0.10 0.10 0.09
CaO 0.54 0.54 0.49
MgO 0.03 0.03 0.03
Na2O 2.83 2.84 2.56
K2O 11.2 11.25 10.13
LOI 0.35 0.00 10.00
Total 99.87 99.99 100.01
Sample (g) 0.3000 0.2700
Li2B4O7 ( g) 3.0000 3.0000 -10
0
10
20
30
40
50
60
0 20 40 60
LO
I X
RF
LOI CHEM
LOI CHEM
Lineær (LOI CHEM)
Fusion Bead Correction for LOI in Various Kinds of Materials
SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 LOI
Summary • By applying the dilution ratio correction, LOI correction and matrix
correction (theoretical alphas), it is possible to obtain a good fitting for calibration curves with wide concentration range for diverse natural rocks and minerals by the fusion method.
• With a few calibration standards, it is necessary to use the de Jongh or Lachance-Traill models, where calibration curves are linear in theory. In this test, de Jongh model was used because some samples contain significant LOI content.
• With a large number of calibration standards, it is possible to use the JIS model, where calibration curves can be quadratic.
• With synthetic fused beads to extend the calibration range, it is possible to obtain a good fitting for calibration curves.
Conclusion • The fusion bead method is useful sample preparation for eliminating
hetrogeneity effects, particle size effects, however chemical
reactions can cause the sample weight to decrease or/and increase
during fusion because of volatilization of H2O, CO2 and oxidation.
• It is possible to correct error factors in fusion for either the FP
method or the Empirical Calibration method.
• It is possible to skip the lengthy independent LOI step in preparing
the samples for fusion, by incorporating the step into the fusion
process and correcting for the associated losses and/or gains via
the correction methods just detailed
• It is possible to weigh catch weights and have them recorded in the
dilution correction.
• The time savings realized by eliminating the LOI step and supporting
varying dilution ratios improves throughput and cuts the analysis