| |
Setting Process Analytical Calibration Factors |
|
| Example: A group of samples was homogenized and split. A portion of each sample was sent to an outside lab for analysis. The other portion was analyzed with the Innov-X analyzer. These were extremely good, samples, very well homogenized and characterized, so the correlation is quite good. Many real world samples may not show correlations that are as good. |
The data are summarized here: |
|
| All the measurements on the Innov-X analyzer are reading approximately 12% too low for Cu and 20+ % for iron. For some customers, this error may not be a problem, for example, certain customers may just be looking to determine “high” or “low” according to a set of criteria, and may not be looking for accuracy. Other customers may want to use the analyzer to get more accurate numbers. In this case, it is necessary to calibrate the analyzer for these samples. |
| Procedure: |
| Plot the data. Innov-X data must be on the X-axis, lab values on the y-axis. (*this order is important) Determine the linear best fit, and both the slope and intercept. |
|
|
| |
| The slope and intercepts for these graphs can be entered directly into the analyzer. In some cases, such as with the copper data, it is sufficient to enter just a correction for the slope as the intercept is almost zero. In others, such as the iron example the slope and intercept should be entered. |
| The software allows a user to enter multiple sets of calibration factors for different applications, or different ore bodies. A group of factors is given a name, and then the factors are entered. The factors can then be selected by name. |
| From Process Analytical mode, |
- Select Options -> Edit Factors
- Enter the administrative password
- Select Add to add a new group of factors
- Provide a name
- Enter factors for each element. Save Changes when done.
- Select File ->Load Factor Set to change the selected group of factors.
|
|
| After adding correction factors, the results are as shown: |
|
Alpha Series |
portable XRF for Cu, Zn, and Pb ore analysis in the mining industry. |
| Overview. |
|
Application. The measurement of metals in ores is important for a variety of applications,ranging from exploration and survey, to final processing. Portable X-ray Fluorescence (XRF) is an effective means to achieve quick, on-the spot results of metal concentrations. In this report, a group of samples, with a wide range of copper, lead and zinc concentrations were measured. Results obtained with the Innov-X Systems Alpha Series portable alloy analyzer were compared with data from laboratory analysis. |
| |
| Hardware. The Alpha Series is a handheld X-Ray Fluorescence Analyzer which uses the latest in XRF technology. A Silver Anode Miniature X-Ray tube combined with a solid state SiPIN detector produces quick and accurate results. The tube operates at 35 KeV, 15uA. Detector resolution is better than 220 eV. The analyzer is controlled by an HP iPAQ pocket computer running Windows CE. |
| Software. Innov-X Systems Materials Analysis software program uses a Fundamental Parameters (FP) algorithm, which automatically takes into account many inter-element issues which can affect XRF results. This FP algorithm can be fine tuned for matrix differences by a series of calibration factors which can be stored in the analyzer. Multiple sets of calibration factors can be saved in the analyzer for easy recall. |
|
| Method. |
| The samples were submitted as fine powders in plastic bags. They were assumed to be homogeneous, and were analyzed, as is, in the plastic bags, with no further sample preparation. This is thought to be a typical style of analysis for samples of this type. Uniformly homogenizing the samples and preparing them in samples cups may slightly improve the accuracy and precision of these measurements. Had the samples not been well ground and homogeneous, additional sample preparation would be recommended. |
| A series of known samples were tested using the analyzer. These data were then compared to laboratory data and a calibration curve was established. The calibration values were stored in the analyzer, and the samples were analyzed for 60 seconds. |
| |
Alpha Series |
provides Cu, Zn, and Pb ore analysis for the mining industry. |
|
| Results. |
| Without a site-specific calibration, the XRF analyzer produced good linearity and reproducibility over two orders of magnitude in concentration – about 0.2% to 20%. Correlation coefficients exceeded 0.95 in all cases. The analysis did show a constant 30% bias over this range with respect to laboratory analysis. |
| This is typical for portable XRF analysis as the low atomic weight elements including O, S, C, Al, Si, Mg and others are not measured. However, given the bias is virtually constant over a large concentration range, a few samples with known concentrations may be used to generate an empirical correction. |
The empirical correction, combined with the fundamental parameters analysis, provides good accuracy over a large concentration range with minimal ore-specific calibration requirements. The
user can apply a simple bias correction empirically, to produce results in very good agreement with the laboratory. |
| Figures 1, 2 and 3 show plots of data obtained with an Innov-X Systems Alpha Series? alloy analyzer after applying the above noted bias correction, versus values reported by a lab. In all cases, there is a very strong correlation between the data obtained with the portable analyzer and the laboratory. |
| The R2 correlation for the linear regression is at least 0.99 for all three data sets. All three data sets encompass a wide range of concentrations, indicating the applicability of the samples. |
| |
|
|
|
