Using Automated Multivariate Statistical Phase Analysis of EDS Spectral Imaging Data for Rapid Evaluation of Smelting Copper Compounds

Non-ferrous metal smelting operations involve refining of each element from highly complex compound material containing major elements such as Pb, Zn and Cu. Having knowledge about the complex morphology of the different compounds in the raw material is critical in the Cu smelting process to optimize the refining efficiency of each major element present.

Currently, raw material evaluation is performed using X-ray diffraction (XRD), inductively-coupled plasma (ICP) and electron probe microanalysis (EPMA). However, average, bulk composition data can only obtained with ICP and XRD.

Performing a quantitative analysis of a complex sample composed of 10 to 20 elements and many different phases using EPMA may require several hours.

In addition, the involvement of highly skilled analysts is a prerequisite of an EPMA analysis to run the instrument and to examine a mixed compound material using electron image contrast.

EDS Multivariate Analytical Methods

Energy-dispersive spectroscopy (EDS) now has improved detection efficiency and shortened acquisition times, thanks to the recent advances in Silicon Drift Detectors (SDD).

The EDS spectral imaging data quality has further improved with powerful peak deconvolution methods, approaching the quality of EPMA.

In addition, the advent of EDS multivariate analytical methods has further simplified the analysis of phase distributions as opposed to just elemental distributions.

This article discusses the phase analysis of Cu-compound raw material using the multivariate statistical analysis of EDS spectral imaging data (Thermo Scientific COMPASS software).

Besides shortened acquisition time (30 minutes) compared to ICP or EPMA, the technique can identify the complicated distribution of the different phases.

Experimental Procedure

The experiment involved evaluating Cu compounds produced in the initial process of copper lead smelting subsequent to the extraction of coarse lead. A field emission scanning electron microscope (FESEM) and EDS were used to analyze the elemental distribution and phase morphology. Reflected light optical microscopy was used to observe the analyzed area (a 230 x 180µm rectangle) for metallic components.

SDD (Thermo Scientific UltraDry detector) was used for EDS data acquisition, during which real-time multivariate statistical phase analysis was also carried out. This was followed by polishing and coating of the sample with carbon for the EDS analysis in the FESEM. Acquisition conditions are listed in the following table:

EDS Analyzer : NORAN System 7
EDS Detector : UltraDry 30 mm2
Accelerating Voltage : 12 kV
Magnification : 400x
Mapping Resolution : 256 ??192
Acquisition Time : 30 minutes
Stores Rate : 19,600 cps
Dead Time
:
22%

Experimental Results and Discussion

The electron image and the cumulative EDS spectrum for the EDS spectral imaging acquisition are presented in Figures 1 and 2.

Different potential peak overlaps of various elemental peaks may be present, according to the cumulative spectrum. These overlap conditions are illustrated in Figure 3.

Secondary electron image

Figure 1. Secondary electron image

The cumulative spectrum for all of the pixels in the data set

Figure 2. The cumulative spectrum for all of the pixels in the data set

Peaks of characteristic X-rays of analyzed elements

Figure 3. Peaks of characteristic X-rays of analyzed elements

The elemental distributions by traditional peak count maps are presented in Figure 4. However, there is no correction of peak overlaps determined by Figure 3 in these maps.

In addition, false element intensities may be provided by these overlaps, affecting the elucidation of the spatial distribution of the elements.

Conventional peak count maps

Figure 4. Conventional peak count maps

The results of the peak deconvoluted quantitative elemental maps are given in Figure 5, showing the real distributions of S, Sb, Ca, Al. This was not clearly observable in the traditional peak count maps owing to the overlapping peaks of other elements.

Quantitative spectral images

Figure 5. Quantitative spectral images

The principal component maps extracted by COMPASS multivariate imaging analysis software are presented in Figure 6.

Figure 7 shows the spectra of the first through the eighth key components and the standardless quantitative results of the spectral data, where each component typically has strong intensities.

The first through eighth component maps of the 16 components extracted by COMPASS multivariate analysis software.

Figure 6. The first through eighth component maps of the 16 components extracted by COMPASS multivariate analysis software.


Principal Component spectra by COMPASS and the standardless quantitative results of the raw data of the area where each component intensity is typically strong.

Figure 7. Principal Component spectra by COMPASS and the standardless quantitative results of the raw data of the area where each component intensity is typically strong.

The results reveal that the Cu compounds studied have the following elemental morphologies and distributions:

  • A Cu-Sn alloy phase is the main phase
  • A Cu-Sn alloy phase is present as an associated phase with the Pb metal phase.
  • A combination phase comprising oxides of Cu-Sn-Pb and Cu-sulfide is present as a nucleus enclosed by Pb-sulfide
  • Easily discerned the glass component phase (C-9) and calcium fluoride phase (C-16), which are highly difficult to be identified by EPMA if their existence was not anticipated
  • Identified S in Pb - a highly difficult task for traditional EDS element mapping, and clearly identified the typical distributions of trace quantities of As, Se, In, Sb included in Cu-Sn alloy

The analysis of these results is helpful in determining and optimizing the processing steps to enhance the refining efficiency of Cu compounds.

The analysis method provided a comprehensive analytical data within 30 minute of data acquisition time, thus ensuring an optimized throughput to the refining process by providing rapid results (Figure 8).

Moreover, the technique has simplified the verification work of the material by allowing for the evaluation of materials generated in each refining process.

Sixteen phases automatically extracted by COMPASS software

Figure 8. Sixteen phases automatically extracted by COMPASS software

Spectral Details of Overlapping Peaks

EDS can collect the X-rays of the whole energy range concurrently and acquire X-ray maps at very low magnifications (below 1000x) without the need for microscope stage scanning. Nevertheless, this sample shows different combinations of elements involving the overlapping of the characteristic EDS X-ray peaks. Hence, up to this point, sample analysis was done with EPMA.

The X-ray counts are simply extracted by the traditional peak counts map (Figure 4) technique in a given energy range of the peak, thus possibly having the X-ray counts of other element peaks which overlap.

The deconvolution method for maps (Figure 5) is the same as that practiced for individual spectra except that it is used to a kernel of pixel spectra for improved statistics. The corrective steps are as follows:

  • Removal of the background (brehmstrahlung) of the EDS spectrum
  • Separation of the contributions of each element from overlapped peaks to get net count maps
  • Application of the appropriate matrix corrections to display atomic or weight percent data

The spectral image results are in good agreement with the EPMA results obtained by using wavelength-dispersive spectroscopy with higher spectral resolution. Nonetheless, the morphology of the chemical compounds cannot be clearly understood from elemental maps. The second principal component was excluded as it is mounting epoxy.

Conclusion

The results clearly demonstrate the advantage of using the multivariate statistical analysis software (COMPASS), which reveals the morphology, chemistry, and distribution of each individual compound produced in the initial process of copper lead smelting, easily and rapidly. This ability helps improving the refining process and achieving cost savings to the smelting process.

This information has been sourced, reviewed and adapted from materials provided by Thermo Fisher Scientific – Materials & Structural Analysis.

For more information on this source, please visit Thermo Fisher Scientific – Materials & Structural Analysis.

Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Submit