Optimizing a Copper Smelting Operation Using Multivariate Statistical Phase Analysis of EDS Data

Copper is refined from complex compounds containing elements such as zinc, lead and other elements. During copper smelting, it is essential to understand the complicated morphology of several compounds in the raw material. This comprehensive understanding of the complex morphology is essential for improvement of the refining efficiency of each element.

Presently, X-ray diffraction (XRD), inductively-coupled plasma (ICP) and electron probe microanalysis (EPMA) are used to evaluate the raw materials. However, these techniques have a number of limitations such as the need for experienced analysts.

In this paper, a copper compound raw material has been analyzed by phase analysis using the multivariate statistical analysis of EDS spectral imaging data (Thermo Scientific COMPASS software). Even though the acquisition time was very brief when compared to ICP or EPMA, the complex distribution of the phases could still be accurately determined.

Experimental Details

After coarse lead is extracted, copper that is obtained in the initial copper lead smelting process is evaluated.

Analysis of element distribution and phase morphology was done using a field emission scanning electron microscope (FESEM) and EDS.

A 230 × 180 µm rectangle, the area to be analyzed was observed by reflected light optical microscopy for metallic components. EDS data acquisition was performed using Thermo Scientific UltraDry detector (SDD) and real time multivariate statistical phase analysis was done during the data acquisition. Polishing and coating of the sample was done for the EDS analysis in the FESEM.

Acquisition Conditions

The table detailing acquisition conditions is provided below:

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

Results and Study

The cumulative EDS spectrum and the electron image is shown in Figures 1 and 2 for the EDS spectral imaging acquisition. The cumulative spectrum shows that there may be a range of potential peak overlaps of several elemental peaks.

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

Figure 3 shows illustrations of some of these overlap conditions. Figure 4 proves the elemental distributions using conventional peak count maps. It is seen that peak overlaps determined by Figure 3 are not corrected in these maps. False element intensities may be provided in these overlaps that will impact the interpretation of the elements’ spatial distribution. The results of the peak deconvoluted quantitative elemental maps are seen in Figure 5. The real distributions of Sb, Ca, Al and S are displayed in figure 5 and it shows that these elemental distributions are not properly seen in the previous peak count maps due to overlapping peaks of other elements.

Peaks of characteristic X-rays of analyzed elements

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

Conventional peak count maps

Figure 4. Conventional peak count maps

Quantitative spectral images

Figure 5. Quantitative spectral images

Figure 6 shows the principal component map extracted by COMPASS multivariate image analysis software, which displays and processes during data acquisition.

The first through eighth component maps of the 16 components extracted by COMPASS multivariate analysis software clearly show the morphologies of the Cu compounds which are difficult to understand thorough element maps alone.

Figure 6. The first through eighth component maps of the 16 components extracted by COMPASS multivariate analysis software clearly show the morphologies of the Cu compounds which are difficult to understand thorough element maps alone.

Figure 7 shows the spectra from first to the eighth principal components and standardless quantitative results of spectral data where intensities of each component are strong.

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

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.

These results prove that the copper compounds evaluated have the morphology and elemental distribution as detailed below:

  • The key phase is a Cu-Sn alloy phase, which exists as an associated phase with the lead metal phase.
  • A combination phase comprising oxides of Cu-Sn-Pb and Cu-sulfide exist as a nucleus, encased by Pb-sulfide.
  • The glass component phase (C-9) and calcium fluoride phase (C-16) were easily differentiated. It was very tough to detect these components by EPMA if their presence was not expected.
  • S in Pb was detected which cannot be easily evaluated using conventional EDS element mapping. The typical distributions of small amounts of As, Se, In, Sb included in Cu-Sn alloy were clearly identified.

By evaluating these results, it is possible to identify and improve the processing steps to enhance efficiency of refining copper compounds. Even though the data acquisition time was 30 minutes, analytical data is complete. These results make sure that an enhanced throughput is ensured to the refining process.

Sixteen phases automatically extracted by COMPASS software

Figure 8. Sixteen phases automatically extracted by COMPASS software

Evaluation of the materials obtained in each refining process, using this analysis technique causes the verification work of the material to be very simple and easy.

Spectral Details of Overlapping Peaks

The main advantage of EDS over EPMA is that the X-rays of the whole energy range can be acquired simultaneously and X-ray maps can be collected at very low magnifications below 1000 x without the need of microscope stage scanning. However, a number of element combinations are seen in the sample where typical X-ray peaks overlap. Hence EPMA has been used for analysis of the samples.

The traditional peak counts map as shown in Figure 4 extracts the X-ray counts in a given energy range of the peaks. Hence it may include X-ray counts of other overlapping element peaks. For instance, the Al-K map has data from the Se-L peak, while the Ca-K, In-L, and Sb-L maps have data from the Sn-L peak. The S-K map and the Pb-M map are very similar. The deconvolution method for maps as shown in Figure 5 is exactly the same as that routine for individual spectra except that it is applied to a kernel of pixel spectra for improved statistical analysis.

The corrective steps are as follows:

  1. Remove the background (brehmstrahlung) of the EDS spectrum.
  2. Separate the element contributions from overlapped peaks and provide net count maps.
  3. Apply the appropriate matrix corrections to display weight or atomic percent data.

The spectral image results are very similar to those collected from the EPMA which uses wavelength-dispersive spectroscopy with higher spectral resolution.

Conclusion

The use of COMPASS, which is a multivariate statistical analysis software, it is possible to quickly and easily determine the distribution, morphology and chemistry of every distinct compound generated in the initial copper lead smelting process. This results in adjustment in the refining process and a 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