Development of a method for automated quantitative analysis of ores using LIBS

Citation
S. Rosenwasser et al., Development of a method for automated quantitative analysis of ores using LIBS, SPECT ACT B, 56(6), 2001, pp. 707-714
Citations number
5
Categorie Soggetti
Spectroscopy /Instrumentation/Analytical Sciences
Journal title
SPECTROCHIMICA ACTA PART B-ATOMIC SPECTROSCOPY
ISSN journal
05848547 → ACNP
Volume
56
Issue
6
Year of publication
2001
Pages
707 - 714
Database
ISI
SICI code
0584-8547(20010629)56:6<707:DOAMFA>2.0.ZU;2-G
Abstract
This paper reports the development of a method for real-time automated quan titative analysis of mineral ores using a commercial laser-induced breakdow n spectroscopy instrument, TRACER (TM) 2100, fitted with a recently develop ed computer controlled auto-sampler. The auto-sampler permits the execution of methods for performing calibrations and analysis of multiple elements o n multiple samples. Furthermore, the analysis is averaged over multiple loc ations on each sample, thus compensating for heterogeneous morphology. The results for phosphate ore are reported here, but similar methods are being developed for a range of ores and minerals. Methods were developed to autom atically perform metallic element calibrations for supplied phosphate ore s amples containing known concentrations of the following minerals: P2O5, CaO , MgO, SiO2 and Al2O3. A spectral line for each desired element was selecte d with respect to the best combination of peak intensity and minimum interf erences from other lines. This is a key step, because of the observed matri x dependence of the technique. The optimum combination of the time interval between the laser firing (plasma formation), signal detection, and the dur ation of the optical detection was then determined for each element, to opt imize spectral line intensity and resolution. The instrument was capable of analyzing the required elements in the phosphate ore samples supplied with 2-4% relative standard deviations for most elements. Calibrations were ach ieved for P, Ca, Mg, Al and Si with linear regression coefficients of 0.985 , 0.980, 0.993, 0.987 and 0.985, respectively. Preparation and analysis tim e for each sample was less than 5 min. (C) 2001 Elsevier Science B.V. All r ights reserved.