Quantitative analyses of glass via laser-induced breakdown spectroscopy in argon C. Gerhard a , J. Hermann b, ⁎, L. Mercadier b , L. Loewenthal a , E. Axente c , C.R. Luculescu c , T. Sarnet b , M. Sentis b , W. Viöl a a Laboratory of Laser and Plasma Technologies, University of Applied Sciences and Arts, Von-Ossietzky-Straße 99, 37085 Göttingen, Germany b LP3, CNRS – Aix–Marseille University, 163 Av. de Luminy, 13288 Marseille, France c Laser–Surface–Plasma Interactions Laboratory, Lasers Department, National Institute for Lasers, Plasma and Radiation Physics, Măgurele, Romania abstract article info Article history: Received 3 April 2014 Accepted 21 July 2014 Available online 27 July 2014 Keywords: Laser-induced breakdown spectroscopy Calibration-free LIBS Local thermodynamic equilibrium Glass analysis Optical glass We demonstrate that elemental analysis of glass with a measurement precision of about 10% can be performed via calibration-free laser-induced breakdown spectroscopy. Therefore, plasma emission spectra recorded during ultraviolet laser ablation of different glasses are compared to the spectral radiance computed for a plasma in local thermodynamic equilibrium. Using an iterative calculation algorithm, we deduce the relative elemental fractions and the plasma properties from the best agreement between measured and computed spectra. The measurement method is validated in two ways. First, the LIBS measurements are performed on fused silica composed of more than 99.9% of SiO 2 . Second, the oxygen fractions measured for heavy flint and barite crown glasses are compared to the values expected from the glass composing oxides. The measured compositions are furthermore compared with those obtained by X-ray photoelectron spectroscopy and energy-dispersive X-ray spectroscopy. It is shown that accurate LIBS analyses require spectra recording with short enough delays between laser pulse and detector gate, when the electron density is larger than 10 17 cm -3 . The results show that laser-induced breakdown spectroscopy based on accurate plasma modeling is suitable for elemental analysis of complex materials such as glasses, with an analytical performance comparable or even better than that obtained with standard techniques. © 2014 Elsevier B.V. All rights reserved. 1. Introduction Glasses are complex materials with a wide variety of chemical compositions [1]. Typically composed of a large number of elements, strong variations of minor and major element concentrations characterize glasses, making their analytical investigations particularly difficult [2]. Standard techniques of glass analysis are energy-dispersive X-ray (EDX) spectroscopy, X-ray fluorescence (XRF) spectroscopy, inductively coupled plasma mass spectrometry (ICP-MS) and inductively coupled plasma atomic emission spectroscopy (ICP-AES) [3–5]. Compared to analysis of alloys, the analytical performance on glasses is typically much lower [6]. Due to the large dispersion of glass compositions, standard samples are generally unavailable and the measurement accuracy is consequently often lower than 10 or 20% [7]. In the last decade, glass analysis via laser-induced breakdown spectroscopy (LIBS) has attracted growing interest. LIBS analyses of various glasses and glassy materials have been investigated in different fields of applications, such as quality control in industrial production [8], material recycling [9,10], art conservation [4,11,12], archaeometry [3] forensic research [13–15] and nuclear waste management [2,16,17]. LIBS was also applied to characterize minor element concentrations in optical glasses [18]. Most of these applications do not require high measurement accuracy, and LIBS takes advantage over the standard techniques because of its unique features: measurements can be performed in real-time, in situ, they are almost non-destructible and do not require particular sample preparation [19]. The main objective of analyzing historical glasses, or glass splinters from a crime scene, is the identification of the glass origin. Thus, the measured chemical composition is compared to a database to classify the glass type. Here, LIBS analyses were mostly performed using calibration curves generated via concentration measurements with the standard techniques of glass analyses [4,5,13]. In some particular applications, standard samples were available and used to calibrate the LIBS measurements [3,20]. The use of standards is efficient when the glass matrix is known and minor or trace elements have to be measured [16,17]. Some authors prepared the standard samples themselves by synthesizing the glass composing oxides with well defined composition ratios [10]. Spectrochimica Acta Part B 101 (2014) 32–45 ⁎ Corresponding author. E-mail address: Hermann@lp3.univ-mrs.fr (J. Hermann). http://dx.doi.org/10.1016/j.sab.2014.07.014 0584-8547/© 2014 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Spectrochimica Acta Part B journal homepage: www.elsevier.com/locate/sab