UNCORRECTED PROOF Talanta xxx (xxxx) xxx-xxx Contents lists available at ScienceDirect Talanta journal homepage: http://ees.elsevier.com Experimental evidence of matrix efects in total-refection X-ray fuorescence analysis: Coke case Alex von Bohlen a , Ramón Fernández-Ruiz b,∗ a Leibniz-Institut für Analytische Wissenschaften ISAS, Bunsen-Kirchhof-Str. 11, 44139, Dortmund, Germany b Servicio Interdepartamental de Investigación (SIdI), Laboratorio de TXRF, Universidad Autónoma de Madrid, 28049, Madrid, Spain ARTICLE INFO Keywords Total refection X-Ray fuorescence TXRF Matrix effects Absorption Coke Beverages Direct analysis ABSTRACT Total-refection X-ray Fluorescence (TXRF) analysis is an established method used in instrumental analytical chemistry. Benefts and drawbacks of the method are well known. They have been described in the past and presented in many papers in the last decades. One of the most impressive features, not common to other X-ray fuorescence methods, is the absence of matrix effects and therefore the use of a simple and reliable quantifca- tion. Nevertheless, in some cases, TXRF suffers from matrix effects if the sample amount exceeds the condition of a transparent flm-like sample or simply if the system is overloaded. In this work, we present the results of TXRF analysis of the different components of beverages summarized under the generic name ‘Coke’. Some of them contain up to 10% of sugars and others only mg-amounts of sweeteners. Different sample pre-treatments were applied in order to get reliable results of trace element contents in Coke. A significant deviation of the concen- tration of elements detected by their characteristic X-ray radiation between 1 and 5 keV (e.g. for P, S, K and Ca) was observed depending on the sample amounts and sample preparation. This behaviour could only be explained by the presence of matrix effects. Transmission curves depending on the residual flm-like sample after drying the beverages were calculated and compared with measurements. The obtained deviations for the different sample preparations and different amounts presented to the spectrometer are in good agreement with the theory. TXRF results obtained after microwave digestion for traditional and light Coke were validated by ICP-MS analyses. 1. Introduction Since the beginning of the regular use of total-refection X-ray fu- orescence (TXRF) as an effective tool of instrumental analytical chem- istry, many applications have been reported. Among others, multi-ele- ment microanalysis and surface analysis have been worked out in the- ory and praxis and were published in more than 4000 original articles [1] in the last decades. On the other hand, the phenomenon of exter- nal total refection of X-rays is well known [2,3]. The advantage of the method is attributed to the geometric arrangement of the exciting X-ray beam with respect to the sample and the intense illumination of it. The arrangement of total refection of the primary radiation leads to an en- hanced signal to noise ratio. The spectral background is reduced dras- tically by a factor of about 1:500. This fact is due to a combination of some effects derived from the interaction of electromagnetic radia- tion with matter. From the practical point of view, one of two prereq- uisites for the successful use of TXRF analysis is the use of a smooth and even surface of a stable solid used as sample carrier and refector as well. The surface deviations, or roughness, of the sample carrier, have to be in this specifc case smaller than λ/15 (with λ = 550 nm). The sec- ond is the micro-amount of sample required for the analysis [4]. Moreover, several other benefts are connected to this geometrical arrangement. One among others is an effective excitation to X-ray fuo- rescence, which in average can be described in a simple way as a dou- ble excitation. The frst excitation is attributed to the incoming X-ray beam and the second to the totally refected beam. However, the cor- rect description of the phenomenon includes a complex interference feld produced by the incoming and the totally refected wave exist- ing just above the sample support as Fig. 1(a) shows (X-ray stand- ing waves, XSW) [5,6]. Additionally, if a very thin sample is used it seems to be transparent for the radiation and a simple way of quantif- cation applying the method of the internal standard [5] becomes pos- sible. This is a real singular characteristic in quantitative X-ray fuores- cence analysis. The method fails if the dimensions of the sample are be- low 0.8 nm. In this case, the excitation to fuorescence by XSW feld will be incomplete and a quantitative result will deliver systematically too ∗ Corresponding author. Universidad Autónoma de Madrid, Facultad de Ciencias, Servicio Interdepartamental de Investigación, Laboratorio de TXRF, Avd. Francisco Tomas y Valiente, 7, Modulo C-13, Cantoblanco, 28049, Madrid, Spain. E-mail address: ramon.fernandez@uam.es (R. Fernández-Ruiz) https://doi.org/10.1016/j.talanta.2019.120562 Received 21 August 2019; Received in revised form 3 November 2019; Accepted 12 November 2019 Available online xxx 0039-9140/© 2019.