Repositorio Institucional de Documentos

2014 94 María del Rosario Flórez García High spectral and spatial resolution for the direct elemental and isotopic analysis of solid samples and complex matrices Departamento Director/es Química Analítica Resano Ezcaray, Martín Vanhaecke, Frank Director/es Tesis Doctoral Autor Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Departamento Director/es María del Rosario Flórez García HIGH SPECTRAL AND SPATIAL RESOLUTION FOR THE DIRECT ELEMENTAL AND ISOTOPIC ANALYSIS OF SOLID SAMPLES AND COMPLEX MATRICES Director/es Química Analítica Resano Ezcaray, Martín Vanhaecke, Frank Tesis Doctoral Autor Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Departamento Director/es Director/es Tesis Doctoral Autor Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA High spectral and spatial resolution for the direct elemental and isotopic analysis of solid samples and complex matrices UNIVERSITY OF ZARAGOZA DEPARTMENT OF ANALYTICAL CHEMISTRY & GHENT UNIVERSITY DEPARTMENT OF ANALYTICAL CHEMISTRY DISSERTATION María del Rosario Flórez García Zaragoza, 2014 The current doctoral thesis is presented as a compendium of the following articles: M. Aramendía, M.R. Flórez, M. Piette, F. Vanhaecke, M. Resano. Al determination in whole blood samples as AlF via high-resolution continuum source graphite furnace molecular absorption spectrometry: potential application to forensic diagnosis of drowning. Journal of Analitycal Atomic Spectrometry 26 (2011) 1964-1973 M. Resano, M.R. Flórez. Direct determination of sulfur in solid samples by means of high-resolution continuum source graphite furnace molecular absorption spectrometry using palladium nanoparticles as chemical modifier. Journal of Analitycal Atomic Spectrometry 27 (2012) 401-412 M.R. Flórez, M. Resano. Direct determination of bromine in plastic materials by means of solid sampling high-resolution continuum source graphite furnace molecular absorption spectrometry. Spectrochimica Acta Part B 88 (2013) 32-39 M. Resano, M.R. Flórez, E. García-Ruiz. Progress in the determination of metalloids and non-metals by means of highresolution continuum source atomic or molecular absorption spectrometry. A critical review. Analytical and Bioanalytical Chemistry 406 (2014) 2239-2259 M. Resano, L. Rello, M. Flórez, M.A. Belarra. On the possibilities of high-resolution continuum source graphite furnace atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple atomic lines. Spectrochimica Acta Part B 66 (2011) 321-328 M. Resano, M.R. Flórez, E. García-Ruiz. High-resolution continuum source atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple lines. A critical review of current possibilities. Spectrochimica Acta Part B 88 (2013) 85-97 R. Evens, K.A.C. De Schamphelaere, L. Balcaen, Y. Wang, K. De Roy, M. Resano, M.R. Flórez, P. Van der Meeren, N. Boon, F. Vanhaecke, C.R. Jansen. Liposomes as an alternative delivery system for investigating dietary metal toxicity to Daphnia magna. Aquatic Toxicology 105 (2011) 661-668 R. Evens, K.A.C. De Schamphelaere, L. Balcaen, Y. Wang, K. De Roy, M. Resano, M.R. Flórez, N. Boon, F. Vanhaecke, C.R. Jansen. The use of liposomes to differentiate between the effects of nickel accumulation and altered food quality in Daphnia magna exposed to dietary nickel. Aquatic Toxicology 109 (2012) 80-89 M.R. Flórez, M. Aramendía, M. Resano, A.C. Lapeña, L. Balcaen, F. Vanhaecke. Isotope ration mapping by means of laser ablationsingle collector-ICP-mass spectrometry: Zn tracer studies in thin sections of Daphnia magna. Journal of Analitycal Atomic Spectrometry 28 (2013) 1005-1015 AGRADECIMIENTOS / ACKNOWLEDGEMENTS Reconozco que, aunque estas sean las primeras hojas, han sido las últimas en las que me he puesto a trabajar escribiendo esta memoria. Si ya es difícil resumir casi cinco años de trabajo experimental en unas pocas hojas, mucho más es encontrar una manera de, con pocas palabras, agradecer tanto y a tantos. Voy a comenzar, como es de justicia, agradeciendo a mis directores de tesis, Martín Resano y Frank Vanhaecke, que han sido el motor de todo esto. Pienso, sinceramente, que no podía haber caído en mejores manos. Martín Resano me acogió en su grupo de investigación sin saber muy bien quién iba a aparecer. Así y todo, desde un principio, me recibió con la confianza y la cercanía que han hecho que todo este trabajo haya salido adelante con mucho esfuerzo pero muy poca angustia. Bajo su dirección y motivación, he conseguido interesarme aún más por la química analítica y he aprendido a desarrollar una mente más práctica y más metodológica. Frank Vanhaecke put his trust in me to be involved in this cotutelle project and has, since then, been welcoming me to his laboratories every summer for my research stays at Ghent University. It has been a great pleasure for me to be so friendly accepted in a representative group in ICPMS and have had the opportunity of getting in touch with the technique with already the best resources. Me gustaría también agradecer especialmente a Maite Aramendía, cuyos conocimientos y experiencia tienen también su huella en mi formación, y cuya amistad y apoyo me han sido de gran ayuda a lo largo de todo el desarrollo de este trabajo. Pero, por supuesto, a lo largo de estos años ha habido muchas personas que de una manera u otra han sido para mí imprescindibles y me gustaría mencionar y agradecer: Al resto de mis compañeros de M.A.R.T.E.: Esperanza García, Miguel Ángel Belarra, Jorge Briceño, Engracia Mozas, Luis Rello, Ana Cris Lapeña y la recientemente abducida, Águeda Cañabate, por su amistad y por formar parte de esta peculiar familia en Zaragoza. Y a todo el departamento de Química Analítica de la Universidad de Zaragoza. I also would like to thank my colleagues at A&MS group and the S12, those who were before and those who are still: Lieve Balcaen (I’ll never forget that I started my research life by your hand), Lara Lobo, Marta Costas, Ana Rua y Eduardo Bolea (“Frank’s Spanish Armada”), Eleonora Balliana and the rest of the people I had the opportunity to meet there during the last three summers: Andrei, Deepti, Winfried, Kris, Lana, Asha, Steven, Veerle, Stepan, Björn, Dmitry, Karen, Karel, Roger, Chantal, Harry, Tine, … I am sure I’m forgetting somebody and I will be very sorry whenever I find out. No puedo olvidarme de mencionar al personal de Inycom: Luis de Miguel, Ramón Esteban, Pilar Sariñena, Manuel León y, por supuesto, Fernando Latorre y Sonia Lizondo, que me ayudaron mucho y con paciencia en mis comienzos a entender más a fondo los entresijos del ContrAA 700. Laboratorios aparte, me gustaría agradecer al personal del C.M.U. Pedro Cerbuna, muy especialmente a Irene, Javier, Mónica y Manu, por toda su ayuda y cariño durante mi estancia en el colegio. Pero sin el constante apoyo y presencia de mis amigos y mi familia, esto se hubiera hecho muy, muy difícil. Muchísimas gracias a mis padres y a mi hermano, que son mi equilibrio y mi soporte más sólido, de cerca y de lejos. Y muchísimas gracias a Luismi, por aguantarme en los momentos fáciles y, sobretodo, llamar a la cordura y seguir ahí durante y después de los difíciles. También se lo debo a mis abuelos, a mis tíos y a mis primos. Al resto de las Vs que, aunque cada vez más desperdigadas, seguimos siendo una red segura. A la nueva generación: Elvira, Nahia, Ane y Arán, por la energía con la que llegan y nos contagian a los que ya la vamos perdiendo un poco. A Irene y Elena, por tantos momentos y tantos lacasitos®. Sólo acordarme de todos vosotros me hace sonreír. Muchas gracias a todos. / Thank you all very much. Contents II.3.3.2. Isotope ration mapping by means of laser ablation-single collector-ICP-mass spectrometry: Zn tracer studies in thin sections of Daphnia magna ...................................................................................................... 98 II.4. CONCLUSIONS ....................................................................................... 109 II.5. REFERENCES ........................................................................................... 111 III. PUBLICATIONS .......................................................................................... 123 Al determination in whole blood samples as AlF via high-resolution continuum source graphite furnace molecular absorption spectrometry: potential application to forensic diagnosis of drowning ........................................... 125 Direct determination of sulfur in solid samples by means of high-resolution continuum source graphite furnace molecular absorption spectrometry using palladium nanoparticles as chemical modifier ................................................ 135 Direct determination of bromine in plastic materials by means of solid sampling high-resolution continuum source graphite furnace molecular absorption spectrometry ........................................................................................... 147 Progress in the determination of metalloids and non-metals by means of high-resolution continuum source atomic or molecular absorption spectrometry. A critical review ................................................................................. 155 On the possibilities of high-resolution continuum source graphite furnace atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple atomic lines .......................................................................... 177 High-resolution continuum source atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple lines. A critical review of current possibilities ................................................................................................... 185 Liposomes as an alternative delivery system for investigating dietary metal toxicity to Daphnia magna ........................................................................................... 199 The use of liposomes to differentiate between the effects of nickel accumulation and altered food quality in Daphnia magna exposed to dietary nickel ......................................................................................................................... 207 Contents Isotope ration mapping by means of laser ablation-single collector-ICPmass spectrometry: Zn tracer studies in thin sections of Daphnia magna .................... 217 GENERAL OUTLINE AND CONCLUSIONS ........................................ 229 General outline and conclusions ...................................................................... 231 Resumen y conclusiones generales .................................................................. 235 Algemeen overzicht en conclusies ................................................................... 241 APPENDIX ............................................................................................................. 247 Thematic unity 1 THEMATIC UNITY The work presented in this dissertation has been developed within the area of expertise and analytical resources of both M.A.R.T.E. (Métodos de Análisis Rápidos con Técnicas Espectroscópicas), research group of the Department of Analytical Chemistry of the University of Zaragoza, and A&MS (Atomic and Mass Spectroscopy), research group of the department of Analytical Chemistry of Ghent University. The work carried out focused on the potential of atomic techniques for the direct analysis of solid samples and complex matrices. Within this scheme, it has been structured in two main blocks of study supported by the compendium of nine publications: i) the exploration of the analytical benefits introduced by high-resolution continuum source atomic absorption spectrometry (HR CS AAS) in terms of spectral resolution; and ii) the improvement of spatial resolution in the direct analysis of solid microsamples. The first block, regarding AAS applications with high spectral resolution, covered the possibilities opened with the introduction of commercially available HR CS AAS instruments for multi-line monitoring and multielement determination as well as the suitability for monitoring broad molecular spectra, which has demonstrated to be particularly beneficial for the determination of non-metals. Regarding the evaluation of direct solid sampling applied to the determination of micro-samples, the benefits of HR CS AAS have been investigated and complemented with an application developed by means of laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) to collect isotopic information in order to produce isotope ratio images with high spatial resolution, which is obviously possible only when using MS. Thematic unity 2 Schematically, the enclosed articles may be assembled as follows: A) Evaluation of the spectral resolution featured by HR CS AAS a.i) Monitoring of non-metals by means of HR CS MAS a.i.1) M. Aramendía, M.R. Flórez, M. Piette, F. Vanhaecke, M. Resano. Al determination in whole blood samples as AlF via highresolution continuum source graphite furnace molecular absorption spectrometry: potential application to forensic diagnosis of drowning. Journal of Analitycal Atomic Spectrometry 26 (2011) 1964-1973 a.i.2) M. Resano, M.R. Flórez. Direct determination of sulfur in solid samples by means of high-resolution continuum source graphite furnace molecular absorption spectrometry using palladium nanoparticles as chemical modifier. Journal of Analitycal Atomic Spectrometry 27 (2012) 401-412 a.i.3) M.R. Flórez, M. Resano. Direct determination of bromine in plastic materials by means of solid sampling high-resolution continuum source graphite furnace molecular absorption spectrometry. Spectrochimica Acta Part B 88 (2013) 32-39 a.i.4) M. Resano, M.R. Flórez, E. García-Ruiz. Progress in the determination of metalloids and non-metals by means of highresolution continuum source atomic or molecular absorption spectrometry. A critical review. Analytical and Bioanalytical Chemistry 406 (2014) 2239-2259 a.ii) Multi-line monitoring a.ii.1) M. Resano, L. Rello, M. Flórez, M.A. Belarra. On the possibilities of high-resolution continuum source graphite furnace atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple atomic lines. Spectrochimica Acta Part B 66 (2011) 321-328 a.ii.2) M. Resano, M.R. Flórez, E. García-Ruiz. High-resolution continuum source atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple lines. A critical review of current possibilities. Spectrochimica Acta Part B 88 (2013) 85-97 Thematic unity 3 B) Spatial resolution in the direct determination of solid micro-samples b.i) Direct determination of Daphnia magna specimens by means of HR CS AAS b.i.1) R. Evens, K.A.C. De Schamphelaere, L. Balcaen, Y. Wang, K. De Roy, M. Resano, M.R. Flórez, P. Van der Meeren, N. Boon, F. Vanhaecke, C.R. Jansen. Liposomes as an alternative delivery system for investigating dietary metal toxicity to Daphnia magna. Aquatic Toxicology 105 (2011) 661-668 b.i.2) R. Evens, K.A.C. De Schamphelaere, L. Balcaen, Y. Wang, K. De Roy, M. Resano, M.R. Flórez, N. Boon, F. Vanhaecke, C.R. Jansen. The use of liposomes to differentiate between the effects of nickel accumulation and altered food quality in Daphnia magna exposed to dietary nickel. Aquatic Toxicology 109 (2012) 80-89 b.ii) Isotope ratio mapping of thin sections of Daphnia magna specimens by means of LA-single collector-ICPMS (LA-SC-ICPMS) M.R. Flórez, M. Aramendía, M. Resano, A.C. Lapeña, L. Balcaen, F. Vanhaecke. Isotope ration mapping by means of laser ablationsingle collector-ICP-mass spectrometry: Zn tracer studies in thin sections of Daphnia magna. Journal of Analitycal Atomic Spectrometry 28 (2013) 1005-1015 I. Introduction Introduction 7 Several studies have been carried out to date regarding the capabilities of HR CS AAS since the development of the first instrumental prototype proposed by Becker-Ross and co-workers [1,2] at the beginning of this last century and, more extensively, since its appearance on the market. Comparative advantages over classic line source AAS (LS AAS) instruments become evident when paying attention to its instrumental setup: I.1. INSTRUMENTAL CHARACTERISTICS OF HIGHRESOLUTION CONTINUUM SOURCE ATOMIC ABSORPTION SPECTROMETRY Figure I.1. Setup of a high-resolution continuum source atomic absorption spectrometer. Flame Graphite furnace or High-resolution monochromator Prism CCD Detector Échelle grating Xenon short-arc lamp Introduction 8 I.1.1. Radiation source The radiation source was replaced by a sole high-energy Xe short arc lamp, working under “hot-spot” mode, that reaches temperatures of about 10,000 K. The intensity gain for most of the elements (at least 1 or 2 orders of magnitude higher) allow better detection limits facing hollow-cathode lamps (HCL). Also advantageous is the fact that with just one lamp a spectral range from the far ultraviolet region (185 nm) to the near infrared (900 nm) is covered, making it possible to explore elements for which no HCL is available. I.1.2. Atomization unit The instrument may be equipped with both flame and/or graphite furnace atomization systems. In this work only the graphite furnace device was employed for all AAS measurements due to its potential to deal with direct solid sampling (SS) analyses. Therefore, all the discussions of this work will be referred to graphite furnace AAS (GFAAS), unless noted otherwise. I.1.3. Monochromator The monochromation system is composed of a 300 mm quartz prism and a 400 mm focal length echelle grating monochromator. This double monochromator is necessary to ensure a resolution power in the range of 140,000 (spectral band width of less than 2 pm for 200 nm), for it to be capable of achieving comparable sensitivities to those of the LS instruments for each element and, therefore, optimum signal-to-noise ratios (S/N). An internal Ne lamp is responsible of an active wavelength-stabilization attaining precision better than ±0.2 pm. Introduction 9 I.1.4. Detection system The greater methodological potential of the instrument however lies in the detection system: that is a charge-coupled device (CCD) array with typically 588 pixels, of which just 200 are used for analytical purposes while the rest remains available for the correction of all the wavelength-independent spectral events recorded in each pixel at the same time (lamp’s flickering, variations in the transmission of the atomizer…). As every pixel is illuminated simultaneously and read out and evaluated independently, the system essentially works with 200 independent detectors. The recorded spectral range varies from 0.2 to 1.0 nm depending on the wavelength. Within that spectral window, just a few pixels (normally 3) are used to measure atomic absorption, so the rest of the 200 pixels can be devoted to correction purposes. Besides, information of not only the line of interest but of its environment is always available, and an important portion of the work presented in this dissertation is based on making the most out of this aspect. I.2. METHODOLOGICAL POSSIBILITIES OF HR CS AAS Of all advantages that the HR CS AAS instrumentation incorporates in comparison with LS AAS (improved signal stability, superior potential for background correction, improved linearity…) [3] this work is partially focused on those regarding the new methodological possibilities derived from the fact that a 200 pixels spectral window is recorded in each measurement. Most specifically the main topics to be covered are: i) Multi-line monitoring of absorption lines for either the same or different analytes Introduction 10 ii) Monitoring of molecular transitions iii) Direct solid sampling. Analysis of microsamples. I.2.1. Multi-line monitoring I.2.1.1. Simultaneous multi-elemental determination One of the main handicaps of LS AAS was its poor potential for carrying out multi-element analyses. HR CS AAS brings up the opportunity to fully exploit the fact that a range of the spectrum is monitored in each measurement and therefore any information gathered within that spectral window may be used for analytical purposes. There is however an important instrumental limitation and that is precisely the very small portion of the spectrum that can be simultaneously monitored: from 0.2-0.3 nm in the far-UV region (where most of the main resonant atomic lines are found) until 0.5 nm approaching the visible region up to almost 1 nm at 800 nm wavelength [4-7]. In this way, the possibilities for developing truly simultaneous multielement determinations are subject to the compliance with the following premises: - The atomic lines of the analytes of interest should be located close enough to fit within the spectral window allowing proper baseline adjustment but separate enough not to pose a spectral interference with each other if, for instance, broadening of one of the lines of interest happens when high amounts of one of the elements monitored are present. - The sensitivity of the lines has to be adequate to the expected content of the corresponding element in the samples under study. That is, the sensitivity ratio between lines of the target elements must be of the Introduction 17 - A very dense and not completely resolved rotational spectrum. Then a quite broad band will be registered. This situation is typical for the UV systems of Al-based or Ga-based molecules [36]. Figure I.2. AlF time-integrated molecular absorbance spectrum obtained vaporizing 500 pg Al in excess of F by means of HR CS GFMAS. - Rotational spectrum evidencing several entirely resolved bands of different sensitivities. This is typically the case for those transitions registered in the visible region of the spectrum, such as the Ca-based molecules. As it was discussed in previous sections, this situation may allow to extend the linear range [37]. Figure I.3. CaBr time-integrated molecular absorbance spectrum obtained vaporizing 100 ng Br in excess of Ca by means of HR CS GFMAS. - A quite dense but resolved rotational spectrum, presenting a great amount of transitions practically as narrow as atomic lines and all Introduction 18 with similar sensitivities. This is for instance the case of CS, NO or PO, already mentioned when talking about background correction. Combining the signal for some of those lines may improve the sensitivity and therefore lead to achieve better LODs [31,38,39]. Figure I.4. CS time-integrated molecular absorbance spectrum obtained vaporizing 50 ng S by means of HR CS GFMAS. A difficulty when dealing with this sort of complex spectra may arise when it comes to fixing the baseline. The dynamic mode applied automatically by the software will most normally set the valleys between the peaks wrongly, particularly when the lines are very closely situated. It is recommended then to set the pixels for the baseline correction manually. A recent implementation in the software included a new baseline correction mode based on an algorithm named IBC (iterative baseline correction). Applying this algorithm results in narrower spectrum lines and a more defined baseline, though also the sensitivity is somewhat reduced. This correction method is especially useful when the pattern of molecular transitions is particularly complex and choosing the suitable correction pixels is not clear. In this way a potential source of error due to the baseline correction can be avoided. A comparison between these three different modes for fixing the baseline is illustrated in Figure I.5. Introduction 19 It constitutes an important part of this doctoral thesis to investigate the possibilities of HR CS GFMAS for the direct determination of non-metals. In this regard, the work will focus not only in applying the most suitable strategy to quantify the spectrum obtained, but also in achieving proper stabilization and formation of the target molecules, during the pyrolysis and vaporization steps, respectively. Figure I.5. CS time-integrated spectrum obtained after vaporization of 50 ng S by means of HR CS GFMAS using different approaches for baseline adjustment: a) correction pixels fixed manually by the analyst under “static” working mode; b) correction pixels automatically set by the software working under “dynamic” working mode; and c) use of IBC mode for baseline correction I.2.3. Direct solid sampling graphite furnace atomic absorption spectrometry applied to the analysis of micro-samples Being able to perform the direct analysis of the target samples offers some significant advantages. For instance, higher sensitivity and speed in every Introduction 20 measurement can be achieved, as well as reducing to the minimum the use of hazardous reagents. Another advantage, particularly important when dealing with ubiquitous or volatile analytes, is that risks of contamination or losses are also highly reduced, being this issue one of the main concerns when carrying out sample digestions. GFAAS has proven all since the 1980s to possess a great potential for SS applications [40], sharing all the advantageous features already pointed out among all SS techniques. However it also shares some weak points. Most of them have been dealt with or even overcame with the successive improvements to the commercially available instruments. For instance, the difficulty in the handling and insertion of the solid samples into the graphite tubes [41] has been overcome with the development of auto-sampling systems equipped with a pair of tweezers for transporting the sample platforms. Automatic versions even include a microbalance for acquiring the sample weigh prior to carry the platform to the graphite tubing. Another drawback accompanying SS techniques is the need for adequate materials for calibration purposes, materials that should matrix match the samples as much as possible. This is no longer critical for GFAAS as several works already demonstrated that with a careful optimization of the furnace conditions and by measuring integrated peak areas, providing isothermal conditions are achieved during the atomization step, it is possible to successfully employ aqueous standards for calibration in most situations [42,43]. However, the most relevant disadvantage is still the large uncertainty that typically affects the results obtained. This detriment to the precision is due to the natural lack of homogeneity of the samples and the fact that very Introduction 21 small amount of sample is tolerated in each platform load (normally around 1 mg) [44]. To deal with this poor precision the only option to date is to increase the number of measurements for each sample, depending on the level of inhomogeneity. In general, monitoring at least 5 solid doses of each sample is rather typical and a relative standard deviation between 5%-15% is considered normal. It is worth mentioning here that employing the median as the estimate of the value of the analyte is often preferred as a way of ruling out the frequently appearing outliers associated with this technique [45]. When dealing with direct analysis of solid samples it is necessary to take into account that the whole matrix will be present within the atomization device and the analyte and the matrix will be simultaneously vaporized, giving rise to potential spectral interferences, e.g. owing to high background signals that may be generated by radiation scattering due to the presence of particles in the atomic vapour or to molecular absorption signals in the analytical region. To minimize this risk of interferences, some approaches are usually applied for establishing of the optimum conditions: i) in a situation where the volatility of the matrix is higher than that of the analyte, a thorough optimization of the pyrolysis stage is essential to eliminate as much matrix as possible; ii) if, on the contrary, the volatility of the analyte is higher, the pyrolysis stage is not critical, and most of the efforts should be devoted to finding an adequate atomization/vaporization temperature to selectively monitor the analyte; and iii) if analyte and matrix show similar volatilities, use of matrix modifiers and/or maintenance of Ar flow during the atomization/vaporization would be necessary to reach one of the previously discussed situations. However, in some applications, temperature program optimization and matrix modifiers cannot completely overcome all spectral interferences. HR Introduction 22 CS AAS had represented a step forward in SS applications due to its superior potential facing background correction, related to the monitoring of a portion of the spectral environment of the analytical line [46,47]. An enhanced detection power, improved signal stability, the possibility of choosing the analytical pixels for the adjustment of the sensitivity and the capability for monitoring non-metals are also advantages of HR CS AAS that can be applied to SS [7]. So far, the fact that only a small amount of sample can be loaded into the sample platform has been regarded as a handicap. However, some interesting applications may derive from this fact, as it is the potential for carrying out homogeneity studies [12,48,49] or the possibility of developing analytical methods for micro-samples [50]. An important part of this thesis work has been focused on the handling and analysis of Daphnia magna specimens, small planktonic crustaceans of size normally around 0.2-5 mm long, used as test-organisms for ecotoxicological studies, each of them weighing just a few tens of micrograms. Bringing this kind of samples into solution for analytical purposes entails the combination of several specimens in order to attain a detectable analytical signal. This process not only represents a potential increase in the risk of contamination but also causes the lost of information regarding the variation among individuals. For most elements, GFAAS offers the possibility to determine total body burdens within each D. magna individually with a suitable sensitivity. As each invertebrate represents an individual sample, the information gathered in each determination is unique and no replicates are possible. It is clear that, in this context, the multi-elemental capabilities of HR CS AAS Introduction 23 technique could be particularly valuable, and so is the possibility for offline data post-processing. The research stays at the A&MS group in the Analytical Chemistry department of Gent University in the context of this PhD had offered the possibility of further exploring the analysis of micro-samples. LA-sector field-ICPMS (LA-SF-ICPMS) allows to perform direct solid analysis for elemental and isotopic determinations with a high spatial resolution (due to the ablation process), a good spectral resolution and a high sensitivity (due to the detection system). Basically, a short laser pulse hits the surface of the sample under study promoting a small amount of sample into a vapour (dry aerosol) that is immediately led to the entrance of the plasma torch. Once inside the plasma, the particles generated in the ablation process are ionized and guided through the mass spectrometer to be adequately discriminated and further amplified when reaching the detection system in order to obtain highly selective and sensitive analytical signals. The parameters of the laser can be appropriately optimized in order to establish a good agreement between sensitivity and spatial resolution, whether lateral or in depth [51]. LA-ICPMS, as all ICPMS-based techniques, presents the capability of providing isotopic information of the samples. In this compendium of works, a study based on the combination of both features (that is, the generation of isotopic spatial information) is presented focused on the mapping of isotope ratios in thin sections of D. magna specimens, as a way of providing spatially-resolved the data for following tracer experiments with stable isotopes. Introduction 24 I.3. REFERENCES [1] U. Heitmann, M. Schütz, H. Becker-Ross, S. Florek, Measurements on the Zeeman-splitting of analytical lines by means of a continuum source graphite furnace atomic absorption spectrometer with a linear charge coupled device array, Spectrochimica Acta Part B: Atomic Spectroscopy 51 (1996) 1095–1105. [2] H. Becker-Ross, S. Florek, U. Heitmann, Observation, identification and correction of structured molecular background by means of continuum source AAS–determination of selenium and arsenic in human urine, Journal of Analytical Atomic Spectrometry 15 (2000) 137–141. [3] B. Welz, High-resolution continuum source AAS: the better way to perform atomic absorption spectrometry, Analytical and Bioanalytical Chemistry 381 (2005) 69–71. [4] J.M. Harnly, The future of atomic absorption spectrometry: a continuum source with a charge coupled array detector, Journal of Analytical Atomic Spectrometry 14 (1999) 137–146. [5] H. Becker-Ross, S. Florek, U. Heitmann, M.D. Huang, M. Okruss, B. Radziuk, Continuum source atomic absorption spectrometry and detector technology: A historical perspective, Spectrochimica Acta Part B: Atomic Spectroscopy 61 (2006) 1015–1030. [6] B. Welz, S. Morés, E. Carasek, M.G.R. Vale, M. Okruss, H. Becker-Ross, High-Resolution Continuum Source Atomic and Molecular Absorption Spectrometry–A Review, Applied Spectroscopy Reviews 45 (2010) 327–354. Introduction 25 [7] M. Resano, E. García-Ruiz, High-resolution continuum source graphite furnace atomic absorption spectrometry: Is it as good as it sounds? A critical review, Analytical and Bioanalytical Chemistry 399 (2011) 323–330. [8] B. Welz, H. Becker-Ross, S. Florek, U. Heitmann, High-Resolution Continuum Source AAS, The Better Way to Do Atomic Absorption Spectrometry, Wiley-VCH, Weinheim, 2005. [9] I.M. Dittert, J.S.A. Silva, R.G.O. Araujo, A.J. Curtius, B. Welz, H. Becker-Ross, Direct and simultaneous determination of Cr and Fe in crude oil using high-resolution continuum source graphite furnace atomic absorption spectrometry, Spectrochimica Acta Part B: Atomic Spectroscopy 64 (2009) 537–543. [10] I.M. Dittert, J.S.A. Silva, R.G.O. Araujo, A.J. Curtius, B. Welz, H. Becker-Ross, Simultaneous determination of cobalt and vanadium in undiluted crude oil using high-resolution continuum source graphite furnace atomic absorption spectrometry, Journal of Analytical Atomic Spectrometry 25 (2010) 590–595. [11] D.P.C. Quadros, E.S. Chaves, F.G. Lepri, D.L.G. Borges, B. Welz, H. Becker-Ross, A.J. Curtius, Evaluation of Brazilian and Venezuelan Crude Oil Samples by Means of the Simultaneous Determination of Ni and V as Their Total and Non-volatile Fractions Using High-Resolution Continuum Source Graphite Furnace Atomic Absorption Spectrometry, Energy & Fuels 24 (2010) 5907–5911. [12] M. Resano, E. Bolea-Fernández, E. Mozas, M.R. Flórez, P. Grinberg, R.E. Sturgeon, Simultaneous determination of Co, Fe, Ni Introduction 26 and Pb in carbon nanotubes by means of solid sampling highresolution continuum source graphite furnace atomic absorption spectrometry, Journal of Analytical Atomic Spectrometry 28 (2013) 657–665. [13] L.M.G. dos Santos, R.G.O. Araujo, B. Welz, S. do C. Jacob, M.G.R. Vale, H. Becker-Ross, Simultaneous determination of Cd and Fe in grain products using direct solid sampling and highresolution continuum source electrothermal atomic absorption spectrometry, Talanta 78 (2009) 577–583. [14] L.M.G. dos Santos, B. Welz, R.G.O. Araujo, S. do C. Jacob, M.G.R. Vale, A. Martens, I.B.G. Martens, H. Becker-Ross, Simultaneous Determination of Cd and Fe in Beans and Soil of Different Regions of Brazil Using High-Resolution Continuum Source Graphite Furnace Atomic Absorption Spectrometry and Direct Solid Sampling, Journal of Agricultural and Food Chemistry 57 (2009) 10089–10094. [15] F. Vignola, D.L.G. Borges, A.J. Curtius, B. Welz, H. Becker-Ross, Simultaneous determination of Cd and Fe in sewage sludge by highresolution continuum source electrothermal atomic absorption spectrometry with slurry sampling, Microchemical Journal 95 (2010) 333–336. [16] A.T. Duarte, M.B. Dessuy, M.G.R. Vale, B. Welz, J.B. de Andrade, Sequential determination of Cd and Cr in biomass samples and their ashes using high-resolution continuum source graphite furnace atomic absorption spectrometry and direct solid sample analysis, Talanta 115 (2013) 55–60. II. Dissertation Dissertation 35 II.1. GENERAL GOALS Our research groups have a long experience in the development of analytical methods for the direct analysis of solid samples using GF-AAS and ICPMS. These methods try to provide analytical information in a more direct and simple way, relying in simple calibration approaches whenever feasible. This PhD continues further this tradition and aims at developing new methods taking advantage of the existing improvements in terms of instrumentation. In this regard, all the determinations attempted (e.g., Al in blood, S in various materials, Br in polymers) present challenges when following the common digestion approaches, such that exploring direct analysis may bring in significant advantages. In addition to this goal of developing straightforward methods, the systematic exploration of the capabilities of a high-resolution continuum source atomic absorption spectrometer is a major objective of this work. This instrumentation provides spectrally resolved information that can help in detecting and correcting for interferences, carry out multi-element determinations or monitor molecular spectrum quantitatively. This second aspect should not be seen as a separate topic, since these characteristics are the ones that might make feasible to develop the direct methods discussed above. Finally, a third goal arises after examination of the benefits and disadvantages of direct SS analysis. These techniques typically offer spatially resolved information, which might be problematic when aiming at achieving a mean value for inhomogenous samples, but can be a very significant advantage when the goal is to obtain information from microsamples. Thus, this work will also explore the potential of HR CS AAS and Dissertation 36 LA-ICPMS to obtain information that can be of ecotoxicological relevance, focusing on analysis of D. magna specimens. Dissertation 37 II.2. INSTRUMENTATION II.2.1. High-resolution continuum source atomic absorption spectrometry - ContrAA 700 All the works described from now on in sections II.3.1, II.3.2 and II.3.3.1 were carried out using a HR CS AAS instrument (ContrAA 700). This instrument, commercially available from Analytik Jena AG (Jena, Germany), is equipped with both flame and graphite furnace atomizers, a Xenon short-arc lamp (GLE, Berlin, Germany) operating in “hot-spot” mode as the radiation source, a high-resolution double echelle monochromator (DEMON) and a linear CCD array detector with 588 pixels, 200 of which are used for monitoring the analytical signal and performing BG corrections, while the rest are used for internal functions such as correcting for fluctuations in the lamp intensity, as further described in the introduction section. The furnace atomizer of the HR CA AAS instrument is a transversely heated graphite tube atomizer, equipped with pyrolytic graphite tubes and platforms for solid sampling and sample introduction, and an automated solid sampling accessory (SSA 600). This solid sampling device allows for automatic weighing and transport of the samples to the furnace by means of a pair of tweezers and incorporates a microbalance with a readability of 1 µg. After first taring the empty sample platform, an appropriate amount of sample is deposited onto the platform to be weighed. Subsequently, all the required chemical modifiers are added to the sample prior to be transferred into the graphite furnace to be subjected to the corresponding optimized temperature program in every case. All these operations are fully controlled by the instrument software, with exception of the sample deposition and the addition of the modifier, which need to be carried out manually. Dissertation 38 Figure I.6. AnalytikJena ContrAA 700 HR CS AAS instrument used throughout this doctoral thesis II.2.2. Laser ablation single collector inductively coupled plasma mass spectrometry - UP193HE laser coupled to Element XR In section II.3.3.2 another instrumentation for the analysis of microsamples was regarded as an implementation on the study of the enhancement of the spatial resolution. For the work presented, a New Wave Research 193 nm wavelength ArF* excimer-based laser ablation system (UP193HE, New Wave Research, CA, USA) was coupled to a sector field-based ICP-MS instrument (Element XR, Thermo Scientific, Bremen, Germany). The LA unit is equipped with an ablation cell mounted on a tri-translational stage. A teardrop-shaped ablation cell [1], built by Axel Gerdes at the University of Frankfurt based on the work of Bleiner and Günther [2], was used throughout the work. This cell has an internal volume of circa 2.5 cm3, which, compared with the standard cell (30 cm3 volume), reduces the washout time significantly, allowing improved time-resolved elemental and isotopic information [3] and, therefore, better spatial resolution. Dissertation 39 A quartz dual pass cyclonic spray chamber with two inlet ports (Elemental Scientific Inc., Omaha, NE, USA) was included in the setup for simultaneous nebulization of a solution with the ablated aerosol (“wet plasma” mode). The aerosols coming from the spray chamber and the laser ablation cell were mixed by means of a Y-shaped connection. The benefits of this setup are discussed in the corresponding methodological section devoted to this study. A new slit system provided by Thermo Fisher Scientific (Bremen, Germany) for the Element XR, which generates flat top peaks at a mass resolution of m/Δm ≈ 2000 (20% ion transmission with respect to that observed at low resolution mode) [4], was used. These mass spectral characteristics are realized via a wider exit slit than the conventional medium resolution slit (offering a resolution of m/Δm ≈ 4000 at 10% ion transmission with respect to low resolution mode), which generates triangular shaped peaks. Figure I.7. A) Thermo Element XR SF-ICPMS instrument and B) New Wave Research UP193HE laser ablation unit used in the work described in the experimental section II.3.3.2 Dissertation 40 II.2.3. Additional instrumentation For the CS molecular absorption monitoring study, a Scanning Electron Microscope (SEM) JEOL (Tokyo, Japan) JSM-6400, configured with an energy dispersive X-ray analyser INCA 300 X-Sight from Oxford Instruments (Abingdon, UK), was used for the examination of the graphite platforms. This instrument was available through the Research Support Services of the University of Zaragoza. For the Zn isotope ratio imaging of D. magna specimens, a Slee (Mainz, Germany) CUT 4060 rotatory microtome was used for cutting the 20 µm sagittal sections of each sample individual. Access to this instrument was possible at the Department of Basic Medical Sciences of Ghent University. Dissertation 41 II.3. METHODOLOGY AND DISCUSSION OF THE RESULTS II.3.1. Monitoring of non-metals II.3.1.1. Al determination in whole blood samples as AlF via highresolution continuum source graphite furnace molecular absorption spectrometry: potential application to forensic diagnosis of drowning Goals The final goal of this work was to assess the validity of Al as a chemical proxy to complement the forensic diagnosis of death-by-drowning in contaminated fresh waters. Diagnosis of true drowning is still one of the most common challenges in forensic pathology [5]. Physiological signs have a limited validity, especially when the body exhibits putrefaction evidences. Additional chemical and biochemical tests are needed in order to offer a reliable way of discern between true drowning and a different lethal cause. Traditionally, the diatom test, based in the finding of diatom algae in a sufficiently high concentration within the organs of the dead body [6] was used as powerful indicator of drowning. However, nowadays some species of these organisms are being gradually extinct in polluted waters, compromising the applicability of this popular test. Therefore, the finding of (bio)chemical markers for drowning diagnosis is still a field of interest in Forensic Science. Monitoring of Al may be an alternative approach worth exploring, as it complies well with the premises for a suitable marker: i) to be present in large amounts in the drowning water and absent or present in very small amounts in the blood of a healthy person (Al levels in the blood of healthy people are usually low, at the 10 µg L-1 level or lower) [7,8], and ii) to enter Dissertation 42 the blood circulation only by the alveolo-capillary membrane during the agony period of the drowning event. From a methodological point of view, a comparison between the direct determination of Al in blood samples by means of AAS and the monitoring of the AlF absorption band after the addition of a fluorinating agent by means of MAS was explored in order to develop a straightforward analytical method enabling the monitoring of Al in suspects of drowning. Methodology Direct determination of Al by means of GFAAS in such a complex matrix as whole blood is not free of hindrances. Occurrence of non-spectral interferences usually implies sample dilution with different reagents [8-10] and even so, standard addition method would still be required for calibration. Moreover, also spectral interferences take place, the most problematic one being the Fe line appearing at 309.278 nm [11], located just between the Al duplet at 309.271/309.284 nm, which is impossible to be separated in time and that overlaps significantly for samples containing more than 0.5 g Fe L-1 (as it is the case for blood samples [12]). This issue, however, can be avoided by measuring in the 396.152 nm line, which offers slightly less sensitivity. The formation of stable Al diatomic molecules and the recording of its molecular transitions has been used in several works particularly for the quantification of non-metals by both flame and graphite furnace MAS [13]. In this work, HR CS MAS is used to further study the absorption spectra of the AlF molecule, although aiming at Al determination. Several blood samples for cadavers found in water media (named as DS-1 to DS-8) along with water collected from the locations where those bodies Dissertation 43 were found were provided for this study. Also blood from corpses not found in water was available to use as control samples, CTRL-1 and CTRL-2. Both groups of samples were stabilized with lithium heparine and were collected around Ghent (Belgium) and its district, where high levels of Al was expected to be found in water accumulations due to the intense industrial activity of the area. With regards to the method validation, a blood reference sample was analysed. Seronorm trace element human whole blood level II was purchased from Sero (Billingstad, Norway), which provides an indicative value for Al content. This sample comes as powder that needs to be reconstituted in purified water. A certified reference material for water (BCR-610 Ground water, high contents) purchased from the Institute of Reference Materials and Measurements (IRMM; Geel, Belgium) was also analysed. Besides, two blood samples collected from Miguel Servet Hospital in Zaragoza (Spain) and stabilized with potassium-EDTA were also available to compare the effects of different matrices on the method. These two blood samples, referred to as LS-1 and LS-2, along with the Seronorm and one of the drowning suspects were also analysed by means of SF-ICPMS, with the Thermo Element XR, after sample digestion with HNO3 and H2O2 heating on a hot plate, for a further validation of the results. For the determination on the atomic line 396.152 nm, samples were diluted a factor of 1:4 with ultrapure water and the standard addition method (of 50, 100 and 200 pg Al added with a diluted aqueous standard) was applied for calibration and five replicate measurement of each point were performed. Dissertation 50 Figure II.4. Comparison of a wavelength-resolved time integrated AlF molecular absorbance signal recorded in the vicinity of 227.4 nm for the vaporization of a similar amount of Al from an aqueous solution and directly from a blood sample under the conditions listed in table II.2. The reproducibility of the MAS method was evaluated by analysing the Seronorm reference sample in four different days under optimized conditions, then the mean of the results for these four different sessions was calculated obtaining a value of 57.5±3.4 µg L-1 (reference value for Seronorm, 58.9±9.9 µg L-1), which was considered as very satisfactory in terms of precision and accuracy. A comparison of the performances of the MAS and AAS techniques was also carried out. While in the case of AAS the need for sample dilution of at least a factor of 1:4 cannot be avoided, the limits of detection for both MAS and AAS are rather similar (1.5 µg L-1 for AAS and 1.8 µg L-1 for MAS), due mainly to the fact that the measurement of the molecular band introduces a higher level of noise to the signal. Regarding the precision, however, better figures are obtained for the MAS method than for the AAS one. Besides, higher temperatures are needed for the AAS technique (optimum atomization temperature of 2500°C), which negatively affects the Dissertation 51 perishable components of the instrument, particularly the sample platform and the graphite tubes, and even under these conditions matrix removal is not complete and a mechanical cleaning of the platform between samples is needed, making the AAS method more prone to contamination problems. In spite of that, results were obtained using both methods for the samples for which SF-ICPMS values were available. All the obtained values are presented in Table II.3, where a good agreement between the results for all the techniques is shown. It is important to stress that MAS was the only technique for which it was not necessary to use standard additions for calibration. Table II.3. Comparison of the Al concentration values obtained for the Seronorm reference material and three blood samples by means of SF-ICPMS, AAS and MAS. Uncertainties are expressed as 95% confidence intervals (calculated as ±ts/√n, where t, s and n have their usual meanings). Reference /μg L-1 SF-ICPMS /μg L-1 (n=3) AAS /μg L-1 (n=5) MAS /μg L-1 (n=5) Seronorm 58.9 ± 9.9 60.1 ± 6.1 56.9 ± 8.1 57.5 ± 3.4a LS-1 - - 46.8 ± 9.2 46.6 ± 5.7 45.0 ± 4.6 LS-2 - - 96.5 ± 10.4 100.3 ± 9.2 96.4 ± 7.6 DS-5 - - 23.7 ± 6.0 24.6 ± 8.8 24.3 ± 5.3 The rest of the blood samples of drowning suspects were analysed just applying the MAS methodology. Results are gathered in Table II.4. The Spearman rank correlation coefficient method [17] reveals a positive correlation between the Al content in the blood of the drowning suspects and the water media in which they were found at a 95% significance level (ρ=0.952; ρc=0.738). This correlation supports the validity of Al as a drowning marker for deaths occurring in water media with high Al concentrations. The number of subjects in this study is however insufficient to extract more sound conclusions from a forensic point of view, since there Dissertation 52 are additional factors affecting the total amount of Al entering the bloodstream that need to be taken into account (e.g., duration of the agony). But in general, from these results it can be inferred that for water samples with concentrations of Al of around 200 µg L-1 (and higher), the concentration of Al in the blood of the drowning bodies would at least double the “basal” level (considering “basal” level a concentration of 10-15 µg L-1 Al, based in the control samples and the samples for which the corresponding water media is “Al-clean”). Table II.4. Al concentration results obtained for the eight samples of blood and water for drowning suspect cases and the two control blood samples by means of HR CS GFMAS. Uncertainties are expressed as 95% confidence intervals (calculated as ±ts/√n, where t, s and n have their usual meanings). Al content in water / μg L-1 Al content in blood / μg L-1 DS-1 2.31 ± 0.24 4.32 ± 0.82 DS-2 412 ± 14 94.3 ± 6.8 DS-3 296 ± 12 37.1 ± 3.7 DS-4 483 ± 18 38.9 ± 3.1 DS-5 195 ± 11 24.3 ± 5.3 DS-6 505 ± 14 133 ± 9 DS-7 14.9 ± 1.0 12.9 ± 2.1 DS-8 30.1 ± 1.7 11.9 ± 1.7 CTRL-1 -- 6.5 ± 1.0 CTRL-2 -- 10.5 ± 1.8 II.3.1.2. Direct determination of S in solid samples by means of highresolution continuum source graphite furnace molecular absorption spectrometry using Pd nanoparticles as chemical modifier Goals The ultimate aim of this work was to develop a direct and straightforward method for the determination of S in a variety of solid materials, relying only in calibration versus aqueous standard solutions. Dissertation 53 S is widely present in nature and essential for all life, being part of the structure of several enzymes, vitamins and all proteins. In the industry is extensively used in the production of many products, such as fertilizers, pesticides and fungicides, food additives, pharmaceutical formulations, depending on its chemical form [18]. Due to the aforementioned spreadability, simple analytical methods for S determination capable of dealing with different types of samples are to be pursued. In this work, the potential of HR CS MAS for the determination of S in direct solid samples, based in the formation of the diatomic CS molecule and the monitoring of its molecular spectra is evaluated. Particular attention was paid to find a suitable combination of chemical modifiers to achieve proper formation and stabilization of the target molecule, as well as a signal independent of the exact chemical form in which S is found in the sample. In this way, calibration with aqueous standards may enable accurate results to be obtained. Methodology Determination of atomic S is, as usual when dealing with most non-metals, not a feasible approach due to the lack of atomic lines within the UV-visible region for this element (in fact, its main resonant line appears at 180.7 nm) for atomic absorption/emission spectrometry techniques [19]. Several sulfide species are known to be formed and are available for MAS evaluation (e.g., CS, GeS, SH, S2, …) [20,21], but most of the previous works are focused on the study of the CS molecule in the region of 258.0 nm due to its superior performance [13]. Although the CS molecular system exhibits a structure of well-defined peaks of similar sensitivity [22], in this work the transition peak appearing Dissertation 54 at 257.958 nm was chosen to be monitored with the central pixel of the spectral window, since no spectral overlaps were previously reported and a better baseline stability was obtained for this peak. A series of solid samples, all of them certified reference materials (CRMs), of very different nature, were selected in order to assess the validity of the method. The materials available were: i) two polyethylene CRMs with different concentration of S, ERM-EC680 and ERM-EC681, produced by the Institute for Reference Materials and Measurements (IRMM, Geel, Belgium) and presented in granular form; ii) Mild Steel 453 and Highspeed steel 482, purchased from the Bureau of Analysed Samples Ltd. (BAS, Newby, UK), available in small chips; iii) two biological reference materials, CRM 129 Elements in hay powder (Community Bureau of Reference, BCR, Geel, Belgium) and 1566a Oyster tissue (National Institute of Standards and Technology, NIST, Gaithersburg, USA), both presented as powders; and iv) NIST 2718 Green petroleum coke, also available in powder form. The samples were cut into fragments of adequate mass making use of a ceramic knife, and deposited onto the platform, along with the corresponding amount of chemical modifier. The platform was subsequently transported to the graphite tube to be subjected to the optimized temperature programme. Table II.5 presents a summary of the instrumental parameter used in the determination. Ru was employed as permanent modifier. The procedure described by Vale et al. [23] for thermal deposition onto the platform furnace was followed. 40 µg of Pd were added in the form of a hydroalcoholic suspension of Pd nanoparticles to properly stabilize S [24]. Dissertation 55 Table II.5. Instrumental parameters used in the determination of S by means of HR CS GFMAS for the promotion and monitor of the CS molecule in gas phase Electronic transition used Δν=0, Χ 1Σ+ → A 1Π of CS molecule Wavelength / nm Main lines of CS in the vicinity of 258 nm (central pixel = 257.958 nm) Number of detector pixels summed per line 3 (~4.3 pm) Ar gas flow during atomization / mL min-1 Stop-flow conditions Chemical modifiers Ru (permanent modifier) + Pd (40 µg) nanoparticles Sample mass / mg 0.5-2.0 (BAS 482, BAS 453) 0.1-0.3 (BCR 129, NIST 1566a) 0.3-1.0 (ERM-EC680) 1.5-3.5 (ERM-EC681) 0.04-0.05 (NIST 2718) Temperature program Step Temperature/ ºC Ramp/ ºC s-1 Hold time/ s Gas flow rate/ L min-1 Drying 110 2 10 2.0 Pyrolysis 800 50 20 2.0 Auto Zero 800 0 5 0.0 Vaporization 2400 2000 5 0.0 Cleaning 2600 1000 5 2.0 Calibration was carried out versus daily prepared aqueous S standard solutions (as Na2SO4). The analytical result was obtained summing the Aint for the 3 central pixels of each one of the transition peaks registered in the vicinity of 258 nm. For each sample, eight determinations (in total) were carried out, and for each determination, five solid replicates were measured. The median was taken as the estimate value of the total S content in the samples. Results and discussion A graphite furnace is particularly suitable for the promotion of the CS molecule due to the obviously high availability of C atoms. This, along with the high stability of the CS bond (714.1 KJ mol-1) [14] makes this molecule very fit for purpose. Dissertation 56 The molecular spectrum of CS (Figure II.5) presents several sharp rotational transitions with a width of just a few picometers, similar to that of an atomic line [25]. The sensitivities for these peaks are quite similar and therefore, while not suitable for linear range expansion, it can be feasible to improve the LOD by combining the signal for some of the most sensitive ones. Figure II.5. Overview of the CS molecular absorption structure around 257.958 nm (central pixel) for the vaporization of 500 ng S added as Na2SO4 solution under the furnace conditions listed in Table II.5. A shows a wavelength resolved time-integrated spectrum, while in B the three-dimensional time and wavelength resolved spectrum is presented. Due to the lack of agreement among the previous works dealing with CS monitoring by means of HR CS MAS [25-27] concerning the use of chemical modifiers, special attention was paid to this issue in the present work. In fact, choosing the adequate/s chemical modifier/s is crucial if proper stabilization of the CS molecule, regardless of the particular chemical form in which S is found, is aimed at. The use of permanent modifiers was already been reported as recommended in the bibliography to obtain a better sensitivity. In this Dissertation 57 work, the permanent modifiers most typically used, W, Zr and Ru, were tested and, although the results were not significantly different, Ru provided somewhat better results in terms of sensitivity (by a 15-20% higher) and better defined temporal profiles. It was observed, by acquiring SEM images, that Ru particles are randomly distributed all across the surface of the graphite sample platforms, in the form of large agglomerates of approximately 0.5-1 µm length, as it is shown in Figure II.6. Figure II.6. SEM images of: A) an overview of the graphite surface of a solid sampling platform with Ru permanent coating (1 000x magnification). The unevenly distributed pale masses correspond to Ru depositions; and B) a closer view of one of the Ru agglomerates (80 000x magnification). In combination with Ru, and based in previous experiences [28], Pd was also tested as chemical modifier in order to interact with S, hopefully breaking its original bonds within the sample structure and helping in stabilizing it until favourable conditions for the promotion of CS in the graphite furnace are reached. For this purpose the chemical form in which Pd was added to the samples was carefully evaluated over four different S species (sodium sulfite, sodium sulfate, thiourea and 3-mercaptopropionic acid) using Ru as permanent modifier in all the experiments. Pyrolysis curves, plotted in Figure II.7, were obtained for 500 ng of S as each of the four S species adding 40 µg of Pd in different ways: i) as Pd nitrate a good stabilization of the less volatile species up to 1200ºC was achieved, while Dissertation 58 the more volatile species are clearly suffering from signal losses; ii) as thermally pre-reduced Pd nitrate, using a prior 1000ºC heating step, the same situation occurs although the differences between less and more volatile species are not as pronounced as they were in the first case; iii) as pre-reduced Pd nitrate by previous addition of an excess of citric acid, only the signal for the most volatile compound (3-mercaptopropionic acid) remains slightly reduced in comparison with the other three species; and iv) as a Pd nanoparticles suspension produced in-house, providing the best results with similar responses for all the four species up to 1000ºC pyrolysis temperature. Figure II.7. Pyrolysis curves obtained for 500 ng of S introduced in different chemical forms when monitoring the 257.958 nm molecular CS line in the presence of Ru as permanent modifier plus 40 µg of Pd, added in different forms: A) as Pd(NO3)2; B) as Pd(NO3)2, but thermally prereduced using a previous 1000ºC pyrolysis step; C) as Pd(NO3)2 together with an excess of citric acid; D) as Pd nanoparticles. 0.00 0.05 0.10 0.15 0.20 0.25 0500 1000 1500 2000 Mercaptopropionic acid Thiourea Na2SO4 Na2SO3 Integrated absorbance / s Temperature / ºC A 0.00 0.05 0.10 0.15 0.20 0.25 0500 1000 1500 2000 Mercaptopropionic acid Thiourea Na2SO4 Na2SO3 Integrated absorbance / s Temperature / ºC B 0.00 0.05 0.10 0.15 0.20 0.25 0500 1000 1500 2000 Mercaptopropionic acid Thiourea Na2SO4 Na2SO3 Integrated absorbance / s Temperature / ºC C 0.00 0.05 0.10 0.15 0.20 0.25 0500 1000 1500 2000 Mercaptopropionic acid Thiourea Na2SO4 Na2SO3 Integrated absorbance / s Temperature / ºC D Dissertation 59 Volynsky and Krivan [29,30] already introduced the principles of the use of colloidal Pd as chemical modifier in graphite furnace techniques, which is the same situation as the current case except for the fact that the particle size used for this work is smaller (of approximately 20 nm size). SEM photographs are presented in Figures II.8 and II.9 comparing the coating of the sample platform with Pd nitrate and Pd nanoparticles after subjecting them to a 600ºC pyrolysis temperature. The addition of nanoparticles provides a homogeneously distributed coating of metallic and highly reactive Pd nanoparticles throughout the bed of the platform (see Figures II.8.A, II.8.B and II.9.B). These particles are available for interaction already during the drying step, offering a large surface due to the small particle size attained. In the case of Pd nitrate, the distribution is not homogeneous but instead it tends to diffuse towards the edges of the platforms (see Figure II.8.C). Moreover a thin layer is formed very prone to crack as the temperature rises during the pyrolysis step (see Figures II.8.D and II.9.C). Figure II.8. SEM images of the sample surface of a graphite platform subjected to a 600°C pyrolysis stage under the following conditions: A and B) with 40 µg Pd previously deposited as an in-house synthesised suspension of Pd nanoparticles (acquired with a 25x and a 10 000x magnification, respectively); C and D) with 40 µg Pd previously deposited as Pd(NO3)2 solution (acquired with a 25x and a 10 000x magnification, respectively). Dissertation 66 II.3.1.3. Direct determination of Br in plastic materials by means of solid sampling high-resolution continuum source graphite furnace molecular absorption spectrometry Goals The aim of this work was to study the potential of HR CS GFMAS for the determination of halogens while developing an analytical method for the direct determination of Br in polymeric samples. This application was chosen because of the growing interest in controlling the toxicity and environmental impact of the great amount of plastic waste generated nowadays. There is a group of additives, particularly used in the production of plastics intended for electronic devices, electrical appliances, clothing and furniture, known as brominated flame retardants (BFRs), whose function is to modify the behaviour of the host materials against fire [31,32]. These sorts of additives are normally organic compounds containing various Br atoms within their molecular structure, and they tend to be bonded to the polymeric material in such a weakly way that they are quite easy to be released into the environment becoming a potential risk to the ecosystem [33-35]. Strict regulations and restrictions have been established then regarding the use of BFRs, leading to the need for robust analytical methods capable of monitoring the presence of these compounds in plastics and ensure whether or not the legal limits are met [36,37]. HR CS GFMAS may offer a simple, fast and sensitive enough approach for the direct determination of Br in plastic samples within such a context based on its potential for solid sampling and the capabilities to monitor Dissertation 67 molecular transitions along a narrow spectral window, as it was discussed in the previous section. Methodology The method developed is based on the formation of the diatomic molecule CaBr in gas phase and the monitoring of its molecular spectra in the vicinity of 625.315 nm. Six reference materials (comprising polyethylene, polypropylene and ABS resin) with different certified contents in Br were analysed in order to validate the method. Table II.7 below provides some information on each one of those materials, such as type of plastic, Br content and physical form. Each plastic material was cut in fragments of appropriate mass according to its Br content using a ceramic knife. After weighing the sample, the required chemical modifiers were added to the platform in order to form and stabilise the CaBr molecule (that is, 30 µg Pd as Pd(NO3)2 solution and 300 µg Ca as CaCO3 solution). The sample platform was then transported to the graphite furnace to be subjected to the optimized temperature program. Dissertation 68 Table II.7. List of polymeric CRMs analysed for Br Availability IRMM IRMM MAT MAT IRMM BAM Institute for Reference Materials and Measurements (IRMM, Geel, Belgium) Modern Analytical Techniques (MAT, Hillsborough, USA) Federal Institute for Materials Research and Testing (BAM, Berlin, Germany) Physical form Granules (aprox. 10 mg weight) Granules (aprox. 10 mg weight) Discs (aprox. 31 mm diameter and 13 mm thickness) Discs (aprox. 31 mm diameter and 13 mm thickness) Granules (aprox. 10 mg weight) Discs (aprox. 40 mm diameter and 6 mm thickness) Certified Br content /mg kg-1 808 ± 19 98 ± 5 1100 ± 44 500 ± 20 2080 ± 70 240 ± 21 Type of material High density polyethylene High density polyethylene High density polyethylene Low density polyethylene Polypropylene Acrylonitrile butadiene styrene ERM-EC680 ERM-EC681 PE-H-11A PE-L-11A ERM-EC591 BAM-H010 Dissertation 69 Table II.8 summarizes the instrumental conditions and temperature program used for this application. Table II.8. Instrumental parameters used in the determination of Br by means of HR CS GFMAS Electronic transition Red system, ∆v=0, X 2Σ+ ! A 2Π (CaBr) Central wavelength 625.315 nm Spectral window 0.73 nm Number of detector pixels summed per line 3 (≈ 11 pm) Chemical modifiers Ca (300 µg) Pd (30 µg) Sample mass range ERM-EC680 0.1 mg - 0.2 mg ERM-EC681 0.7 mg - 1.0 mg PE-H-11A 0.1 mg - 0.2 mg PE-L-11A 0.1 mg - 0.4 mg ERM-EC591 0.1 mg - 0.15 mg BAM-H010 0.3 mg - 1.0 mg Optimizing the amount of Ca was required in order to ensure complete interaction with all Br species present in the samples and therefore maximize the sensitivity. Is has to be taken into consideration that Ca forms stable diatomic molecules with most non-metals and association of Ca with F and Cl gives rise to more stable mono-halides than CaBr. Therefore, if a material with high contents of F and/or Cl is to be determined, a direct method without any prior sample treatment, like the one optimized in this work, would not be suitable. Anyway, 300 µg Ca has proven to be a Temperature program Step Temperature/ °C Ramp/ °C s-1 Hold time/ s Gas flow rate/ L min-1 Drying 90 5 20 2.0 Drying 120 5 30 2.0 Pyrolysis 1000 50 30 2.0 Auto Zero 1000 0 5 0.0 Vaporization 2100 3000 6 0.0 Cleaning 2500 1000 4 2.0 Dissertation 70 sufficient amount to ensure a complete reaction when 100 ng of Br or lower amounts are present. Pd was used as chemical modifier in order to stabilize the analyte during the pyrolysis step and best results were obtained when added in masses above 10 µg for a 100 ng Br solution. Paying attention to the temperature program, the CaBr molecule allows for pyrolysis temperatures up to 1400-1500°C. After this temperature, the signal drop is rather abrupt, and so is the growth for the signal when monitoring the vaporization temperature, reaching a maximum at 19002200°C. This is a very positive situation, where a pyrolysis temperature high enough to allow most matrix removal can be used and, at the same time, a rather mild vaporization temperature can be applied during the measurement, resulting in a beneficial effect over the lifetime of all perishable components of the furnace. Calibration was carried out versus daily prepared aqueous Br standard solutions. The analytical result was obtained summing the Aint for the 3 central pixels of the transition peak registered at 625.315 nm. Each sample was analysed in a different session. For each sample, seven determinations (in total) were carried out, and for each determination, three solid replicates were measured. The median was taken as the estimate value of the total Br content in the samples. Results and discussion As it was already mentioned in the introduction chapter, the main atomic absorption lines of Br lie within the vacuum-UV region. The principal line is located at 148.845 nm [38]. Because it is not possible to measure in such short wavelength with the instrumentation currently available, it is Dissertation 71 necessary to search for an alternative. Just a couple of previous works were published using HR CS AAS for studying the absorption spectra of Brbased molecules. Huang et al. [39] made a general study comparing the formation and absorption structures of Br binary compounds with Al, Sr and Ca, further focusing in the AlBr and CaBr molecules. Limburg and Einax [40] optimized an analytical method for the determination of Br as CaBr in water solutions and organic solvents. In this work, and based on those previous studies, Ca was preferred over Al for mainly two reasons: i) CaBr is less volatile than AlBr [39]. In monitoring CaBr, it is possible to use pyrolysis temperatures up to 1400°C-1500°C (up to 900°C in the case of AlBr), assuring a better removal of the matrix, and this is particularly important when the direct analysis of solid samples is intended; and ii) the resolved structure of its spectrum opposite to the broad band of AlBr. CaBr has its main resonance system between 625 and 628 nm, consisting on a structure of fine rotational lines that are well resolved and show a maximum peak at 625.315 nm [22]. This kind of molecular absorption structure greatly simplifies the selection of pixels for baseline correction, providing the method with a suitable robustness. In Figure II.14 a typical absorption profile for CaBr is depicted. In this work, only the four more sensitive peaks were taken into account to be studied in detail. These transitions appearing at 625.315 nm, 625.223 nm, 625.135 nm and 625.058 nm are close and distant enough in sensitivity to make it feasible to explore the possibility of improving the LOD as well as expanding the linear range [41,42]. Indeed, if the intensity of all four main lines is combined, it is possible to reduce the LOD by a factor of three (1.8 Dissertation 72 µg g-1 Br in comparison to 5.4 µg g-1, attainable when monitoring only the most sensitive peak). Figure II.14. Overview of the CaBr molecular absorption structure around 624.982 nm (central pixel) for the vaporization of 100 ng Br, plus 300 µg Ca and 30 µg Pd under the furnace conditions listed in Table II.8. A shows a wavelength resolved timeintegrated spectrum, while in B the three-dimensional time and wavelength resolved spectrum is presented. On the other hand, the difference in sensitivity between the four lines allows for the expansion of the linear range along almost three orders of magnitude, as it is presented in Table II.9. Of course, making use of the signal for less sensitive peaks would increase the linear range even further, but that would be of no interest for the current application. 624.614 624.798 624.982 625.166 625.350 Wavelength / nm 0.000 0.020 0.040 0.060 0.080 0.100 0.120 Integrated absorbance / s A 624.614 624.798 624.982 625.166 625.350 Wavelength / nm 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Time / s 0.00 0.05 0.10 0.15 0.20 Absorbance B Dissertation 73 Table II.9. Analytical features for the four main CaBr rotational transitions monitored under the conditions summarized in Table II.8. Characteristic mass values were calculated as 0.0044 divided by the slope of the calibration curve. LODs values were calculated as three times the standard deviation of the blank (n=10) divided by the slope of the calibration curve. Wavelength/ nm m0 / ng LOD / ng Linear range R2 625.315 1.32 5.4 Up to 300 ng 0.999 625.223 1.77 7.3 Up to 300 ng 0.999 625.135 2.42 12.7 Up to 700 ng 0.996 625.058 3.00 18.2 50 – 2000 ng 0.997 It is worth mentioning that an alternative to increase the linear range always available when using HR CS AAS is the possibility to select side pixels, instead of using the central pixel plus the two adjacent ones (CP ± 1), which is the option recommended for achieving the best LOD [42-44]. This strategy to expand linearity has been successfully used in previous works [42,45,46], but seems to be limited in this particular case by i) the proximity of the adjacent lines and ii) the asymmetry of the peaks, as shown in Figure II.15. This aspect is relevant because if the signal is symmetrical, as it is most often the case for atomic lines, by using two side pixels located at the same distance from the CP (e.g. +3 and −3), the overall signal may remain constant even if there are some spectral drifts during the measurements (the increase in one pixel would be balanced by the decrease in the other one). However, that is clearly not the case for CaBr lines, as the signal becomes very asymmetric when moving more than 2 pixels away from the central pixel. Thus, monitoring side pixels is not recommended in this case. Fortunately, the availability of less sensitive CaBr makes it easier to accommodate the sensitivity of the method to that of the sample and that should be the preferred solution for monitoring high Br amounts. Dissertation 74 Figure II.15. Wavelength resolved time-integrated absorbance spectrum obtained by vaporization of approximately 250 ng of Br (0.310 mg of sample ERM-EC680). The signal only shows the region where the main CaBr molecular absorption lines are located. The signal recorded by every individual detector pixel is highlighted (blue dots). Monitoring several absorption peaks is also an advantage for better assessing the reliability of the results, permitting to directly detect any spectral overlap affecting one of the peaks. Therefore, the determination of the plastic samples was carried out in all of the four main sensitive lines of the CaBr molecular system following the methodological procedure described before. It is important to emphasize that in these optimal conditions the signals obtained for the solid samples and for aqueous standard solutions were always comparable in terms of peak area. Figure II.16 shows an example of the typical temporal signal profile found for polymers and for solutions. Dissertation 75 Figure II.16. Comparison of a time-resolved molecular absorbance signal recorded at the 625.315 nm CaBr transition (sum of three central pixels) after vaporizing a similar amount of Br from an aqueous solution and directly from a polyethylene material under the conditions listed in Table II.8. The results obtained for the CRMs listed in Table II.7 and for each of the absorption peaks listed in Table II.9 are displayed in Table II.10. In all of the cases, a good agreement with the reference values was found. Anyhow, best results are obtained when combining the values of the four transitions studied, which brings about also the best precision (RSD values between 3 and 7 %). It can be concluded that HR CS GFMAS offers a fast and simple method for determining total Br content in polymers, which is sensitive enough to comply with regulations. This method can be useful for screening purposes with the idea of filtering out most of the samples prior to be subjected to more time-consuming techniques that may selectively analyse the target brominated compounds in each of those in which the total Br content is high enough to be of concern. Dissertation 82 significantly different it is not advisable to sum their signals, as the less sensitive ones will moderately improve the analytical signal while contributing to increase the noise level. However, in this last situation an easy expansion of the linearity is feasible, as each of the lines should cover a very different linear range. Two clear examples of these situations were already presented when dealing with molecular absorption structures (CS and CaBr, representing i) and ii) cases, respectively). In this work the Ni triplet appearing in the 234.6 nm region is evaluated in order to further illustrate the situation in which atomic lines with different sensitivities are at disposal. Figure II.19 displays the spectrum recorded from 234.533 nm to 234.792 nm, where three lines of Ni appear (at 234.554, 234.663 and 234.751 nm) with a representation of their linear ranges. Table II.13 summarizes the analytical features of each of these lines. Figure II.19. Wavelength-resolved time-integrated spectrum showing the Ni triplet in the vicinity of 234.6 nm of 10 ng Ni added as an elemental aqueous solution by means of HR CS AAS. The linear range of each of the three lines is also shown. 20 to 2000 pg 500 to 50000 pg 200 to 20000 pg 234.533 234.598 234.662 234.727 234.792 Wavelength / nm -0.05 0.10 0.25 0.40 0.55 0.70 0.85 Absorbance Wednesday, 23 March 2011 Dissertation 83 Table II.13. Analytical features of the three lines of the Ni triplet found in the vicinity of 234.6 nm Wavelength/ nm m0/pg LOD/pg Linear range/pg R2 234.554 16 11 Up to 2,000 0.998 234.663 260 210 Up to 50,000 0.999 234.751 140 160 Up to 20,000 0.999 The three lines exhibit very different sensitivities and therefore very different linear behaviour, permitting to cover a range from 20 to 50,000 pg. In this way, the line that fits best for every sample can be chosen a posteriori without modifying the experimental conditions or repeating any measurements. Concerning the multi-elemental capabilities of the HR CS AAS technique, it was also previously discussed how, when lines of different elements lie close enough to be recorded within the same spectral window, two situations may be encountered: i) that the elements of interest show similar volatilities; and ii) that the elements show very different thermal behaviour. In this work, both situations were studied and illustrated with examples. In the situation where the volatilities are similar it is rather simple to optimize a temperature program that suits all the elements for an almost simultaneous atomization. The neighbouring of 352.5 nm, where an atomic line for Ni (352.454 nm, with a m0 of 30 pg), a duplet for Fe (352.604/352.617 nm, with m0 of 3 and 27 ng, respectively) and a line for Co (352.685 nm, with a m0 of 75 pg) are found, was further studied in this work. Direct solid determination of a biological CRM, NIST 1566a Oyster tissue, was investigated under the instrumental parameters listed in Table II.11. It is important to stress that there is an adequate correspondence between the sensitivities of these lines and the analyte contents in this sample, which is a requisite for simultaneous multi-element analysis using Dissertation 84 HR CS GFAAS. The organic sample matrix is easy to remove at moderate pyrolysis temperatures, avoiding analyte losses as well as minimizing potential matrix interferences, which makes this kind of sample ideal for SS HR CS GFAAS analysis. A three-dimensional spectrum recorded for one of the sample replicates is presented in Figure II.20. Figure II.20. Three-dimensional timeand wavelength-resolved absorbance spectrum obtained for the simultaneous determination of Co (352.685 nm), Fe (352.604/352.617 nm) and Ni (352.454 nm) in 2.795 mg of NIST SRM 1566a Oyster tissue by means of SS HR CS GFAAS under the instrumental conditions listed in Table II.11. As can be seen, well-defined unimodal signal profiles were obtained for all analytes, which were actually very similar both in terms of shape and, most important, in terms of sensitivity to those obtained for aqueous standard solutions, thus enabling simple calibration just by constructing the curve with aqueous standard solutions. The results for this determination are presented in Table II.14, which proved the validity of the method to obtain accurate multi-element values in a fast and simple way. Dissertation 85 Table II.14. Results obtained for the simultaneous multi-element determination of Co, Fe and Ni (n=5) in NIST SRM 1566a Oyster Tissue by means of solid sampling HR CS GFAAS. Uncertainties are expressed as 95% confidence intervals (calculated as ±ts/√n, where t, s and n have their usual meanings) Wavelength / nm Analyte Result / µg g-1 Certified value / µg g-1 352.454 Ni 2.08 ± 0.20 2.25 ± 0.45 352.604 Fe 557 ± 23 539 ± 32 352.617 Fe 567 ± 31 539 ± 32 352.685 Co 0.59 ± 0.06 0.57 ± 0.11 When the elements of interest show very different volatilities, two possibilities can be considered. If a simultaneous determination is preferred, compromise conditions of temperature and chemical modifier/s have to be employed. However, in this case a sequential approach is also possible programing a different atomization step for each of the analytes [47-50]. Moreover, also the spectral region can be modified for each atomization step, and so are the Ar flow conditions and the chemical modifier/s, allowing for the determination of each analyte in its optimum conditions. To compare these two approaches, a biological reference material, BCR 679 White cabbage was analysed for Cd and Ni contents. In the region of 228.9 nm, two lines that are sufficiently close to be recorded simultaneously appear, corresponding to the main Cd atomic line (228.802 nm) and a secondary Ni atomic line (228.998 nm). Compromise furnace conditions were developed in order to simultaneously monitor both signals, using an atomization temperature high enough to atomize both elements but without decreasing too much the sensitivity for the most volatile one (Cd). In addition, the amount of chemical modifier was also optimized, paying special attention to the stabilization of Cd during the pyrolysis. The conditions finally used are summarized in Table II.12(a) and the threedimensional spectrum obtained is presented in Figure II.21. Dissertation 86 Figure II.21. Three-dimensional timeand wavelength-resolved absorbance spectrum obtained for the simultaneous determination of Cd (228.802 nm) and Ni (228.998 nm) in 0.453 mg of BCR CRM 679 White cabbage by means of solid sampling HR CS GFAAS under the instrumental conditions listed in Table II.12(a). However, by selecting two different atomic absorption lines for Ni and Cd for this sample and the optimum conditions for each one of them it is possible to subject the sample to two temperature programs consecutively and obtain sequential results for Cd and Ni (the optimum atomization temperature for Cd would be suitable for Ni pyrolysis), but from the same sample replicate. Conditions for the analysis are presented in Table II.12(b) and an illustrative bi-dimensional plot of both signals in time is deployed in Figure II.22. Results obtained for both approaches along with the certified reference values are presented in Table II.15 for comparison purposes. An excellent agreement with the expected values is reported in both situations. Again, it was feasible to carry out the calibration with aqueous standard solutions. Dissertation 87 Figure II.22. Time profile for the sequential and selective atomization of Cd and Ni in 0.266 mg of BCR CRM 679 White cabbage by means of SS HR CS GFAAS under the instrumental conditions listed in Table II.12(b). Table II.15. Comparative results obtained for the simultaneous (using the conditions listed in Table II.12(a)) and sequential (using the conditions listed in II.12(b)) multielement determination of Cd and Ni (n=5) in BCR CRM 679 White cabbage by means of SS HR CS GFAAS. Uncertainties are expressed as 95% confidence intervals (calculated as ±ts/√n, where t, s and n have their usual meanings) Analyte Simultaneous method/µg g-1 Sequential method/µg g-1 Certified value / µg g-1 Cd 1.69 ± 0.13 1.60 ± 0.14 1.66 ± 0.07 Ni 27.3 ± 1.7 27.6 ± 1.9 27.0 ± 0.8 Figure II.23 presents an example of the sequential situation further expanded when a highly volatile element (Hg) wants to be determined along with an element with medium volatility (Zn) and two of low volatility (Cr and Mn). The last two elements show closely adjacent lines such that they are simultaneously monitored. In summary, although limited when comparing with other multi-element techniques, the commercially available HR CS AAS instrumentation offers a potential possibility for performing multi-line monitoring and direct Dissertation 88 multi-element analysis, which represents a key improvement in the trajectory of AAS techniques. Figure II.23. Sequential and selective atomization of Hg (5 ng), Zn (20 ng) and Cr (40 pg) added as a multi-element aqueous standard solution by means of HR CS GFAAS. The signal of Mn, due to the use of KMnO4 as chemical modifier for Hg, was divided by a factor of 5 to keep the same scale II.3.2.2. High-resolution continuum source atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple lines. A critical review of current possibilities This review work critically examines the potential of HR CS AAS for multi-line monitoring, discussing in detail the possible strategies to develop multi-element methods, considering the requirements and limitations of the technique. Moreover, the article also discusses other advantages deriving from multiline monitoring, such as: i) the expansion of the linear range by measuring multiplets; ii) the possible improvements in the LOD and in precision by summing the signals from different lines of the same element or molecule; iii) the possibilities to correct for matrix effects by selecting an IS; and iv) the accurate mathematical correction of spectral overlaps by simultaneous Dissertation 89 monitoring of other lines of the interfering molecule or element that are free of interferences. All these aspects are illustrated with examples found in the literature, stressing their importance for the straightforward analysis of solid samples and complex materials. Finally, the differences found when the flame is used as an atomization unit, thus resulting in quasi-stable signals, are also addressed. Since the majority of these aspects have been already commented extensively in the introduction and in the previous section, they will not be repeated herein. II.3.3. Microsampling II.3.3.1. Simultaneous direct determination of Ni and Cd in Daphnia magna specimens by means of high-resolution continuum source graphite furnace atomic absorption spectrometry for eco-toxicological assessment purposes Goals This work is part of a broader cooperation study with the Laboratory of Environmental Toxicology and the Laboratory of Microbial Ecology and Technology of Ghent University. In this work, the capabilities of SS HR CS GFAAS for the direct determination of the total Ni body burden in individual specimens of D. magna (small aquatic invertebrates of approx. 2 mm length that are widely used in eco-toxicological research) were investigated. Out of the interpretation of the obtained results from the eco-toxicological point of view two papers were published. A first one is based in the study of the potential of using metal-contaminated liposomes as an alternative delivery system of dietary metals rather than the usually employed metal- Dissertation 90 exposed algae. A reduction in the content of some of the essential nutrients (e.g. fatty acids) has been reported for the latter to be influenced by this metal exposure, resulting in equivocal results. A second paper was presented establishing an experiment for the quantification of the contribution of the aforementioned nutritional quality shifts to the toxic effects when metal-contaminated algae are used as food vectors. While the toxicity experiment was only focused on Ni exposure and only the concerning results were published in the aforementioned articles, basal concentrations of Cd were also of interest in view of future eco-toxicological studies based on Cd-stress [51,52]. In this doctoral thesis only the analytical aspects of the method developed will be further discussed, stressing the suitability of SS HR CS GFAAS for the direct determination of unique individual specimens of small size. Methodology An analytical multi-element method was developed in order to monitor the Ni body content (analyte of interest) as well as Cd basal concentrations in a batch of 192 D. magna specimens divided in different lots. The members of each lot were subjected to different dietary and waterborne conditions. The spectral analytical window was then centered on 228.899 nm so to properly monitor both Ni and Cd atomic lines. The spectral region has been already described in Section II.3.2.1 to illustrate the possibilities of HR CS AAS for simultaneous multi-element determinations when dealing with analytes if very different volatilities and applied to the analysis of BCR CRM 679 White cabbage. The furnace conditions for the determination of Cd and Ni in D. magna samples were slightly modified, though. As the main interest was focused on Ni, a maximum sensitivity had to be assured for this Dissertation 91 element. To this aim, atomization temperature of 2400°C was preferred and additional Ar gas flow in this step was suppressed. Pyrolysis stage was fixed in 700°C. Instrumental conditions are listed in Table II.16. Table II.16. Instrumental parameters used for the simultaneous multi-elemental determination of Cd and Ni in D. magna by means of SS HR CS GFAAS Analyte Cd Ni Wavelength / nm 228.802 228.998 Number of pixels summed 3 (≈3.7 pm) 3 (≈3.7 pm) Chemical modifier 1 µg Pd (added as Pd(NO3)2 solution) Temperature program Step Temperature/ºC Ramp/ºC s-1 Time/s Gas flow rate/L min-1 Drying 150 5 35 2.0 Pyrolysis 700 50 30 2.0 Auto-zero 700 0 5 0.0 Atomization 2400 1500 10 0.0 Cleaning 2600 500 4 2.0 Pd modifier was needed in order to stabilize Cd during the pyrolysis. However, it is worth pointing out here that the 10 µL Pd(NO3)2 solution also fulfils the role of preventing the ejection of the sample from the furnace during the drying step due to the sweeping effect of the Ar flow over the very light invertebrate individuals, as already recommended in a previous work [45]. D. magna individuals were collected after being subjected to a 21-day toxicity test. Performance [53] and culture conditions can be found elsewhere [54]. Individuals were exposed to two different diet combinations: - 10:1 w/w of control algae and Ni-exposed liposomes at 5 different levels (variations should only be observed for Ni content). Dissertation 98 II.3.3.2. Isotope ratio mapping by means of laser ablation-single collectorICP-mass spectrometry: Zn tracer studies in thin sections of Daphnia magna Goals Taking advantage of the possibility of working with different instrumental setups offered by this co-tutelage work, LA-SC-ICPMS was also studied as a complementary technique for microsampling. In this work, two important features of LA-ICPMS were combined: i) its capability for providing isotopic information [55], and ii) the high spatial resolution reported for elemental mapping [56]. While both characteristics are increasingly addressed in the literature in a separate way, very few studies have reported on the use of LA-ICPMS for acquiring isotope ratio mappings [57], and none based on the use of more easily accessible SCICPMS In order to evaluate the capabilities of LA-SC-ICPMS for filling this methodological gap, the use of this technique for developing a tracer experiment in D. magna individuals exposed to increased dissolved Zn concentrations with different isotopic composition was optimized so to offer a way for presenting the data as Zn isotopic ratio images of thin sections. The development of this methodological approach included: i) the evaluation of the performance of a new entrance/exit slit combination available for the Thermo Element XR instrument capable of providing flattopped spectral peaks and better enhanced ion transmission when operating at higher mass resolution (m/∆m≈2000) in comparison with the conventional medium resolution one (m/∆m≈4000, triangular shaped peaks); ii) the use of a wet plasma set-up for improving plasma robustness; iii) thorough optimization of the ablation conditions and data acquisition Dissertation 99 parameters in order to attain the best precision possible; and iv) adequate data treatment. Methodology After culturing D. magna specimens from a monoclonal in-house laboratory population under the normal dietary and waterborne conditions, that may be found elsewhere [58], 21 days old individuals were collected and exposed to a water medium fortified with a Zn spike, isotopically enriched in 64Zn and 67Zn for a total Zn concentration of 1 mg L-1. The isotopic composition of the Zn spike was determined to be 64Zn (50.52%), 66Zn(2.12%), 67Zn(44.79%), 68Zn(2.52%) and 70Zn(0.05%) [59]. The Zn exposure lasted for 24 hours without any further feeding or aeration, then the individuals were properly rinsed and dehydrated to be embedded in cubes filled with molten paraffin for, once cold, being cut in 20 µm thickness sagittal sections using a rotary microtome. The fine sections were then adhered to glass plates for ablation. Measurement conditions and instrumental parameters used for sample ablation are summarized in Table II.20. “Wet plasma” conditions were achieved by simultaneously nebulizing a 5 µg L-1 Cu isotopic standard solution (NIST SRM 976, National Institute of Standards and Technology, MD, USA) with the dry aerosol generated in the ablation. In this way, Cu isotope ratio was monitored and used as internal standard for mass bias correction. NIST SRM 612 Trace elements in glass reference material was used for optimizing the instrumental settings at the beginning of every measurement session so to maximize the Zn signal intensity, to obtain good peak shapes and attain a stable U/Th ratio close to unity, minimizing the formation of oxides and doubly-charged ions [60]. Dissertation 100 Table II.20. Instrumental settings and data acquisition parameters for the ablation of thin sections of D. magna individuals by means of LA-SF-ICPMS LA unit parameters: New Wave Research UP193HE Wavelength 193 nm Cell volume ≈ 2.5 cm3 Spot size 30 μm Repetition rate 10 Hz Scanning speed 60 μm s-1 Energy density 1.8-2.1 J cm-2 Laser warm-up time 10 s ICPMS instrumental parameters: Thermo Element XR Mass resolution, m/Δm 2000 RF power 1120 W Auxiliary gas flow rate (Ar) 0.70 L min-1 Nebulizer gas flow rate (Ar) 0.75 L min-1 Carrier (He) gas flow rate through cell 0.60 L min-1 Data acquisition parameters Scanning mode E-Scan Settling time 1 ms Samples per peak 50 Mass window 10% Sample time 3 ms Segment duration (total dwell time per nuclide) 15 ms Runs/passes 600/1 Nuclides monitored 63Cu+, 64Zn+, 65Cu+, 66Zn+, 67Zn+, 68Zn+ Concerning data handling, rejection of spiky signals (about 0.5% of the acquired points or less) was done manually by spotting the signals exceeding the level of the 10 adjacent points by more than 3 times their standard deviation. On the other hand, the first 10 seconds of laser’s warmup were considered gas blank signal and subtracted from every data point. If the Zn net signal laid beneath the blank value plus three times the standard deviation for the gas blank, the Zn concentration was considered too low for providing reliable isotope ratios and these values were disregarded for constructing the final isotope ratio image. Dissertation 101 The isotope ratios considered for image composition were 66Zn/64Zn, 68Zn/64Zn, 66Zn/67Zn and 68Zn/67Zn. A point-by-point mass bias correction was applied for each isotope ratio by applying the empirical strategy proposed by Longerich et al. [61] and improved by Maréchal et al. [62] and Woodhead [63]. Estimation of the K-factor was done by the exponential model of Russell et al. [64]. Results and discussion During the whole process of optimization for developing a method to acquire Zn isotope ratio images in such thin sections of a D. magna specimen particular attention was paid to three main aspects [65]: i) maximization of sensitivity; ii) reduction of the occurrence of spiky signals; and iii) improvement of precision by trying to achieve quasi-simultaneous measurement conditions. The parameters affecting these factors are all interrelated and, therefore, the optimization requires to find compromise conditions in most cases. It is firstly necessary to point out that the Zn concentration in the samples is expected to be quite low (approximately 80 µg g-1 in base level populations) [45]. Moreover, Zn presents several potential spectral interferences. In this type of samples the most problematic ones are due to the formation of polyatomic species based on P and Ca (mainly present in the exoskeleton of the crustaceans) and on S and Cl (essential elements for all living organisms). For this reason, a sector field instrument capable of working under medium resolution conditions would be required [66,67]. The use of a conventional medium resolution slit (m/∆m≈4000) completely overcome these interferences but at the same time entails a significant drop in sensitivity as well as triangular-shaped peaks which also negatively affects the precision of the isotope ratio measurements. To cope with these Dissertation 102 detrimental effects a new medium resolution slit system based on a wider exit slit was used in this work, providing a 50% decrease in resolution (though sufficient for finding an interference-free region in the centre of the peak, as it is shown in Figure II.26) but flat-topped peaks (leading to higher precision) and enhanced ion transmission efficiency (leading to higher sensitivity, approx. a factor of 2). Regarding the laser parameters for the ablation, the spot size of the laser beam had to be optimized in order to achieve an adequate compromise between sensitivity and spatial resolution considering the small dimensions of the samples under study. Also the scan speed needed to be adjusted, in accordance with the laser energy and repetition rate, so to minimize sample acquisition time without sacrificing the sensitivity and spatial resolution achieved. In this case, 30 µm spot size resulted in a good agreement between resolution and signal intensity, with a 10 Hz repetition rate, a laser fluence of 1.8-2.1 J cm-2 (for ablating the total thickness of the sample without inflicting damage to the glass) and a 60 µm s-1 scan speed (for a reasonable total acquisition time). A teardrop ablation cell [1-3] with a reduce volume of 2.5 cm3 was used in this work to reduce as much as possible the required washout time and therefore, minimize the signal expansion. As for the achievement of more robust plasma conditions, wet plasma conditions are known to provide higher robustness towards matrix effects [65,68] and more stable signals with a significantly reduced amount of spikes, as shown in Figure II.27. In this work, a setup was developed based on an experimental setting presented by O’Connor et al. [68]. A Y-shaped joint was used to connect the liquid aerosol coming from a dual-pass spray chamber to the stream of the dry aerosol coming from the ablation cell, Dissertation 103 where both merge and flow toward the plasma torch. A diagram giving an outline of this disposition is presented in Figure II.28. Figure II.26. Mass spectra in the vicinity of 64Zn and 68Zn when simultaneously nebulizing high concentrations of Cl and S, using different mass resolution slit systems. A) Mass spectrum in the vicinity of 64Zn for a solution containing 3 µg L-1 Zn and 10 mg L-1 S using a conventional medium resolution slit system (m/Δm ≈ 4000). B) Mass spectrum in the vicinity of 64Zn for a solution containing 3 µg L-1 Zn and 10 mg L-1 S using a new medium resolution slit system (flat-topped peaks, m/Δm ≈ 2000). C) Mass spectrum in the vicinity of 68Zn for a solution containing 3 µg L-1 Zn and 4.25 g L-1 Cl using a conventional medium resolution slit system (m/Δm ≈ 4000). D) Mass spectrum in the vicinity of 68Zn for a solution containing 3 µg L-1 Zn and 4.25 g L-1 Cl using a new medium resolution slit system (flat-topped peaks, m/Δm ≈ 2000). . Dissertation 104 Figure II.27. Comparison between the 64Zn signals obtained upon ablation of NIST SRM 612 glass (400 µm, 20 Hz, 10 µm s-1, 10 J cm2) measured in both dry and wet (through nebulization of a 0.14 M HNO3 solution) plasma conditions. Figure II.28. Wet plasma setup used for the acquisition of Zn isotope ratio images This approach permitted to nebulize a Cu isotopic standard solution for the generation of the wet plasma and be able to use it afterward for mass bias correction purposes in data treatment. On the other hand, in order to attain a situation as close as possible to the ideally simultaneous detection of the isotopes monitored, the mass Dissertation 105 spectrometer acquisition parameters had to be studied, particularly paying attention to the nuclides monitored (they should be as close as possible), the settling time and the total dwell time. An internal standard (Cu) was selected that allowed for scanning of all nuclides only by changing the accelerating voltage, under a fixed magnet mass. In this way, the settling time could be set at 1 ms, the minimum allowed by the instrument software. To establish an adequate total dwell time, an homogeneous material for Zn (NIST SRM 612 glass) [69] was measured under the same conditions of mass window (central 10% of the peak), samples per peak (50) and settling time (1 ms) but varying the time spent at each sample from 1 ms (5 ms dwell time per nuclide) to 16 ms (83 ms dwell time per nuclide) in order to evaluate its influence on the measurement precision. Data points were averaged every approximately 500 ms, so to obtain 30 x 30 µm square pixels, and the point-by-point RSD was compared for every isotope ratio measured. Both medium resolution slits systems were tested for this optimization, as illustrated in Figure II.29. It was concluded that the one providing flat-topped peaks leaded to improved isotope ratio precisions in all cases. Moreover, best precision was obtained when using from 15 ms to 27 ms acquisition time (5% RSD or better for the most favourable ratios, 66Zn/64Zn and 68Zn/64Zn). Under these optimized conditions the images were generated ablating 30 µm vertically adjacent rasters. Approximately 50 scan lines of about 50 s duration each were needed to cover the whole sample section. As the ablation moved at 60 µm s-1 with a 10 Hz repetition rate, a pace of 6 µm per laser shot (0.1 s) was established. Measuring 50 samples per peak and measuring just in the central 10% portion of the nominal peak, 5 points per nuclide were registered every 0.5 s (30 µm spot size), which means 1 point every 0.1 s. That means that a signal intensity value for every target nuclide Dissertation 106 was actually obtained for every laser shot. With every laser shot, only 6 µm of new material are sampled, together with 24 µm of an already ablated area. In this way, the moving average of 5 data points was used for processing the collected data under these conditions, as it perfectly simulate the ablation process. Figure II.29. Measurement precision (%RSD) for Zn isotope ratios as a function of the total dwell time selected per nuclide, obtained for the ablation of NIST SRM 612 glass (200 µm, 20 Hz, 10 µm s-1, 10 J cm2) under wet plasma conditions, using A) a new medium resolution slit system (flat-topped peaks, m/Δm ≈ 2000) and B) a conventional medium resolution slit system (triangular peaks, m/Δm ≈ 4000). The number between brackets next to each dwell time value indicates the number of points averaged for obtaining a single isotope ratio value every 500 ms. Natural values for the isotope ratios showed in the charts: 66Zn/64Zn=0.564, 68Zn/64Zn=0.375, 67Zn/66Zn=0.146, 68Zn/67Zn=4.567. Figure II.30.A shows the 64Zn body distribution image of a D. magna specimen exposed to 64Zn-enriched water medium. It can be observed how Zn preferentially accumulates in the eye, the gastrointestinal tract, the gills and the egg/embryo, as it was expected from previous findings presented in the literature [70]. Figures II.30.B and II.30.C show the 66Zn/64Zn and Dissertation 107 68Zn/64Zn isotope ratio images, respectively, obtained after transforming the elemental images for the same 20 µm section of a D. magna. Figure II.30. Isotope mapping of 64Zn (A), 66Zn/64Zn isotope ratio (B) and 68Zn/64Zn isotope ratio (C) distribution over a 20 µm section of a D. magna specimen exposed to a 64Zn-enriched water medium obtained by LA-SC-ICPMS Average isotope ratios were calculated for the different mentioned regions for both control and exposed D. magna individuals. The analytical uncertainty was estimated by ablating 4 contiguous sections of the same specimen in each case (control and exposed). Results are illustrated in Figure II.31. From this figure it can be seen that in the control specimen no significant difference was encountered between the body areas of accumulation, with isotope ratio values close to the expected for Zn of isotopic natural abundance (66Zn/64Zn=0.564 and 68Zn/64Zn=0.375). However, exposed specimens record isotope ratios close to natural values only in the Dissertation 114 [23] M.G.R. Vale, M.M. Silva, B. Welz, R. Nowka, Control of spectral and non-spectral interferences in the determination of thallium in river and marine sediments using solid sampling electrothermal atomic absorption spectrometry, Journal of Analytical Atomic Spectrometry 17 (2002) 38–45. [24] A.B. Volynsky, Mechanisms of action of platinum group modifiers in electrothermal atomic absorption spectrometry, Spectrochimica Acta Part B: Atomic Spectroscopy 55 (2000) 103–150. [25] Z. Kowalewska, Feasibility of high-resolution continuum source molecular absorption spectrometry in flame and furnace for sulphur determination in petroleum products, Spectrochimica Acta Part B: Atomic Spectroscopy 66 (2011) 546–556. [26] U. Heitmann, H. Becker-Ross, S. Florek, M.-D. Huang, M. Okruss, Determination of non-metals via molecular absorption using highresolution continuum source absorption spectrometry and graphite furnace atomization, Journal of Analytical Atomic Spectrometry 21 (2006) 1314–1320. [27] H.S. Ferreira, F.G. Lepri, B. Welz, E. Carasek, M.-D. Huang, Determination of sulfur in biological samples using high-resolution molecular absorption spectrometry in a graphite furnace with direct solid sampling, Journal of Analytical Atomic Spectrometry 25 (2010) 1039–1045. [28] M. Resano, M. Verstraete, F. Vanhaecke, L. Moens, J. Claessens, Direct determination of sulfur in Bisphenol A at ultratrace levels by means of solid sampling-electrothermal vaporization-ICP-MS, Journal of Analytical Atomic Spectrometry 16 (2001) 793–800. Dissertation 115 [29] A.B. Volynsky, V. Krivan, Comparison of various forms of palladium used as chemical modifiers for the determination of selenium by electrothermal atomic absorption spectrometry, Journal of Analytical Atomic Spectrometry 11 (1996) 159–164. [30] A.B. Volynsky, V. Krivan, Colloidal palladium–a promising chemical modifier for electrothermal atomic absorption spectrometry, Spectrochimica Acta Part B: Atomic Spectroscopy 52 (1997) 1293–1304. [31] H. Zweifel, R.D. Maier, M. Schiller (Eds.), Plastics Additives Handbook, 6th ed., Hanser Publications, Cincinnati, OH, 2009. [32] E. Eljarrat, D. Barceló (Eds.), Brominated Flame Retardants, The handbook of environmental chemistry 16, Springer-Verlag, Berlin Heidelberg, 2011. [33] M. Alaee, R.J. Wenning, The significance of brominated flame retardants in the environment: current understanding, issues and challenges, Chemosphere 46 (2002) 579–582. [34] P.O. Darnerud, Toxic effects of brominated flame retardants in man and in wildlife, Environment International 29 (2003) 841–853. [35] L.S. Birnbaum, D.F. Staskal, Brominated flame retardants: cause for concern? Environmental Health Perspectives 112 (2004) 9–17. [36] Directive 2002/95/EC of the European Parliament and of the Council of 27 January 2003 on the restriction of the use of certain hazardous substances in electrical and electronic equipment (RoHS), Official Journal of the European Union (2003), L 37/19–L 37/23. Dissertation 116 [37] Commission Decision of 18 August 2005 amending Directive 2002/95/EC of the European Parliament and of the Council for the purpose of establishing the maximum concentration values for certain hazardous substances in electrical and electronic equipment (2005/618/EC), Official Journal of the European Union (2005), L 214/65. [38] NIST Atomic Spectra Database Lines Form, http://physics.nist.gov/PhysRefData/ASD/lines_form.html [39] M.-D. Huang, H. Becker-Ross, S. Florek, U. Heitmann, M. Okruss, High-resolution continuum source electrothermal absorption spectrometry of AlBr and CaBr for the determination of bromine, Spectrochimica Acta Part B: Atomic Spectroscopy 63 (2008) 566– 570. [40] T. Limburg, J.W. Einax, Determination of bromine using highresolution continuum source molecular absorption spectrometry in a graphite furnace, Microchemical Journal 107 (2013) 31–36. [41] M. Resano, J. Briceño, M.A. Belarra, Direct determination of phosphorus in biological samples using a solid sampling-high resolution-continuum source electrothermal spectrometer: comparison of atomic and molecular absorption spectrometry, Journal of Analytical Atomic Spectrometry 24 (2009) 1343–1354. [42] U. Heitmann, B. Welz, D.L.G. Borges, F.G. Lepri, Feasibility of peak volume, side pixel and multiple peak registration in highresolution continuum source atomic absorption spectrometry, Spectrochimica Acta Part B: Atomic Spectroscopy 62 (2007) 1222– 1230. Dissertation 117 [43] M. Resano, J. Briceño, M.A. Belarra, Direct determination of Hg in polymers by solid sampling-graphite furnace atomic absorption spectrometry. A comparison of the performance of line source and continuum source instrumentation, Spectrochimica Acta Part B: Atomic Spectroscopy 64 (2009) 520–529. [44] M. Aramendía, M.R. Flórez, M. Piette, F. Vanhaecke, M. Resano, Al determination in whole blood samples as AlF via high-resolution continuum source graphite furnace molecular absorption spectrometry: potential application to forensic diagnosis of drowning, Journal of Analytical Atomic Spectrometry 26 (2011) 1964–1973. [45] J. Briceño, M.A. Belarra, K.A.C. De Schamphelaere, S. Vanblaere, C.R. Janssen, F. Vanhaecke, M. Resano, Direct determination of Zn in individual Daphnia magna specimens by means of solid sampling high-resolution continuum source graphite furnace atomic absorption spectrometry, Journal of Analytical Atomic Spectrometry 25 (2010) 503–510. [46] M. Resano, A.C. Lapeña, M.A. Belarra, Potential of solid sampling high-resolution continuum source graphite furnace atomic absorption spectrometry to monitor the Ag body burden in individual Daphnia magna specimens exposed to Ag nanoparticles, Analytical Methods 5 (2013) 1130–1139. [47] F. Vignola, D.L.G. Borges, A.J. Curtius, B. Welz, H. Becker-Ross, Simultaneous determination of Cd and Fe in sewage sludge by highresolution continuum source electrothermal atomic absorption spectrometry with slurry sampling, Microchemical Journal 95 (2010) 333–336. Dissertation 118 [48] L.M.G. dos Santos, B. Welz, R.G.O. Araujo, S. do C. Jacob, M.G.R. Vale, A. Martens, I.B.G. Martens, H. Becker-Ross, Simultaneous Determination of Cd and Fe in Beans and Soil of Different Regions of Brazil Using High-Resolution Continuum Source Graphite Furnace Atomic Absorption Spectrometry and Direct Solid Sampling, Journal of Agricultural and Food Chemistry 57 (2009) 10089–10094. [49] L.M.G. dos Santos, R.G.O. Araujo, B. Welz, S. do C. Jacob, M.G.R. Vale, H. Becker-Ross, Simultaneous determination of Cd and Fe in grain products using direct solid sampling and highresolution continuum source electrothermal atomic absorption spectrometry, Talanta 78 (2009) 577–583. [50] B. Welz, S. Morés, E. Carasek, M.G.R. Vale, M. Okruss, H. Becker-Ross, High-Resolution Continuum Source Atomic and Molecular Absorption Spectrometry–A Review, Applied Spectroscopy Reviews 45 (2010) 327–354. [51] M. Messiaen, C.R. Janssen, O. Thas, K.A.C. De Schamphelaere, The potential for adaptation in a natural Daphnia magna population: broad and narrow-sense heritability of net reproductive rate under Cd stress at two temperatures, Ecotoxicology 21 (2012) 1899–1910. [52] M. Messiaen, C.R. Janssen, L. De Meester, K.A.C. De Schamphelaere, The initial tolerance to sub-lethal Cd exposure is the same among ten naïve pond populations of Daphnia magna, but their micro-evolutionary potential to develop resistance is very different, Aquatic Toxicology 144-145 (2013) 322–331. [53] Organisation for Economic Co-operation Development (OECD) Dissertation 119 guidelines for testing of chemicals. Guideline 211: Daphnia magna reproduction test, OECD Publication Service, Paris, 2012. [54] R. Evens, K.A.C. De Schamphelaere, C.R. Janssen, The effects of dietary nickel exposure on growth and reproduction of Daphnia magna, Aquatic Toxicology 94 (2009) 138–144. [55] J.S. Becker, State-of-the-art and progress in precise and accurate isotope ratio measurements by ICP-MS and LA-ICP-MS, Journal of Analytical Atomic Spectrometry 17 (2002) 1172–1185. [56] I. Konz, B. Fernández, M. Luisa Fernández, R. Pereiro, A. SanzMedel, Laser ablation ICP-MS for quantitative biomedical applications, Analytical and Bioanalytical Chemistry 403 (2012) 2113–2125. [57] D.S. Urgast, S. Hill, I.-S. Kwun, J.H. Beattie, H. Goenaga-Infante, J. Feldmann, Zinc isotope ratio imaging of rat brain thin sections from stable isotope tracer studies by LA-MC-ICP-MS, Metallomics 4 (2012) 1057–1063. [58] B.T.A. Muyssen, K.A.C. De Schamphelaere, C.R. Janssen, Mechanisms of chronic waterborne Zn toxicity in Daphnia magna, Aquatic Toxicology 77 (2006) 393–401. [59] F. Vanhaecke, L. Balcaen, D. Malinovsky, Use of single-collector and multi-collector ICP-mass spectrometry for isotopic analysis, Journal of Analytical Atomic Spectrometry 24 (2009) 863–886. [60] B. Hattendorf, C. Latkoczy, D. Günther, Peer Reviewed: Laser Ablation-ICPMS, Analytical Chemistry 75 (2003) 341A–347A. Dissertation 120 [61] H.P. Longerich, B.J. Fryer, D.F. Strong, Determination of lead isotope ratios by inductively coupled plasma-mass spectrometry (ICP-MS), Spectrochimica Acta Part B: Atomic Spectroscopy 42 (1987) 39–48. [62] C.N. Maréchal, P. Télouk, F. Albarède, Precise analysis of copper and zinc isotopic compositions by plasma-source mass spectrometry, Chemical Geology 156 (1999) 251–273. [63] J. Woodhead, A simple method for obtaining highly accurate Pb isotope data by MC-ICP-MS, Journal of Analytical Atomic Spectrometry 17 (2002) 1381–1385. [64] W.A. Russell, D.A. Papanastassiou, T.A. Tombrello, Ca isotope fractionation on the Earth and other solar system materials, Geochimica et Cosmochimica Acta 42 (1978) 1075–1090. [65] M. Aramendía, M. Resano, F. Vanhaecke, Isotope ratio determination by laser ablation-single collector-inductively coupled plasma-mass spectrometry. General capabilities and possibilities for improvement, Journal of Analytical Atomic Spectrometry 25 (2010) 390–404. [66] L. Moens, P. Verrept, R. Dams, U. Greb, G. Jung, B. Laser, New high-resolution inductively coupled plasma mass spectrometry technology applied for the determination of V, Fe, Cu, Zn and Ag in human serum, Journal of Analytical Atomic Spectrometry 9 (1994) 1075–1078. [67] F. Vanhaecke, L. Moens, Overcoming spectral overlap in isotopic analysis via singleand multi-collector ICP-mass spectrometry, Dissertation 121 Analytical and Bioanalytical Chemistry 378 (2004) 232–240. [68] C. O’Connor, B.L. Sharp, P. Evans, On-line additions of aqueous standards for calibration of laser ablation inductively coupled plasma mass spectrometry: theory and comparison of wet and dry plasma conditions, Journal of Analytical Atomic Spectrometry 21 (2006) 556–565. [69] S.M. Eggins, J.M.G. Shelley, Compositional Heterogeneity in NIST SRM 610-617 Glasses, Geostandards Newsletter 26 (2002) 269–286. [70] B. De Samber, R. Evens, K. De Schamphelaere, G. Silversmit, B. Masschaele, T. Schoonjans, B. Vekemans, C.R. Janssen, L. Van Hoorebeke, I. Szalóki, F. Vanhaecke, G. Falkenberg, L. Vincze, A combination of synchrotron and laboratory X-ray techniques for studying tissue-specific trace level metal distributions in Daphnia magna, Journal of Analytical Atomic Spectrometry 23 (2008) 829– 839. III. Publications Publications 125 Al determination in whole blood samples as AlF via high-resolution continuum source graphite furnace molecular absorption spectrometry: potential application to forensic diagnosis of drowning Maite Aramend! ıa,* a Mar! ıa R. Fl! orez, b Michel Piette, c Frank Vanhaecke d and Mart! ın Resano* b Received 17th June 2011, Accepted 9th August 2011 DOI: 10.1039/c1ja10183h In this work, a new methodology for direct determination of Al in whole blood samples by means of high-resolution continuum source graphite furnace molecular absorption spectrometry has been developed, based on the formation of the AlF diatomic molecule in the graphite furnace and the subsequent monitoring of its molecular absorption. The proposed methodology provides an alternative method to conventional atomic absorption, solving most of the problems related to the latter technique, particularly matrix effects, providing a straightforward alternative for blood analysis. The addition of NH 4 F$HF, which is required for promotion of the AlF molecule, was found to improve sample matrix removal for whole blood samples, whether they contain EDTA or heparin as anticoagulant agents. Besides minimizing residues in the graphite platform, this circumstance enabled the use of aqueous standards to build a calibration curve, avoiding the need for the cumbersome method of standard additions, while not affecting significantly detection capabilities (1.8 mgL !1 LOD). The method developed was also used for exploring the possibilities of Al as a chemical marker assisting forensic diagnosis of death-by-drowning. For this purpose, a set of samples (water and blood) obtained from 8 drowning suspects and two controls were analysed for their Al levels. Although additional studies with a large number of samples would be needed in order to draw definitive conclusions from a forensic point of view, a positive correlation between Al concentration in the drowning water and Al concentration in the blood of drowning suspects was found, supporting the validity of Al as a marker for drowning diagnosis. 1 Introduction Drowning is the second leading cause of death from unintentional injury, and accounts for more than half a million deaths annually worldwide. 1 However, the post-mortem diagnosis of drowning continues to be one of the most difficult in forensic pathology, especially when it comes to differentiate between ‘‘true drowned’’ bodies and corpses recovered out of water but with a different cause of death. 2 In fact, in spite of the extensive discussions about the topic found in the medico-legal literature of the last 20 years, 3,4 the ideal test for this purpose still needs to be established. 4 Besides the physiopathological alterations induced by this picture, diagnosis of death-by-drowning is usually supported by additional tests (none of them definitive), based on the fact that the water and, thus, the particles, microorganisms and chemical substances it contains, will enter the blood circulation by different processes at the moment of death. As a result, detection and/or quantification of these water components in the blood and/or certain parts of the dead bodies can help in arriving at a reliable diagnosis of drowning. The diatom test, for instance, looks for the presence of diatom algae in different organs of the body that are irrigated with water-contaminated blood during the agony, such as brain, liver or kidneys. 5 The applicability of this popular method is, however, compromised due to the gradual extermination of diatom algae in contaminated waters, and that is why finding alternative (bio)chemical markers is still a hot research topic in forensic science. Any candidate marker should fulfil three requirements: (i) the marker should have the possibility to pass the alveolo-capillary membrane to get into the blood circulation in the agony period; (ii) it should be present in large amounts in the drowning water while being absent or present in very small amounts in the blood/body of healthy persons; and (iii) there should be no a Centro Universitario de la Defensa-Academia General Militar de Zaragoza, Carretera de Huesca s/n, 50090 Zaragoza, Spain. E-mail: [email protected] b University of Zaragoza, Faculty of Sciences, Department of Analytical Chemistry, Pedro Cerbuna 12, Zaragoza, E-50009, Spain c Department of Forensic Medicine, Ghent University, J. Kluyskensstraat 29, 9000 Ghent, Belgium d Department of Analytical Chemistry, Ghent University, Krijgslaan 281S12, 9000 Ghent, Belgium 1964 | J. Anal. At. Spectrom., 2011, 26, 1964–1973 This journal is ªThe Royal Society of Chemistry 2011 Dynamic Article LinksC < JAAS Cite this: J. Anal. At. Spectrom., 2011, 26, 1964 www.rsc.org/jaas PAPER Published on 06 September 2011. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:36:15. View Article Online / Journal Homepage / Table of Contents for this issue Al determination in whole blood samples as AlF via HR CS AAS 132 possibility for evaluating and correcting for the influence of these interferences a posteriori. 3.3.2. Analysis of the blood samples for method validation. Once the working conditions had been optimized, analysis of the blood samples mentioned before was carried out. Under the working conditions described in the previous sections, very similar signal profiles for the Al aqueous standards and the blood samples could be obtained, as seen in Fig. 5. This enables the use of direct external calibration for quantification purposes, making the use of cumbersome standard addition methodology unnecessary. Although sample dilution is not strictly necessary, better repeatability can be obtained if samples are diluted 1 : 4 with ultrapure water, especially for the real samples that contain heparin or EDTA to prevent coagulation, probably due to a better interaction of the chemical modifier with the sample. For samples containing low amounts of Al, however, direct analysis of the undiluted samples is still possible and offers acceptable results. Results for the analysis of the validation samples (after 1 : 4 dilution), obtained following the experimental procedure described in Section 2.2.3, are summarized in Table 3. As seen in this table, good agreement between the values obtained with the proposed methodology and the reference values was obtained in all cases, and adequate figures in terms of result precision and accuracy were obtained for this working methodology in all of the cases. In order to check method reproducibility, the Seronorm sample was analysed in four different days following the aforementioned experimental procedure, and the final result displayed in Table 3 was calculated as the mean of these four determinations. As seen in the table, very satisfactory results were also obtained in this regard. 3.4. Analysis of the blood samples by HR-CS GFAAS and comparison with the MAS method Analysis of the same samples (Seronorm, DS-5, LS-1 and LS-2) was carried out by means of SF-ICPMS, as described in Section 2.2.4, to obtain truly independent values. Furthermore, analysis by means of HR-CS GFAAS using the atomic line at 396.152 nm was also carried out to compare the performances of the MAS and AAS techniques. The optimized temperature program used in this case is shown in Table 2. The main differences with the program used for the analysis based on the use of the molecular band lie in the atomization (or vaporization) and drying steps. In the case of the atomization/vaporization, the use of the atomic line requires higher temperatures (2500 !C) than that of the molecular band, which could result in a reduced lifetime for the graphite tubes in the former case. A three-step drying stage was also necessary in this case in order to avoid spilling of the sample at the beginning of the pyrolysis step, although the length of the stages could be somewhat reduced. As previously commented, the influence of non-spectral interferences is severe for the direct determination of Al in whole blood, and sample dilution was needed to reduce matrix effects and residue build-up in the graphite furnace. In spite of this fact, matrix effects could not be fully avoided, even for 1 : 9 dilution factors, and the use of standard additions for calibration was necessary. Implementation of this measure for dealing with matrix effects did not avoid the need for dilution, as acceptable results could only be obtained for 1 : 4 dilution factors or above. This, self-evidently, has a negative influence on the minimum Al concentration values that can be determined in blood samples by using this method. In spite of this fact, and even if the MAS method permitted the direct analysis of the undiluted blood samples, the detection capabilities of both methodologies are very similar, with LODs of 1.5 mgL "1 for the AAS method and 1.8 mgL "1 for the MAS method, respectively. In reality, sensitivity for the atomic line and the molecular band are similar, but evaluation and measurement of the molecular band introduces more noise. As a consequence, the LOD for the MAS method is poorer than that of the AAS method, although the need for dilution in the latter case compensates for this difference. Besides, matrix removal was not complete with the AAS technique and mechanical removal of the residues had to be carried out after everydetermination, which obviously increases the risk for contamination. For the rest, and as shown in Table 3, good agreement with the results obtained for the molecular band and the reference values available were also obtained in this case for the validation samples, although with poorer precision figures than those obtained with the MAS technique (as could be expected for the standard addition methodology). Considering the total analysis time, performance of the MAS technique is also better than that of the AAS technique due to the need for standard addition calibration in the latter case, which represents a significant decrease in sample throughput as detailed in Section 2.2.3. Summarizing, while both techniques offer rather similar figures of merit regarding detection capabilities, the MAS technique solves the existing problems related to matrix effects, residue build-up and calibration observed for the AAS technique, which results in a methodology less prone to contamination and much easier to use, and a remarkable improvement in sample throughput. 3.5. Suitability of the determination of Al levels in blood and water samples for diagnosis of drowning Finally, analysis of the drowning suspect samples was carried out with the AlF-MAS methodology described in the previous Fig. 5 Comparison of wavelength resolved time-integrated absorbance spectra for the vaporization of 10 mL of a 50 mgL "1 Al aqueous solution and 10 mL of a blood sample containing 46.6 mgL "1 Al and EDTA, using the working conditions summarized in Table 1. This journal is ªThe Royal Society of Chemistry 2011 J. Anal. At. Spectrom., 2011, 26, 1964–1973 | 1971 Published on 06 September 2011. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:36:15. View Article Online Publications 133 sections. A series of 10 blood samples were analysed, 8 obtained from drowning suspects and 2 from bodies recovered under different circumstances and used as controls. 8 water samples taken from the water media where the drowning suspects had been found were also analysed following the methodology described in Section 2.2.3. Results for all of these analyses are gathered in Table 4. From these data, some general ideas about the utility of the proposed methodology for drowning diagnosis can be extracted. First of all, application of the Spearman rank correlation coefficient method 44 reveals a positive correlation between the Al content in water and Al content in blood of the drowning suspects at a 95% significance level (r¼0.952; r critical ¼0.738), supporting the validity of Al as a drowning marker. Establishing an Al threshold concentration in blood for indicating a clear drowning diagnosis is, however, quite risky at this point, although some indications can be given in this regard. Firstly, and as revealed by the results obtained for samples CTRL-1, CTRL-2, DS-1, DS-7 and DS-8, Al levels up to 10–15 mgL "1 found in the blood samples can be considered as ‘‘normal’’ or ‘‘basal’’ levels. Translating this Al result into clear answers for drowning diagnosis obviously depends on the Al content found in the water media: if low Al concentration values are found in the drowning media (like for cases DS-1, DS-7 and DS-8), the validity of Al as a drowning marker is obviously inexistent. On the other hand, if high Al concentrations are found in the water, Al values below this limit should indicate a cause of death different from drowning. In this regard, determining an Al concentration threshold in the water high enough to induce a clear increase in the ‘‘basal’’ Al levels in blood is difficult, as additional factors such as, e.g., duration of the agony would certainly have a significant influence on the total amount of Al entering the bloodstream. This idea is supported by the fact that, as observed when comparing cases DS-2 and DS-4, a higher Al level in the water media does not immediately result in a higher Al level in the blood of the suspect. However, it is also true that for all water samples showing 200 mg Al per litre or higher, an Al blood level at least twice as high as the ‘‘basal’’ level has been found. From all of the above, it seems clear that the method proposed has some potential as a valid marker for drowning diagnosis, although an extensive study investigating a large number of suspects and controls would be needed in order to extract clearer conclusions from a forensic point of view. Carrying out this study is probably worth trying, considering the few diagnostic tools that forensic doctors have at their disposal when dealing with death by drowning. 4 Conclusions The use of HR-CS GFMAS for determination of Al in blood samples via formation of the AlF molecule in the graphite furnace and monitoring of its molecular absorption provides an alternative method to conventional atomic absorption, resolving most of the problems related to the latter technique. Addition of NH 4 F$HF required for promotion of the AlF molecule has the advantage of improving sample matrix removal and minimizing residues in the graphite platform that, otherwise, need to be removed mechanically increasing the risk for contamination. As a result, with this method it is possible to construct the calibration curve with aqueous standards, avoiding the need for the cumbersome method of standard additions, while detection capabilities are essentially maintained (1.8 mgL "1 vs. 1.5 mgL "1 LODs for the MAS and AAS methods, respectively). Application of this cost-effective and relatively easy to use methodology for Al determination in the blood of drowning suspects, on the other hand, has shown some potential to be used as a valid marker for drowning diagnosis, as a positive correlation between Al concentration in drowning water and Al concentration in the blood of drowning suspects seems to be present. Additional studies with a large number of corpses and controls would be, however, needed in order to draw more definitive conclusions from a forensic point of view. Acknowledgements This work has been funded by the Spanish Ministry of Science and Innovation (Project CTQ2009-08606) and the Aragon Government (Departamento de Ciencia, Tecnolog! ıa y Universidad del Gobierno de Arag! on y Fondo Social Europeo and Fundaci! on ARAID and ‘‘Obra Social de IberCaja’’, with additional support from Inycom). Maite Aramend! ıa thanks the FWO-Vlaanderen for her postdoctoral grant. Mar! ıa del Rosario Fl! orez thanks the Spanish Ministry of Science and Innovation for her doctoral grant. Table 3 Comparison of the reference and determined Al concentration values for the validation samples. SF-ICPMS, AAS and MAS results were obtained following the experimental procedures included in Sections 2.2.3 and 2.2.4. Uncertainties are expressed as 95% confidence intervals Reference/ mgL "1 SF-ICPMS/ mgL "1 (n¼3) AAS/ mgL "1 (n¼5) MAS/ mgL "1 (n¼5) Seronorm 58.9 #9.9 60.1 #6.1 56.9 #8.1 57.5 #3.4 a LS-1 — 46.8 #9.2 46.6 #5.7 45.0 #4.6 LS-2 — 96.5 #10.4 100.3 #9.2 96.4 #7.6 DS-5 — 23.7 #6.0 24.6 #8.8 24.3 #5.3 a This result was calculated as the mean of four different determinations carried out in four different days. Table 4 Al concentration levels found in the blood and drowning water of eight drowning suspects collected in the area of Gent (DS-1 to DS-8) plus two control corpses collected in the same geographical area under other circumstances different from drowning (CTRL-1 and -2). Uncertainties are expressed as 95% confidence intervals (3 and 5 replicates were measured for water and blood samples, respectively) Al content in water/mgL "1 Al content in blood/mgL "1 DS-1 2.31 #0.24 4.32 #0.82 DS-2 412 #14 94.3 #6.8 DS-3 296 #12 37.1 #3.7 DS-4 483 #18 38.9 #3.1 DS-5 195 #11 24.3 #5.3 DS-6 505 #14 133 #9 DS-7 14.9 #1.0 12.9 #2.1 DS-8 30.1 #1.7 11.9 #1.7 CTRL-1 — 6.5 #1.0 CTRL-2 — 10.5 #1.8 1972 | J. Anal. At. Spectrom., 2011, 26, 1964–1973 This journal is ªThe Royal Society of Chemistry 2011 Published on 06 September 2011. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:36:15. View Article Online Al determination in whole blood samples as AlF via HR CS AAS 134 Notes and references 1 M. D. P! erez-C! arceles, A. Sib! on, M. L. Gil del Castillo, M. A. Vizcaya, E. Osuna, T. Casas, J. L. Romero and A. Luna, Biol. Trace Elem. Res., 2008, 126, 27–37. 2 M. H. A. Piette and E. A. De Letter, Forensic Sci. Int., 2006, 163, 1–9. 3 J. Timperman, Forensic Sci., 1972, 1, 397–409. 4 B. Ludes and P. Fornes, Drowning, in Forensic Medicine: Clinical and Pathological Aspects, ed. J. Payne-James, A. Busuttil and W. Smock, Greenwich Medical Media Ltd., London, San Francisco, 2003, pp. 247–257. 5 J. H€ urlimann, P. Feer, F. Elber, K. Niederberger, R. Dirnhofer and D. Wyler, Int. J. Leg. Med., 2000, 114, 6–14. 6 C. Yu-Chuan, D. Zhao-Ke and Z. Jia-Zhen, Forensic Sci. Int., 1990, 46, 289–294. 7 B.-L. Zhu, K. Ishida, M. Taniguchi, L. Quan, S. Oritani, K. Tsuda, Y. Kamikodai, M. Q. Fujita and H. Maeda, Leg. Med., 2003, 5, S298–S301. 8 M. Piette, B. Desmet and R. Dams, Sci. Total Environ., 1994, 141, 269–273. 9 J. E. Azparren, E. Perucha, P. Mart! ınez, R. Mu~ noz and G. Vallejo, Forensic Sci. Int., 2007, 168, 138–142. 10 J. E. Azparren, A. Fernandez-Rodriguez and G. Vallejo, Forensic Sci. Int., 2003, 137, 55–59. 11 K. G. Ljunggren, V. Lidums and B. Sj€ ogren, Br. J. Ind. Med., 1991, 48, 106–109. 12 J.-P. Goull! e, L. Mahieu, J. Castermant, N. Neveu, L. Bonneau, G. Lain! e, D. Bouige and C. Lacroix, Forensic Sci. Int., 2005, 153, 39–44. 13 C. T. Driscoll and W. D. Schecher, Environ. Geochem. Health, 1990, 12, 28–49. 14 E. M. Skelly and F. T. DiStefano, Appl. Spectrosc., 1988, 42, 1302– 1306. 15 I. Rodushkin, E. Engstr€ om and D. C. Baxter, Anal. Bioanal. Chem., 2010, 396, 365–377. 16 S. Salomon, P. Giamarchi, A. Le Bihan, H. Becker-Ross and U. Heitmann, Spectrochim. Acta, Part B, 2000, 55, 1337– 1350. 17 S. Tang, P. J. Parsons and W. Slavin, Spectrochim. Acta, Part B, 1997, 52, 1351–1365. 18 D. Bohrer, U. Heitmann, M. Huang, H. Becker-Ross, S. Florek, B. Welz and D. Bertagnolli, Spectrochim. Acta, Part B, 2007, 62, 1012–1018. 19 I. Rodushkin, F. € Odman and S. Branth, Fresenius J. Anal. Chem., 1999, 364, 338–346. 20 P. C. D’Haese, F. L. Van de Vyver, F. A. de Wolff and M. E. De Broe, Clin. Chem. (Washington, D. C.), 1985, 31, 24–29. 21 W. Frech, A. Cedergren, C. Cederberg and J. Vessman, Clin. Chem. (Washington, D. C.), 1982, 28, 2259–2263. 22 B. Welz, H. Becker-Ross, S. Florek and U. Heitmann, High Resolution Continuum Source AAS-the Better Way to do Atomic Absorption Spectrometry, Wiley-VCH, Weinheim, 2005. 23 B. Welz, Anal. Bioanal. Chem., 2005, 381, 69–71. 24 B. Welz, D. L. G. Borges, F. G. Lepri, M. G. R. Vale and U. Heitmann, Spectrochim. Acta, Part B, 2007, 62, 873–883. 25 M. Resano, F. Vanhaecke and M. T. C. de Loos-Vollebregt, J. Anal. At. Spectrom., 2008, 23, 1450–1475. 26 H. Becker-Ross, S. Florek and U. Heitmann, J. Anal. At. Spectrom., 2000, 15, 137–142. 27 M. Resano and E. Garc! ıa-Ruiz, Anal. Bioanal. Chem., 2011, 399, 323– 330. 28 M. Resano, L. Rello, M. Fl! orez and M. A. Belarra, Spectrochim. Acta, Part B, 2011, 66, 321–328. 29 B. Welz, S. Mor! es, E. Carasek, M. G. R. Vale, M. Okruss and H. Becker-Ross, Appl. Spectrosc. Rev., 2010, 45, 327–354. 30 B. Welz, F. G. Lepri, R. G. O. Araujo, S. L. C. Ferreira, M. D. Huang, M. Okruss and H. Becker-Ross, Anal. Chim. Acta, 2009, 647, 137–148. 31 U. Heitmann, M. Sch€ utz, H. Becker-Ross and S. Florek, Spectrochim. Acta, Part B, 1996, 51, 1095–1105. 32 M. G. R. Vale, N. Oleszczuk and W. N. L. dos Santos, Appl. Spectrosc. Rev., 2006, 41, 377–400. 33 CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC Press, Boca Raton, 74th edn, 1994. 34 M. Resano, M. Aramend! ıa and F. Vanhaecke, J. Anal. At. Spectrom., 2006, 21, 1036–1044. 35 U. Heitmann, B. Welz, D. L. G. Borges and F. G. Lepri, Spectrochim. Acta, Part B, 2007, 62, 1222–1230. 36 M. Resano, J. Brice~ no and M. A. Belarra, Spectrochim. Acta, Part B, 2009, 64, 520–529. 37 M. Resano, J. Brice~ no and M. A. Belarra, J. Anal. At. Spectrom., 2009, 24, 1343–1354. 38 J. Brice~ no, M. A. Belarra, K. A. C. De Schamphelaere, S. Vanblaere, C. R. Janssen, F. Vanhaecke and M. Resano, J. Anal. At. Spectrom., 2010, 25, 503–510. 39 K. Tsunoda, K. Fujiwara and K. Fuwa, Anal. Chem., 1977, 49, 2035– 2039. 40 M. D. Huang, H. Becker-Ross, S. Florek, U. Heitmann and M. Okruss, Spectrochim. Acta, Part B, 2006, 61, 572–578. 41 Plasus SpecLine Database, version 2.1, Plasus, K€ oningsbrunn, Germany. 42 M. Resano, E. Mozas, C. Crespo, J. Brice~ no, J. del Campo Menoyo and M. A. Belarra, J. Anal. At. Spectrom., 2010, 25, 1864–1873. 43 R. G. O. Araujo, B. Welz, F. Vignola and H. Becker-Ross, Talanta, 2009, 80, 846–852. 44 J. N. Miller and J. C. Miller, Statistics and Chemometrics for Analytical Chemistry, Pearson Education Limited, Essex, England, 5th edn, 2005. This journal is ªThe Royal Society of Chemistry 2011 J. Anal. At. Spectrom., 2011, 26, 1964–1973 | 1973 Published on 06 September 2011. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:36:15. View Article Online Publications 135 Direct determination of sulfur in solid samples by means of high-resolution continuum source graphite furnace molecular absorption spectrometry using palladium nanoparticles as chemical modifier Mart! ın Resano*and Mar! ıa R. Fl! orez Received 4th November 2011, Accepted 16th December 2011 DOI: 10.1039/c2ja10322b This work investigates the potential benefits of using Pd nanoparticles, in combination with Ru as permanent modifier, for sulfur monitoring as CS by means of high-resolution continuum source graphite furnace molecular absorption spectrometry. Upon heating, these small (approx. 20 nm diameter) Pd nanoparticles are evenly distributed over the platform surface, offering a larger surface for interaction with the analyte during the drying and pyrolysis steps. In this way, a more efficient stabilization of sulfur species can be achieved. Furthermore, a similar analytical response is obtained regardless of the chemical form in which sulfur species are originally found, thus making it easier to develop quantitative analytical methods for this analyte. When using these modifiers, and under optimized working conditions, it is possible to use this technique for direct analysis of different types of solid samples (biological, petroleum coke, polyethylene and steel CRMs), thus circumventing the traditional drawbacks associated with sample digestion when sulfur determination is aimed at. Accurate results are obtained using only the central CS line found at 257.958 nm, with precision values in the 5–10% RSD range (14 ng characteristic mass; 9 ng limit of detection). Moreover, the combined use of the main six CS lines available in the spectral area simultaneously monitored by the detector permits to further improve precision to 3–5% RSD, while decreasing the limit of detection down to 3 ng (the characteristic mass is also 3 ng), which represents a relative limit of detection of approx. 1 mgg !1 , as calculated for the sample with the lowest sulfur content. In all cases, straightforward calibration with aqueous standards was proved to be feasible, and the sample throughput is of 3–4 samples per hour (5 replicates per sample). 1. Introduction Sulfur is an abundant element on earth’s crust (0.044%) and can be considered as one of the main building blocks for life in our planet, owing to its presence in proteins. Furthermore, sulfur, in different chemical forms, is an important raw material for a wide variety of industries and products, being currently used in fertilizers (mainly as sulfate), fungicide and pesticide formulations (as elemental sulfur), pharmaceuticals (organic sulfur), as an additive of foodstuffs (as SO 2 ), among many other applications. 1 Given the importance of this element, and considering its extensive use in many fields, the development of fast and straightforward methods for its determination in various types of samples is of great interest. However, for various reasons, such as lack of atomic lines in the UV-visible region when opting for atomic absorption or atomic emission spectrometry (AAS or AES), 2 or spectral overlaps when using inductively coupled plasma mass spectrometry (ICP-MS), 3 sulfur is not simple to determine at trace levels even with the most sensitive of the atomic spectrometry techniques. Still today, cumbersome and not very selective spectrophotometric methods are widely used, in addition to the so-called dedicated analyzers, 4 which are costly and often specific for particular types of samples. Regarding the use of graphite furnace (GF) atomic absorption spectrometers, despite the lack of suitable atomic lines in the spectral region that is typically accessible, the determination of sulfur was already evaluated in the 80s based on the formation of different sulfide species (e.g., GeS or CS) and the monitoring of their molecular absorption using a hollow cathode lamp (HCL) of another element that emitted light in the spectral area of interest, as reported by Dittrich and Vorberg 5 and Tittarelli and Lavorato. 6 These early applications are reviewed in a recent work. 7 Obviously, the performance of such technique could not be very satisfactory, owing to the poor suitability of the HCL source and the low resolution of ‘‘classic’’ AAS, which resulted in insurmountable problems when dealing with molecular sharp lines showing structured background. This situation has changed very significantly upon arrival of high-resolution continuum source (HR CS) GFAAS University of Zaragoza, Faculty of Sciences, Department of Analytical Chemistry, Pedro Cerbuna 12, Zaragoza, E-50009, Spain. E-mail: [email protected] This journal is ªThe Royal Society of Chemistry 2012 J. Anal. At. Spectrom., 2012, 27, 401–412 | 401 Dynamic Article LinksC < JAAS Cite this: J. Anal. At. Spectrom., 2012, 27, 401 www.rsc.org/jaas PAPER Published on 26 January 2012. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:37:13. View Article Online / Journal Homepage / Table of Contents for this issue Direct determination of S in solid samples by means of HR CS GFMAS 136 instrumentation, which has opened up new possibilities in this field. 8–12 In addition to offering several advantages for the measurement of atomic absorption, this type of instrumentation permits the reliable quantitative monitoring of the molecular absorption of diatomic molecules, owing to the continuous nature and the high intensity of the radiation source (Xe lamp), and the excellent spectral resolution provided (better than 2 pm at 200 nm). In this way, new methods for the determination of sulfur, phosphorus and the halogens using graphite furnace molecular absorption spectrometry (GFMAS) have been proposed, as reviewed by Welz et al. 7 This GFMAS technique can also offer some advantages for the determination of metals, in some particular situations. 13 However, despite these new possibilities, one of the main reasons to opt for the use of graphite furnace-based techniques is the potential to carry out direct analysis of solid samples, thus improving sensitivity and expeditiousness. This aspect is even more relevant in the case of sulfur determination, as the ubiquity of the element, the high volatility of many of its compounds, and the obvious impossibility to use sulfuric acid as reagent to help dissolve the sample, make it difficult to obtain reliable results when digestion approaches are attempted. In particular, the potential for suffering from analyte losses can be very serious, and it is not surprising that protocols for sulfur determination often include many steps (e.g., reduction to hydrogen sulfide, trapping of the latter in a solution by precipitation, and redissolution of the formed compound) to minimize this risk, 14 at the cost of sample throughput. The potential of GFMAS has, however, been very seldom evaluated for direct analysis of solid samples. The only two works reporting on this topic so far have investigated the determination of phosphorus 15 and, precisely, of sulfur, 16 in biological samples, in both cases. There may be reasons for this limited number of works, as the process of formation of suitable diatomic molecules from a solid sample is, a priori, more complex than the process of atomization. Moreover, a review of the few works devoted to sulfur monitoring by GFMAS 17 clearly indicates remaining problems for achieving proper stabilization of this element, as well as for obtaining an analytical signal that is independent of the chemical form in which the analyte is found. It is clear that more work is required to expand the field of applications for this element, as will be discussed in Section 3. The aim of this work is to further investigate the potential of GFMAS for sulfur monitoring, evaluating more efficient ways to add the chemical modifier (e.g., Pd as nanoparticles), in an attempt to develop suitable methods for the direct determination of sulfur in very different types of solid samples, namely biological materials, petroleum coke, polyethylene and steel. The ultimate goal is always to develop simple procedures relying only on the use of aqueous standards for calibration. 2. Experimental 2.1. Instrumentation All the experiments in this work were carried out using a HR CS AAS, ContrAA 700, commercially available from Analytik Jena AG (Jena, Germany) and equipped with both graphite furnace and flame atomizers. The optical system comprises a xenon short-arc lamp (GLE, Berlin, Germany) operating in ‘‘hot-spot’’ mode as the radiation source, a high-resolution double echelle monochromator (DEMON) and a linear CCD array detector with 588 pixels, 200 of which are used for analytical purposes (monitoring of the analytical signal and BG correction), while the rest are used for internal functions, such as correcting for fluctuations in the lamp intensity. More details on this type of instrumentation can be found elsewhere. 18,19 This HR CS AAS instrument is equipped with a transversely heated graphite tube atomizer and both solid sampling and liquid sampling autosamplers. The solid sampling device (SSA 600) incorporates a microbalance with a readability of 1 mg. 20 Pyrolytic graphite tubes were used in all experiments. Solid samples were introduced using solid sampling graphite platforms. A transmission electron microscope JEOL (Tokyo, Japan) JEM 2000 FX II was used for estimating the size of the nanoparticles, and a Scanning Electron Microscope JEOL JSM-6400, configured with an energy dispersive X-ray analyzer INCA 300 X-Sight from Oxford Instruments (Abingdon, UK), was used for examination of the graphite platforms. 2.2. Reagents and standards 2.2.1. Reagents and chemical modifiers. Purified water was obtained from a milli-Q system (Millipore, Billerica, USA). Pd, Ru and Zr solutions were prepared from commercially available 1gL !1 single-element standards (Merck, Darmstadt, Germany), by appropriate dilution with 0.14 mol L !1 HNO 3 . Citric acid (3 g in 100 mL) and Na 2 WO 4 2H 2 O (7.8 g in 100 mL) were also prepared from the solid reagents (Merck) dissolved in milli-Q water. 14 mol L !1 HNO 3 was purchased from Merck. All the reagents were of analytical grade purity or higher. Pd nanoparticles were prepared as follows: 11 mL of a 10 g L !1 standard solution of Pd (as Pd(NO 3 ) 2 ) were mixed with 20 mL of an aqueous solution of polyvinylpyrrolidone (PVP, M w z 58 000, Alfa Aesar, Karlsruhe, Germany), which was prepared dissolving 264 mg of PVP in 100 mL of milli-Q water, and with 25 mL of ethanol (grade 90%, Alfa Aesar, Karlsruhe). The mixture was stirred and allowed to reflux for 3 hours under air resulting in a dark brown solution. This procedure is described in more detail elsewhere. 21 2.2.2. Standards. S aqueous solutions (100 mg L !1 ) were prepared by dissolving in milli-Q water suitable amounts of Na 2 SO 4 , Na 2 SO 3 , thiourea (Panreac, Barcelona, Spain) and 3mercaptopropionic acid (Alfa Aesar, Karlsruhe, Germany). A 1 gL !1 S solution (S as NH 4 SO 4 ) was also purchased from SPEX CertiPrep (Metuchen, USA). All the reagents were of analytical grade purity or higher. 2.3. Samples In order to assess the validity of the method developed, four different certified reference materials (CRMs) were used for analysis. ERM-EC680 and ERM-EC681 are polyethylene CRMs produced by BCR (Community Bureau of Reference, Geel, Belgium). These samples are available in granular form, with each granule weighing 8–13 mg. CRM Mild Steel 453 and Highspeed steel 482 were available from the Bureau of Analysed 402 | J. Anal. At. Spectrom., 2012, 27, 401–412 This journal is ªThe Royal Society of Chemistry 2012 Published on 26 January 2012. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:37:13. View Article Online Publications 137 Samples Ltd. (BAS, Newby, UK). These samples are available in small chips, with each chip weighing 10–13 mg. The rest of the CRMs investigated were available as fine powders: BCR CRM 129 Elements in Hay powder, as well as NIST (National Institute of Standards and Technology, Gaithersburg, USA) Standard Reference Materials 1566a Oyster tissue and 2718 Green petroleum coke. 2.4. Procedure for determination of S by solid sampling HR CS GFMAS 2.4.1. Impregnation of the tube and platform with Ru. Thermal deposition was chosen as the method for coating the sample platforms with Ruthenium as a permanent modifier, following the procedure described by Vale et al. 22 In this procedure, 10 repetitive injections of 40 mLofa1gL !1 standard solution of Ru were carried out, and after each injection the temperature program shown in Table 1 was applied. A single extra injection, followed by the full temperature program used for the impregnation, was also applied after every run of 20 sample replicates, in order to avoid coating loses that may result in sensitivity variations during analyses. 2.4.2. Sample analysis. No sample pretreatment was needed, other than cutting the samples into fragments of appropriate mass (see Table 2) using a ceramic knife, when necessary (polyethylene and steel materials). The solid sampling device used allows for automatic weighing and transport of the samples into the furnace. The empty platform was first transported to the microbalance using a pair of tweezers. After taring, an appropriate amount of sample was deposited onto the platform and weighed. The corresponding amount of the chemical modifier was added afterwards. The platform was then transferred to the graphite furnace and subsequently subjected to the temperature program. All these operations were fully controlled by the computer except for the deposition of the sample and the modifier, which were carried out manually. The operating conditions are summarized in Table 2. For every line evaluated, the values obtained for three detector pixels (the central pixel plus the adjacent ones, CP "1), corresponding to a spectral interval of 4.3 pm, were summed. The calibration was carried out using 10 mL of an S aqueous solution of the appropriate concentration and 20 mL of the Pd nanoparticle suspension as modifier (40 mg of Pd), deposited with a micropipette onto the solid sampling platform. For every determination, five solid samples were analyzed and the median of the five results was taken as representative value. 23 Integrated absorbance (A int ) was selected as the measurement mode in all the circumstances. 3. Results and discussion 3.1. Sulfur monitoring by means of graphite furnace molecular absorption spectrometry 3.1.1. Wavelength selection. As discussed in the Introduction, the use of HR CS AAS instrumentation, either with a flame atomizer or with a graphite furnace, also permits the monitoring of molecular lines, opening new possibilities for sulfur determination. In particular, of the different options available (e.g., CS, SH, S 2 ,etc.), there seems to be an agreement in the literature on the superior performance obtained when monitoring the spectrum of the CS molecule around 258.0 nm, which corresponds with the Dy¼0 vibrational sequence of the electronic transition X 1 S + /A 1 P. 17 This spectrum shows many sharp rotational lines, of a width similar to that typically observed for atomic lines (a few picometers), which are characterized by similar sensitivity, even though the most sensitive ‘‘lines’’ have been reported to be found exactly at 257.593 nm, 257.958 nm and 258.056 nm. 24–29 In addition to the good sensitivity, selection of CS is also preferable owing to the high stability of the CS bond (714.1 kJ mol !1 ), which makes formation of this molecule to be highly favored in a graphite furnace, 17,25 where carbon is obviously highly available. Of these lines mentioned above, different authors have used several of them in their works. 16,17,27 In our case, centering the spectra around the strong line of 257.958 nm was decided, since potential spectral overlaps with an Fe line have been reported for the other most sensitive line of 258.056 nm, 16 which may hamper its use, in particular, for analysis of steel, as will be discussed later Table 1 Temperature program used for the thermal deposition of Ru as a permanent modifier in the solid sample platforms. In total, 10 injections of modifier solution were carried out. However, after each of the first 9 injections, only stages from 1 to 5 were applied. The complete temperature program was only applied after the last injection 22 Stage Temperature/$C Ramp/ $Cs !1 Hold time/s Gas flow rate/L min !1 (1) Drying 100 7 10 2.0 (2) Drying 130 7 40 2.0 (3) Drying 160 100 60 2.0 (4) Pyrolysis 1000 100 20 2.0 (5) Pyrolysis 1400 100 5 0.1 (6) Pyrolysis 2000 100 5 0.1 Table 2 Instrumental parameters used in the determination of sulfur by means of HR CS GFMAS Electronic transition used Dy¼0, X 1 S + /A 1 Pof CS molecule Wavelength/nm Main lines of CS in the vicinity of 258 nm (central pixel ¼257.958 nm) Number of detector pixels summed per line 3(%4.3 pm) Ar gas flow during atomization/mL min !1 Stop-flow conditions Chemical modifiers Ru (permanent modifier) + Pd (40 mg) nanoparticles Sample mass/mg 0.5–2.0 (BAS 482, BAS 453) 0.1–0.3 (BCR 129, NIST 1566a) 0.3–1.0 (ERM-EC680) 1.5–3.5 (ERM-EC681) 0.04–0.05 (NIST 2718) Temperature program Step Temperature/$C Ramp/$Cs !1 Hold time/s Drying 110 2 10 Pyrolysis 800 50 20 Auto Zero 800 0 5 Vaporization 2400 2000 5 Cleaning 2600 1000 5 This journal is ªThe Royal Society of Chemistry 2012 J. Anal. At. Spectrom., 2012, 27, 401–412 | 403 Published on 26 January 2012. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:37:13. View Article Online Direct determination of S in solid samples by means of HR CS GFMAS 138 on. Besides, better baseline stability was obtained in this spectral area in our working conditions. An example of the spectrum obtained for vaporization of a sulfur solution with the instrument used in this work is shown in Fig. 1a, where this sharp rotational structure mentioned before can be clearly appreciated. Fig. 1b shows the temporal evolution of the signal, which, as expected, is exactly the same for all the ‘‘lines’’ monitored, confirming its correspondence with the same CS molecule. Finally, it can also be mentioned that a recent work published by Jim et al., using a home-made continuum source AAS instrument (characterized by lower resolution but a wider spectral window in comparison with the instrument used in the current work), has preferred the monitoring of a broad absorption band, not reported before and located between 200 and 210 nm, for S determination in coal using the slurry technique. 30 The attribution of this band to a particular species is not evident, but it is suggested to be related with both C and S, and its signal is strengthened by the fact that a filter furnace atomizer (offering a porous graphite structure and, thus, a large C surface for C and S interaction) was used in that work. Since the filter furnace device does not really permit direct solid sampling analysis, this approach will not be investigated in the present work. 3.1.2. Selection of chemical modifier. While regarding the selection of the best molecule for sulfur monitoring and of the spectral area of most interest there is a good agreement among researchers, the same can hardly be said concerning the use of chemical modifier(s), as discussed in more detail by Kowalewska. 17 If we focus on the three only works reporting on CS monitoring by HR CS GFMAS to date, Heitmann et al. recommended the use of methane during pyrolysis and addition of Ca for S stabilization, 27 while Ferreira et al. did not see any change when adding methane and demonstrated superior performance of Pd in comparison with Ca, 16 particularly when analyzing solid samples. In both cases, however, addition of Zr as permanent modifier was recommended. Ferreira et al. stated that this could be possibly explained because otherwise sulfur species that are in direct contact with graphite might be intercalated in its layered structure, ultimately not giving rise to CS but to the very volatile CS 2 . 16 Finally, Kowalewska opted for the combination, Pd + Mg in organic forms, for analysis of petroleum. 17 The conclusion of these previous works is that the level of stabilization obtained is still not sufficient. In fact, a very dissimilar performance was reported for different sulfur species introduced as aqueous standards, by Ferreira et al. 16 Furthermore, Kowalewska reported that light petroleum products containing volatile sulfur compounds could not be accurately analyzed. 17 It appears that, while the arrival of HR CS AAS has indeed made it feasible to accurately monitor the sharp rotational structure of CS electronic transitions opening new ways for S determination, achieving proper stabilization of sulfur in a graphite furnace regardless of the chemical form in which sulfur is present still remains a challenge, as already discussed in early works with line source MAS. 31,32 This is a situation that can only be solved by using more effective chemical modifiers. This aspect was investigated in more detail in the current work. In agreement with the works discussed before, we also saw the benefit of using a permanent modifier to obtain more sensitivity. Zr, W and Ru were tested for this purpose, and best results were actually obtained with Ru, which is a typical permanent modifier, 33–35 but was not reported before for S monitoring. The difference of using Ru compared with the use of W or Zr is not very significant, but approximately 15–20% more sensitivity was obtained with Ru, and peaks were better defined, showing 30– 40% larger peak heights and, thus, less tailing. Ru was therefore used in all further experiments. In addition to permanent modification with Ru, Pd was also tested as chemical modifier. This combination of a permanent modifier plus Pd has provided satisfactory results in similar situations. 15 Moreover, previous results of sulfur monitoring based on graphite furnace techniques (electrothermal vaporization coupled to ICP-MS 36 and ICP-OES 37 ) have demonstrated a strong tendency for Pd to interact with sulfur. This aspect may be critical in the current situation in which the goal is to monitor diatomic species, which is somewhat more complicated than traditional atomic measurements. The basic idea is to add a modifier that may attract sulfur, thus helping in breaking the original sulfur bonds early in the process, and that stabilizes sulfur long enough until it can be released under furnace Fig. 1 (A) Wavelength resolved time-integrated absorbance spectrum obtained after the vaporization of 500 ng of sulfur (as Na 2 SO 4 ) recorded when monitoring the spectral region in the vicinity of 258 nm by means of high-resolution continuum source graphite furnace molecular absorption spectrometry under the conditions described in Table 2. (B) Time and wavelength resolved absorbance spectrum obtained after the vaporization of 500 ng of sulfur (as Na 2 SO 4 ) in the same conditions as (A). 404 | J. Anal. At. Spectrom., 2012, 27, 401–412 This journal is ªThe Royal Society of Chemistry 2012 Published on 26 January 2012. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:37:13. View Article Online Publications 139 conditions that are favorable for the formation of the CS molecule. In this way, not only pyrolysis temperatures high enough to remove most of matrix components can be used, but also the analytical response may be normalized regardless of the chemical form in which sulfur is introduced. Still, for optimum performance, the chemical form in which Pd is added may be critical, as well as adding an amount sufficient for efficient interaction (at least a 5/1 Pd S !1 molar ratio, according to our results). To compare the different responses obtained as a function of the way in which Pd is added, pyrolysis curves for four different sulfur species (sodium sulfite, sodium sulfate, thiourea and 3-mercaptopropionic acid) were monitored using Ru as chemical modifier, and adding the same amount of Pd (40 mg) in different ways. The results obtained are shown in Fig. 2. As expected, addition of Pd as nitrate does not seem to be the most efficient approach. As can be seen in Fig. 2a, while this form seems to stabilize well the inorganic species (which are less volatile: Na 2 SO 4 boiling point is 1429 "C, while Na 2 SO 3 decomposes at 593 "C) up to 1200 "C, it does not offer the same performance for the more volatile organic species (thiourea decomposes at 182 "C, while the boiling point of 3-mercaptopropionic acid is 110 "C), for which a lower signal is clearly obtained even at low pyrolysis values. It is well known that reduction of Pd can help in achieving a better performance, since in that way metallic Pd is available earlier in the process, and that form is believed to be the most active one. 15,38–40 This reduction can be traditionally achieved either thermally (preheating a Pd solution) or by addition of a reductant, such as citric acid. As shown in Fig. 2b and c, the situation improves when using such approaches. In particular, the addition of citric acid seems very promising, as only the signal for the most volatile sulfur species is slightly reduced compared with the rest. In this particular case, it could be hypothesized that the addition of citric acid might not only help in producing metallic Pd earlier, but it might also release some active carbon 41 during the pyrolysis, thus further helping in the formation of CS. 16 However, the use of citric acid in combination with thermally prereduced Pd did not provide any improvement in the performance of this modifier, thus rendering this hypothesis as very unlikely. There is however another possibility for providing metallic Pd that could be of interest in this context, and that is the addition of Pd nanoparticles. Contrary to the current situation in many scientific fields, the use of nanoparticles as modifiers in graphite furnace techniques has been reported very seldom. Besides, the nanoparticles investigated to date (naked Ag 42 and Au 43 nanoparticles) correspond with elements that are not considered as the most efficient modifiers, and were probably chosen mainly because their nanoparticles are widely available. Thus, no significant advantages have been observed in those works, other that the low blanks achieved, since in terms of analyte stabilization the values reported (e.g., maximum pyrolysis temperature without losses) are comparable with those found for the most usual modifiers. There are, however, reasons to believe that the use of Pd nanoparticles may be more promising in the current case. Pd nanoparticles can be easily produced and stabilized in the presence of a suitable agent (for instance, polyvinylpyrrolidone) in hydroalcoholic media, and, thus, can be directly added in Fig. 2 Pyrolysis curves obtained for 500 ng of sulfur introduced in different chemical forms when monitoring the 257.958 nm CS molecular line (sum of 3 central pixels) in the presence of Ru as permanent modifier plus 40 mg of Pd, added in different forms: (A) as Pd(NO 3 ) 2 ; (B) as Pd(NO 3 ) 2 , but thermally prereduced using a prior 1000 "C pyrolysis step; (C) as Pd(NO 3 ) 2 together with an excess of citric acid; (D) as Pd nanoparticles, produced as described in Section 2.2.1. This journal is ªThe Royal Society of Chemistry 2012 J. Anal. At. Spectrom., 2012, 27, 401–412 | 405 Published on 26 January 2012. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:37:13. View Article Online Direct determination of S in solid samples by means of HR CS GFMAS 140 solution (although the term suspension is more proper). The concept behind their use is similar to that of colloidal Pd, introduced by Volynsky and Krivan, 44,45 except for the particle size, which is smaller in the current case. This is a critical aspect, because the smaller the particle size, the higher the surface that is available for interaction with the analyte, as demonstrated by the improved performance obtained when using nanoparticles as catalysts in synthesis processes. 46,47 To sum up, a priori, the addition of Pd nanoparticles should provide a better performance because metallic and highly reactive Pd nanoparticles would be available for interaction already during the drying step. This approach was tested in the current work. Pd nanoparticles of approx. 20 nm size (see Fig. 3), as evaluated by TEM monitoring, were synthetized as described in Section 2.2.1., 21 and 20 mL of this suspension (55/45 water/ethanol mixture) were added together with the sulfur aqueous standard, in the same way as it would be done with any Pd aqueous solution. As shown in Fig. 2d, this approach provided the best results, permitting to obtain very similar responses for all the species investigated up to pyrolysis of 1000 !C, which should be sufficient for analysis of many types of samples. Thus, these nanoparticles were used as chemical modifiers for further experiments. In order to further explore the reasons behind the improved performance of Pd when added as nanoparticles, examination of several graphite platforms after heating at various temperatures was carried out using Scanning Electron Microscopy (SEM). First of all, it was observed that Ru particles are randomly distributed across the platform, in the form of large agglomerates (approx. 0.5–1 mm length). Concerning Pd, when added as Pd (NO 3 ) 2 , it was confirmed that it is not evenly distributed over the platform surface. It has already been described in the literature that this modifier tends to migrate towards the edges of the platform either during drying or pyrolysis. 48,49 In this case, it could be seen already after drying at 200 !C that Pd it is mostly found in the walls of the concave platform (boat) used in this work. In addition to this, in the area in which Pd is present, it appears as a thin layer covering the graphite (see Fig. 4b), a layer that shows some cracks that increase with higher temperatures. As stated before, these aspects are already well-known. In fact, different ways to improve the homogeneity of Pd distribution have been discussed in the literature. The use of ascorbic acid or other reductants may help, but still the distribution of Pd agglomerates has been reported to be uneven. 38 Addition of Mg (NO 3 ) 2 (ref. 48) or electrodeposition 49,50 have been reported to be more successful concerning the homogeneity of the Pd distribution, but Pd is still found in the form of relatively large agglomerates of approx. 1 mm diameter. 50 Fig. 3 Transmission electron microscopy image (50 000"magnification) of the Pd nanoparticles produced as described in Section 2.2.1. Fig. 4 Scanning electron microscopy images of: (a) an empty graphite platform (50 000"magnification), after pyrolysis at 600 !C; (b) a graphite platform (50 000"magnification) in which 40 mg Pd has been deposited as Pd(NO 3 ) 2 solution, after pyrolysis at 600 !C; (c) a graphite platform (10 000" magnification) in which 40 mg Pd has been deposited as Pd nanoparticles, produced as described in Section 2.2.1., after pyrolysis at 600 !C; (d) a graphite platform (50 000"magnification) in which 40 mg Pd has been deposited as Pd nanoparticles, produced as described in Section 2.2.1., after pyrolysis at 600 !C. 406 | J. Anal. At. Spectrom., 2012, 27, 401–412 This journal is ªThe Royal Society of Chemistry 2012 Published on 26 January 2012. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:37:13. View Article Online Publications 141 In contrast with this situation, when Pd is added from a nanoparticle suspension, it was observed that a very homogeneous distribution was achieved after drying. Moreover, Pd was found in the form of nanoparticles of approx. 20 nm. This form was maintained during the pyrolysis (even though the average size tends to increase a bit with temperature). Fig. 4c and d illustrate this point showing a graphite platform after pyrolysis at 600 !C, where the even distribution of spherical Pd nanoparticles can be appreciated. It should be noted that the large and bright particle that appears in the right center of Fig. 4c corresponds to a Ru agglomerate. Finally, after increasing the temperature up to 2400 !C (vaporization temperature), no indication of remaining Pd was found in the graphite furnace. It can be concluded that the ultimate goal of attaining very small droplets of metallic Pd homogeneously distributed all over the platform, as discussed in the literature many times, 38,48–50 seems to be better achieved when adding this modifier as a nanoparticle suspension. In this way, a larger surface for interaction with the analyte is available during drying and pyrolysis. Moreover, Pd is removed efficiently during the vaporization, thus minimizing potential memory effects. The nanoparticle suspension showed good stability (no evidence of agglomeration) during the period in which the experiments were carried out (approx. six months) for a concentration range between 0.5 g L "1 and 2.0 g L "1 . Obtaining higher concentration proved more challenging, but it was not really necessary for the current (and for most) applications. These nanoparticles are simple to synthesize and use, and, thus, they may offer some potential advantages for other analytes besides sulfur, particularly for elements that owing to their high volatility and/or complex chemistry are hard to stabilize in a graphite furnace. Further work will be directed to their preparation in other alcoholic media (e.g., propanol), since it has been reported that achieving even smaller particle sizes (and, thus, perhaps a better performance) may be possible in that way. 21 3.2. Analytical performance of the procedure proposed After selection of the chemical modifier (Ru as permanent plus Pd nanoparticles), the rest of the parameters were optimized. As can be seen in Fig. 5a, best peak area values are obtained for vaporization temperatures of 2200 !C. However, the signal is best defined (less tailing) if a higher vaporization is used, as can be deduced from the peak height values (see Fig. 5b). As a compromise, a value of 2400 !C was selected for CS vaporization. Good linearity was obtained under these working conditions for an interval that extended from 50 to 2500 ng S, as shown in Fig. 6a, which is a wide range for GFMAS, and this is an important factor for direct solid sampling analysis, owing to the Fig. 5 Vaporization curves obtained for 500 ng of sulfur introduced in different chemical forms when monitoring the 257.958 nm CS molecular line (sum of 3 central pixels) in the presence of Ru as permanent modifier plus 40 mg of Pd, added as Pd nanoparticles. Both the integrated absorbance (A) and the maximum peak height (B) values are shown. Fig. 6 (A) Linearity observed when monitoring the 257.958 nm CS molecular line under the conditions shown in Table 2. (B) Integrated area as a function of the mass of sulfur for the 12 most sensitive CS molecular lines shown in Fig. 1a, measured under the conditions listed in Table 2. This journal is ªThe Royal Society of Chemistry 2012 J. Anal. At. Spectrom., 2012, 27, 401–412 | 407 Published on 26 January 2012. Downloaded by UNIVERSIDAD DE ZARAGOZA on 25/03/2014 15:37:13. View Article Online Appendix 251 Journal impact factor 2012 (JCR ISI): 3.155 Subject categories (Journal ranking): Analytical Chemistry (17/75, Q1), Spectroscopy (7/43, Q1) 7. High-resolution continuum source atomic absorption spectrometry for the simultaneous or sequential monitoring of multiple lines. A critical review of current possibilities M. Resano, M.R. Flórez, E. García-Ruiz Spectrochimica Acta Part B 88 (2013) 85-97 Journal impact factor 2012 (JCR ISI): 3.141 Subject categories (Journal ranking): Spectroscopy (8/43, Q1) 8. Direct determination of bromine in plastic materials by means of solid sampling high-resolution continuum source graphite furnace molecular absorption spectrometry M.R. Flórez, M. Resano Spectrochimica Acta Part B 88 (2013) 32-39 Journal impact factor 2012 (JCR ISI): 3.141 Subject categories (Journal ranking): Spectroscopy (8/43, Q1) 9. Progress in the determination of metalloids and non-metals by means of high-resolution continuum source atomic or molecular absorption spectrometry. A critical review M. Resano, M.R. Flórez, E. García-Ruiz Analytical and Bioanalytical Chemistry 406 (2014) 2239-2259 Journal impact factor 2012 (JCR ISI): 3.659 Subject categories (Journal ranking): Analytical Chemistry (9/75, Q1), Biochemical Research Methods (18/75, Q1) Appendix 252 Justification of the contribution of the PhD student The PhD student, Ms. María del Rosario Flórez García, has carried out the whole experimental work gathered in the current doctoral thesis and has actively collaborated in the experimental design, evaluation of results and realization of the previously presented publications.