Non-spectral Matrix Effects Research Articles

On the Possibility of Reducing the Systematic and Random Errors of Atomic Emission Spectral Analysis Using Multiline Calibration

The main systematic and random errors of atomic emission spectral analysis and various procedures used for their reduction are considered: classical internal standard approach, non-spectral matrix effects, the use of several spectral lines of the analyte and internal standard with and without weight coefficients, signal drift, instability of sample injection and conditions of spectra excitation. The necessity of incorporating those methods into the software of the devices intended for atomic emission spectral analysis with various sources of spectrum excitation, including inductively coupled plasma is demonstrated. An algorithm and stand-alone program for optimization of the calibration curve based on multiline recording of the spectral lines of matrix elements and analyte, internal standards, solvent, and discharge atmosphere, capable of implementing the above methods of error reduction both in calibration and analysis are developed.

Determination of Cd in food using an electrothermal vaporization capacitively coupled plasma microtorch optical emission microspectrometer: Compliance with European legislation and comparison with graphite furnace atomic absorption spectrometry

A precise, accurate and sensitive method based on electrothermal vaporization capacitively coupled plasma microtorch optical emission spectrometry for Cd determination in various dietary products was developed. The instrumentation is inexpensive as it includes a low power (15 W) and low argon consumption (150 ml min−1) microplasma connected to a small-sized Rh electrothermal vaporization device and a low resolution microspectrometer. The sample was subjected to microwave assisted digestion in HNO3 – H2O2 mixture and Cd was determined in 10 μl sample without preconcentration. The standard addition method was used to compensate for the non-spectral matrix effects. The new method was in-house validated by analyzing certified reference materials and test dietary samples for which the Cd maximum level was established in Commission Regulation 488/2014/EU. The figures of merit of the proposed method and the experimental results were compared to those obtained in graphite furnace atomic absorption spectrometry as standardized method and discussed in relation with the requirements in the Commission Decision 2002/657/EC and Commission Regulations 2011/836/EU and 2007/333/EC. The limit of detection in fresh sample (0.001–0.009 mg kg−1), limit of quantification (0.002–0.018 mg kg−1), recovery (103 ± 15%), trueness (−3.0)–(+7.3)% and precision (0.8–14.0%) were similar to those found in the reference method and comply with the minimal criteria imposed to a method to be used in official laboratories for Cd determination in food. Precision for Cd determination close to maximum admitted levels or higher was of 0.8–6.0%, while for lower concentration in the range 4.0–14.0%. The Bland and Altman statistical test allowed successful evaluation of the new method as an alternative to the well established graphite-furnace atomic absorption spectrometry. The proposed method has analytical potential and is easy to run, while the full miniaturized instrumentation could be considered as a useful tool for the future as an alternative to the traditional instrumentation.

Application of low-cost electrothermal vaporization capacitively coupled plasma microtorch optical emission spectrometry for simultaneous determination of Cd and Pb in environmental samples

The study discusses the applicability and analytical capability of a laboratory miniaturized system including a small-sized electrothermal vaporizer with Rh filament and a low power (15W) and low Ar consumption (150mlmin−1) capacitively coupled plasma microtorch interfaced with a low resolution microspectrometer for the simultaneous determination of Cd and Pb in environmental samples by optical emission spectrometry. Ten-microliter samples were dried (100°C, 80s) and vaporized (1500°C, 6s) by filament heating, then sample vapor was introduced into plasma via an Ar stream and 20 successive emission episodes were recorded with 500ms integration time per episode. The operating parameters of the electrothermal vaporization device and plasma were optimized in order to achieve the simultaneous determination of the two elements with best analytical performance. The method was verified by analyzing CRMs of soil, and then applied to real samples digested in aqua regia. The standard addition method was used to compensate for the non-spectral matrix effects on the analytical signal. Under the optimized conditions the limits of detection were 130±35μgkg−1 (13pg) for Cd at 228.802nm and 2100±175μgkg−1 (210pg) for Pb at 368.346nm. Recovery was in the range 98±18% for Cd and 103±11% for Pb for 95% confidence interval, while the relative standard deviation of repeatability was 4–12% (n=3). The investigated instrumentation has analytical potential and prototyping perspective to be used for toxic element determination in environmental samples.

Testing the limits of micro-scale analyses of Si stable isotopes by femtosecond laser ablation multicollector inductively coupled plasma mass spectrometry with application to rock weathering

An analytical protocol for accurate in-situ Si stable isotope analysis has been established on a new second-generation custom-built femtosecond laser ablation system. The laser was coupled to a multicollector inductively coupled plasma mass spectrometer (fsLA-MC-ICP-MS). We investigated the influence of laser parameters such as spot size, laser focussing, energy density and repetition rate, and ICP-MS operating conditions such as ICP mass load, spectral and non-spectral matrix effects, signal intensities, and data processing on precision and accuracy of Si isotope ratios. We found that stable and reproducible ICP conditions were obtained by using He as aerosol carrier gas mixed with Ar/H2O before entering the plasma. Precise δ29Si and δ30Si values (better than ±0.23‰, 2SD) can be obtained if the area ablated is at least 50×50μm; or, alternatively, for the analysis of geometric features down to the width of the laser spot (about 20μm) if an equivalent area is covered. Larger areas can be analysed by rastering the laser beam, whereas small single spot analyses reduce the attainable precision of δ30Si to ca. ±0.6‰, 2SD, for <30μm diameter spots. It was found that focussing the laser beam beneath the sample surface with energy densities between 1 and 3.8J/cm2 yields optimal analytical conditions for all materials investigated here. Using pure quartz (NIST 8546 aka. NBS-28) as measurement standard for calibration (standard-sample-bracketing) did result in accurate and precise data of international reference materials and samples covering a wide range in chemical compositions (Si single crystal IRMM-017, basaltic glasses KL2-G, BHVO-2G and BHVO-2, andesitic glass ML3B-G, rhyolitic glass ATHO-G, diopside glass JER, soda-lime glasses NIST SRM 612 and 610, San Carlos olivine). No composition-dependent matrix effect was discernible within uncertainties of the method.The method was applied to investigate the Si isotope signature of rock weathering at the micro-scale in a corestone sampled from a highly weathered roadcut profile in the tropical Highlands of Sri Lanka. The results show that secondary weathering products accumulated in cracks and grain boundaries are isotopically lighter than their unweathered plagioclase host, consistent with isotopically heavy dissolved Si found in rivers.

O-TOF-ICP-MS analysis of rare earth elements, noble elements, uranium and thorium in river-relating species

The determination of rare earth elements (REEs), Au, Pt, Ir, Pd, Th and U in various river species was performed by the orthogonal time-of-flight inductively coupled plasma mass spectrometry (o-TOF-ICP-MS). The method working conditions were optimised in order to minimise the presence and possible spectral interferences of oxides. Ratios MO+/M+ as well as interference of light REE and Ba oxides/hydroxides with high REEs were evaluated and confirmed to be insignificant. Using the internal standard Re, non-spectral matrix effects (originally decreasing of intensities up to 15%) were overcome and recoveries were found from 92 to 105% for all matrices analysed. For solutions, limits of detection (3σ) were 0.14–0.82 for REEs, Th, U and Y, 1.18 for La, 4.3–5.6 for Au, Pt, Ir and Pd 11 for Sc (all in ng L−1). The Principal component analysis was used for classification of samples according to their places of origin successfully. The o-TOF-ICP-MS was proved to be a very sensitive and suitable technique for bio-monitoring purposes and was employed in the analysis of biota samples (fish, insect, profiles, benthal growths) originated from five different places in the river Elbe (Czech Republic).

Unmodified manganese ferrite nanoparticles as a new sorbent for solid-phase extraction of trace metal–APDC complexes followed by inductively coupled plasma mass spectrometry analysis

The applicability of MnFe2O4 nanoparticles as a new sorbent for group pre-concentration of V, Co, Ni, Cu, Zn, As, Se, Cd and Pb was investigated and compared with that of magnetite nanoparticles. A solid-phase extraction (SPE) of target analytes based on sorption of their hydrophobic complexes with ammonium pyrrolidine dithiocarbamate (APDC) on the surface of unmodified nanoparticles (NPs) was optimized. Magnetic NPs with retained metal complexes were easily separated from the bulk solution by a permanent magnet applied for 5 min. Analyte restoration in the final solution was accomplished by heating with 0.5 mL of 7 mol L−1 nitric acid. The obtained solutions were suitable for continuous nebulization in ICP-MS. Spectral and non-spectral matrix effects for urine analysis (direct and after SPE) were studied and adequate calibration strategies are suggested. Under optimized conditions the magnetically assisted SPE procedure enables enrichment of target analytes by factors of between 7.4 and 10, with linear dynamic ranges of 1–100 μg L−1 for V, Co, Ni, Cd, Pb and 10–1000 μg L−1 for Zn, As, Se and method detection limits in the interval 0.01–0.7 μg L−1. The relative standard deviations (RSD%, for 10 μg L−1 V, Co, Ni, Cd, Pb and for 100 μg L−1 Zn, As, Se, n = 6) were less than 5.5%. The accuracy of the method was evaluated by analysing urine certified reference material Seronorm™ Trace Elements Urine 201205 and obtained recoveries were in the range 85–109%. For correct determination of As and Se in urine, a preliminary microwave sample treatment with a mixture HNO3 + H2O2 was needed, but it led to a worsening of method detection limits. The developed method was successfully applied for analysis of human urine.

Controlling the mass bias introduced by anionic and organic matrices in silicon isotopic measurements by MC-ICP-MS

One of the most widely used sample preparation methods for Si isotopic analyses (δ30Si and δ29Si) is based on cationic chromatography, which does not remove anions from samples. Although it was first thought that the presence of anions in natural concentrations does not distort the isotopic analyses, it has recently been shown that the presence of sulfate can induce a significant shift in isotopic ratio measurements above SO42−/Si ratios (wt) of 0.02. Here, we show that dissolved organic matter can also induce a major Si isotopic bias when analysing river waters. To overcome these non-spectral matrix effects we propose fast and reliable ways, tested on natural freshwater and rock digestion solutions. The sulfate matrix effect is solved by adding to both sample and bracketing standard sulfuric acid in large excess compared to the naturally occurring SO42−. The organic matrix is mineralized by the combined action of UV-C and ozone. We also provide the first δ30Si signature measurements of two common geostandards: SGR-1 (δ30Si = +0.03‰) and FeR-1 (δ30Si = −0.20‰).

Low power capacitively coupled plasma microtorch for simultaneous multielemental determination by atomic emission using microspectrometers

A new, low power capacitively coupled plasma microtorch (30 W, 13.56 MHz, 0.5 L min −1 Ar) was investigated in conjunction with commercially available microspectrometers for the simultaneous multielemental analysis by atomic emission spectrometry of liquid samples without desolvation. Emission spectrum is simpler than in ICP-AES, the resonance lines are the most intense, so that a microspectrometer with FWHM of at least 1.5 nm is satisfactory for the record. The deviation from the Boltzmann distribution for Fe I has demonstrated the departure from the LTE in plasma. The non-spectral matrix effects of Li, Na, K, Ca and Mg on analytes emission are depressive and depend on matrix volatility, ionization potential of the interferent and analyte excitation energy. The detection limits (μg mL −1) are in the range 0.003 (Li) and 1.5 (Mn). The use of the standard additions method allowed the simultaneous determination of elements in environmental certified reference materials with overall recovery of 95 ± 10% and relative standard deviation of 1.7–8.2%. Compared to the traditional ICP, the microtorch has a simple construction, runs at low argon flow and can be integrated in a portable system suitable for in-situ simultaneous determination of elements.

Preliminary investigation of a medium power argon radiofrequency capacitively coupled plasma as atomization cell in atomic fluorescence spectrometry of cadmium

The single ring electrode radiofrequency capacitively coupled plasma torch (SRTr.f.CCP) operated at 275 W, 27.12 MHz and Ar flow rate below 0.7 l min −1 was investigated for the first time as atomization cell in atomic fluorescence spectrometry (AFS) using electrodeless discharge lamps (EDL) as primary radiation source and charged coupled devices as detector. The signal to background ratio (SBR) and limit of detection for Cd determination by EDL-SRTr.f.CCP-AFS were compared to those obtained in atomic emission spectrometry using the same plasma torch. The detection limit in fluorescence was 4.3 ng ml −1 Cd compared to 65 ng ml −1 and 40 ng ml −1 reported in r.f.CCP-atomic emission (AES) equipped with single or double ring electrode. The lower detection limit in EDL-SRTr.f.CCP-AFS is due to a much better SBR in fluorescence. The limit of detection was also compared to those in atomic fluorescence with inductively coupled plasma (0.4 ng ml −1), microwave plasma torch (0.25 ng ml −1) and air–acetylene flame (8 ng ml −1). The influence of light-scattering through the plasma and the secondary reflection of the primary radiation on the wall of the quartz tube on the analytical performance are discussed. The non-spectral matrix effects of Ca, Mg and easily ionized elements are much lower in EDL-SRTr.f.CCP-AFS compared to SRTr.f.CCP-AES. The new technique was applied in the determination of Cd in contaminated soils, industrial hazardous waste (0.4–370 mg kg −1) and water (113 μg l −1) with repeatability of 4–8% and reproducibility in the range of 5–12%, similar to those in ICP-AES. The results were checked by the analysis of a soil and water CRM with a recovery degree of 97 ± 9% and 98 ± 4%, for a confidence limit of 95%. The present EDL-SRTr.f.CCP-AFS is a promising technique for Cd determination in environmental samples.

Influence of non-spectral matrix effects on the accuracy of Pb isotope ratio measurement by MC-ICP-MS: implications for the external normalization method of instrumental mass bias correction

Ca, Al, Fe and Mg are identified as the dominant inorganic matrix components in Pb purified by anion exchange chromatography from a wide compositional range of silicate samples. Doping experiments demonstrate that these elements cause non-spectral matrix effects that result in inaccuracy of mass bias-corrected MC-ICP-MS Pb isotope ratios ranging from −150 to +120 ppm per amu. Although small, these inaccuracies are significant at the level of precision required for application in mantle geochemistry today. The inaccuracy results from matrix induced per mil level variation in the Tl to Pb mass bias ratio (δf(Tl)/f(Pb)‰). In matrix-free standards, δf(Tl)/f(Pb)‰ varies from −0.30 to +0.21‰, depending on run conditions, whereas the mean δf(Tl)/f(Pb) in matrix experiments varies from −1.45‰ with an Fe matrix to 7.11‰ with a Ca matrix. The matrix also causes increased Pb and Tl sensitivity. Signal enhancement plateaus out at high matrix levels. This finite response to the matrix can be used as an advantage. With a common matrix of Mg added to both doped Pb(Tl) standards and their bracketing un-doped standards, variation in f(Tl)/f(Pb) is greatly reduced, resulting in external reproducibility of combined doped and un-doped Pb isotope ratios (30–61 ppm per amu) similar to single session external reproducibility of pure standards (20–60 ppm per amu). Addition of a common matrix to samples and standards may provide a viable alternative to thorough purification in the effort to reduce or eliminate non-spectral matrix effects in MC-IPC-MS isotope ratio measurements.

Towards the development of a fossil bone geochemical standard: An inter-laboratory study

Ten international laboratories participated in an inter-laboratory comparison of a fossil bone composite with the objective of producing a matrix and structure-matched reference material for studies of the bio-mineralization of ancient fossil bone. We report the major and trace element compositions of the fossil bone composite, using in-situ method as well as various wet chemical digestion techniques. For major element concentrations, the intra-laboratory analytical precision (%RSD r) ranges from 7 to 18%, with higher percentages for Ti and K. The %RSD r are smaller than the inter-laboratory analytical precision (%RSD R; <15–30%). Trace element concentrations vary by ∼5 orders of magnitude (0.1 mg kg −1 for Th to 10,000 mg kg −1 for Ba). The intra-laboratory analytical precision %RSD r varies between 8 and 45%. The reproducibility values (%RSD R) range from 13 to <50%, although extreme value >100% was found for the high field strength elements (Hf, Th, Zr, Nb). The rare earth element (REE) concentrations, which vary over 3 orders of magnitude, have %RSD r and %RSD R values at 8–15% and 20–32%, respectively. However, the REE patterns (which are very important for paleo-environmental, taphonomic and paleo-oceanographic analyses) are much more consistent. These data suggest that the complex and unpredictable nature of the mineralogical and chemical composition of fossil bone makes it difficult to set-up and calibrate analytical instruments using conventional standards, and may result in non-spectral matrix effects. We propose an analytical protocol that can be employed in future inter-laboratory studies to produce a certified fossil bone geochemical standard.

Determination of chromium in whole blood by DRC–ICP–MS: spectral and non-spectral interferences

Quantification of chromium in whole blood has been performed by ICP-quadrupole MS. The spectrometer was equipped with a dynamic reaction cell (DRC) with ammonia as reaction gas. The rejection parameter q (RPq) of the DRC and the flow rate of ammonia (NH3) were optimized and set at 0.7 and 0.6 mL min(-1), respectively. Blood was diluted 1:51 (v/v) with an aqueous solution containing 0.1 mg L(-1) NH4OH, 0.1 g L(-1) EDTA, 5 mg L(-1) n-butanol, and 0.1 per thousand Triton X100. Non-spectral matrix effects observed when using the DRC were confirmed by use of vanadium. External calibration with blank and standard solutions prepared in purified water led to biased results for quality control samples. Standard addition calibration was therefore used and its validity verified. By comparing the slopes and calculating residues, it was proved that the plot obtained with standard additions and the plot obtained from blood samples of different concentrations were aligned down to 0.05 microg L(-1) after dilution.

Influence of the operating conditions and of the optical transition on non-spectral matrix effects in inductively coupled plasma-atomic emission spectrometry

In order to study in ICP-AES, the influence of the plasma operating conditions, power and carrier gas flow rate, and of the optical transition on non-spectral matrix interferences, line-rich elements such as Mn, Cr and Cu have been selected. Selection of a large pool of lines was possible because of the use of multichannel solid-state detection. An axially viewed plasma was used. Matrices were K, Na, Li, Ca and Mg. Matrix effect was evaluated by comparing the signals for test elements in water. Use of robust conditions led to an almost flat response, while non-robust conditions led to a significant scattering of the signal changes. In the case of Mn, the z 7P Mn multiplet was exemplified as it contains not only the most Mn sensitive line, the Mn II 257.610 nm resonance line, but also the 259.372 and 260.568 nm resonance lines, and the non-resonant Mn II 343.897 nm line. Even under robust conditions, the non-resonant line exhibited a different behavior. The difference with the other resonance lines was reduced by using an axially viewed ICP with a large injector id, or suppressed by using a radially viewed ICP. In the case of Cr, the z 6D Cr II multiplet was selected as it contains three resonant lines linked to the a 6S fundamental, and other non-resonant lines. The behavior was identical under robust conditions, while an abnormal behavior was observed for the Cr II 334.78 nm line under non-robust conditions, depending on the extent of these non-robust conditions. Cu was an interesting element as ionic lines lie in the energy sum range 15.96–16.26 eV, i.e. slightly above the Ar ionization energy. It was shown that, under robust conditions, the line behavior was not similar although the energy range was small. Moreover, this behavior was depending on the ICP system used for the experiment. It was concluded that not only the magnitude of matrix effects depends on the operating conditions but also may depend on the optical transition, illustrating the complexity of these effects.

Application of internal standardization in ICP-QMS through discrete sample introduction methodologies

Non-spectral matrix effects in ICP-MS are often corrected by the method of internal standardisation. In the case of discrete sampling a common algorithm refers to the ratio of either maxima or areas of both analyte and internal standard peaks. The current work describes a protocol for internal standardisation across the transient peak based on the calculation of analyte-to-internal standard ratio for every scan sweep. Thus, a new converted function, Every Sweep Internal Standardisation (ESIS) ratio versus time, is obtained. One required merit is a dynamic estimation of variable matrix influence. Three discrete sampling systems coupled to an ICP-QMS instrument have been studied (i.e., flow injection, air segmented discrete introduction and injection into an air carrier). Two internal standard introduction modes (i.e., continuous and pulse) have been evaluated. For continuous mode, the ESIS function has a profile similar to the sample injection peak while for the pulse mode it has approximately constant values only within a limited time interval. The ESIS zone of stability for the tested sampling systems has been critically evaluated. By ESIS protocol lower RSDs have been found than the common internal standardisation. The matrix correction efficiency achieved with ESIS ranged between 96%–105%, 94%–106% and 93%–110% for an alcoholic beverage (alcohol content up to 40%), high salt content solution (NaCl up to 3%) and moss sample, respectively.

誘導結合プラズマ質量分析法による岩石試料の微量元素の定量分析

For quantitative analyses of trace elements in rock samples by inductively coupled plasma-mass spectrometry (ICP-MS), it is necessary to accommodate instrumental drift and irregularities in detector linearity. Interferences of oxide molecular ions and non-spectral matrix effects are also taken into consideration to establish an appropriate analytical method. Relatively rare trace elements in rock samples, such as rare earth elements, are determined by alternatively analyzing a 10-ppb standard solution and 1000 times diluted rock solutions containing a 10-ppb Indium internal standard. Analytical results for both samples and standards are normalized to the internal standard, which enables minimization of instrumental drift and correction of uncertainty in dilution factors. Using the internal standard, matrix effect is effectively canceled out at the dilutions factor. Determinations of elements present in relatively high concentrations, such as Rb, Sr, Y and Zr, is achieved by increasing dilution to achieve acceptable detector linearity. This further suppresses matrix effects originating from concomitant major elements.

Trace analysis of impurities in sol-gel prepared BaTiO 3 -powders with ICP-MS

An ICP-MS method for the determination of ultra-trace impurities of 21 elements (Be, Al, Ca, Cr, Mn, Co, Ni, Sr, Zr, Ag Cd, In, Sn, La, Ce, Er, Hf, W, Pb, Bi, U) at pg/g to ng/g level in BaTiO3-powders and precursors prepared by different sol-gel-methods was developed. The non-spectral matrix effects like suppression of the analyte signal and spectral effects like the formation of polyatomic ions like MO+, MAr+, MCl+ and M2+ which interfere with isotopes in isobar overlaps was investigated. To correct for these polyatomic ions a “blank solution” with the composition of the sample matrix was measured and the data were subtracted from the results of the sample; a standard addition method for calibration and 45Sc, 89Y, 103Rh and 141Pr as internal standards to compensate the matrix effects were used and improvements in accuracy and precision were shown. The stability of the instrument and the detection limits in the presence of the barium and titanium matrix were established. In BaTiO3 different amounts of trace impurities were detected in the μg/g to ng/g range. The main impurities are strontium and calcium. The source of the impurities are mainly the educts but it is shown that contamination during the synthesis process is also possible.

Direct inductively coupled plasma mass spectrometric determination of cadmium in urine with special attention to matrix effect correction

In the direct ICP-MS determination of Cd in urine, non-spectral matrix effects caused by high concentrations of sodium chloride can play an important role. In the present study, matrix effects were reduced by an optimum choice of the instrumental conditions and of the internal standard element, in this case Rh. Residual matrix effects were corrected with two procedures: the first was based on the signal of35Cl and the second on the matrix effect experienced by the internal standard element. Average recovery with either procedure both of a Cd spike added to the samples and of the concentration found for a certified reference material was between 99 and 101% with a relative standard deviation of 2.3%. The correction factors in the correction equations were calculated from the recoveries of a limited number of recovery experiments. Sample preparation consisted of a 10-fold dilution of the sample with 1% HNO3. 114Cd was found to be the best isotope with a detection limit of 8 ng l–1 in a non-diluted urine sample. The Rh matrix effect-based procedure can most probably also be used to correct for matrix effects in other sample types.