Fast Method of Multi-Elemental Analysis of Stream Sediment Samples by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) with prior Single-Step Microwave-Assisted Digestion

A aplicação de procedimentos de digestão por micro-ondas em uma única etapa foi investigada visando obter um método rápido para a análise multielementar em amostras de sedimentos de córrego empregando a espectrometria de massas por plasma indutivamente acoplado (SF-ICP-MS). O material de referência certificado de sedimento de córrego produzido pelo Centro Nacional de Análise para Ferro e Aço, China (NCS CD 73309), foi selecionado como referência. Os reagentes empregados e avaliados para os procedimentos de digestão e extração foram: HNO3, mistura de HNO3/HCl (9:1 v/v), mistura de HCl/HNO3 (água régia, 3:1 v/v) e mistura de HNO3/HCl/HF (8:1:1 v/v/v). Ag, Al, As, Ba, Be, Bi, Cd, Co, Cr, Cu, Fe, K, Li, Mn, Mo, Na, Ni, Pb, Sb, Sn, Sr, Ti, Tl, V e Zn foram avaliados, sendo que a maioria puderam ser determinados quando somente HNO3 foi empregado para a digestão da amostra de sedimento de córrego. Para a determinação de Sb, o uso da água régia foi necessário. Berílio, K, Na, Ti e Tl foram fácil e satisfatoriamente isolados das matrizes ambientais quando água régia foi empregada para a digestão da amostra, embora os valores de recuperação para esses elementos tenham sido melhores quando a mistura HNO3/HCl/HF (8:1:1 v/v/v) foi empregada.


Introduction
Inductively coupled plasma-mass spectrometry is a powerful technique for determination of the trace elements in various environmental samples.This technique usually requires a transformation of a sample into solution.Therefore, investigation of the digestion methods makes an important step of the validation (optimization) of analytical procedure.More precisely, the results of a digestion procedure depend on various factors such as matrix of the material to be analyzed, reagents applied for decomposition, temperature, duration of the digestion procedure, etc.
Most of the papers published so far deal with the analyses of only a few elements in stream sediment materials at the same time and do not provide the information on the possibilities of simultaneous analysis of larger number of elements.Thus Sastre and co-workers 1 Vol.22, No. 3, 2011   described the determination of Cd, Cu, Pb and Zn in various environmental materials among which is stream sediment reference material.The authors undertook two multi-step microwave-assisted decomposition procedures: one was total digestion procedure which involved HNO 3 , HF, and HClO 4 , but the other procedure was based on use of aqua regia.Total digestion procedure yielded higer recoveries of the analyzed metals.Pöykiö et al. 2 reported on the analysis of Cr, Cu, and Ni in stream sediment as well as in marine sediment reference material.The procedure described in their work implies a microwave-assisted decomposition of the materials using HNO 3 , HCl, HF as well as H 3 BO 3 .Based on recovery values, Pöykiö et al. 2 determined Cr in higher concentration, while Cu and Ni in lower concentrations than certified ones.
Microwave-assisted digestion procedures for sample preparation are widely applied for decomposition of various types of samples from the environment, for instance soil, 3,4 coal, 5 various biological materials, 6 atmospheric aerosols, 7 etc.The survey of the literature can be summarized as follows: different authors -different digestion procedures.
A method for the transfer of the chemical elements from a sediment sample into solution should be chosen regarding the goals of the investigation, especially when an assessment of the anthropogenic impact on the sediment metal and metalloids content have to be done.That is because too strong digestion condition (e.g., when HF is applied) leads to an over-digestion of a sample with subsequentely erroneous conclusion about eventual environmental risks. 8It is well known that mobility and potential toxicity depend on the metal-to-sediment binding forms.For most of the environmental testings, digestion of the stream sediment samples using only nitric acid or aqua regia yields acceptable results. 4On the other hand, the selection of the digestion parameters may depend on the chemical properties of the investigated element(s).For instance, Al or Sb are elements whose recoveries strongly depend on the acid applied as it will be shown later here.
In this investigation, four series of digestion of the CRM stream sediment (NCS DC 73309), China National Analysis Center for Iron and Steel, Beijing, China, 2004 have been performed.In each series different acidic reagent was applied.The digestion of the sediment samples was carried out by closed-vessels microwave-assisted sample preparation system and the content of the selected metals and metalloids in the prepared solutions was determined by sector field mass spectrometry with inductively coupled plasma (SF-ICP-MS).The main objective of this investigation is to establish how a selection of an acid/acid mixture employed for digestion affects the recovery of the selected metals and metalloids from stream sediment samples.

Materials and methods
Certified reference material stream sediment (NCS DC 73309, also known as GBW 07311), China National Analysis Center for Iron and Steel, Beijing, China, 2004 has been used for this validation.
Mass calibration of the SF-ICP-MS was performed using multi-elemental solution that contained B, Ba Co, Fe, Ga, In, K, Li, Lu, Na, Rh, Sc, Tl, U, and Y in the concentration of 1 µg L -1 (5% v/v HNO 3 ).
Microwave digestion system Multiwave 3000, Anton Paar (Graz, Austria) with high pressure Teflon vessels and the rotor XF-100 was used for wet digestion of the stream sediment CRM samples.Each solid sample (ca.0.10 g) was put in the pre-cleaned vessel and mixed with the reagent(s).Volume of the acid/acid mixture was 10 mL.Since the XF-100 rotor accepts 8 vessels, seven replicas of the CRM stream sediment sample along with one procedural blank have been digested at the same time.The rotor was prepared according to the manufacturer's instructions and the parameters of the digestion procedure have been programmed as follows: T max = 230 °C, P max = 1400 W; t (ramp) = 20 min, t (heating at max. temperature) = 20 min, t (ventilation) = 20 min.The selected heating temperature and power are the maximum for this model of rotor.Since previous studies as well as author's preliminary tests revealed that even small variation of the applied digestion temperature has significant influence on the extraction recoveries for some elements, it has been decided to perform the investigation by heating the reaction mixtures at maximum available temperature, i.e. 230 °C.
When instrumental digestion routine has been finished, each solution was transferred into a plastic 100 mL volumetric flask (Nalgene, Rochester, NY, USA) and diluted to 100 mL.An aliquot of the solution (ca. 10 mL) was centrifuged.Centrifuge (Sigma, Sain Louis, MO, USA) was adjusted to 3000 rotations min -1 ; 20 min.After centrifugation, the 10-fold dilution (namely, 1 mL of sample solution was diluted to 10 mL solution) and addition of the internal standard (In 1 µg L -1 ) was made prior to measurement on the SF-ICP-MS.
Teflon vessels of the microwave system were accurately cleaned after each digestion procedure.So, after removing the sample solutions, each vessel was rinsed with distilled-deionized water and dried under laminar air flow.Then, 10 mL of the acid reagent that was used in the precedent digestion procedure was instilled in each vessel.Finally, somewhat shortened digestion-like instrumental procedure (T max = 230 °C, P max = 1400 W; t (ramp) = 10 min, t (heating at max. temperature) = 10 min, t (ventilation) = 20 min) was applied in order to remove remaining metal traces from the Teflon vessels.After that, the vessels were carefully (5-7 times) rinsed with water and dried under laminated air at the ambient temperature.
All elements (Ag, Al, As, Ba, Be, Bi, Cd, Co, Cr, Cu, Fe, K, Li, Mn, Mo, Na, Ni, Pb, Sb, Sn, Sr, Ti, Tl, V, and Zn) were determined by SF-ICP-MS (Element 2, Thermo Finnigan, Bremen, Germany).The analyzed solutions as well as standard solutions into plasma were transported by the auto-sampler (ASX 510, Cetac Technologies, Omaha, NE, USA) and sample introduction kit consisting of a conical nebulizer (Thermo Finnigan) and Scott-type glass spray chamber (Thermo Finnigan).Before starting the analysis, mass calibration was performed and the measuring conditions of the instrument were optimized (ion sample depth, ion lens setting, auxiliary gas flow, and coolant gas flow) in order to obtain maximum sensitivity.
It is important to pay attention to the selection of the isotopes of the analyzed elements as well as to the instrumental resolution which was employed for measuring certain elements.Regarding to the selection of the isotopes, it is known that one should choose the most abundant isotope of an element whenever it is possible.When possible, more than one isotope of given element should be measured; one for quantification and the other for control of possible interferences.In this study, this strategy was adopted for Cd, Mo and Sb.An isotope cannot be chosen for measuring when subject to a spectral interference.In addition, the isotopes are classified according to the instrumental resolutions in a manner which ensures separation of the analyzed isotope from interfering ions.It is well known that the sensitivity of the instrument at the medium resolution (MR) is just 10% of the sensitivity at the low resolution (LR), while at the high resolution mode (HR) its sensitivity decreases to only 1% of the sensitivity at LR. Therefore, the rule that governs the selection of instrumental resolution for measurement of some isotope is: "Whenever it is possible one should choose the lowest resolution possible, but higher resolution may be chosen only if necessary."In this research, certain isotopes were measured at the medium resolution mode (MR): 47 Ti, 51 V, 52 Cr, 55 Mn, 59 Co, 60 Ni, 63 Cu, 66 Zn, and 86 Sr, because it is well known that these isotopes at low resolution would suffer from interferences such as 12 75 As ( 40 Ar 35 Cl). 9 Table 1 summarizes the details of the instrumental operating conditions and the data acquisition parameters.Calibration curves were obtained by applying external standardization with series of standard solutions; 0 (i.e.blank) and multi-element standard solutions of 1.000 and 10.00 mg L -1 as well as standard solutions of Ag (1.000 and 10.00 mg L -1 ) and standard solution containing Na and K at the mg L -1 level.Method for calculation of the calibration curve was the weighted regression.High linearity of the calibration curves was obtained; correlation range (r 2 ) for calibration curves was 0.9850-1.000.One of the important characteristics of the SF-ICP-MS technique is a linear response in the wide range of concentrations.Therefore this technique allows undertaking simultaneous measurements of different chemical elements which in certain samples are present in various concentrations and it guarantees satisfactory results.

Procedural blank solutions and detection limits
Procedural blank solutions were prepared in the same way as sample solutions; equal volumes of an acid or acid mixture, which have been used for sample digestion, were microwave-assisted heated in the same manner that was applied for the sediment samples digestion.The measured concentrations (mg L -1 ) of the elements in the procedural blank were subtracted from the concentration values measured in the sediment CRM solutions.Table 2 shows the average metal and metalloid concentration values (µg L -1 ) that were measured in procedural blank solutions as well as in the common blank solutions (1% v/v HNO 3 , s.p., 1% v/v HCl, s.p.).The results shown in the Table 2 reveal that the average procedural blank readings, for all measured elements, are far below concentration values measured in the solutions of the stream sediment.The values given in the Table 2 suggest that higher values of some elements (such as Al, K, Na) in the blank solutions are the consequence of the higher concentration of those elements in the stream sediment and their subsequent, somewhat increased, retention in the laboratory equipment.On the basis of the procedural blank values, the values of detection limits (LOD) have been calculated for this method.
The LOD shown in Table 3 represent the lowest concentration value (in mg kg -1 ) of an element that can be quantified in a stream sediment sample by performance

Results and Discussion
This paper studied analytical procedures for determination of selected metals and metalloids in stream sediment with prior microwave-assisted digestion procedure using four different acidic media.The main purpose of the work is the comparison of the extraction abilities of the acid or acid mixtures and the possibility of their application in fast multi-elemental analyses of the stream sediment samples.The results of the trace elements determination in the stream sediment certified reference material, obtained after the analyses of seven replicates are shown in Table 4.
Investigation of the extraction of the metals and metalloids from various types of solid samples at different extraction conditions is important, because such results show how a change of the reagent and/or conditions of sample digestion (temperature, use of open or closed vessels) can greatly affect the recovery of an analyte.When a scientist decides to determine a metal content in stream sediment sample, the extraction conditions have to be chosen concerning the goal of the investigation.Namely, if geological characterization of the sample is to be performed i.e., if total digestion of the samples is required -what should be undertaken is an aggressive extraction procedure.On the other hand, determination of biologically available metal content should be made using milder digestion or extraction conditions. 1,10he results shown in Table 4 have been obtained after analysis of seven replicates, which is significantly more than the number of replicates usually employed in environmental studies.It can be noticed that recovery of the greater part of the selected metals and metalloids increased when aqua regia or mixture HNO 3 /HCl/HF (8:1:1 v/v/v) was applied instead of HNO 3 alone.The recovery values of some metals (Be, K, Na, Ti, and Tl) were largely increased by employing hydrofluoric acid along with hydrochloric and nitric acid, because partial digestion of the silicate matrix took place.The main exception to this pattern is Al: in solutions that contained HF, Al was determined in lower concentration due to precipitation of aluminium fluoride.It is important to mention that low recovery values obtained for K and Na as well as for Al in HNO 3 and HNO 3 /HCl mixtures are due to binding of these metals in silicate core of the stream sediment.The recovery of Sb was significantly increased in aqua regia in comparison with the digestion procedures undertaken by nitric acid or by mixture of nitric and hydrochloric acid (9:1 v/v).It is known that in nitric acid Sb transforms into mixture of insoluble oxides (Sb 2 O 3 , Sb 2 O 5 , Sb 4 O 4 (OH) 2 (NO 3 ) 2 ), 11,12 but in the presence of hydrochloric acid Sb changes into chloro-complexes 11 and in the presence of hydrofluoric acid Sb exists as [SbF 6 ] -. 11 These facts about analysis of Sb in the solid samples have to be taken into consideration; Sb is an element of increasing environmental scientists' interest mainly due to anthropogenic emission of this metalloid in the environment. 13The results of the analyses of Sn show that recovery values of this element are low in all four applied digestion reagents.Since, according to the author's previous experience, a recovery of Sn usually greatly increases when stronger reagent for digestion is applied, the observed weak increase of the recovery of Sn could be explained in terms of the nature of Sn binding in the tested stream sediment reference material.Also, high certified value of the Sn content (370 mg kg -1 ) and RSD of the Sn analyses support this hypothesis.Furthermore, the results obtained by Meeravali and Kumar 14 suggest that the amount of a stream sediment sample could affect Sn recovery.The values obtained for Ba reveal constancy and for Sr a moderate increasing of the recovery when stronger reagent is applied.However, even when the strongest reagent, among investigated acidic reagents, was used (i.e., HNO 3 /HCl/HF, 8:1:1 v/v/v), the recovery of Ba was 71% and for Sr was only 61%.This observation could be explained both in terms of the nature of Sr binding in the stream sediment and low solubility of strontium fluoride. 12The other one earth-alkaline element selected for analysis was Be.Its recoveries observed in the cases of the digestion processes involving HNO 3 alone or HNO 3 and HCl were quite low.This can be explained in terms of Be chemistry: in HNO 3 , Be tends to form covalent compound such as Be(NO 3 ) 2 that is scarcely soluble 12 or Be 4 O(NO 3 ) 6 which is volatile. 11On the other hand, in the presence of HF, Be forms BeF 2 , which is quite soluble; that is why the recovery value of Be obtained when the digestion was carried out with mixture HNO 3 /HCl/HF (8:1:1 v/v/v) is high (> 100%). 12The other metal selected in this work, which shows similar pattern of the recovery is Ti.In the acidic solutions, such as HNO 3 or HCl, Ti forms insoluble oxo-complexes, but reaction of Ti and HF yields soluble fluoro-complex, [TiF 6 ] 2-. 11,12Finally, the results of the determination of Tl can also be explained in terms of Tl chemistry.That is, in HNO 3 , Tl forms insoluble oxides.
In the presence of HCl, Tl forms TlCl, which is scarcely soluble.In the reaction of Tl and HF, TlF is formed; TlF is well soluble and that results in high recovery of Tl when HF is involved in digestion procedure. 11he relative standard deviation values (RSD) were found to be about 5% or less for most of the elements in all four procedures.The RSD values for some elements that were measured in low concentrations in solutions (such as Sb, Tl) were above 10%.It is important to notice a high RSD value, 110%, for Al in the HNO 3 /HCl/HF digestate.This can be attributed to the precipitation of AlF 3 in acid medium and its irregular distribution through the solutions.When comparing the RSD values obtained in this study with those calculated using certified values, it can be noticed that in some cases in this investigation significantly higher RSD values have been obtained.For instance, RSD of the determination of Sb in HNO 3 and HNO 3 /HCl (9:1 v/v) leachates are 19 and 27%, respectively, while the certified one is 4.1%.Similar situation occurred with Sn (14 vs. 6%) and Tl (12 vs. 6.9%), both in HNO 3 leachate.Regarding the analyses of two major elements (K and Na), the RSD decreases upon applying stronger digestion condition.Thus, for example, K was determined with the RSD of 11% in HNO 3 leachate, while that in HNO 3 /HCl/HF (8:1:1 v/v/v) leachate was 4.9%.The certified RSD for K is 11%.According to certificate for stream sediment used in this study, the RSD value for Na is 3.4%.Similar to the case of K, also for Na, RSD obtained in the procedure concerning the use of HNO 3 for digestion was 9.9%, while, on the other hand, in the HNO 3 /HCl/HF (8:1:1 v/v/v) leachate Na was determined with RSD of 3.9%.On the other hand, for some of the elements, lower RSD values than those calculated from the certified values have been obtained in this study.For Ag the RSD values obtained in this study vary from 3.2 to 5.7%, while 6.3% is calculated from the certified values.The most important case is Cd: RSD calculated from certified values is almost 22%, while RSD in four series of the experiments described in this paper ranged from 4.0 to 5.2%.
Literature survey reveals that data on similar experiments on the reference material used in this study are quite rare.As mentioned previously, Sastre et al. 1 published the results of the determination of Cd, Cu, Pb, and Zn.The recoveries of all four analyzed elements were somewhat above 100%, but also decomposition medium was quite strong: 70% v/v HNO 3 , 40% v/v HF, 70% v/v HClO 4 , as well as 30% m/m H 2 O 2 .Valverde 15 undertook wet digestion of the same certified reference material, using aqua regia with reflux during 16 h.This author measured only As by ICP-MS and found it in a concentration of 190.6 mg kg -1 , 101% recovery. 15tream sediment reference material NCS DC 73309 was used also by Lima et al. 16 for method optimization of As, Cd, and Pb determination by electrothermal atomic absorption spectrometry (ETAAS), with prior microwave-assisted digestion of solid samples.The mixture for digestion was composed as follows: 4 mL HNO 3 , 6 mL HCl, as well as 10 mL HF, and maximum temperature was 160 °C.Lima et al. 16 undertook a series of experiments which implicated use of seven different permanent modifiers.Therefore the authors obtained a series of seven different results for each element that has been measured.Thus, they determined the recoveries of As, Cd, and Pb in ranges 80-101%, 78-101%, and 86-101%, respectively.
Consequently, one can conclude that elemental concentrations determined in a certain material from the environment, such as stream sediment, are very sensitive to several experimental parameters.Therefore, a comparison of the numerical values is possible when at least one group of parameters, in different experiments that are to be compared, is the same (for instance temperature and duration of digestion, digestion medium or matrix of the analyzed material).One, but not the only one, criterion for estimating the effectiveness of a digestion method is the recovery value obtained for the analyzed element.It is necessary to consider that low recovery does not necessarily imply unsuitability of a method, because the recovery is affected by at least four factors: nature of metal-to-sediment binding, chemical properties of the analyzed element, extraction reagent(s) used, and the temperature of the extraction procedure.On the other hand, a numerically low extraction value does not imply a low concentration of an element and a problem of reaching the method detection limit, i.e. analysis of some element can be characterized by low recovery and, at the same time, relatively high concentration in sediment sample and, subsequently, in digestate.Namely, low recovery in such case does not imply an analytical problem.
Taking into consideration the results presented in this paper, it can be concluded that the method employing aqua regia as a reagent for wet digestion is the most favorable for simple and fast, ecologically oriented, multi-elemental analysis of stream sediments, for two main reasons: (i) aqua regia is a reagent that is not so strong to provide a complete dissolution of a stream sediment samples, and such an analysis generally should not lead to misconclusions about potential ecological risks due to the biologically available metal fraction present in the stream sediment samples, and (ii) aqua regia allows chemical change of a large part of the investigated elements into the chemical forms which are soluble in acidic medium and therefore C 35 Cl, 35 Cl 16 O, 40 Ar 12 C, 39 K 16 O, 40 Ar 16 O 1 H, 44 Ca 16 O, 40 Ar 23 Na, 40 Ar 26 Mg, and 75 As 11 B, respectively.The elements measured at the high resolution mode (HR) and the interferences (in brackets) that one should avoid are: 23 Na ( 7 Li 16 O), 27 Al ( 13 C 14 N), 39 K ( 23 Na 16 O), 56 Fe ( 40 Ar 16 O; 40 Ca 16 O), and

Table 2 .
Measured metal concentrations (µg L -1 ) in procedural blank solutions and in calibration blank solution

Table 3 .
LOD of metals in solid environmental samples obtained by SF-ICP-MS after microwave-assisted digestion of samples in

Table 4 .
Determination of metals and semi-metals in the stream sediment CRM by SF-ICP-MS after microwave-assisted digestion with various reagents (mean ± standard deviation, n = 8)