Determination of cadmium, copper, mercury, lead and ...

Microchemical Journal 128 (2016) 198–207

Contents lists available at ScienceDirect

Microchemical Journal journal homepage: www.elsevier.com/locate/microc

Determination of cadmium, copper, mercury, lead and zinc mass fractions in marine sediment by isotope dilution inductively coupled plasma mass spectrometry applied as a reference method Irena Wysocka a,b, Emilia Vassileva a,⁎ a b

International Atomic Energy Agency, Environmental Laboratories, 4 Quai Antoine 1er, 98000, Monaco Polish Geological Institute-National Research Institute, Rakowiecka 4, 00-975 Warsaw, Poland

a r t i c l e

i n f o

Article history: Received 8 April 2016 Received in revised form 5 May 2016 Accepted 5 May 2016 Available online 07 May 2016 Keywords: ID ICP-MS Trace elements Marine sediment Reference measurements Validation Uncertainty

a b s t r a c t Isotope Dilution Inductively Coupled Plasma Mass Spectrometry (ID ICP-MS) has been applied for the determination of the total mass fractions of five trace elements (Cd, Cu, Hg, Pb and Zn) in marine sediment candidate reference material IAEA-458. Because of the rather complex matrix of the sample and the expected spectral interferences, special care was taken for the validation of the applied method, particularly for its measurement step. Reference isotopic measurements were carried out by quadrupole inductively coupled plasma mass spectrometer (ICP-QMS) or by sector field inductively coupled plasma mass spectrometer (ICP-SFMS). Comparative studies were performed with the aim to examine the degree of equivalence between employed techniques in terms of the high requirements for the quality of reference measurement data. In order to reduce the number of analytical steps multiple spiking approach was applied. The measurements were performed after a microwave digestion of the blend samples assuring complete isotopic equilibration. Cd was determined after matrix separation by ICP-QMS or ICP-SFMS at low resolution mode. Additionally Cd isotopic ratios were measured by ICP-QMS operated in collision reaction mode, without separation of the matrix. ICPSFMS at medium resolution and ICP-QMS methods were used for Zn and Cu determinations. Pb isotopic ratios were measured by ICP-QMS operated in standard mode. Because of the low Hg content in the sample Hg isotopic ratios measurements were carried out only by ICP-SFMS. The entire ID ICP-MS measurement process was described by mathematical equations and the combined uncertainty was estimated. All factors influencing the final results and their uncertainties were systematically investigated. This included procedural blank, moisture content in the sediment material and parameters affecting the blend ratio measurements (instrumental background, spectral interferences, dead time and mass discrimination effects as well as the repeatability of measured isotopic ratios). The excellent agreement between mass fraction values determined by different measurement techniques proved that all potential problems coming from the complex sediment matrix and the spectral interferences were solved. The consistency of the obtained results confirmed that ICP-QMS can be the method of choice even for the reference measurements, when the element content in a sample is sufficiently high and all steps in the analytical procedure are well described and understood. The modeling of the analytical process resulted in adequate validation of measurement procedure, establishing traceability of the measurement results and estimating the realistic final expanded uncertainty. The combined uncertainties associated to the obtained mass fractions were considerably small (2% for Cd, Cu, Zn and Pb and to 4.9% in the case of Hg, k = 2) and in line with the primary assumption to the reference measurements. The developed procedure has been successfully applied for the characterisation of the IAEA-458 candidate reference material as well as used in the calculation of the assigned values in the frame of wide world IAEA-458 interlaboratory comparisons study. © 2016 Elsevier B.V. All rights reserved.

1. Introduction

⁎ Corresponding author. E-mail address: [email protected] (E. Vassileva).

http://dx.doi.org/10.1016/j.microc.2016.05.002 0026-265X/© 2016 Elsevier B.V. All rights reserved.

In the recent years, much attention has been paid to the chemical composition of the marine sediments in the coastal region due to the deterioration of oceanic ecosystems. The industrial and urban activities

I. Wysocka, E. Vassileva / Microchemical Journal 128 (2016) 198–207

have significantly contributed to the increase of element contamination (e.g. As, Cd, Cu, Cr, Hg, Mn, Ni, Pb, Sn, Zn) of marine environment and have directly influenced the costal ecosystems. The major sources of pollution in the sea waters include municipal wastewater treatment and disposal, urban solid waste disposal, release of harmful concentration of nutrients, storage, transportation and disposal of hazardous waste and activities contributing to the destruction of the coastline and coastal habitats [1–9]. Marine sediments, especially estuarine and coastal, have been used to evaluate water quality because of a higher stability and lower variability of the sediment compared to the water column [5]. However, heavy metals are not fixed permanently in the sediment and the variation of the physico-chemical characteristics of the water column (pH, salinity, temperature, redox potential and the concentration of different organic ligands) can release part of the metal content trapped into the sediment to the water column making them available to living organisms [4,6]. Furthermore, the sediments integrate the concentration of pollutants throughout time and, therefore, this can be useful to study the historic evolution of contamination and to predict its future effects [10–12]. The vertical distribution of some trace elements in sediments can in fact be regarded as an historical record of environmental changes. Various studies have shown that trace element concentration profiles in sediments can be effectively used to assess paleo contamination events with the preindustrial concentration usually considered as the background [13]. Marine sediments are very important for any comprehensive marine monitoring program and the accurate determination of metal mass fractions in sediment cores is an issue of great environment significance. However, the determination of trace elements in marine sediment involves complex analytical methodology and still needs development of relevant analytical protocols. The application of appropriate sediment Certified Reference Materials (CRMs) or Reference Materials (RMs) can help to overcome some methodological problems in the procedures applied by the designated environmental monitoring laboratories [14,15]. For the enforcement of regulations, and for studies at an international level, for example to assess the effects of pollutants on the environment, it is essential that measurements obtained in different laboratories on different occasions are comparable. Comparability can be achieved by making measurements traceable to a common system of a reference and providing measurement uncertainty statement to demonstrate the result's reliability [16–18]. The reliability and comparability of the analytical results, produced in this context are crucial points for management of the environment, taking decisions and meaningful actions in the remediation policy. The element determination in marine matrices typically relies on experimental processes using analytical instruments with a great level of complexity. Despite the improvements in methodologies and instrumentation, unanticipated problems with analysis of complex samples often become evident when the results are obtained by different analytical methods or different laboratories. The simplest way to verify the results for analytical laboratories is to use suitable standards, e.g. the reference materials for method validation, which already offer uncertainty and traceability of their stated quantities. By using reference materials for method validation, laboratories can demonstrate that their measurement results are traceable, or in other words that they are globally comparable [17,18,19]. The preparation of CRMs requires the application of the analytical reference procedures enabling to obtain low expanded uncertainty on the characterization component. The uncertainty budget of a CRM usually consists of three main components, which are contributions from the homogeneity, the stability and the characterization studies. This requires the application of analytical reference procedures enabling expanded uncertainties on the characterization component of the total uncertainty of 1–3% or preferably even below [20,21]. The calibration technique which is suited best for the quantification of trace element mass fractions is undoubtedly isotope dilution mass spectrometry (ID ICP-MS) [22,23].

199

The isotope dilution is based on a change of an element isotopic composition by adding an isotopic standard containing the same element, but with a distinct isotopic composition, and measuring the magnitude of the change induced. One of the significant advantages over other approaches, is that the analyte recovery does not have to be quantitative providing that isotopic equilibration has been achieved between all of the analyte and added spike material. Isotope dilution offers the possibility to determine major to trace element mass fractions of elements in any matrix, with superior accuracy and precision compared to the external calibration method and is often used for certified reference material characterization [22–26]. In this respect, a sector field mass spectrometer is preferable to quadrupole-based instruments because the resulting flat-topped peak shape obtained at low resolution mode enables a precise isotope intensity measurement to be achieved [26]. It should be noted that, as in the case for other calibration strategies, isotope dilution cannot compensate for random contamination which may occur during sample workup and an intensive blank monitoring is thus always necessary [27]. Moreover, problem of possible spectral interferences should be considered and the appropriate way of their correction has to be applied during ICP-MS measurements [24,25]. The development and validation of an ID ICP-MS based reference procedure for the quantification of Cd, Cu, Hg, Pb and Zn in marine sediment with the objective of achieving an uncertainty target on final results of 1.5–3.0% (k = 2) and SI traceable values are described in this study. A systematic assessment of all factors influencing the measurement results as sample-spike isotopic equilibrium, homogeneity study, blend ratio measurements, efficiency of the sample digestion procedure, possible interferences, matrix effects etc. was done throughout the present study. Modeling of the entire measurement process and the use of reference materials, relates each of obtained results to SI units of the mole or the kilogram [19].

2. Experimental part 2.1. Chemicals and materials High quality deionized water from Milli-Q system (Millipore, Bedford, MA, USA) was used throughout this work. Ultra-pure 70% HNO3 (Ultrex®, T. T. Baker, Phillipsburg, NJ, USA), 30.5% H2O2 (p.a. from Merck, Darmstadt, Germany), 40% HF (Suprapur®, Merck, Darmstadt, Germany) and 36% HCl (Ultrex®, T. T. Baker, Phillipsburg, NJ, USA) were used for sample digestion. Only new lab ware material (bottles, vessels, tips, syringes etc.) was employed and it was pre-cleaned thoroughly following a procedure described elsewhere in detail [28]. In order to avoid risk of memory effects from previous experiments, digestion vessels were submitted to an appropriate cleaning procedure. To reduce the risks of airborne contamination all sample processing steps were performed in the clean chemical laboratory (class b 100). The following isotopic reference materials were used to spike the samples: IRMM-622 (111Cd), IRMM-632 (65Cu), IRMM-654 (68Zn), ERM-AE640 (202Hg) from the European Commission, Institute for Reference Materials and Measurements, Geel, Belgium (IRMM) and NIST SRM-991 (206Pb) from the National Institute for Standards and Technologies, USA. For mass discrimination corrections during ICP-MS measurements single element standard solutions of cadmium and copper with natural isotopic composition (from Merck, Darmstadt, Germany) or certified isotopic reference materials (Hg from ERM-AE639, Pb from NIST SRM-981 and Zn from IRMM-3702) were applied. Working standard solutions were prepared gravimetrically by appropriate dilution of stock standard solutions. The natural isotopic compositions for Cd, Hg, Cu and Zn were taken from IUPAC tables [29]. The anion exchanger AG1X-8 (Bio-Rad, Hercules, USA) was used for matrix separation. AG 1X8 was converted from the chloride form to the nitrate form by shaking it with 2% nitric acid. The ion exchanger was left to

200


settle and the acid was decanted off. This procedure was repeated a few times to complete the conversion and to purify the ion-exchange resin.

2.2. IAEA 458 sample 2.2.1. Sample preparation A sample of 34 kg of sediment was delivered by the Korean Ocean Research and Development Institute. The freeze-dried material was milled to a powder in a grinder Retsch SM 200 (Retsch, Haan, Germany). The powder was then sieved through a set of sieves (Fritsch, Idar Oberstein, Germany) and the fraction of 26 μm was collected. The sieved material with a particle size of less than 26 μm was further homogenized. The homogeneity was performed by mixing the material in a stainless steel rotating homogenizer Moritz ERM-BB124 (Moritz, Chatou, France) for 14 days at a temperature of 20 °C (+/− 2°), and relative humidity of 50%. After checking for the homogeneity of the sample material, aliquots of about 30 g were packed into pre-cleaned brown borosilicate glass bottles with polyethylene screw caps and then sealed in plastic bags.

2.2.2. Homogeneity study Extensive homogeneity tests were carried out on this material in order to ensure its suitability as a reference sample and to estimate the uncertainty associated with its homogeneity. In total, 10 bottles were selected using random stratified sampling of the whole batch. Care was taken to ensure that the order of measurements did not correspond to the filling sequence of the bottles. Three subsamples from each bottle were analyzed for their total element mass fractions. The withinbottle homogeneity was assessed by 15 replicate determinations of the content of investigated trace elements in one bottle. The measurements were performed by solid sampling atomic absorption spectrometry under repeatability conditions, and in a randomized way, in order to be able to separate a potential analytical drift from a trend in the filling sequence. The determination of the total mercury was done in solid subsamples with solid mercury analyzer. All methods used for homogeneity studies were previously validated.

2.2.3. Stability study Three sets of five bottles each were stored in the dark at different temperatures, −20 °C, +20 °C and +60 °C, just after the bottling process and kept at described conditions over a period of 2 years. One isochronous study over 6 weeks was applied in order to evaluate the short-term stability of the materials during transport, and one isochronous study over 24 months, to evaluate the stability during storage. The obtained results were compared with the results from samples kept at − 20 °C during this period (−20 °C is considered as the reference temperature). The obtained results show sufficient homogeneity and stability. More details concerning the preparation of candidate reference material, homogeneity and stability testing are given elsewhere [30].

2.2.4. Moisture determination The correction for water content in marine sediment material was applied. The moisture content was determined on the basis of 3 sample replicates. The sediment material was taken from three different bottles. Measurements of water content were performed for about 0.5 g of powdered sample that was put into 3 aluminum vessels and then dried at 105 ± 2 °C for 24 h in an oven. Then weighing and repeated drying was performed until constant mass was attained (0.0002 g difference between two successive weighs). Each weighing was carried out after the sample reached thermal equilibrium at room temperature in a desiccator. The moisture content was calculated by subtracting the sample mass before and after drying. The average moisture content was calculated and used for element mass fraction correction.

2.3. Instrumentation The blends and working standard solutions were prepared gravimetrically by appropriate dilution of the stock standard solutions using an analytical balance (Mettler Toledo, Switzerland). Digestion of the marine sediment samples was performed in a closed microwave system (Mars-X CEM, USA) equipped with a carousel holding 12 digestion Teflon vessels. The isotope ratio measurements were carried out with a quadrupole inductively coupled plasma mass spectrometer (XSeries 2 Thermo Scientific, Bremen, Germany) or a sector field high resolution inductively coupled plasma mass spectrometer (Attom Nu Instruments Ltd., Wrexham, UK). Both instruments were equipped with a Micromist nebulizer (0.2 mL/min, Glass Expansion, Australia) and a cyclonic spray chamber cooled by Peltier cooling system (Glass Expansion, Australia). The instrument conditions were daily checked and adjusted for optimum sensitivity and repeatability of the isotope ratio measurements. The quadrupole inductively coupled plasma mass spectrometer with a collision cell for interference reduction was employed. Mixture of pure hydrogen and helium (Air Liquide, France) was used as the collision gas. Servo operating mode and fast scan ion optics were used for all isotope ratio measurements carried out with ICP-SFMS. The fast scan ion optics was used for fast peak jumping between selected masses across a mass range of total width of 40% of the static mass at which the magnet is parked. The optimized instrumental parameters for ICP-QMS and ICP-SFMS are listed in Table 1.

2.4. Sample preparation Typically 0.1 g of the sediment subsample was placed into a microwave vessel and spiked directly with isotopic reference solutions. Weighing of the subsamples and addition of the spike aliquots were performed exclusively according to metrological gravimetric principles in the clean and humidity controlled area using substitution measurements against operational mass standards. The evaluation of the added amount reference solution was a compromise between several factors including results from preliminary measurements, the characteristics of the spike materials, the final uncertainty [31], sufficiently high counting rate and dead time effects. Within these conditions, the calculated optimum blend ratios were as follow: 0.107 for n(110Cd)/ n(111Cd), 0.891 for n(65Cu)/n(63Cu), 0.254 for n(200Hg)/n(202Hg), 0.137 for n(208Pb)/n(206Pb) and 0.652 for n(66Zn)/n(68Zn) with corresponding error magnification factors 2.5, 0.8, 1.6, 0.7 and 2.8 respectively. In order to reduce the number of analytical steps and the amounts of the chemical reagents used, a multiple spiking was applied. For the isotopic ratio measurement, 3 blend solutions were prepared, using sediment subsamples from three different bottles. The procedure applied for the preparation of the blend solutions was the following: about 0.1 g sediment sample, adequate volumes of the spike solutions, 4 mL of HNO 3 , 0.1 mL of HCl, and 1 mL of HF were subsequently added into each Teflon vessel followed by samples digestion in the closed microwave system. The 5-steps microwave digestion program was applied: I) 5 min ramp time from 0 to 300 W and 5 min hold time; II) 2 min ramp time to 600 W and 2 min hold time; III) 2 min ramp time to 1200 W and 10 min hold time; IV) 2 min ramp time to 600 W and 15 min hold time; V) 20 min cooling time. After sample digestion Teflon vessels were placed on a ceramic heating plate and evaporated to near dryness. The final residues were dissolved in 0.14 mol·L − 1 HNO 3 , then quantitatively transferred to 50 mL PE bottles and finally stored at 4°C. The blend solutions were further diluted with 0.14 mol·L − 1 HNO 3 prior to the ICP-MS measurements. Four procedural blanks were digested together with the sediment subsamples.


201

Table 1 The optimized instrumental parameters for the ICP-QMS and ICP-SFMS methods. a. Optimized ICP-QMS instrumental parameters for the isotopic ratio determinations (XSeries 2, Thermo Scientific) in standard and collision cell mode. Parameter

Cd

Measurement mode Plasma gas flow, L·min−1 Nebuliser gas flow, L·min−1 Auxiliary gas flow, L·min−1 RF power, W Sensitivity for 1 ppb In, cps Background on mass 220, cps Dead time, ns Number of sweeps/replicate Number of replicates Dwell time per amu, ms Sample uptake, mL/min Oxide formation CeO+/Ce+, %

Cu

Pb

Zn

Cd

Cu

standard

standard

standard

standard

collision cell

collision cell

13.5

13.5

13.5

13.5

13.5

13.5

0.91

0.91

0.90

0.91

0.89

0.89

1.05

1.00

1.00

1.05

0.95

0.95

1200

1200

1200

1200

1200

1200

N100,000

N100,000 N100,000

N100,000

N20,000

N20,000

b2

b2

b2

b2

b2

b2

39

68

43

68

39

68 300

150

300

200

100

150

3

3

3

3

3

3

75

100

100

100

75

100

0.2

0.2

0.2

0.2

0.2

0.2

b3

b3

b3

b3

b0.2

b0.2

105

201 64 105 Pd, 110Cd, 111Cd, 112Cd, 113Cd, 95Mo, 63Cu, Hg, 204Pb, 206Pb, Zn, 66Zn, 67Zn, Pd, 110Cd, 111Cd, 112Cd, 113Cd, Measured isotopes 120 65 207 68 95 Sn, 90Zr Cu Pb, 208Pb Zn, 70Zn Mo, 120Sn, 90Zr b. Optimized ICP-SFMS instrumental parameters for the isotopic ratio determinations at low and medium resolutions (Attom, Nu Instruments).

63 65

Cu, Cu

Parameter

Cd

Hg

Cu

Zn

Resolution Plasma gas flow, L·min−1 Nebuliser gas flow, psi Auxiliary gas flow, L·min−1 RF power, W Sensitivity for 1 ppb In, cps Background on mass 220, cps Number of sweeps Dead time, ns Number of replicates Dwell time per isotope, ms Acquisition mode Sample uptake, mL/min Oxide formation CeO+/Ce+, % Measured isotopes

300 (low) 13.5 31.2 0.81 1300 N1000000 b2 2000 16 3 1000 peak jumping 0.2 b1 105 Pd, 110Cd, 111Cd, 112Cd, 113Cd, 95Mo, 120Sn, 90Zr

300 (low) 13.5 31.0 0.80 1300 N1,000,000 b2 2000 15 3 1000 peak jumping 0.2 b1 200 Hg, 201Hg, 202Hg

4000 (medium) 13.5 30.6 0.82 1300 N100,000 b2 2000 15 3 1000 peak jumping 0.2 b1 63 Cu,65Cu

4000 (medium) 13.5 30.6 0.84 1300 N100,000 b2 2000 15 3 1000 peak jumping 0.2 b1 64 Zn, 66Zn, 67Zn, 68Zn, 70Zn

2.5. Chromatographic separation of Cd In order to minimize the polyatomic interferences coming from the presence of Mo and Zr in the sediment sample, matrix separation in the blend samples was applied before cadmium measurements. Potentially all cadmium isotopes (110Cd, 111Cd, 112Cd, 113Cd and 114Cd) can have interferences deriving from the molybdenum oxides (94Mo16O, 95 Mo16O, 96Mo16O, 97Mo16O and 98Mo16O). 110Cd and 112Cd can be additionally influenced by zirconium oxides (94Zr16O, 96Zr16O). The matrix

Table 2 ICP-MS measurement modes applied (+) for element mass fractions determination in the marine sediment sample. ICP-QMS (XSeries 2) Element

Cd Cu Hg Pb Zn

separation was performed by anion exchange chromatography and as a result spectral interferences were significantly minimized. The major matrix cations (Na, K, Mg and Ca) passed through the column without interaction with the anion exchanger, whereas the Cd-chloride complexes were retained on the column and later on eluted with diluted HNO3. For the matrix separation, 0.5 mL of 35.5% HCl was mixed with the aliquot of the sample and then uploaded on the column (3 × 0.8 cm) filled with the anion exchanger AG1-X8 in the nitrate form. The column was rinsed first with 10 mL 2.5% HCl, then with 10 mL 0.25% HCl. The elution of the analyzed fractions was done with 12 mL 2% HNO3. The first 7 mL of eluted fraction was discarded, and the last 5 mL of eluted fraction was collected and the Cd, Mo, Zr, Sn and Pd isotopes were measured by ICP-MS. 2.6. ICP-MS measurements

ICP-SFMS (ATTOM)

Standard mode

Collison cell mode

Low resolution

Medium resolution

+(after separation) + − + +

+ + − − −

+(after separation) − + − −

− + − − +

The selection of the isotopes to be measured in the blend samples was done with respect to availability of spike materials, abundance of the isotopes and possible spectral interferences in the ICP-MS measurement step. The Cd isotopes 110Cd, 111Cd, 112Cd and 113Cd were measured simultaneously in the blends, as the comparison of the results of the ID ICPMS calculations using different pairs of isotopes can help revealing the

202


presence of bias due to spectroscopic interferences. Together with Cd isotopes the intensities on the mass numbers 90 (Zr), 95 (Mo), 105 (Pd) and 120 (Sn) were also monitored in the blend solutions independently on the strategy applied. The ratios n(200Hg)/n(202Hg) and n(201Hg)/n(202Hg) were selected for Hg mass fraction determination as all isotopes were free of significant spectral interferences and the measurements were done by more sensitive ICP-SFMS method because of low mercury content in the sediment sample. In the case of copper to evaluate the possible polyatomic interferences coming from the ArNa+ formation (m/z = 63) in the plasma the measurements were performed at medium mass resolution or collision cell mode. The pair of 65Cu and 63Cu was the only possibility to determine copper isotopic ratios in the blends. For determination of lead mass fractions in the blend samples the ratio n(208Pb)/n(206Pb) was used. In order to determine lead isotopic composition in the IAEA-458 sediment material all lead isotopes were measured independently in the digested samples. To avoid the interferences from 64Ni on 64Zn the isotope 66Zn was chosen for isotopic ratio determination. The ratio n(66Zn)/n(68Zn) was further used in the ID ICP-MS equation. A careful rinsing followed by a check of the instrumental background was performed prior to every sample measurement in order to monitor sample to sample memory effects and correct for them, if necessary. The instrumentation and summary of the applied measurement modes for selected isotopic ratios of Cd, Cu, Hg, Pb and Zn in the blend samples is presented in Table 2. The signal intensities per replicate were corrected for dead time, instrumental background and possible interferences prior to calculating an average isotopic ratio and its relative standard deviation. The dead time value and its associated standard uncertainty for every element were determined according to correction methods described by Nelms et al. [32]. All isotopic ratios were corrected for mass discrimination effect. Phenomenon of mass discrimination was corrected by bracketing measurements every two samples with an isotopic reference solution (certified isotopic reference material or natural-like isotopic composition standard). The mass discrimination effects were evaluated by measuring the signal intensities of the isotopes of interest then calculating the isotopic ratios and finally calculating the K-factors from the ratio of the certified and the measured values. The mass discrimination correction was straightforward and resulted from the multiplication of the every measured blend ratio by average K-factor measured shortly before and after the blend. The procedural blanks, which were prepared in the same way as the blends were measured on the separate day and were determined by external calibration with 3 different concentrations of element standards (Cd, Cu, Hg, Pb and Zn) prepared by dilution of multi-element Merck standard.

2.7. ID ICP-MS calculation and uncertainty estimation

differentiation described by Kragten [35], was used for the uncertainty calculations. The uncertainties for the spike materials were given on the certificates. The sediment sample was assumed to contain natural isotopic composition of Cd, Cu, Hg and Zn and isotopic composition with associated uncertainties was taken from IUPAC [29]. The lead isotopic composition was determined experimentally by ICP-MS measurements.

Table 3 Equations used for calculation of Cd, Hg Cu, Pb, Hg and Zn mass fractions in the sediment samples.

cx ¼

my mx

cy

Ry −K b Rb δDT K b Rb −Rx δDT

0 n 1 X Rxi C B B i¼1 C δBG δINT C− Bl : 1 B BX C n mx 1−W δBG δINT @ A Ryi i¼1

Kb ¼

RIUPAC ðICRMÞ RIUPAC ðICRMÞ þ =2 Ri δDT δBG δINT Riþ1 δDT δBG δINT

15 X

ðcx Þi

i¼1

C¼

15

Rb ¼

Ia1 Ia2

δWB

corr corr

IBG

corr

¼ Iraw −I BG

IDT

corr

¼

IINT

corr

Iraw 1−Iraw τ

¼ Iraw −IINT

Parameter

Index

x

Isotopes measured for determination of isotopic ratio of element a Blend of sediment and spike solution Correction for instrumental background Blends Isotopic certified reference material Interference Not corrected intensity of isotope Within bottle homogeneity Sediment sample

y

Spike solution

A

Abundance value

a1,a2

Bl

Absolute amount of the element measured in the procedural blank

b

c, C

Amount content, mol·g−1

BG

I

In the present study the set of equations described in Table 3, representing the ID ICP-MS mathematical model was used to calculate Cd, Cu, Hg, Pb and Zn mass fractions in marine sediment sample. The value for each parameter in the described equations was obtained either by a measurement, mathematical calculation or the certificates and it had an associated standard uncertainty. The individual uncertainty components attached to all identified experimental steps involved in the determination were combined together according to ISO guidelines [33]. All uncertainties indicated are expanded uncertainty Uc = k uc where uc is a combined standard uncertainty and k is a coverage factor equal to 2. Combined standard uncertainties on the obtained reference values were obtained by propagating together individual uncertainty components. A dedicated software program [34], based on the numerical method of

−1

i

m

Measured intensity, count·s Mass discrimination correction term Mass, g

INT

M

Molar mass, g·mol−1

raw

n

Number of moles, mol

WB

Kb

R Certified isotopic ratio Ri, Determined isotopic ratio Ri+1 Moisture content of sediment W sample, % τ Dead time, s Unity multiplicative factors δ carrying uncertainty associated to corrections for various effects γ Mass fraction, g·g−1

ICRM

INT_corr DT_corr

Correction for spectral interference Correction for dead time


3. Results and discussions 3.1. Sample preparation The main drawback of microwave methods is the risk of incomplete decomposition of sediment sample matrix when the digestion step (acid mixture, temperature program etc.) is not fully optimized. Analysis of sediment samples can cause particular problems because of difficulties in achieving complete digestion. Incomplete digestion, can be the reason for strong matrix effects visible as the molecular interferences, signal suppression or enhancement. Harsh digestion conditions are achievable with the high pressure systems or with a two-stage microwave-assisted digestion procedure. Incomplete digestion of a sediment material could result in silicate residues when evaporating to dryness, demonstrating remaining matrix substances. The residual matrix can disturb a chromatographic separation and create the matrix effects during isotope ratio determinations. Therefore, carefully developed sample preparation procedure is a key issue for the reference measurements. Several methods of sediment sample decomposition involving various combinations of acids and microwave digestion have been described in the scientific literature [36,37]. In the present study the closed microwave system was applied and the sediment samples were dissolved in the mixture of HNO3, HF and HCl. The use of hydrofluoric acid followed by evaporation near to dryness resulted in less solid residues in the digests, which is advantageous for the following ICP-MS measurements. The addition of small volumes of HCl to the acid mixture improved recovery for Hg due to the formation of less volatile compound of mercury — HgCl2. The important advantages of a microwave digestion include minimizing contamination, lower reagent and sample consumption, reduction of losses of volatile species and additionally decreasing in analysis time. 3.2. Isotope ratio measurements and mass fractions determination It is well known that spectral interferences can significantly obstruct isotopic determination done by ICP-MS. Many isotopes of analytical interest suffer from spectral interferences (isobaric or polyatomic ions) especially when the samples with a complex matrix are analyzed [38, 39,40]. With the aim to overcome such interferences a few solutions can be employed during ICP-MS measurement step. The simplest one, but not always possible, is the selection of interference-free isotopes for analysis. Of course, it is applicable only to poly-isotopic elements and limited to the isotopes with sufficiently high abundance, especially when the concentration of ultra-trace element is an object of determination. The second solution, which allows efficiently overcoming many isobaric interferences, is the use of mathematical equations for interferences correction. In many cases the limiting factor for the application of mathematical corrections is the relationship between the analyte and the interfering ion content, which cannot be changed by dilution. The content of the analyte and the spectral interference can vary from sample to sample even when the samples are expected to be of “the same type”. In the situation when the level of analyte of interest is about quantification limit and the intensity signal of interfering ion is hundreds or thousands times higher application of mathematical correction might not solve the problem of spectral interferences. More recently the advancement in ICP-MS instruments equipped with a collision and/or reaction cell has eliminated or at least minimized the problem of polyatomic interferences observed during determination of some elements [40]. In the case of application of collision/reaction cell the interferences are removed in two ways: either the polyatomic interfering ions are converted to harmless noninterfering species or the analyte is converted into another ion which is not subject to interference by the coexisting ions. One more option is the use of a double focusing magnetic/electrostatic sector instruments which can operate in different resolution

203

modes (low, medium and high). A higher resolution mode allows differentiating the analyte of interest and the interference but reduces the sensitivity, approximately from a relative 100% ions transmission at the lowest to about 1–2% transmission at the highest. However, the resolution offered by ICP-SFMS technique is about 10,000, but it is still not enough to solve the problem of the spectral interferences for some elements. Alternative strategy to solve the problem of spectral interferences is the application of chemical separation. This strategy is not directly related to the instrument performance, but its application can significantly improve the isotopic ratio measurements. Besides some disadvantages (time consuming procedure, possible contamination source and additional reagents usage) chemical separation can be very effective tool for removing interfering ions, especially when they occur in the sample in the amounts much higher than an analyte. The separation step might allow removing the matrix and pre-concentrating the analyte which is very desirable when ultra-trace amounts of elements are determined by ID ICP-MS. The application of one or more of the above mentioned approaches to solve spectral interferences problem depends on the nature of analyte of interest and its content in the sample, the sample matrix and the type of available instrumentation. A successful strategy requires a full understanding of a technique used and detailed knowledge of the sample matrices. The application of one or several approaches to solve the problems with spectral interferences in the present study will be further discussed. 3.2.1. Cadmium Cadmium determination by ICP-MS can easily be affected by isobaric and polyatomic interferences caused by the presence of tin, indium or palladium and from the formation of molybdenum and zirconium oxides during the measurement step. The preliminary investigation has shown that the marine sediment, used in this study, contains zirconium, molybdenum and tin in amounts significantly higher than cadmium and they cause spectral interferences on cadmium isotopes measurement. Due to rather low cadmium mass fraction (about 0.5 mg/kg) and much higher mass fraction of Zr, Mo, and Sn (180 mg/kg, 6 mg/kg and 5 mg/kg, respectively) determination of cadmium in this sediment sample has been recognized as the most challenging task. Elimination or at least minimizing the isobaric and polyatomic interferences originated from the presence of Sn, Mo and Zr was achieved either by matrix separation performed before ICP-MS measurements in standard mode or by the application of collision cell (CC) without matrix separation. Unfortunately, conducting Cd measurements in high resolution mode does not solve the problem with spectral interferences, because the mass resolution below 10,000 is not sufficient for separation of Cd isotopes from molybdenum and zirconium oxides. The developed chromatographic protocol enabled to reduce the concentration of Mo and Zr in the collected fraction by the factor about 8 and 2000. After the separation step cadmium isotopes were corrected only for molybdenum oxides interferences. Moreover, the separation protocol allowed to pre-concentrate Cd by a factor of 2, which was important for the further improvement of the isotopic ratio precision determined in standard mode. The generic disadvantage of using the quadrupole inductively coupled plasma mass spectrometer in collision cell mode was the decrease of sensitivity of about 5 times in comparison to the standard measurement mode. The intensities of cadmium isotopes measured by collision cell after correction for interferences were only about 1000 cps at the mass 110 and about 21,000 cps at the mass 111. If the intensities of cadmium isotopes in blends were not mathematically corrected for 94Zr16O+, 94Mo16O+, and 95Mo16O+ formation, the results calculated for 110Cd/111Cd would have been about 20% higher even after application of collision cell measurement mode. Because of low content of the cadmium in the sediment sample the isotopic ratios of selected Cd isotopes were also determined with the ICP-SFMS at low mass resolution. As shown in Table 4 despite of the

204


Table 4 Comparison of the total element mass fractions in the IAEA-458 marine sediment sample obtained by different mass spectrometry techniques and measurement modes. Element

Measurement technique/mode

Element mass fraction, mg·kg−1 (k = 2)

Cd Cd Cd Cu Cu Cu Hg Pb Zn Zn

ICP-QMS ICP-CC-QMS ICP-SFMS ICP-QMS ICP-CC-QMS ICP-SFMS ICP-SFMS ICP-QMS ICP-QMS ICP-SFMS

0.491 ± 0.008 0.490 ± 0.010 0.491 ± 0.008 48.71 ± 0.44 48.63 ± 0.43 48.80 ± 0.54 0.0469 ± 0.0023 38.2 ± 0.8 157.7 ± 2.8 157.4 ± 3.5

Molar masses used: Cd 112.411 (8) g·mol−1, Cu 63.546 (1) g·mol−1, Hg 200.60 (47) g·mol−1, Pb 207.215 (50) g·mol−1, Zn 65.378 (5) g·mol−1.

difference in intensities obtained with various analytical strategies and ICP-MS techniques, the final results were in very good agreement. This finding shows that above described spectral interferences on Cd isotopes were successfully corrected. 3.2.2. Copper It has been found in the literature that polyatomic interferences (e.g. 40 Ar23Na+, 40Ca23Na+, 36Ar27Al+, 38Ar27Al+) can influence the copper isotopic measurements. To evaluate the impact of the potential interferences the copper isotopic ratios were determined by ICP-QMS, ICP-CCQMS and ICP-SFMS. The isotopic ratio precision in the blends and the solution used for the correction of mass discrimination effect was below 0.3% in all measurements modes. The signal intensities in the diluted blend solutions were comparable when the measurements were carried out with ICP-CC-QMS and ICP-SFMS mass spectrometry techniques at medium mass resolution (R = 4000), assuming that this resolution will be sufficient to separate 63Cu from the interference caused by 40 Ar23Na+ and 40Ca23Na+. The obtained results with the different instruments and measurement modes were in very good agreement within their expanded uncertainties (Table 4), confirming that there was no influence of 40Ar23Na+ polyatomic ions on copper isotopic ratio determination in the sediment sample. Therefore the results obtained by ICP-QMS were selected for further comparison. 3.2.3. Mercury Because of the low content of mercury in the investigated sediment sample the selection of the measurement method was mainly based on the instrument sensitivity. The isotopic pair of 201Hg/202Hg was selected for ID ICP-MS calculations. Both isotopes were free of significant spectral interferences and no mathematical corrections were applied. Mercury isotopic measurements were done by ICP-SFMS, which offered very good sensitivity; however, very low content of Hg in sediment digests still strongly influenced the isotopic ratio repeatability. The repeatability of the n(201Hg)/n(202Hg) isotopic ratios was around 1.7%, about 2–4 times higher than for the other elements. The memory effect for mercury was checked and the rinsing time carefully optimized. It was found out that 10 min rinsing time between the standard and sample analysis is sufficient to decrease mercury background to the initial instrumental background level. The higher instrumental background and poorer isotopic ratio measurement repeatability increased expanded uncertainty for Hg mass fraction in comparison to the other elements. 3.2.4. Lead Due to the relatively high concentration of lead in the investigated sediment sample, lead isotopic measurements were done only with ICP-QMS. The problem of isobaric interference originated from 204Hg was solved by application of the mathematical correction. The isotopic pair of 208Pb/206Pb was selected for ID-MS calculations.

The Pb isotopic composition in the sediment sample was determined by three independent measurements of the isotope ratios of 204 Pb/206Pb, 207Pb/206Pb and 208Pb/206Pb in the digested and diluted sediment sample. As the isotope 204Pb suffers from isobaric interference caused by 204Hg, the 201Hg signal was monitored during the isotopic composition measurement and used to correct the 204Pb intensity which latter was applied for 204Pb/206Pb isotope ratio calculations. The signal intensity for 204Pb was about 250 times higher than the signal of 204Hg in the blend samples. The atomic mass of lead in the sample was calculated to be 207.215 ± 0.001 g·mol−1. 3.2.5. Zinc Zinc was the most abundant element in the analyzed marine sediment, however similarly to copper might suffer from spectral interferences (e.g. 43Ca23Na+, 44Ca23Na+, 36Ar27Al+). Zinc isotopic ratios were measured in the blend samples after 50-times dilution by ICP-QMS and ICP-SFMS (R = 4000) methods without matrix separation. To avoid the interferences from 64Ni on mass 64Zn the other isotope, 66Zn was chosen for isotopic ratio measurements. The ratio n(66Zn)/ n(68Zn) was further used in the ID ICP-MS equation. The isotopic ratio measurement precisions in the blends and the solution used for the correction of mass discrimination effect were below 0.8% in all measurement modes. The total Zn mass fractions and their expanded uncertainties obtained in this study are presented in Table 4. The results obtained by ICP-QMS and ICP-SFMS analysis were in very good agreement with the relative difference between them below 0.2%. Because ICP-SFMS instruments are less common in the analytical laboratories and more complex for the operation, the value derived from ICP-QMS analysis was used as a reference for Zn mass fraction. 3.3. Comparison of the results and their uncertainties The complete protocol deployed for the determination of the Cd, Hg, Cu and Zn mass fractions in the marine sediment sample combined a sample decomposition stage carried out in closed microwave system with “direct” ID ICP-MS measurements. The mass fractions of the investigated trace elements were calculated with the set of equations (Table 3) representing mathematical model of the applied analytical procedure. The individual uncertainty components associated to the corrections for moisture content, procedural blank, homogeneity and stability of the sediment sample, but also to those of the isotope ratio measurement results (instrumental background, spectral interferences, dead time and mass discrimination effects, as well as the repeatability of isotopic ratios measurements) were propagated together. In that way, the combined uncertainty statements (as detailed in Table 3) went far beyond the simple repeatability calculations and reflected the understanding of the measurement process. It is evident that the total uncertainty budget directly enables the analyst to source the main contributors of uncertainty and figure the contribution of the single parameters to the total uncertainty of the adequate element mass fraction in the marine sediment sample. The isotope ratios of the blends were measured as described in the experimental section. From these isotope ratios, the weighing and the spike data, the element mass fractions were calculated using Eq. (1) as presented in Table 3. The standard uncertainty for weighing, procedural blanks, the instrumental background and the moisture content were the minor contributors to the total uncertainty of determined elements. The procedural blank samples for every element followed the same analytical procedure as marine sediment sample. The values obtained for procedural blanks have shown that it was no contamination problem for any of the determined analytes. The mass fractions in the procedural blanks were below 0.5% of the total mass fraction of of any element determined in the marine sediment sample.


Additionally, it was important that uncertainties on IAEA reference values cover for possible lack of within bottle homogeneity. Therefore the combined uncertainty on IAEA-458 values included contribution from uncertainty on homogeneity. Each of the obtained result was multiplied by a unity factor caring uncertainty equal to the within bottle variance (Eq. (3) from Table 3). The estimated expanded uncertainties for Cd, Cu, Pb and Zn were below or equal 2.0%, only in case of mercury it was 4.9%. The main uncertainty contributions to the total uncertainty for all investigated elements are shown on Fig. 1. As can be seen from Fig. 1 the main contributors to the total uncertainty on Cd were the uncertainty on IUPAC Cd isotopic composition (between 29% and 44% of the total uncertainty), uncertainty on 110 Cd/111Cd ratio in the isotopic reference material IRMM-622 (between 25% and 37%) and the uncertainty on blend isotope ratio measurements (between 7% and 30%). The last contributor was visibly higher when

205

ICP-CC-QMS method was used for measurements and it is associated with very low cadmium intensities in the blend samples. As can be seen in Table 4 all analytical protocols applied for Cd isotopic ratio measurements lead to similar results with low expanded uncertainties (Table 5). The expanded uncertainty for Cd mass fraction of 2.1% (k = 2) was obtained when collision cell mode was applied, whereas in the case of ICP-SFMS measurement it was 1.7% (k = 2). Although, the differences in Cd signal intensities in the blends were significant when measured by various mass spectrometry methods, the isotopic ratio precisions were not visibly different. The isotopic ratio measurement repeatability was below 0.5% in all measurements approaches. The main contributor to the total uncertainty on Cu mass fraction determination by ICP-QMS was the uncertainty on 65Cu concentration of the isotopic reference material IRMM 632 (60.4%), followed by the

Fig. 1. Main contributors to the expanded uncertainty for Cd (1a), Cu (1b) and Zn (1c) mass fractions in the IAEA-458 marine sediment.

206


Table 5 Uncertainty budget for the determination of Cd, Cu, Hg, Pb and Zn in marine sediment sample (IAEA-458); s.u. — standard uncertainty, IR — isotope ratio, IC — isotopic composition. Element (measurement mode)

Cd (ICP-QMS)

Cu (ICP-QMS)

Hg (ICP-SFMS)

Pb (ICP-QMS)

Zn (ICP-QMS)

Mass fraction, mg·kg−1 Relative expanded uncertainty (k = 2)

0.491 1.70%

48.71 0.90%

0.0469 4.90%

38.2 2.04%

157.7 1.80%

Main uncertainty contributions, % s.u. on the IUPAC element IC s.u. on the IR in spike CRM s.u. on the blend IR measurements s.u. on the correction for procedural blank s.u. on the correction for mass discrimination effects s.u. on the correction for moisture content s.u. on the isotope concentration in spike CRM s.u. on the sample weighing s.u. on the correction for dead time effects s.u. on the correction for instrumental background others

42.8 36.8 10.9 3.0 2.5 1.2 1.1 1.0 0.3 – 0.4

– – 4.8 – 2.7 4.5 83.2 2.7 1.2 – 0.9

12.2 – 62.8 – 11.9 – 4.2 – 6.2 1.6 1.1

– – 1.1 – – – – 0.6 96.3 0.9 1.1

25.7 – 52.3 1.3 18.4 1.2 – 0.6 0.2 0.1 0.2

uncertainty on the correction for the dead time effects (16.2%) and uncertainty on the correction for mass discrimination effects (9.6%). The expanded uncertainty was below 0.6% for all procedures applied and the obtained results were in very good agreement. The expanded uncertainty for copper mass fraction was below 0.6% for all procedures applied and there was no significant difference within stated uncertainties between all results (Table 5). The main contributor to the total uncertainty on Cu mass fraction determined by ICP-SFMS was the uncertainty on the spike (65Cu) concentration of the isotopic reference material IRMM-632 (60.4%) followed by the uncertainty on the correction for the dead time effects (16.2%) and uncertainty on the correction for mass discrimination effects (9.6%). In case of ICP-QMS determinations the main contributor to the total uncertainty on Cu mass fraction was the uncertainty on 65Cu concentration of the isotopic reference material IRMM-632 (83.2%) followed by the uncertainty on the blend isotope ratio measurements (4.8%). The correction for the isotopic ratio measurements repeatability was the most important contributor (62.8%) to the total uncertainty of mercury determination by ID ICP-MS method. Also the uncertainty on the IUPAC Hg isotopic composition (12.2%) and the correction for the mass discrimination effect (11.9%) were significant factors contributing to the total uncertainty of measurement values. These three factors gave almost 87% to the total uncertainty. The instrumental background and memory effect were also contributing to the combined uncertainty for Hg mass fraction in marine sediments with almost 8% (Table 5). The main contributor to the expanded uncertainty on the determination of total Zn mass fraction by ICP-QMS was the uncertainty on blend ratio measurements (52.3% contribution to the total uncertainty budget), followed by the uncertainty contribution on Zn isotopic composition of the isotopic standard IRMM-3702 (25.7%), uncertainty on the correction for the dead time effects and uncertainty on the correction for mass discrimination effects. The correction for the dead time effects was the most important contributor (96.3%) to the total uncertainty of lead determination and can be directly related to the relatively high signal intensities of lead isotopes (about 1,000,000 cps for 206Pb) in the blend solutions.

3.4. Validation of the measurement procedure and traceability of measurement results

Table 6 Mass fractions for Cd, Hg Cu, Pb and Zn in the marine sediment sample (with the expanded uncertainty, k = 2) obtained in this study and IAEA-458 certified values.

3.5. Conclusions

Element mass fraction, mg·kg−1, (k = 2) Element Cd Cu Hg Pb Zn

ID ICP-MS

IAEA-458 certified values

0.491 ± 0.008 48.71 ± 0.44 0.0469 ± 0.0023 38.2 ± 0.8 157.7 ± 2.8

0.493 ± 0.033 48.1 ± 3.1 0.0437 ± 0.0025 35.5 ± 2.0 154 ± 7

The recommendations of the ISO/IEC 17025 guide on “General requirements for the competence of testing and calibration laboratories” for validation of the measurement process were followed during present study [16]. While uncertainty estimation in and of itself is a validation of an experimental protocol, the ISO guide 17025 also includes the use of a standard reference material for calibration, comparison of results between methods, a systematic assessment of factors influencing the result, and participation in inter-laboratory comparisons as approaches by which validation may be achieved. All of these requirements were undertaken and are presented here. The entire ID ICP-MS measurement process was described by mathematical modeling showed in Table 3. Combined uncertainty budget was calculated for every result. All factors influencing the final results and isotopic equilibrium were systematically investigated. This included the procedural blank (contamination), the moisture content in the sediment material and all factors affecting the blend ratio measurements (instrumental background, spectral interferences, the dead time effect, the mass discrimination and repeatability). The individual uncertainty contributions attached to all identified experimental steps involved in the measurements were combined together according to ISO guidelines. As it is shown in the Table 6 the mass fraction values obtained in the present study were not significantly different within stated uncertainty from the certified values for Cd, Cu, Hg, Pb, Zn achieved in the IAEA-458 certification process [30]. The good agreement between the results additionally validated the analytical procedure we applied for determination of the element mass fractions in the costal sediment reference material. The traceability chain, linking the element mass fraction to the SI units is evidenced in the ID ICP-MS mathematical model given in Table 3. This model, together with the associated equations, subcalculations or the references to the certified values, relates each calculation parameter to the SI units of the mole and the kilogram.

The applied analytical protocol allowed determining the mass fractions of Cd, Cu, Hg, Pb and Zn in the marine sediment with relatively low expanded uncertainties (2%–4.9%). The isotopic measurements for Cd, Cu and Zn were explicitly validated by applying at least two different mass spectrometry techniques or two different measurement modes and using the certified isotopic reference materials for the spikes and the mass discrimination corrections. The excellent agreement between the obtained results showed that all potential problems coming from


the complex sediment matrix and the spectral interferences were solved. The consistency of the results obtained by applying both mass spectrometry techniques confirmed that ICP-QMS can be the method of choice for the reference measurements when the element content in a sample is sufficient for obtaining a good ion counting statistics. The comparable values of the expanded uncertainties estimated for ICP-QMS and ICP-SFMS measurements were additional proofs for the similarity of the measurement performances of both techniques (Table 5) applied in the present study. The reference values have been derived from the implementation of the experimental process (i.e. from sample preparation through the instrumental analysis and the data evaluation steps) entirely described under the form of mathematical equations, and establishing the functional relationship between a measurand and the input quantities. This transparency allowed adequate validation of the analytical procedure, establishing the traceability and comparability of the measurement results and a realistic estimation of the final expanded uncertainty. The expanded uncertainties associated to the values were according to the primary objective of the reference measurements. The good agreement of the achieved results with the certified values of IAEA-458 CRM, derived from its certification campaign, additionally validated developed in this study analytical procedure. Acknowledgments The Agency is grateful for the support provided to its Environment Laboratories by the Government of the Principality of Monaco. References [1] A. Mashiatullah, M.Z. Chaudhary, N. Ahmad, T. Javed, A. Ghaffar, Metal pollution and ecological risk assessment in marine sediments of Karachi Coast, Pakistan, Environ. Mont. Assess. 185 (2013) 1555–1565. [2] L.M. Buruaem, M.A. Hortellani, J.E. Sarkis, L.V. Costa-Lotufo, D.M.S. Abessa, Contamination of port zone sediments by metals from large marine ecosystems of Brazil, Mar. Pollut. Bull. 64 (2012) 479–488. [3] M.S. Reddy, S. Basha, V.G.S. Kumar, H.V. Joshi, Distribution, enrichment and accumulation of heavy metals in coastal sediments of Alang–Sosiya ship scrapping yard, India, Mar. Pollut. Bull. 48 (2004) 378–402. [4] M. Fujita, Y. Ide, D. Sato, P.S. Kench, Y. Kuwahara, H. Yokoki, H. Kayanne, Heavy metal contamination of coastal lagoon sediments: Fongafale islet, Funafuti Atoll, Tuvalu, Chemosphere 95 (2014) 628–634. [5] A. Landajo, G. Arana, A. De Diego, N. Etxebarria, O. Zuloaga, D. Amouroux, Analysis of heavy metal distribution in superficial estuarine sediments (estuary of Bilbao, Basque Country) by open-focused microwave-assisted extraction and ICP-OES, Chemosphere 56 (2004) 1033–1041. [6] P. Wright, C.F. Mason, Spatial and seasonal variation in heavy metals in the sediments and biota of two adjacent estuaries, the Orwell and the Stour, in eastern England, Sci. Total Environ. 226 (1999) 139–156. [7] A. Buccolieri, G. Buccolieri, N. Cardellicchio, A. Dell'Atti, A.D. Leo, A. Maci, Heavy metals in marine sediments of Taranto Gulf (Ionian Sea, Southern Italy), Mar. Chem. 99 (2006) 227–235. [8] J. Morillo, J. Usero, I. Gracia, Heavy metal distribution in marine sediments from the southwest coast of Spain, Chemosphere 55 (2004) 431–442. [9] D. Daby, Coastal pollution and potential biomonitors of metals in Mauritius, Water Air Soil Pollut. 174 (2006) 63–91. [10] N.J. Valette-Silver, The use of sediment cores to reconstruct historical trends in contamination of estuarine and coastal sediments, Estuaries 16 (1993) 577–588. [11] A.B. Cundya, I.W. Croudace, A. Cearretac, M.J. Irabiend, Reconstructing historical trends in metal input in heavily-disturbed, contaminated estuaries: studies from Bilbao, Southampton Water and Sicily, Appl. Geochem. 18 (2003) 311–325. [12] A.C. Ruiz-Fernandez, C. Hillaire-Marcel, F. Paez-Osuna, B. Ghaleb, M. Soto-Jimenez, Historical trends of metal pollution recorded in the sediments, Appl. Geochem. 18 (2003) 577–588. [13] M. Mil-Homensa, J. Blumb, J. Canárioc, M. Caetanoc, A.M. Costaa, S.M. Lebreirod, M. Trancosoe, T. Richterf, H. de Stigterf, M. Johnsonb, V. Brancoc, R. Cesárioc, F. Mouroe, M. Mateuse, W. Boerf, Z. Meloe, Tracing anthropogenic Hg and Pb input using stable Hg and Pb isotope ratios in sediments of the central Portuguese margin, Chem. Geol. 336 (2013) 62–71.

207

[14] M. Beck, P. Böning, U. Schückel, T. Stiehl, B. Schnetger, J. Rullkötter, H.J. Brumsack, Consistent assessment of trace metal contamination in surface sediments and suspended particulate matter: a case study from the Jade Bay in NW Germany, Mar. Pollut. Bull. 70 (2013) 100–111. [15] G.F. Birch, M.A. Olmos, Sediment-bound heavy metals as indicators of human influence and biological risk in coastal water bodies, ICES J. Mar. Sci. 65 (2008) 1407–1413. [16] International Organization for Standarization, ISO/IEC 17025, General Requirements for the Competence of Testing and Calibration Laboratories, Geneva, 2005. [17] B. Magnusson, U. Ornemark (Eds.), Eurachem, The fitness for purpose of analytical methods — a laboratory guide to method validation and related topics, Guide, 2014, www.eurachem.org (ISBN 978-91-87461-59-0). [18] S.L.R. Ellison, B. King, M. Rösslein, M. Salit, A. Williams (Eds.), Eurachem, Traceability in chemical measurement. A guide to achieving comparable results in chemical measurement, Guide, 2003, www.eurachem.org. [19] P. De Bievre, P.D.P. Taylor, Traceability to the si of amount of substance measurements: from ignoring to realizing, a chemist's view, Metrologia 34 (1997) 67–75. [20] International Organization for Standardization, ISO Guide 35:2006, Reference Materials — General and Statistical Principles for Certification, ISO, Geneva, 2006. [21] A. van der Veen, T.P.J. Linsinger, H. Schimmel, A. Lamberty, J. Pauwels, Uncertainty calculations in the certification of reference materials 4. Characterisation and certification, Accred. Qual. Assur. 6 (2001) 290–294. [22] P. Klingbeil, J. Vogl, W. Pritzkow, G. Riebe, J. Müller, Comparative studies on the certification of reference materials by ICPMS and TIMS using isotope dilution procedures, Anal. Chem. 73 (2001) 1881–1888. [23] J. Vogl, Characterisation of reference materials by isotope dilution mass spectrometry, J. Anal. At. Spectrom. 22 (2007) 475–492. [24] K.E. Murphy, S.E. Long, R.D. Vocke, On the certification of cadmium at trace and ultratrace levels in standard reference materials using ID ICP-MS, Anal. Bioanal. Chem. 387 (2007) 2453–2461. [25] K. Inagaki, A. Takatsu, T. Kuroiwa, A. Nakama, S. Eyama, K. Chiba, K. Okamoto, Certified sediment reference materials for trace element analysis from the National Metrology Institute of Japan (NMIJ), Anal. Bioanal. Chem. 378 (2004) 1271–1276. [26] O. Hanousek, L. Rottmann, T. Prohaska, Mass RESOLUTION, in: T. Prohaska, J. Irrgeher, A. Zitek, N. Jakubowski (Eds.), Sector Field Mass Spectrometry for Elemental and Isotopic Analysis, The Royal Society of Chemistry, Cambrige 2015, pp. 97–106. [27] M. Sargent, C. Harrington, R. Harte (Eds.), Guidelines for achieving high accuracy in isotope dilution mass spectrometry (IDMS), Royal Society of Chemistry, Cambrige, 2002. [28] E. Vassileva, C.R. Quétel, I. Petrov, Certification of the Cu and Cd amount contents in artificial food digest using isotope dilution inductively coupled plasma mass spectrometry for pilot study 13 of the Comité Consultatif pour la Quantité de Matière, Spectrochim. Acta B 58 (2003) 1553–1565. [29] M. Berglund, M.E. Wieser, Isotopic composition of the elements 2009 (IUPAC technical report), Pure Appl. Chem. 83 (2011) 397–410. [30] Certification of Trace Element Mass Fractions in IAEA-458 Marine Sediment Sample, IAEA Analytical Quality in Nuclear Applications No. IAEA/AQ/31, IAEA, Vienna, 2013. [31] P. De Bièvre, Trace element analysis in biological specimens, in: R.F. Herber, M. Stoeppler (Eds.), Isotope Dilution Mass Spectrometry (IDMS), Elsevier, Amsterdam 1994, pp. 169–183. [32] S. Nelms, C. Quétel, T. Prohaska, P. Taylor, Evaluation of detector dead time calculation models for ICP-MS, J. Anal. At. Spectrom. 16 (2001) 333–338. [33] International Standards Organization, Guide to Expression of Uncertainty in Measurement, ISO, Geneva, 1993. [34] GUM Workbench, The Software Tool for the Expression of Uncertainty in Measurement, Metrodata GmbH, D-79639 Grenzach-Wyhlen, Germany, 2000. [35] J. Kragten, Calculating standard deviations and confidence intervals with universally applicable spreadsheet technique, Analyst 119 (1994) 2161–2165. [36] K. Inagaki, A. Takatsu, A. Uchiumi, A. Nakama, K. Okamoto, Determination of trace elements in sediment by isotope dilution inductively coupled plasma mass spectrometry using co-precipitate separation technique, Anal. Sci. 17 (2001) (supplement, i991). [37] K. Inagaki, A. Takatsu, A. Nakama, S. Eyama, T. Yarita, K. Okamoto, K. Chiba, Determination of cadmium in sediment by isotope dilution inductively coupled plasma mass spectrometry using co-precipitate separation technique, J. Anal. At. Spectrom. 16 (2001) 1370–1374. [38] S. Révillon, D. Hureau-Mazaudier, Improvements in digestion protocols for trace element and isotope determinations in stream and lake sediment reference materials (JSd-1,JSd-2, JSd-3, JLk-1 and LKSD-1), Geostand. Geoanal. Res. 33 (2009) 397–413. [39] J.W. McLaren, D. Beauchemin, S.S. Berman, Application of isotope dilution inductively coupled plasma mass spectrometry to the analysis of marine sediments, Anal. Chem. 59 (1987) 610–613. [40] A. Makishima, H. Kitagawa, E. Nakamura, Simultaneous determination of Cd, In, Tl and Bi by isotope dilution-internal standardisation ICP-QMS with corrections using externally measured MoO+/Mo+ ratios, Geostand. Geoanal. Res. 35 (2011) 57–67.