Ultrasensitive and Fast Voltammetric Determination of Iron in

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

Ultrasensitive and fast voltammetric determination of iron in seawater by atmospheric oxygen catalysis in 500 µl samples Salvatore Caprara, Luis Miguel Laglera, and Damiano Monticelli Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.5b01239 • Publication Date (Web): 26 May 2015 Downloaded from http://pubs.acs.org on June 1, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 10

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 10

Ultrasensitive and fast voltammetric determination of iron in seawater by atmospheric oxygen catalysis in 500 µl samples Salvatore Caprara,‡ Luis M. Laglera † and Damiano Monticelli‡* ‡

Dipartimento di Scienza e Alta Tecnologia, Università degli Studi dell’Insubria, via Valleggio 11, 22100 Como, Italy

†

FI-TRACE, Departamento de Química, Universidad de las Islas Baleares, Palma, Balearic Islands 07122, Spain

ABSTRACT: A new method based on adsorptive cathodic stripping voltammetry with catalytic enhancement for the determination of total dissolved iron in seawater is reported. It was demonstrated that iron detection at the ultratrace level (0.1 nM) may be achieved in small samples (500 µL) with high sensitivity, no need for purging, no added oxidant and a limit of detection of 5 pM. The proposed method is based on the adsorption of the complex Fe/2,3dihydroxynaphthalene (DHN) exploiting the catalytic effect of atmospheric oxygen. As opposite to the original method (Obata and van den Berg, Anal. Chem. 2001, 73, 2522-2528), atmospheric oxygen dissolved in solution replaced bromate ions in the oxidation of the iron complex: removing bromate reduces the blank level and avoid the use of a carcinogenic species. Moreover, the new method is based on a recently introduced hardware that enables the determinations to be performed in 500 µL samples. The analysis were carried out on buffered samples (pH 8.15, HEPPS 0.01 M), 10 µM DHN and iron quantified by the standard addition method. The sensitivity is 49 nA nM-1 min-1 with 30 s deposition time and the LOD is equal to 5 pM. As a result, the whole procedure for the quantification of iron in one sample requires around 7.5 minutes. The new method was validated via analysis on two reference samples (SAFe S and SAFe D2) with low iron content collected in the North Pacific Ocean.

INTRODUCTION Iron detection in seawater at the ultratrace level is one of the issues at the forefront of elemental analytical chemistry. Due to the low solubility of iron(III), the common redox form at the pH and ionic strength of seawater, iron solubility in organic-free seawater is in the narrow range 10-100 pM 1 with natural concentrations increased by organic complexation to the 0.1-10 nM range 2 . Dissolved iron analysis challenges the analytical chemist to reach extreme detection capabilities, prevent the high risk of sample contamination and tackle the issue of a high salinity matrix. Facing this challenge was encouraged by fundamental research in oceanography based on the oligonutrient role of iron. Low iron concentrations consistently found in HNLC (High Nutrient Low Chlorophill) areas led to the formulation of the “iron hypothesis”, the possibility that iron is the limiting agent for primary productivity (algal growth) in vast areas of the oceans3. This limitation has important consequences on the regulation of the global climate via a reduction of the ability of the oceanic phytoplankton to sink atmospheric CO2 and export a part of it to the deep ocean in a process referred to as the biological pump 4.

Up to now, only cathodic stripping voltammetry (CSV) is able to detect natural concentrations of iron in seawater at the ultratrace level without requiring to a pretreatment step. Mass spectrometric methods require sample preconcentration and matrix modification whereas chemiluminescent and spectrophotometric methods require of sample preconcentration and modification/homogenization of the redox state of iron (see 5 for methods based on preconcentration and mass spectrometry detection and 6 for chemiluminescence and spectrophotometric methods: reviews on iron determination in seawater in 7,8). Accordingly, several methods have been developed based on CSV, with9,10 and without11,12 catalytic enhancement of the signal. Despite the great advantage of being easily adaptable to onboard operation in oceanographic cruises (an antivibration support only is required), these methods did not find extensive application and their use seems to be abandoned in favor of preconcentration procedures coupled to spectrophotometric, chemiluninescent (both easily adaptable to onboard operation) or mass spectrometry detection. As an example, the two extensive intercalibration exercises concerning iron determination in oceanic waters featured a limited number of methods


Page 3 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


based on cathodic stripping: one out of 24 for the IRONAGES exercise13 and four out of 21 for the SAFe intercalibration14 (regarding the latter, the updated version of the reference value features only one method based on CSV out of 29, see http://es.ucsc.edu/~kbruland/GeotracesSaFe/kwbGeotrac esSaFe.html). Further evidence comes from a recent (2012) survey including >13000 dissolved iron data15: 62% were obtained by chemiluminescence, 20% by spectrophotometry, 8% by ICPMS, 6.5% by GFAAS and 3.5% by CSV (if papers instead of number of measurements are used for the calculation, the data change to 54%, 12%, 11%, 12%, 11%, respectively). Recent advancements have nevertheless greatly simplified the application of CSV protocols and detection capabilities have been greatly enhanced by a careful evaluation of the experimental parameters. Dissolved total concentrations were reliably obtained without UV irradiation in open oceanic waters, firstly by acidification to pH 1.7 and a short, 30 s, microwave treatment in 200516, subsequently further simplified in 2013 to acidification only and allowing a few hours time before analysis 17. Improvement of detection limits on the determination of iron concentrations by CSV have been achieved by addition of oxidants (bromate or hydrogen peroxide) that create a chemical catalysis by quick oxidation of the freshly produced Fe(II) before it can diffuse away from the electrode 18 9. Purging time and the need to add an oxidant as the catalytic agent could be eliminated by using dissolved oxygen as the catalytic enhancer via electrode generation of H2O210. The latter procedure used salycilaldoxime (SA) as electroactive ligand and led to a further simplification of the methodology although it was developed for speciation analysis and suffers from a relatively low sensitivity (reported LODs 24-30 pM for a 300 s deposition time10).

In the present paper, we took advantage of all of these recent innovations to define a fast and reliable method for iron determination at the ultratrace level in oceanic waters. The method is based in the catalytic CSV determination of the Fe-DHN complex in the presence of oxygen at the atmospheric partial pressure. Extensive optimization allowed a detection limit of 5 pM with a deposition time of 30 s and 90 s only of equilibration time. The method was validated by the analysis of ultratrace level interlaboratory standards of dissolved iron in oceanic waters. EXPERIMENTAL SECTION Apparatus The measurements were performed on a 757 VA stand (Metrohm) equipped with a three electrode configuration: a mercury hanging drop electrode (0.12 mm2 drop area), a graphite rod as a counter electrode and a reference Ag/AgCl 3M KCl reference electrode. A 2 mL quartz sample cell was used in the present work: details on its construction are reported elsewhere19. The analyses were carried out in an open system under a laminar flow hood (Asalair 1200 FLO), to guarantee an as fast as possible achievement of equilibrium between atmospheric and dissolved oxygen. This system differs from the closed one19 in that four holes are opened in the plastic support of the small cell (see Figure 1). The small volume cell latter could be as well supported from the bottom and the plastic black support completely removed, but we preferred to keep this configuration for practical reasons.

The sample volume requirement of the standard configuration of CSV equipments, 10 mL, has also been recently reduced twenty-fold to 500 µL by a careful design of the sample holder19. This is a clear advantage for systems where the sample available is constrained (interstitial waters, plankton culture monitoring, fractionation subsamples, etc) and implies a smaller utilization of reagents, as opposed to analytical methods requiring pretreatment which necessitate sample volumes of 12-500 mL 8. Detection capabilities, the key factor for iron determination in seawater, were also significantly -

increased for the Fe-DHN/BrO3 system by a fine tuning of the experimental conditions17, leading to a LOD of 5 pM. All of these referred improvements combined could lead to a strongly reduced sample manipulation (no UV pretreatment, fewer added reagents), smaller sample volumes, reduced analysis time (no UV pretreatment, no purge period), higher sensitivity and strongly reduced sample requirement (new sample holder design).

Figure 1 - Images of the new instrumental cell: 1) view through an hole, the positioning of the three electrodes system is visible; 2) view from above of the new cell inserted in its stand; 3) overview of the whole system.

Effect of the partial pressure of oxygen During the optimization step, different oxygen/nitrogen ratios were employed to assess the effect of oxygen on the analytical signal. O2 and N2 pure gases from gas cylinders were mixed by a home made system:



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

different O2 percentage in N2 (80% to 20%) were obtained by mixing adequate fluxes of the two gases as measured by a digital flow meter. The mix was subsequently fed to the purge line of the polarograph. The small cell was placed inside a standard, closed glass cell to ensure that the headspace was filled and the solution saturated with the O2/N2 mixture . Reagents Ultrapure water produced by a Millipore MilliQ A10 system was used throughout (18.2 MΩ cm conductivity, 3 ppb TOC). Pure nitric acid (VWR, Ultrapure NORMATOM for trace metal analysis), hydrochloric acids (Fluka, TraceSELECT® ultra) and ammonia (Fluka, TraceSELECT® ultra) were employed. All plastic materials were cleaned by successive soaking in detergent (one week) and twice in 2% HNO3 (one week each), with ultrapure water rinsing in between steps and prior to usage. A 0.2 M buffer solution was prepared by dissolving the adequate amount of solid HEPPS (3-[4-(2-hydroxyethyl)1-piperazinyl]propanesulfonic acid) (Sigma, ≥ 99.5 %) and adding solid sodium hydroxide (Sigma-Aldrich, 99.99 % trace metal basis) to a final pH of 8.15: the buffer was purified by two equilibrations with a colloidal solution of MnO2 9. The use of NaOH instead of ammonia resulted necessary to obtain reliable results in small volumes: peaks lowered and shifted towards positive potential due to the slow evaporation of ammonia and consequent continuous slight decrease in pH (this trend in peak height and position as a function of pH is consistent with previous observations). This effect is due to the high surface to volume ratio of the sample and the peculiar geometry of the cell (see Figure 1). Seven buffers at different pH (7.12, 7.41, 7.69, 7.89, 8.30, 8.60, 8.93) were prepared for pH dependence experiments by addition of adequate volumes of NaOH or HCl solutions to the HEPPS buffer solution already described. These buffer solutions were purified by equilibration with Chelex 100® (Fluka) before use as the added NaOH solution caused a small iron contamination. A 2,3-dihydroxynaphthalene (DHN) solution (2 mM) was prepared dissolving DHN (Aldrich, ≥ 98.0 %) in ultrapure water. A 10 mM DHN solution was also prepared to test the effect of DHN concentration on sensitivity and LOD. A diluted standard solution, 5 nM in iron, was prepared from a 1000 mg/L standard from Fluka and acidified to pH 3 by concentrated HCl. A 1 M solution of NaOH was prepared by dissolving solid sodium hydroxide (Sigma-Aldrich, 99.99 % trace metal basis) and purified by three equilibrations with a colloidal solution of MnO2. Limits of detection under different conditions (oxygen percentage, square wave frequency, DHN concentration and pH) were calculated applying the IUPAC

Page 4 of 10

recommendation as 3.3•sblank/sensitivity20, where sblank is the standard deviation of the blank signal. The samples used for validation were the deep and surface reference waters collected during the SAFe (Sampling and Analysis of Fe) program14. Consensus values are as follows: SAFe S 0.097±0.043 nM and SAFe D2 0.91±0.17 nM14. Analytical procedure for the determination of the iron total concentration in seawater All the sample manipulations and the voltammetric determinations were performed in a laminar flow hood. The small cell was filled with a blank (500 µL of ultrapure water, 10 mM HEPPS and 10 µM DHN) in between analyses to avoid carry over due to adsorption on the cell wall. The required amount of acidified sample (500 µL per intended replicate) was transferred to a 60 mL bottle. 10 µL of the 1 M NaOH solution were added per 500 µL of transferred sample to neutralize it. This procedure ensured that the acidic sample is not transferred into the small cell, as it was observed that significant changes in the pH of the solution contained in the small cell caused erratic results. 500 µL of the neutralized sample were pipetted into the voltammetric cell. Ultraviolet digestion was not necessary prior to analysis, as already demonstrated in a previous paper17. A 25 µL aliquot of HEPPS pH buffer (final concentration 0.01 M, pH 8.15) and 50 µL of the DHN solution (final concentration 10 µM) were subsequently added. The solution was stirred for 90 sec, and three replicate measurements were registered. The measurement was repeated after a standard addition of iron (III) of appropriate concentration to calibrate the sensitivity. The instrumental parameters used for the determination are listed in Table 1: the optimization of the sweep mode and square wave frequency is discussed in the following sections, whereas the optimization of the deposition potential lead to the same value (-0.1V) as reported in the literature 9,17. Table 1 - Instrumental parameters for the square wave sweep. Instrumental parameter Purging time (s)

0

Stirring time (s)

90

Deposition potential (V)

-0.1

Equilibration time (s)

10

Deposition time (s)

30

Start potential (V)

- 0.35

End Potential (V)

- 0.75

Voltage step (V)

0.005

Amplitude (V)

0.05


Page 5 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 10

Frequency (Hz)

The initial 90 s stirring time was the minimum time needed to achieve a stable signal: such an equilibration time may be needed to allow the DHN to outcompete the natural ligands (samples were not UV digested) or to reach the equilibrium oxygen concentration in the sample. Blank levels were measured for each analysis batch and subtracted from sample concentration: contribution to the blank came from the pH buffer mainly (80 %) and from the sodium hydroxide solution (20 %), giving a stable blank value of 86±1.5 pM (n=5). RESULTS AND DISCUSSION Dihydroxynaphthalene was used as the ligand in this work because it gives the best result in terms of signal to noise ratio9,17. Data are presented for both low ionic strength media (ultrapure water) and seawater: although the focus is on seawater analysis, the application to low ionic strength samples may be of interest21, although real freshwater samples were not analyzed in the present work. Oxygen catalytic effect The first evidence of a catalytic effect of oxygen on the reduction of the Fe-DHN complex was observed when the solution containing DHN and the pH buffer was not purged with nitrogen in the absence of bromate; a CSV wave was present and replicates showed a decreasing signal due to the voltammetric stand continuously purging the headspace with nitrogen thus decreasing the oxygen present in solution (purging of the cell headspace is an unavoidable feature of the Metrohm stand). This observation led to the discovery of the catalytic effect of oxygen and the design of the open cell to warrant a continuous flux of atmospheric oxygen to the sample surface. The catalytic effect caused by oxygen was demonstrated by blowing mixtures of nitrogen and oxygen with increasing oxygen percentages in the cell headspace and measuring the resulting sensitivities (see Figure 2).

Figure 2 - Effect of percentage of oxygen on the voltammetric signal in seawater and ultrapure water. The voltammetric cell contained 500 µL of ultrapure water or seawater, 1 nM of iron, 10 mM of HEPPS buffer and 2 µM of DHN, the scan was preceded by 30 sec of deposition time at - 0.1 V. RSD% of the sensitivities were between 2.5% and 5%.

Increasing the percentage of O2 caused the sensitivity to grow to very high values for both the matrices: sensitivities of 160 and 90 nA/nM min-1 for ultrapure water and seawater, respectively, were measured in 80% oxygen. The sensitivity was around twice in ultrapure water with respect to seawater (1.9 on average). The almost linear relationship between O2 volume percentage and sensitivity is a good indication of the catalytic effect of oxygen in this system10,22. Increasing oxygen percentage, i.e. the concentration of the catalysts, is the main way to increase sensitivity, as is always true for catalytic systems in cathodic stripping voltammetry22. Nevertheless, no cost in terms of blank level is paid when oxygen partial pressure is raised, marking a difference with the usually employed oxidizing salts that significantly contribute to blank levels. Taking advantage of this very high sensitivities would result in an increase in the complexity of the apparatus, as an oxygen cylinder and a gas flow regulation system would be needed, which may be an issue during field analysis. Experimental evidence suggests that hydrogen peroxide generated by oxygen reduction during the deposition time at -0.1V is responsible for the catalytic effect, as already suggested for the analogous method based on salicylaldoxime10. The voltammetric peak is actually at the positive potential end, -0.58 V, of the peak due to the reduction of H2O2 to H2O, whereas O2 reduction to H2O2 occur at much more positive potentials. The iron complex acts as a catalyst in this reaction, as it facilitate the electron transfer among the electrode and the H2O2 molecule. The latter is not a catalyst as it is the final electron acceptor: the usual representation of the oxidant being the catalyst that reoxidizes the freshly formed reduced species of the electroactive metal complex is not the correct description of the mechanism 22. Hydrogen peroxide catalytic effect Once established the catalytic effect caused by oxygen, hydrogen peroxide was also tested. The latter was already demonstrated to catalytically enhance the signal in the original paper on iron determination in seawater employing DHN9. The results of these experiments were reported in Figure 3: sensitivity, as in the case of oxygen, linearly increases with hydrogen peroxide concentration, in accordance with the results of Obata et al.9. Nevertheless, the signal became unstable as the H2O2 concentration was raised, leading to a deterioration of the LOD at hydrogen peroxide concentrations higher than 1 mM. The best performances were obtained with a concentration of hydrogen peroxide equal to 1 mM: sensitivity 34 nA nM-1 min-1 and LOD 36 pM (30 s



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 10

deposition time). Besides its possible use for analytical purposes, the enhancement of the analytical signal caused by hydrogen peroxide is a further indication, although indirect, that it is responsible for the catalytic effect in the iron DHN system.

Figure 4 - Comparison between different square wave frequencies in seawater. The sensitivity and the value of the LOD were calculated for each frequency (0.5 nM Fe, 10 mM HEPPS, 10 µM DHN, 30 s deposition time).

Figure 3 - Effect of variation of H2O2 concentration in ultrapure water. Same experimental conditions as Figure 2.

OPTIMISATION OF INSTRUMENTAL PARAMETERS The new method based on catalytic cathodic stripping voltammetry of DHN iron complex in the presence of oxygen was thoroughly optimized. Seawater samples were not UV digested during the optimization stage as UV digestion was not required during analysis. Scan mode Linear scan, differential pulse (DP) and square wave (SqW) scan modes were investigated: preliminary tests showed a very high, around 200 nA, baseline using a linear scan, whereas DP and SqW showed a much lower baseline (40-50 nA). The best result in terms of signal to noise ratio was obtained with the square wave modulation. The frequency of the square wave was subsequently optimized in the range between 1 Hz and 500 Hz. The signal increased up to 10 Hz, then it remained constant up to 25 Hz and finally decreased down to be undetectable at the highest investigated frequencies because of a huge increase of the baseline up to around 500 nA. The range between 5 and 50 Hz was better investigated to assess the best value of this parameter: four different frequency were chosen (5, 10, 25 and 50 Hz). Sensitivities and LODs were calculated for each frequency value (Figure 4). Sensitivity and LOD showed the same behavior in seawater (Figure 4) and ultrapure water (Figure S - 1). The 10 Hz modulation was the best choice in terms of LOD as it benefits from a high sensitivity (50 nA/nM min-1) and a limited baseline level (around 40 nA at the iron peak maximum). Accordingly, all the following analyses were carried out using a square wave sweep with a 10 Hz modulation.

DHN concentration The concentration of the added ligand is known to exert a great effect on the analytical signal: accordingly, its effect was systematically investigated in both ultrapure water and seawater. Figure 5 shows that the sensitivity increased almost linearly with DHN concentration up to 500 µM, decreasing thereafter (note the logarithmic scale for DHN concentration). The measured limits of detection followed a peculiar trend: they had a minimum at 2 and 10 µM DHN for ultrapure water and seawater, respectively, and raised steadily for higher DHN concentrations. This trend is due to the signal continuously increasing when DHN concentrations higher than 10 µM are used. This variability is partially offset by a giant sensitivity (up to 555 nA/nM min-1 in ultrapure water, 229 nA/nM min-1 in seawater) at very high DHN concentrations, but the S/N ratio is nevertheless decreasing. These experimental conditions, although ensuring the highest sensitivity ever registered in CSV, are not suitable for analytical purposed as the signal is constantly raising leading to a very low reproducibility. External pollution is a possible reason for such behavior. All the experiments were carried out under a laminar flow hood, but the measuring cell has four big holes (see Figure 1), needed to assurance the oxygen equilibrium between the solution and the atmosphere, greatly facilitating the pollution of the samples. An addition possible source of contamination is through the porous ceramic frit that separates the reference electrode salt bridge and the sample: a potential way out for this problem would be the use of a pseudo reference electrode, such as the one introduced in 23. Experiments carried out in the 10 mL cell with the classical glass salt bridge and the polytetrafluoroethylene salt bridge with the commercial porous ceramic frit showed that the first would cancel a slow signal increment detected with the polytetrafluoroethylene salt bridge. For the same reason it


Page 7 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


has been suggested to fill the reference salt bridge with clean seawater 24. It is interesting noting that such a behavior, i.e. increasing peak current, was observed with high DHN concentrations only, as this solutions may be more easily contaminated. Given these experimental results, a ligand concentration equal to 2 µM for the analysis of low ionic strength sample is proposed: these experimental conditions allow a limit of detection of 10 pM and a sensitivity of 49 nA nM-1 min-1. A higher DHN concentration, 10 µM, ensured the best limit of detection for seawater: 5 pM for 30 s deposition time with a sensitivity of 49 nA nM-1 min-1. The higher DHN concentration required for seawater to achieve the same sensitivity as in ultrapure water, may be easily explained with the need to outcompete major ions (especially calcium) which may compete with iron for complexation with DHN. Regarding the different sensitivity registered in ultrapure water and seawater, a first factor is oxygen solubility: oxygen is about 25 % more soluble in freshwater than in sea water (35 psu) at 20 °C (6.35 vs 5.17 ml/L of oxygen for 0 and 35 psu, respectively25), leading to a parallel increase in sensitivity (see Figure 2, S vs O2 %). A second important factor is the presence of natural organic matter in the seawater sample as the procedure does not require UV digestion prior to analysis. Natural organic matter has already been demonstrated to cause a reduction in sensitivity in stripping voltammetry 26. Differences in the electrode-solution double layer composition, in iron speciation and electron transfer kinetics may also play a role, although such a deep knowledge of the system is presently not at hand.

enhancement in the presence of O2 or H2O2, has been achieved with a standard 10 mL cell in the same conditions and with a different ligand (salicylaldoxime), but the sensitivity values reported here were never achieved (see the section Comparison with other CSV methods). As discussed in that section, we think this feature must be due to the easy oxygen diffusion into the reduced sample volume. pH The effect of pH on sensitivity and LOD is reported in Figure 6 for ultrapure water: as explained in the experimental section, eight pH buffers were prepared at different pH by the addition of sodium hydroxide or hydrochloric acid and purified to avoid any contamination introduced by NaOH or HCl solutions. In ultrapure water, the fivefold increase in sensitivity in the pH range 7-9 was analogous to the trend previously reported by some of us in the presence of bromate as the oxidizing agent17. Nevertheless, this increase in sensitivity was compensated by a steady decreasing of the signal at pH higher than 8.15 which led to an increase in the LOD values (see Figure S - 2). The signal was instead stable for around 20 min at pH 8.15 and 7.9: this stability window is adequate for a two standard addition procedure with five replicates, if 30 s deposition time is used.

Figure 6 - Effect of pH on sensitivity and LOD in ultrapure water. The voltammetric cell contained 500 µL of ultrapure water, 0.5 nM of iron, 10 mM of HEPPS buffer and 2 µM of DHN, the scan was preceded by 30 sec of deposition time at 0.1 V.

Figure 5 - Variation caused by the concentration of DHN in terms of sensitivity and LOD, in ultrapure water and seawater. All the solutions were 10 mM HEPES and 0.5 nM Fe.

This very high sensitivity is a peculiar feature of the cell configuration used: the same effect, i.e. signal

Accordingly, the optimal pH for the new method is between 7.90 and 8.15. A much better stability was found in seawater (pH 8.15) compared to ultrapure water: a relative standard deviation of 3% was registered for 30 repeated measurements (30 min) without any clear trend. The optimum conditions matching the range of natural seawater pH is extremely convenient as this method with minimum modifications could be used for determining



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 10

Table 2 - Comparison with other CSV methods. LOD, sensitivity, deposition time and ligand concentration are reported. CSV method

Ligand concentration

Dep. time

LOD

Sensitivity -1

Ref. -1

µM

s

pM

nA nM min

Fe/TAC

10

300

100

4.3

11

Fe/SA

27.5

600

10

16

27

Fe/NN

20

60

90

3.1

12

20

30

80

6.2

28

-

20

300

13

7.8

9

Fe/DHN/BrO3 (pH 8.7)

-

30

90

5

34

17

Fe/SA/O2

5

300

24–30

4.4

10

Fe/DHN/O2

10

30

5

49

this study

Fe/NN/BrO3

-

Fe/DHN/BrO3 (pH 8.0)

the organic speciation of iron in seawater by competitive ligand equilibration – cathodic stripping voltammetry, CLE – CSV. Figures of merit Under optimized conditions, the limit of detection (LOD) in ultrapure water and seawater were 10 pM (2 µM DHN) and 5 pM (10 µM DHN), respectively, with a deposition time of 30s: sensitivity close to 50 nA nM-1 min-1 was registered in the two matrices. Preliminary experiments showed that sensitivity may be further increased by increasing the deposition time: sensitivity increased up to a deposition time of 120 seconds in seawater and leveled thereafter (Figure S3). The limit of linearity (LOL) at a deposition time of 30 s was estimated as 0.6 nM in ultrapure water and 1 nM in seawater: this range may be easily expanded by reducing the deposition time if needed (Figure S - 3). Analysis of reference samples Two Pacific Ocean samples14 with consensus values after interlaboratory determination of the iron concentrations were analyzed to validate this new method: the higher concentration one was diluted to keep its concentration at the ultratrace level (Table 2). A very good agreement was observed: all the experimental data did not show statistically significant difference from the standard ones (two tailed t test, p = 0.05). Table 3 - Results of cathodic stripping voltammetry of Pacific Ocean samples with consensus value for ultratrace level of iron. Sample

Dilution

Consensus value Found nM

SAFe S

none

0.097 ± 0.043

SAFe D2

1:10 1:2

0.91 ± 0.17

n*

0.096 ± 0.005 6 1.00 ± 0.03

5

1.02 ± 0.02

3

Comparison with other CSV methods Table 3 reports the values of LOD and sensitivity for previous research using CSV. Obata et al.9 in the work introducing DHN as a ligand for iron detection in seawater reported a LOD equal to 13 pM for the system Fe/DHN/BrO3-, Laglera et al.17 with their optimization lowered this value down to 5 pM. This study maintains the value of LOD at 5 pM but using only 30 sec of deposition time (3 fold shorter). Moreover, the sensitivity was improved 1.5 fold, up to around 50 nA nM-1 min-1. A five hundred microliters sample size was used in the present study as opposed to 10 mL in all the other CSV methods. The value of 5 pM is up to date the best LOD value for CSV methods: a better value (2 pM) was achieved only by the method based on Mg(OH)2 coprecipitation followed by ICP-MS determination29 which requires a 50 mL sample. The very high sensitivity registered in this study may be due to several factors. We think the main one is the possibility for oxygen to fast diffuse to the electrode because of the limited, around 1 mm, solution thickness needed to reach the electrode from the gas phase above the sample. This allows a much higher oxygen flux during the deposition step, resulting in a higher hydrogen peroxide concentration in the diffusion layer and finally in a higher catalytic enhancement of the signal (see Figure 3). Moreover, the more oxygen is reduced, the higher the resulting pH in the diffusion layer, as this reaction consumes hydrogen ions 30: a higher pH was demonstrated to increase sensitivity in the bulk (see Figure 6) and could as well cause an increase in sensitivity when the complex experience a more alkaline pH in the diffusion layer. Support for this hypothesis came from preliminary experiments in the 10 mL cell conducted in the same conditions presented here: atmospheric oxygen gave a catalytic enhancement of the signal, but the resulting sensitivity was three times smaller. CONCLUSIONS

*number of replicated analyses


Page 9 of 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


A new method for iron determination in seawater at the ultratrace level was proposed and validated. The method is based on CSV of the complex formed by dihydroxynaphthalene and iron with catalytic enhancement by dissolved oxygen. It enables the quantification of iron in a sample volume of 500 µL down to a limit of detection of 5 pM employing 30 s of deposition time. Such a low LOD is due to a very high sensitivity (50 nA nM-1 min-1), a moderate and stable baseline (around 25 nA) in conjunction to a good signal stability. The proposed method also eliminates the use of bromate, a suspected carcinogen, which was previously employed as an oxidant to achieve the catalytic enhancement. Although the detection capabilities were demonstrated to be suitable for the determination of the lowest iron concentrations in seawater, directions for their further enhancement may be devised. In particular, beside increasing the deposition time, raising the oxygen partial pressure seems a strategy at hand, although at cost of an increase in system complexity. Alternatively, raising the ligand concentration from 10 µM to 500 µM lead to an impressive increase in sensitivity, although the deterioration in signal stability limited the gain in signal to noise ratio. A cleaner environment (a clean room) and the use of a pseudoreference electrode free from liquid junctions could stabilize the signal as contamination was proposed to be the cause of the raising signal observed at high ligand concentrations. A better understanding of the intimate mechanism of the electrode reaction and the role of oxygen diffusion in solution would provide solid ground for this method. Ongoing research aims at measuring the heterogeneous transfer rate from the electrode to the iron complex and, subsequently, from the complex to hydrogen peroxide: molecular modeling is expected to help in appreciating mechanicistic aspects. Moreover, the proposed mechanism for the high sensitivity, i.e. the fast oxygen diffusion in the sample due to the high surface to volume ratio, is being assessed. The possibility to apply this method for the speciation analysis of iron by the competitive ligand equilibration procedure (CLE-CSV) is under study, although at present the time stability of the signal has been verified for short periods of time only. Finally, the proposed method awaits an extensive application, possibly in the field, to thoroughly assess its analytical capabilities.

ASSOCIATED CONTENT Supporting Information. Figures S1-S3 including additional experimental results.

AUTHOR INFORMATION Corresponding Author

Author Contributions The manuscript was written through contributions of all authors.

REFERENCES (1) Liu, X. W.; Millero, F. J. Mar. Chem. 2002, 77, 43. (2) Johnson, K. S.; Michael Gordon, R.; Coale, K. H. Mar. Chem. 1997, 57, 137. (3) Martin, J. H. Paleoceanography 1990, 5, 1. (4) Longhurst, A. R.; Harrison, W. G. Progress in Oceanography 1989, 22, 47. (5) Biller, D. V.; Bruland, K. W. Marine Chemistry 2012, 130–131, 12. (6) Bowie, A. R.; Sedwick, P. N.; Worsfold, P. J. Limnology and Oceanography: Methods 2004, 2, 42. (7) Achterberg, E. P.; Holland, T. W.; Bowie, A. R.; Mantoura, R. F. C.; Worsfold, P. J. Analytica Chimica Acta 2001, 442, 1. (8) Worsfold, P. J.; Lohan, M. C.; Ussher, S. J.; Bowie, A. R. Mar. Chem. 2014, 166, 25. (9) Obata, H.; van den Berg, C. M. G. Analytical Chemistry 2001, 73, 2522. (10) Abualhaija, M. M.; van den Berg, C. M. G. Marine Chemistry 2014, 164, 60. (11) Croot, P. L.; Johansson, M. Electroanalysis 2000, 12, 565. (12) Gledhill, M.; van den Berg, C. M. G. Mar. Chem. 1995, 50, 51. (13) Bowie, A. R.; Achterberg, E. P.; Croot, P. L.; De Baar, H. J. W.; Laan, P.; Moffett, J. W.; Ussher, S.; Worsfold, P. J. Mar. Chem. 2006, 98, 81. (14) Johnson, K. S.; Elrod, V.; Fitzwater, S.; Plant, J.; Boyle, E.; Bergquist, B.; Bruland, K.; Aguilar-Islas, A.; Buck, K.; Lohan, M.; Smith, G. J.; Sohst, B.; Coale, K.; Gordon, M.; Tanner, S.; Measures, C.; Moffett, J.; Barbeau, K.; King, A.; Bowie, A.; Chase, Z.; Cullen, J.; Laan, P.; Landing, W.; Mendez, J.; Milne, A.; Obata, H.; Doi, T.; Ossiander, L.; Sarthou, G.; Sedwick, P.; Van den Berg, S.; Laglera-Baquer, L.; Wu, J.-f.; Cai, Y. Eos, Transactions American Geophysical Union 2007, 88, 131. (15) Tagliabue, A.; Mtshali, T.; Aumont, O.; Bowie, A. R.; Klunder, M. B.; Roychoudhury, A. N.; Swart, S. Biogeosciences 2012, 9, 2333. (16) Bruland, K. W.; Rue, E. L.; Smith, G. J.; DiTullio, G. R. Marine Chemistry 2005, 93, 81. (17) Laglera, L. M.; Santos-Echeandia, J.; Caprara, S.; Monticelli, D. Analytical Chemistry 2013, 85, 2486. (18) Yokoi, K.; van den Berg, C. M. G. Electroanalysis 1992, 4, 65. (19) Monticelli, D.; Laglera, L. M.; Caprara, S. Talanta 2014, 128, 273. (20) Currie, L. A. Pure and Applied Chemistry 1995, 67, 1699. (21) Nagai, T.; Imai, A.; Matsushige, K.; Fukushima, T. Aquatic Microbial Ecology 2006, 44, 231. (22) Bobrowski, A.; Zarȩbski, J. Electroanalysis 2000, 12, 1177. (23) Beati, A. A. G. F.; Reis, R. M.; Rocha, R. S.; Lanza, M. R. V. Industrial & Engineering Chemistry Research 2012, 51, 5367. (24) van den Berg, C. M. G. Anal. Chem. 2006, 78, 156. (25) Weiss, R. Deep Sea Research and Oceanographic Abstracts 1970, 17, 721.

* Damiano Monticelli, [email protected]; tel. +390312386427



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(26) Louis, Y.; Cmuk, P.; Omanović, D.; Garnier, C.; Lenoble, V.; Mounier, S.; Pižeta, I. Anal. Chim. Acta 2008, 606, 37. (27) Rue, E. L.; Bruland, K. W. Mar. Chem. 1995, 50, 117. (28) Aldrich, A. P.; van den Berg, C. M. G. Electroanalysis 1998, 10, 369.

Page 10 of 10

(29) Wu, J. Marine Chemistry 2007, 103, 370. (30) Aguilar, D.; Galceran, J.; Companys, E.; Puy, J.; Parat, C.; Authier, L.; Potin-Gautier, M. Phys. Chem. Chem. Phys. 2013, 15, 17510.


Ultrasensitive and Fast Voltammetric Determination of Iron in

Recommend Documents