Subscriber access provided by KING MONGKUT'S UNIVERSITY OF TECHNOLOGY THONBURI (UniNet)

Article

Exhaustive Thin Layer Cyclic Voltammetry for Absolute Multianalyte Halide Detection Maria Cuartero, Gastón A. Crespo, Majid Ghahraman Afshar, and Eric Bakker Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/ac503344f • Publication Date (Web): 15 Oct 2014 Downloaded from http://pubs.acs.org on October 18, 2014

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 31

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

Exhaustive Thin Layer Cyclic Voltammetry for Absolute Multianalyte Halide Detection Maria Cuartero, Gastón A. Crespo, Majid Ghahraman Afshar and Eric Bakker* Department of Inorganic and Analytical Chemistry, University of Geneva, Quai Ernest-Ansermet 30, CH1211 Geneva, Switzerland.

ACS Paragon Plus Environment

1

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

Page 2 of 31

ABSTRACT Water analysis is one of the greatest challenges in the field of environmental analysis. In particular, seawater analysis is often difficult because of large amount of NaCl may mask the determination of other ions, i.e., nutrients, halides, and carbonate species. We demonstrate here the use of thin layer samples controlled by cyclic voltammetry to analyze water samples for chloride, bromide and iodide. The fabrication of a microfluidic electrochemical cell based on an Ag/AgX wire (working electrode) inserted in a tubular Nafion membrane is described, which confines the sample solution layer to less than 15 µm. By increasing the applied potential, halide ions present in the thin layer sample (X-) are electrodeposited on the working electrode as AgX, while their respective counterions are transported across the permselective membrane to an outer solution. Thin layer cyclic voltammetry allows us to obtain separated peaks in mixed samples of these three halides, finding a linear relationship between the halide concentration and the corresponding peak area from about 10-5 to 0.1 M for bromide and iodide and from 10-4 to 0.6 M for chloride. This technique was successfully applied for the halide analysis in tap, mineral and river water as well as seawater. The proposed methodology is absolute and potentially calibrationfree, as evidenced by an observed 2.5% RSD cell to cell reproducibility and independence of operating temperature.

KEYWORDS: thin layer, halide detection, argentometry, and cyclic voltammetry.

ACS Paragon Plus Environment

2

Page 3 of 31

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

INTRODUCTION Low cost microfluidic thin layer samples controlled by dynamic electrochemistry techniques have been put forward recently by our group and others in order to establish a series of robust and reliable sensors for environmental and clinical analysis.1-4 Initial reports were more focused on fundamental studies, where cations such as potassium and calcium were measured in the range of 0.01-0.1 mM.5-8 The integrated charge (coulometry) that arises from the cations transported through the ion-selective membrane by an applied potential of appropriate amplitude was found to be linear with concentration. Subsequent work was centered on instrumental improvements and characterization of the lower limit of detection.9,10 Even more recently, the polyions protamine/heparin were also successful determined by this methodology in the relevant therapeutic range.11-14 Using a similar thin layer configuration, but replacing the polypropylene ion selective membrane by a tubular commercial Nafion membrane (40 cm), a desalinator unit was developed.15 Nafion has been used in several electrochemical devices as a permselective cation-exchange membrane.16,17 In our case, and as a result of the thin layer concept, this interesting system allows one to reduce the chloride concentration in seawater (or other halides) down to mM levels in the entire injected sample plug. Eliminating the strongest interfering anion in seawater (0.6 M NaCl), the remaining anionic species may be more easily measured by coupling the desalinator unit with a complementary readout of choice located downstream. Instead of physically detaching the desalination and readout electrochemistry as suggested above, the same cell, even shorter (4 cm), may be used for measuring halides, and in fact to resolve mixtures of halides. The halides chloride, bromide and iodide are routinely measured by ion-chromatography18 or volumetric titration.19 As a consequence of many years of scientific efforts, the first methodology is today very reliable, but disadvantages typical of bench-top devices remain. This includes the cost of analysis (pumps, columns, injector, quality of solvents, etc.), the lack of portability, a time consuming sample preparation (filtering at 0.22 mm pore size), and the need for trained personnel to operate the equipment.

ACS Paragon Plus Environment

3

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

Page 4 of 31

Additionally, for applications where in-situ or on site measurements are required, this separation methodology may not be the most suitable choice, also because a simple halide analysis takes about 15 min without considering the external calibration curve. The second methodology is based on titrimetric analysis by adding AgNO3 solution to a sample aliquot, accompanied by visual (formation of the colorful precipitate), potentiometric (Ag based selective membrane) or conductometric sensors as endpoint detectors. Even though argentometric analysis allows one to measure individual halides at relatively high concentrations (higher than 10 mM), a mixture of three halides is difficult to resolve since coprecipitation of silver halides may take place. In addition, of course, this type of procedure still requires sampling and the splitting into aliquots and does not lend itself to detections in situ. In view of providing an alternative route for measuring halides in environmental samples, we explore here the exhaustive electrochemical conversion of a ca. 15-µm thick aqueous layer. As a proof of concept, slow scan cyclic voltammetry is used to oxidize a silver element as plated silver halide salts. Importantly, the use of a thin layer sample enhances some electrochemical characteristics with respect to experiments with a high volume to surface ratio,20-22 including chemical resolution (for three halides), sensitivity of the peaks (height and peak width) and diminution of the peak separation, which renders the system highly reproducible between determinations. The results are accompanied with numerical simulations to help understand the electrochemical basis of this system.

ACS Paragon Plus Environment

4

Page 5 of 31

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

THEORY The proposed model assumes a bare silver (Ag) wire in contact with a confined volume of NaCl. When an anodic sweep potential of appropriate magnitude is applied to the cell, a surface layer of Ag is oxidized (Eq 1), resulting in the formation of a solid AgCl according to the solubility constant (Eq 2). Accordingly, chloride is first consumed at the metal electrode surface and consequently in the entire thin aqueous layer. At the same time, the counter ion of chloride migrates across the Nafion membrane that contacts the other side of the thin sample layer. We assume that the observed current is limited by the diffusion of chloride. Although [AgCl2]- is known to be the dominant dissolved complex species for up to 1 M chloride concentrations,23 the solubility of silver chloride remains comparatively small and the formation of AgCl is assumed to be complete in this calculation. It is also assumed that other processes such as adsorption or charge transfer occur much faster than ion diffusion in aqueous solution. The electrochemical conversion process is based on the following reduction and precipitation reactions:

,

,

,

o Ered = 0.799 V , Ag + / Ag

(1)

K s = 1.8 ×10−10

(2)

o E red , AgCl / Ag = 0.222 V

(3)

For solving the differential partial equations, we define a one-dimensional space (x) that evolves in a time (t) grid with the two respective finite elements (∆x, ∆t). The metal electrode is placed at position 0 and the Nafion membrane at element xmax. The forward scan is performed in a defined interval of time (0≦ t ≦ tmax) and the backward scan between (tmax ≦ t ≦ 2 tmax), which modulates the applied potential starting from the initial potential, Ecell(0), as follows:

Ecell (t) = Ecell (0) + υt

0 ≦t ≦tmax

ACS Paragon Plus Environment

(4)

5

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

Ecell (t) = Ecell (0) + 2υtmax − υ t

Page 6 of 31

tmax ≦t ≦2 tmax

(5)

Re-writing the Eqn. 3 with the Nernst equation, we reach Eqn. 6, where cCl(0,t) corresponds to the concentration of chloride at the electrode surface:

0 EAgCl /Ag (t) = EAgCl /Ag −

RT ln cCl (0,t) F

(6)

The boundary condition for the first element (0,t) is obtained by solving Eqn. 6 for cCl(0,t).

cCl (0,t) = exp(−

F 0 (E(t) − EAgCl /Ag )) RT

(7)

The initial concentration at t=0 for the entire thin layer sample is equal to the bulk concentration:

cX (x, 0) = cCl*

(8)

0 ≦ x ≦ xmax

The continuity equation is used to compute the concentration changes with time:

 ∂cCl (t)   ∂J Cl (x)    = −   ∂t ∂x  t x

(9)

where J indicates the flux of Cl-. This equation is linearized with finite elements to calculate the concentration elements for all times as established:24

c(x,t) = c(x,t − 1) + Daq

∆t {c(x + 1,t − 1) − 2c(x,t − 1) + c(x − 1,t − 1)} ∆x 2

(10)

The concentration for the last position in x (xmax) is calculated by a reflection of the concentration gradient as follows:

c(xmax ,t) = c(xmax ,t −1) + 2Daq

∆t c(xmax ,t −1) − c(xmax −1,t −1) ∆x 2

{

}

ACS Paragon Plus Environment

(11)

6

Page 7 of 31

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

Because the concentration of chloride varies in the entire thin layer as a function of time, the inner phase boundary potential variation at the Nafion membrane (potential at Nafion/thin layer sample interface, see Figure 1) is also considered by the Nernst equation for this permselective membrane. We assume that the sodium concentration is equal to the chloride concentration and substitute the two concentrations to give:

Emem (t) = A +

RT ln cCl (xmax ,t) F

(12)

where A includes all constant potential contributions, including the (indifferent) outer concentration of salt and the potential at the reference electrode in the outer solution (see Figure 1). The cell potential is described by Eq 13 as the standard reduction potential of silver chloride minus the membrane potential, with all other potential contributions of the cell again included in term A of Eqn. 12:

Ecell (t) = E AgCl/Ag (t) − Emem (t)

(13)

The current density is proportional to the concentration gradient multiplied by the diffusion coefficient in the first two elements as follows:

 c(0,t) − c(1,t)  j = −FDaq   ∆x  

(14)

This current density is plotted against the cell potential from Eqn. 13, which results in the desired thin layer voltammogram driven by a coupled precipitation/ion transport process.

ACS Paragon Plus Environment

7

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

Page 8 of 31

EXPERIMENTAL SECTION Reagent, Materials and equipment. Aqueous solutions were prepared by dissolving the appropriate sodium salts (Sigma Aldrich) in deionized water (>18 MΩ/m). The TT-020 Perma Pure Nafion tubing (o.d. × i.d. = 0.51 × 0.36 mm) was purchased from Dr. Marino Mueller AG (Zurich, Switzerland). Silver wire of 0.5 mm diameter (99.9% trace metals basis) and PEEK tubing (o.d. × i.d. = 1.56 × 0.76 mm) were from Sigma-Aldrich. HPLC connectors were obtained from Supelco. Cyclic voltammograms were obtained with PGSTAT 302N (Metrohm Autolab, Utrecht, The Netherlands) controlled by a personal computer using Nova 1.8 software (supplied by Autolab) in the dynamic electrochemical measurement mode. The experiments were carried out by filling the microfluidic cell with appropriate solutions using an ISMATEC peristaltic pump (Glattbrugg, Switzerland) at a flow rate of 25 µL min−1. In the case of water sample analysis, a Metrohm 761 compact IC chromatograph with anionexchanging column (6.1006.520 Metrosep A Supp 5) was used as a reference method. The eluent was a solution composed of 1 mM NaHCO3 + 3.2 mM Na2CO3, along with 50 mM H2SO4 for regeneration of the anion suppressor (0.8 mL min-1). Preparation of the cell. Working electrode: The silver wire of 0.5 mm diameter was drawn through a jeweler drawing plate until a 0.3 mm diameter. The wire was cleaned with acetone and after with deionized water. A piece of 15 mm length of this wire was oxidized electrochemically in a solution of either 1M NaI or 1M HCl, as specified, for 3 h at a constant anodic current of 0.4 mA cm-2. The coated electrode was washed with deionized water. Finally, a 30 mm piece of PEEK tubing was glued with epoxy resin to the distal end of the wire with a depth of about 40 mm. Counter reference electrode: A 20 mm piece of 0.64 mm diameter silver wire was coated with AgI, following the same procedure described above.

ACS Paragon Plus Environment

8

Page 9 of 31

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

Preparation of the membrane: The Nafion tubing was cut to a length of 100 mm. Two 40 mm long pieces of silver wire of 0.35 mm diameter were inserted into both ends of the tube at about 35 mm depth to prevent the tube endings of the membrane from clogging or collapsing in the process of attaching the connectors. To provide leak-proof connectors in the cell, both ends of the tubing were glued into 30 mm long pieces of the PEEK tubing with epoxy resin, delimiting the length of the Nafion membrane to 40 mm (see Figure S1, Supporting Information). After drying the epoxy (45 min at 60°C in the oven), the two silver pieces were pulled out. Finally, the ends of the PEEK tubing were cut flat with a tubing cutter. Cell Assembly: The working electrode was carefully inserted into the membrane, forming a sample chamber. Theoretically the sample layer thickness corresponds to 15 µm. The tips of the membrane and the electrode fitted with the PEEK tube allowed us to use standard HPLC ¼ in. connectors, with 1/8 in. nuts and 1/16 in. ferrules. A Plexiglas T- connector was used to assembly the cell similar to as described in previous earlier works.11 The counter-reference electrode was placed around the Nafion membrane (see Figure S1). Cyclic voltammetry measurements. Cyclic voltammetry experiments were carried out in stopped flow mode and at room temperature (23°C). The entire procedure was programmed in Nova software. After filling the void between membrane and the working electrode with the sample solution, the pump was stopped and the electrochemistry procedure was applied (forward and backward signals were recorded). The sample was refreshed by pumping a new solution for a period of 120-s. The cell was found to work for at least three weeks in continuous operation.

ACS Paragon Plus Environment

9

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

Page 10 of 31

RESULTS AND DISCUSSION The aim of this work consists in the establishment of an electrochemical protocol based on a thin aqueous layer of small (0.6 µL) volume for measuring a mixture of halides in real samples. Figure 1a shows a 100-fold magnified scheme of the microfluidic cell, drawn to scale. The thin layer sample is defined by the space between the oxidized Ag wire and the tubular Nafion membrane, resulting in a spacing of 15 µm. Figure 1b illustrates the electrochemical mechanism that is thought to take place when a linear sweep potential is applied to the electrochemical cell. The potential is applied between the inner silver element, considered here the working electrode, and another silver/silver halide element placed in the outer compartment that operates as a reference/counter electrode. Typically, both (inner and outer) compartments are filled with a sodium halide solution. In the outer compartment, a specified concentration of sodium halide was used as outer solution to achieve a defined cell potential. An anodic potential scan induces the oxidation of the silver wire and a resulting current that is limited by the consumption of the halides in the thin sample layer. As a result, a silver halide precipitate is formed on the electrode surface. The electrode becomes visibly black if a chloride solution is used, suggesting AgCl formation. Simultaneously, to fulfill the charge balance condition, the counter ion (i.e., Na+) must be transported from the thin layer across the perm-selective Nafion membrane to the outer solution. Therefore, at an optimal applied potential, chloride and sodium should both be depleted from the sample layer according to thin layer theory (see below). Subsequently, a backward sweep potential is applied that aims to re-reduce the oxidized AgX deposited in the previous scan and to prepare the electrode for the following analysis. In initial attempts to fabricate the microfluidic cell, both Ag and Ag/AgCl wires were characterized as working electrodes in terms of their electrochemical performance. Figure 2a compares two cyclic voltammograms for an initially symmetrical cell of 0.1M NaCl inside and out and for either a bare Ag wire and a Ag/AgCl wire in contact with the sample compartment. Both oxidation peaks appear at the same potential. During the backward scan, an additional process appears to occur at the bare silver

ACS Paragon Plus Environment

10

Page 11 of 31

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

electrode at more negative potentials (-163 mV). In the latter case, the Nafion membrane becomes unexpectedly blue in some regions, indicating the deposition of silver halide particles in the membrane.2527

Chloride ions are known to form soluble complexes with the free Ag+, mainly [AgCl2]-,23 which may

diffuse away from the electrode surface. Owing to the fact that the chloride concentration diminishes at the Nafion membrane surface, those complexes may in turn precipitate as silver chloride. Previous experience suggests that an Ag/AgCl element may perhaps be most suited to prevent membrane contamination and undesired reactions at the silver electrode. However, further experiments using alternating chloride and iodide solutions in the sample showed a residual memory effect of the oxidized wire that is explained by the preference AgI>AgBr>AgCl (with the solubility products 8.3 x 1017

, 5.4 x 10-13 and 1.8 x 10-10 for AgI, AgBr and AgCl, respectively) deposited on the Ag/AgCl electrode.

The adsorption of iodide at AgCl-based electrodes is well indeed established.28 As a consequence, significant irreproducibility was observed in subsequent scans (see Figure S2, Supporting Information). To overcome this, we decided to use Ag/AgI wires as working and counter/reference elements, see Experimental for details. Using this configuration the mentioned problems were successfully addressed (see Figure 2b). Furthermore, temperature changes of 25-60° C did not significantly influence the integrated charge of the oxidized and reduced peaks for a symmetrical system of 0.1 M NaCl concentration, achieving a % RSD of less than 0.8% (data not shown). Having established the appropriate fluidic set-up for detecting halides, the analytical characteristics were established in artificial samples (NaCl, NaBr and NaI in deionized water). Figure 3a, b and c show cyclic voltammograms for NaCl, NaBr and NaI (at 0.1 M), respectively, in a symmetrical configuration and at different scan rates ranging from 10 mV s-1 to 100 mV s-1. The separation between the oxidized and the reduced peaks was found to increase significantly with increasing scan rate, although the reversibility of the process is maintained. At a scan rate of 10 mV/s and detecting chloride, for example, the integrated charge from the forward and backward peaks were found as 0.3614 C and 0.3629 C, respectively. The peak separations were 283 mV and 420 mV for 10 mV s-1 and 100 mV s-1 for the uncorrected

ACS Paragon Plus Environment

11

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

Page 12 of 31

voltammograms. A linear relationship was found between peak intensity and scan rate, confirming thin layer behavior. Any deviation from linearity could be due to a deviation from thin layer behavior because of the heterogeneous distribution of the internal space between membrane and silver wire (the sample layer thickness is not homogeneous along the entire tubing length). Therefore, a mixed response may need to be considered (peak current as a function of υ and υ0.5). Nevertheless, the observed linearity suggests that the current remains controlled by diffusion kinetics in the aqueous phase. Consequently, the contribution of other process to the total current (i.e., adsorption, nucleation, charge transfer, etc.) may be neglected based on this observation. Analyzing Figure 3 in more detail, one suspects a significant ohmic drop that could be subtracted from the cyclic voltammograms. The data in Figure 3 were corrected for the ohmic contribution to the applied potential, using a resistance value of 30 ohm, as confirmed by electrochemistry impedance spectroscopy, see Figure S3. As an example, Figure 4 shows the corrected data for NaCl from Figure 3a. Subtracting the iR drop contribution, the peak separation was found to be less pronounced (200 mV) and became independent of scan rate. As the iR drop correction does not modify the peak height, the same linearity of peak current with scan rate is conserved. A control experiment at high volume to surface ratios (classical three electrode configuration) showed a very poor peak resolution with an extremely large iR drop (Figure S4). Independent of any ohmic drop correction, all peaks showed pronounced tails. Numerical simulations confirmed the observed behavior. Figure 5 exhibits theoretical voltammograms at different scan rates for 1 M NaCl. In this simulation only the forward peak is shown, which corresponds well to experimental observations. The backward peak is not shown owing to the fact that in our simple model, the current tends to infinite. In practice, the amount of oxidized species that can be re-reduced is limited and given by the total turnover during the anodic scan. For information, simulated concentration profiles for forward and backward scans are included in Figure S5.

ACS Paragon Plus Environment

12

Page 13 of 31

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

Calibration curves for Cl-, Br- and I- were separately performed using cyclic voltammetry as readout (see Figure 6a, b and c). Figure 6a shows the obtained cyclic voltammograms for different concentrations of chloride (from 10-6 to 0.6 M) using a fixed 0.1M concentration of NaCl in the outside solution. The integrated charge followed a linear behavior as a function of the chloride concentration in the thin layer sample from 10-4 to 0.6 M. The corresponding calibration curve is shown in Figure 6d for three replicates, exhibiting a charge of (3.67 ± 0.03) cCl mC M-1 + (0.014±0.008) mC). The lower limit of detection for this configuration was found as 10-4 M. This value is affected by the selected outer solution concentration. On the one hand, decreasing the concentration of the outer solution down to 10-4 M NaCl, peaks are only observed from 0.1 M to 0.6 M concentration. Moreover, irreproducible peaks were obtained due to poor ion transport (migration) in the outer solution containing counter/reference electrode. On the other hand, the use of 0.6 M NaCl as outer solution (Figure S6) resulted in a non-linear calibration, suggesting a significant transmembrane sodium chloride flux.15 In conclusion, a compromise between lowest limit of detection and widest linear range needed to be chosen. After careful optimization, a concentration of 0.1 M was selected for most experiments, resulting in a detection limit of around 10-4 M (0.2 mA residual current). This limit of detection is given by the contamination of the thin layer by an inward electrolyte flux. Figures 6b and 6c exhibit the cyclic voltammograms obtained for different concentrations of bromide and iodide respectively. For these halides, linear correlations between peak area and concentration were also obtained (Figures 6d). The linear range was from 10-4 to 0.6 M for bromide and from 10-5 to 0.6 M for iodide. The linear regressions were, for bromide, charge = (3.81±0.04) cBr mC M1

+(0.017±0.010) mC and, for iodide, charge = (3.49±0.04) cI mC M-1+(0.026±0.001) mC. This concept was extended to the resolution of a mixture of the three halides, which is impossible to

achieve with potentiometric AgX membrane electrodes since the precipitation of silver iodide always dominates. The outer solution composition was re-evaluated for this purpose. Figure 7a shows the voltammograms of each individual halide using 0.1 M NaCl as outer solution, with observed peak positions for Cl-, Br- and I- at 169, 20 and -274 mV, respectively. A mixture of the three halides was

ACS Paragon Plus Environment

13

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

Page 14 of 31

subsequently measured at the same scan rate (10 mV s-1). After this qualitative experiment each peak in the mixture can be assigned to its respective halide. Similar experiments were performed in 0.1 M NaBr and NaI solutions (Figure 7b and c respectively). As expected, a shift in the potential domain is observed when the outer solution nature is varied. This shift is a result of the variation of the reference potential. Figure 7d displays the voltammograms for a mixed halide sample obtained at different scan rates and using an outer solution of 0.1 M NaCl, NaBr and NaI. The resolution between the three peaks improved with decreasing scan rates. Thin layer behavior was again confirmed by the linear relationship between the peak current and the scan rate for each halide. To obtain an accurate charge for the overlapped bromide and chloride peaks, a deconvolution Gaussian treatment was applied to the raw voltammograms, see Figure S7 in Supporting Information. With this procedure (experimental configuration and subsequent data treatment) the obtained charges from the oxidized peaks were extremely reproducible from scan to scan with a single cell, at less than 2% RSD, see Figure S8, and even employing different electrochemical cells, at 2.5% RSD, N=6. This suggests that the approach is highly promising as a calibration-free methodology. Figure 8a shows the cyclic voltammograms obtained for different halide concentrations using an outer solution of either 0.1 M NaCl, NaBr and NaI, trying to minimize inward fluxes of the three halides to the sample solutions. Additionally, the response to chloride in a fixed Br- and I- background is visualized in Figure S9, see Supporting Information. This experiment was designed in view of an analytical application of the fluidic cell for determining real water samples that contain high amounts of Cl- compared to other halides. Linear calibration curves were obtained for the three halides (Figure S9), giving charge = (1.48±0.03) cCl mC M-1 + (0.064±0.007) mC for chloride, charge = (1.74±0.05) cBr mC M-1 + (0.034±0.003) mC for bromide and charge = (1.88±0.02) cI mC M-1 + (0.002±0.001) mC for iodide, with limit of detections of 6.4 x 10-5, 3.4 x 10-5 and 2.1 x 10-5 M for chloride, bromide and iodide respectively. The thin layer cyclic voltammetric cell was applied to water sample analysis. Cyclic voltammograms of seawater, river, tap and mineral water were obtained using a mixture of 0.1 M NaCl, NaBr and NaI

ACS Paragon Plus Environment

14

Page 15 of 31

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

solution in the outer solution (Figure S10 and Table S1). We noted that iodide and bromide peaks appeared in some water samples where such elements should not be observed, as in tap water. It is understood here that inward fluxes are responsible for these residual signals. When samples were measured using just 0.1 M NaCl as outer solution the obtained cyclic voltammograms did not show these mentioned peaks (Figure 9). The amount of each halide presents in the sample was again determined by external calibration (see Figure S9c for the calibration graph), giving charge = (2.28±0.01) cCl mC M-1 + (0.038±0.008) mC for chloride, charge = (2.38±0.03) cBr mC M-1 + (0.016±0.010) mC for bromide and charge = (2.07±0.03) cI mC M-1 + (0.038±0.010) mC for iodide, with limit of detections of 6.0 x 10-5, 2.5 x 10-5 and 2.5 x 10-5 M for chloride, bromide and iodide respectively. Overall, either outer solution is adequate to quantify the amount of iodide, bromide and chloride in water samples, but the use of 0.1 M NaCl as outer solution is more convenient for samples that contain reduced levels of bromide and iodide as is typical for most real world samples. To validate these results, comparative values were obtained by ion chromatography (IC) as reference method. The concentrations observed by our system correspond well to the ones obtained by IC (Table 1). As expected, the most abundant halide in all samples was chloride. We note that chloride levels in seawater can be determined in the coulometric detection cell without dilution, which is not possible with IC owing to the saturation of the column. Bromide was also measured in sea, mineral and tap water at low ppm values with the current methodology, see Table 1.

ACS Paragon Plus Environment

15

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

Page 16 of 31

CONCLUSIONS A thin layer cell based on a silver iodide working electrode and a tubular Nafion membrane was fabricated and cyclic voltammograms were obtained for chloride, bromide and iodide in mixed samples. Thin layer cyclic voltammetry allows us to obtain separated peaks for the three halides, exhibiting a linear relationship between halide concentration and the corresponding peak area from about 10-5 to 0.1 M for bromide and iodide and from 10-4 to 0.6 M for chloride. The technique was successfully applied for the halide analysis of tap, mineral and river water as well as seawater. The key advantage of this technique over previous methodologies is a total conversion of the halide salts, resulting in an absolute measurement methodology that is found to be both temperature independent and highly reproducible, even from cell to cell. In contrast to earlier work on coulometric ion sensing, slow scan cyclic voltammetry has the added benefit of providing peak shaped responses that can be better separated and analyzed than the monotonous current decays observed so far with simple potential steps. We believe that this methodology forms a very promising basis also for future work on coulometric ion sensors with ionophore-based membranes and films, where the approach can be extended to absolute multianalyte analysis.

ACKNOWLEDGMENT The authors thank the Swiss National Science Foundation and the European Union (FP7-GA 614002SCHeMA project) for supporting this research.

ACS Paragon Plus Environment

16

Page 17 of 31

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

ASSOCIATED CONTENT Supporting Information. Additional information as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding Author *[email protected]

ACS Paragon Plus Environment

17

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

Page 18 of 31

REFERENCES (1) Crespo, G. A.; Bakker, E. RSC Advances 2013, 3, 25461-25474. (2) Yoshizumi, A.; Uehara, A.; Kasuno, M.; Kitatsuji, Y.; Yoshida, Z.; Kihara, S. J. Electroanal. Chem. 2005, 581, 275-283. (3) Amemiya, S.; Kim, Y.; Ishimatsu, R.; Kabagambe, B. Anal. Bioanal. Chem. 2011, 399, 571-579. (4) Kasuno, M.; Fujimoto, K.; Kakitani, Y.; Matsushita, T.; Kihara, S. J. Electroanal. Chem. 2011, 651, 111-117. (5) Dorokhin, D.; Crespo, G. A.; Afshar, M. G.; Bakker, E. Analyst 2014, 139, 48-51. (6) Sohail, M.; De Marco, R.; Lamb, K.; Bakker, E. Anal. Chim. Acta 2012, 744, 39-44. (7) Grygolowicz-Pawlak, E.; Bakker, E. Electrochim. Acta 2011, 56, 10359-10363. (8) Bakker, E. Anal. Chem. 2011, 83, 486-493. (9) Shvarev, A.; Neel, B.; Bakker, E. Anal. Chem. 2012, 84, 8038-8044. (10) Grygolowicz-Pawlak, E.; Numnuam, A.; Thavarungkul, P.; Kanatharana, P.; Bakker, E. Anal. Chem. 2012, 84, 1327-1335. (11) Crespo, G. A.; Afshar, M. G.; Dorokhin, D.; Bakker, E. Anal. Chem. 2014, 86, 1357-1360. (12) Kihara, S.; Kasuno, M. Analyt. Sci. 2011, 27, 1-11. (13) Rodgers, P. J.; Jing, P.; Kim, Y.; Amemiya, S. J. Am. Chem. Soc. 2008, 130, 7436-7442. (14) Langmaier, J.; Olsak, J.; Samcova, E.; Samec, Z.; Trojanek, A. Electroanalysis 2006, 18, 1329-1338. (15) Grygolowicz-Pawlak, E.; Sohail, M.; Pawlak, M.; Neel, B.; Shvarev, A.; de Marco, R.; Bakker, E. Anal. Chem. 2012, 84, 6158-6165. (16) Grygolowicz-Pawlak, E.; Crespo, G. A.; Afshar, M. G.; Mistlberger, G.; Bakker, E. Anal. Chem. 2013, 85, 6208-6212. (17) Wahono, N.; Qin, S.; Oomen, P.; Cremers, T. I. F.; de Vries, M. G.; Westerink, B. H. C. Biosens. Bioelectron. 2012, 33, 260-266. (18) Michalski, R.; Jablonska, M.; Szopa, S.; Lyko, A. Crit. Rev. Anal. Chem. 2011, 41, 133-150.

ACS Paragon Plus Environment

18

Page 19 of 31

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

(19) Harris, D. C. Quantitative Chemical Analysis; W.H Freeman and Company: New York, 2007; Vol. 7. (20) Hassan, H. H.; Ibrahim, M. A. M.; Abd El Rehim, S. S.; Amin, M. A. Int. J. Electrochem. Sci. 2010, 5, 278-294. (21) Birss, V. I.; Wright, G. A. Electrochim. Acta 1982, 27, 1439-1443. (22) Birss, V. I.; Wright, G. A. Electrochim. Acta 1982, 27, 1429-1437. (23) Fritz, J. J. J. Sol. Chem. 1985, 14, 865-879. (24) Morf, W. E.; Pretsch, E.; De Rooij, N. F. J. Electroanal. Chem. 2007, 602, 43-54. (25) Domenech, B.; Munoz, M.; Muraviev, D. N.; Macanas, J. Chem. Comm. 2014, 50, 4693-4695. (26) Liu, J.; Wang, Z.; Liu, F. D.; Kane, A. B.; Hurt, R. H. Acs Nano 2012, 6, 9887-9899. (27) Xing, S.; Xu, H.; Chen, J.; Shi, G.; Jin, L. J. Electroanal. Chem. 2011, 652, 60-65. (28) Harsanyi, E. G.; Toth, K.; Polos, L.; Pungor, E. Ana.l Chem. 1982, 54, 1094-1097.

ACS Paragon Plus Environment

19

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

Page 20 of 31

Figures Captions

Figure 1.

a) Scheme of the electrochemical tubular cell used in this work (magnification, 100). d1=0.3 mm (Ag bare), d2=0.33 mm (Ag/AgX), d3=0.36 mm (Thin layer + Ag/AgX, d4=0.51 mm. Nafion + thin layer + Ag/AgX, l1=40 mm length of the cell b) Illustration of the electrochemical mechanism upon application of a cathodic potential to the cell. X-, halide, R- cation exchanger in the Nafion membrane, WE: working electrode, CE/RE: counter-reference electrode (coiled Ag/AgX)

Figure 2.

a) Comparison between bare silver and Ag/AgCl wires for 0.1 M NaCl (inner and outer solution), scan rate=10 mV/s. b) Cyclic voltammograms obtained for 0.1 M NaCl, 0.1 M NaBr and 0.1 M NaI using a Ag/AgCl working electrode. Identical inner and outer solutions, scan rate=10 mV/s.

Figure 3.

Cyclic voltammograms at different scan rates for: a) 0.1 M NaCl , b) 0.1 M NaBr and c) 0.1 M NaI (identical inner and outer solution composition) at scan rates of 100, 75, 50, 20 and 10 mV/s, from top to bottom. Inset: peak height vs. scan rate. Inner element: Ag/AgI.

Figure 4.

Ohmic drop correction for the data shown in Figure 3a (Eapp-iR = E, with R=30 Ω).

Figure 5.

Numerical simulation of cyclic voltammograms at different scan rates (100, 75, 50, 20 and 10 mV/s, from top to bottom). Parameters: Daq =10-7 dm2s-1, ∆t=2ms, ∆x=2 µm.

Figure 6.

Obtained voltammograms at different concentrations of a) NaCl, b) NaBr and c) NaI (0.6, 0.5, 0.4, 0.3, 0.2, 0.1, 0.06, 0.03, 0,01, 5.10-3, 1.10-3, 5.10-4, 1.10-4 10-5 M from top to bottom) with 0.1 M NaCl, 0.1 M NaBr and 0.1 M NaI as outer solution, respectively. d) Observed calibration curves for chloride (circle), bromide (triangle) and iodide (square). (Scan rate=10 mV/s).

Figure 7.

Cyclic voltammograms of individual halides (solid line) 0.1 M NaCl, 0.1 M NaBr, 0.1 M NaI and a mixture (dashed line) of 0.1 M NaCl, NaBr and NaCl using different outer solutions: a) 0.1 M NaCl, b) 0.1 M NaBr and c) 0.1 M NaI (10 mV/s scan rate). d) Voltammograms of a symmetrical configuration (a mixture of 0.1 M NaCl, NaBr and NaI ) at different scan rates (100, 75, 50, 20, 10, 5 and 2 mV/s from top to bottom). Inset: confirmation of thin layer behavior (ip vs scan rate).

Figure 8.

Observed voltammograms for different halide concentrations in equimolar mixtures.

Figure 9.

Thin layer voltammograms of a) mineral, tap and river water and b) seawater. Scan rate: 5 mVs-1. Outer solution: 0.1 M NaCl.

ACS Paragon Plus Environment

20

Page 21 of 31

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

Table 1 – Water samples analyzed by the proposed thin layer methodology. Samples Sea Tap Mineral River

Chloride / ppm 21483±60 (21400a) 13±1 (13) 17±1 (16) 6±1 (6)

Proposed methodology Bromide / ppm 17±1 (nqb) 3±1 (

Exhaustive thin-layer cyclic voltammetry for absolute multianalyte halide detection.

Water analysis is one of the greatest challenges in the field of environmental analysis. In particular, seawater analysis is often difficult because a...
748KB Sizes 0 Downloads 6 Views