Critical Review—Electrocatalytic Sensors for Arsenic Oxo Species

As noted above, the primary focus of electroanalytical methodology for the arsenic has been on the oxidation of As(III). In the absence of a catalyst, this process does not occur at an analytically useful rate, as exemplified by its absence at a bare carbon electrode. Common approaches to alleviate this limitation include the use of a Pt electrode and the modification of carbon electrode by a stable film of a redox mediator. At Pt, Catherino⁴⁷ hypothesized a role of active oxygen in the oxidation pathway, which caused the formation of As(V) to overlap the oxidation of water. The mediated oxidation also involves overlapping electrode reactions in that generation of the oxidized form of the mediator that subsequently reacts with As(III) is a basic requirement. In both cases, the development of an electroanalytical method requires a means of correction for a potentially high background current in the measurement of that due to the oxidation of As(III).

Exacerbating the problem of the substantial background current is its relationship to noise in the measurement; that is, a simple subtraction does not solve the problem. Electrochemical current measurements are limited either by flicker or by shot noise. In the former case, the noise is directly proportional to current, and in the latter case, it is proportional to the square root of the current. In experiments on concentrations at and below the μg l⁻¹ level, the presence of noise from an overlapping faradaic process can obviate statistically meaningful measurement of the analytical signal unless some means of mitigating the noise that is related to the magnitude of the background current is employed. For electrode reactions not complicated by the need for a catalyst or a coupled chemical reaction, the background is due to charging of the double layer upon a change of applied potential. With the growing interest in utilization of nanomaterials in electrochemical sensors, which are characterized by high active surface areas and good permeability, it is reasonable to expect sizeable capacitive background currents that can affect reproducibility and complicate analytical determinations at low concentrations. Under such conditions, enhancement of the signal-to-noise level of analytical currents can be achieved through application of pulse voltammetric approaches⁴⁸ (instrumentation-based improvement) and/or by combining the As-preconcentration step with the electrocatalytic current enhancement effects (chemical sensing solution). In the latter case, the feasibility of As(V) preconcentration through deposition and agglomeration, followed by electrocatalytic reduction on platinum nanoparticles is a significant consideration.³⁸

Pulse voltammetry where a time delay prior to measurement of the analytical current improves the signal-to-background ratio in that the capacitive (charging) current decays exponentially whereas the analytical (Faradaic) current decays with the square root of time. With complications such as those encountered in the oxidation of As(III), this approach is limited in utility. However, extending the scope of pulse voltammetry to incorporate a three-step scheme allows the determination of As(III) at a Pt electrode in acidic solution.⁴⁸ The steps, sequentially, are to potentials where adsorbed hydroxyl radicals are generated, platinum oxide is formed, and a pristine Pt surface is restored. Oxidation of As(III) occurs where the adsorbed hydroxyl is present. In this manner a detection limit of 5 μmol L⁻¹ is obtained. Employing this analytical method to the design of a sensor where selectivity is a factor, as discussed below, needs to consider that at the potential where the electrode is activated (E-range, 0.7–1.1 V vs SCE in 0.1 mol L⁻¹ H₂SO₄) numerous organic and inorganic species are oxidized.

Recently, nanoparticles of platinum and related metals have been used in the electroanalytical determination of As(III).^26,29 With Pt nanoparticles, the results were consistent with those of Williams and Johnson,⁴⁸ namely that active oxygen designated as PtOH was a key to the oxidation of As(III).²⁹ A layer-by-layer assembly of poly(diallyldimethylammonium chloride) and citrate-capped Au nanoparticles on a gold electrodes gave a linear calibration curve for As(III) down to 20 μmol l⁻¹.⁴⁹ From the shape of the current-voltage curve, the analyte was adsorbed on the surface prior to its oxidation. Of importance to adapting this electrode to a sensor is that it showed a resistance to poisoning. This result, although based on limited data, is reasonable in that there is the possibility of protection of the electrochemically active sites afforded by the outer layer of polymer.

An alternative to utilizing a transient active surface is to coat an electrode with a stable mediator. An example of a modified electrode for the determination of As(III) is carbon coated with a thin layer of ruthenium oxide that is stabilized by cyano crosslinks, RuO/CNO.^24,50 This film is formed by cyclic voltammetry of a Ru(CN)₆⁴⁻, Ru³⁺ mixture at pH 2. The diminution of the cathodic peak current and the marked enhancement of the corresponding anodic peak is consisted with a mediated process whereby the As(III) is oxidized by the electrochemically generated Ru(IV) at 0.8 V vs SCE.²⁴ Figure 3 illustrates both (A) the voltammetric oxidation of As(III) and (B) the possibility of its long-term steady-state monitoring under amperometric conditions. The involvement of active oxygen is suggested by comparison to the behavior of As(III) at Pt. Evidence of this mediation pathway is more apparent from studies on the oxidation of sulfhydryl compounds and of thiocyanate. For example, bulk electrolysis of cysteine at a RuO/CNO-coated electrode yielded cysteic acid rather than a dimer, which would passivate the surface, and the analogous experiment with SCN^- yielded sulfate.⁵¹

Zoom In Zoom Out Reset image size

As discussed previously, anodic stripping voltammetry provides improved detection limits for As(III). For example, this analyte was preconcentrated by constant potential electrolysis as an arsenic film on a working electrode of Pt nanoparticles on glassy carbon.²⁹ When the oxidation of this deposit was by linear potential scan voltammetry, a detection limit of 2.1 ± 0.05 μg As L⁻¹ was achieved;²⁹ however, copper significantly interfered because of Cu₂As₃ formation during the deposition.^26,29,52 The interference of copper notwithstanding, improved detection limit using Au nanoparticles on carbon nanotubes as the working electrode was reported.⁵³ Recently Sonkoue et al.⁵⁴ developed a simple method for modifying a gold electrode and applied it to the anodic stripping voltammetric determination of As(III) in river water. Interference by copper and chloride limits the utility of this system although a detection limit of 1.4 μg As L⁻¹ in a standard sample was achieved.

Overall, several methods for the determination of As(III) with detection limits in the 10–50 μg l⁻¹ range by oxidation using potentiostatic measurements at mediator-modified electrodes or about an order-of-magnitude lower detection limit by pulse stripping voltammetry at nanoparticle-modified electrodes have been reported. Of these general approaches, stripping voltammetry, while readily employed in an operator-controlled laboratory procedure, is less amenable to incorporation in a stand-alone sensor.

Development of an electroanalytical method for As(V) is challenging because its reduction is more difficult to achieve than either the oxidation or reduction of As(III). Although sometimes reported as electrochemically inert, As(V) can reduced at a high over-potentials in acid solution^38,55–58 and in neutral solution in the presence of manganese.^59,60 For example, in neutral solution, when manganese (most likely in a form of various manganese oxides) was present on a gold electrode surface, an arsenic film was deposited at −1.3 V vs Ag/AgCl/3 mol KCl L⁻¹.⁵⁹ The stripping step was its oxidation to As(III). With a 180 s deposition time, a limit of detection of 0.2 nmol As(V) L⁻¹ was reported.⁵⁹ Also noteworthy is the report of the reduction of As(V) under potentiostatic conditions, −0.15 V vs SCE, at a glassy carbon electrode modified with poly(aniline-co–o–aminophenol).⁶¹ The reported detection limit was 0.5 μmol As L⁻¹.⁶¹

We recently investigated the mechanism of As(V) reduction in 0.5 mol l⁻¹ H₂SO₄ at a glassy carbon electrode modified with Pt nanoparticles.³⁸ In terms of developing an analytical method, prior to its reduction As(V) is adsorbed to bare Pt surfaces of the nanoparticles. Evidence was presented showing that a structural reorganization of the resulting film occurs whereby a polynuclear species is formed. The electrochemical reduction of the film is catalyzed by Pt° and PtH. The electrocatalytic reduction of As(V) proceeds within the deposited metaarsenite-type polymeric chains rather than in bulk of the solution. As often is seen when the deposit is a film, the correlation of the current for the reduction of the As(V) deposit is nonlinear. Also somewhat complicating the analytical approach is that, unless subjected to voltammetric potential cycling in the range from ca. 1 to 0 V (vs RHE), which facilitates agglomeration, the film does not survive medium transfer. Nevertheless, concentration down to about 2 μmol As(V) L⁻¹ were detected in the preliminary analytical study.

A bridge to a standalone sensor for inorganic arsenic may be based on electrochemical reduction to arsine.^62–66 Various forms of carbon, highly ordered pyrolytic graphite for example,⁶⁵ have been used as the cathode. The potentials for reduction of As(III) and As(V) differ markedly, which is advantageous for speciation studies. In contrast, an Al cathode can be operated to provide similar efficiency for these species but suffers from low stability in the required acidic solution.⁶⁵ The actual measurements of the generated arsine were spectroscopic. However, arsine has been electrochemically oxidized.^67,68

In summary, the development of sensors rather than detectors always faces the challenges of detection limit and especially selectivity. As discussed above, in the case of arsenic species the use of various nanoparticles for catalysis is of particular importance to sensitivity but can exacerbate selectivity issues because the promotion of electron transfer by mediation increases the number of electroactive species at a given potential. Recent studies offer potential routes to alleviate this problem, as described below.

Several methods are available to explore the topic of selectivity. Applying an overlayer of an anion exchange polymer to the working electrode can reject one group of common interferents, namely metal cations. Quaternized poly(vinylpyridine), qPVP, is a representative example. A platinum electrode modified by adsorption of iodide and coated with a qPVP film served as a voltammetric sensor for nitrite.⁶⁹ However, it is important to note that As(III) forms a neutral oxo species via protonation over a wide range of acidity. A more common overlayer, Nafion^®, is widely used for measurements where its cation-exchange properties need to be explored. These materials not only reject ions of charge sign opposite to that of the analyte but also block passivation of the electrode by adsorption of macromolecules in the sample and serving in some cases as a host for the catalytic species. Such protected electrodes may be especially well suited to use in conjunction with cell platforms made as thin layers such as by inkjet printing.⁷⁰ A promising approach involves the use of gels as the receiving phase,^41,71 particularly a gel loaded with nanoparticles.⁴¹ An electrode modified with a gel layer impregnated with catalytic and adsorbing nanoparticles could possibly address both sensitivity and selectivity issues. In this regard, such separation methods have been used for arsenic speciation.³⁹

Because arsenic oxo species can be reduced to arsine gas, an air-gap route to selectivity is available. Indeed, electrochemical generation of arsine has been used in conjunction with atomic spectroscopy for the determination of arsenic.⁷² The initial design dates to the Clark oxygen electrode⁷³ where a traditional electrochemical cell was separated from a sample by a gas permeable membrane. Several design variations are possible when an air-gap is used. A planar cell coated with a room-temperature ionic liquid of low volatility comprises a gas sensor.⁷⁴ A solid electrolyte such as a silica sol-gel that hosts reference, counter and working electrodes has been used as an amperometric gas sensor.⁷⁵ Doping the sol-gel with salts comprising a humidistat makes the response humidity independent over a wide range,⁷⁶ and the demonstration that at a working electrode protruding into the gas phase the electrochemical reaction take place at the three-phase boundary rather than requiring partitioning into the sol-gel accounts for the rapid response time.⁷⁷ The historic concept of the Clark oxygen electrode can also be extended to sensing and differentiation of arsenic species through application of ion-exchange membranes, rather than the neutral one, as it was demonstrated with so called voltammetric ion selective electrode for chromium(VI) utilizing the anion-exchange membrane over-coating the planar three-electrode set-up with the electrocatalytic modified working electrode.⁷⁸ The ion exchange properties and selectivity in monitoring of arsenates (particularly of As(III) oxo species) are strongly dependent on pH.⁷⁹ Nevertheless, it should be remembered that protons are needed to induce surface reactions of the noble metal catalysts (including Pt), and the high concentration of protons (acidic conditions) seems to be crucial during the reduction of As(V) (direct involvement of H⁺ in the reduction process).³⁸

It is well established that distribution of arsenic in water and natural attenuation of As pollution are strongly affected by adsorptive interactions of As(III) and As(V) with mineral species, e.g. of Fe and Al oxides and clay contents of soils.⁴² Having in mind sorption mechanisms and coprecipitation phenomena occurring on surfaces of iron oxides and oxy-hydroxides, new preconcentration methods can be developed and utilized for sensing of arsenic pollutants. A combination of catalytic metal nanoparticles with certain metal oxide supports, e.g. zirconia, which is stable and positively charged and able to attract anions in acid media^80,81 could provide means of preconcentration and the enhanced electroanalytical sensing. Furthermore, the so called solid-state voltammetric approaches,^82–84 which utilize ultramicroelectrodes and permit operation in the absence of external electrolyte phase, could be used to sense arsenite and arsenate species as well as their conjugate acids sorbed onto oxides or polymer gels. Among possible arrangements for solid-state type measurements (Fig. 4), there have been historically proposed three electrode systems consisting of ultramicrodisk electrode (e.g. Pt. Au), silver pseudoreference, and Pt counter electrode. The electrodes can immobilized within a plane Fig. 4a (over-coated with an investigated gel-type material that can preconcentrate As); or the material analyzed can exist between the ultramicrodisk working electrode and the planar assembly (Fig. 4b) consisting of silver disk (pseudoreference) and glassy carbon ring (counter electrode). In practice, any large surface area electrode can serve as a pseudoreference for preliminary sensing (Fig. 4b). Care must be exercised to provide the oxygen-free atmosphere in the presence or absence of contact with the analyzed aqueous solution (for simplicity not shown in Fig. 4b). Alternatively, miniaturized sensing devices based on interdigitated arrays could be considered both under stationary and flow (e.g. chromatographic) conditions.^85,86

Zoom In Zoom Out Reset image size

The concept of generation of arsine gas through reduction of highly inert arsenate species can be compared with the strategies concerning electroreduction, for example, of CO₂ in aqueous solutions where the competitive hydrogen evolution reaction (HER) is a complicating side-reaction.^86–88 Thus the efficient electroreductions (arsenates, CO₂, nitrites, bromates) should be, in principle, facilitated at the proton-depleted electrochemical interfaces. On the other hand, the electroreduction mechanisms do require protons. Based on our preliminary results with tungsten oxide nanostructures capable of sorption of hydrogen in the case of CO₂ reduction,⁸⁹ it is reasonable to explore electrocatalytic surfaces utilizing e.g. various transition metal oxides⁹² that, intrinsically, can be more reactive toward arsine generation rather than hydrogen evolution. In other words, the As(V)-to-arsine conversion research may be pursued in acidic and moderately acidic media that to date have been barely explored. The reduction or selective conversion of arsenates to arsine can be envisioned by both conventional-electrochemical and visible-light-induced photoelectrochemical means. In this respect, special attention should be paid to the development of robust semiconducting materials⁹⁰ and to recent achievements in the area of photoelectrochemical conversion of carbon dioxide.^91–93

Advancing the design of mediators and catalysts may also enhance selectivity. In this regard, it is well known that mediation and catalysis involving two steps, such as electron and oxygen transfer, occurs in series.^97,98 It is reasonable to expect that electrocatalytic interfaces for sensing inorganic arsenic oxo species will comprise, in addition to catalytic noble metal nanoparticles, nanostructured transition metal oxides (e.g. of ruthenium, vanadium, tungsten or molybdenum), either in simple or derivatized forms.^94,95,96 Having in mind the observation that successful electrocatalytic reduction of As(V) requires adsorption or deposition,³⁸ followed by agglomeration, of arsenate species on surfaces of noble metal nanoparticles, application of high surface area metal oxide supports or carriers may lead to enhancement of the analytical current. To restore clean Pt surface, the arsenic deposits must be removed through application of either negative potentials, e.g. −0.5 V (vs RHE) where vigorous hydrogen evolution occurs, or positive potentials, e.g. 1.4 V (vs RHE) where higher Pt oxides are formed. In the latter case, the process is completed by stepping the potential to a value where the oxide is reduced. The fact that As(V) is preoncentrated at the positive potentials where PtO exists³⁸ could make the whole approach more selective: under such conditions, most of adsorbates originating from organic interferences are effectively oxidized and desorbed (at PtO) whereas typical inorganic interferents,⁴⁹ such as Cl⁻, NO₃⁻, Cu²⁺, Pb²⁺ Zn²⁺ are neither deposited nor oxidized. Manipulation of the composition of the mediator/catalytic system may comprise a route to separating the redox potential of the analyte from those of major interferents.⁸⁸

The continued investigation of the catalysis of the electrochemical reduction of As(V) is important.^16,38 In this regard, characterization of potential mediators and investigation of the role of the base electrode material⁶⁵ are important. Because sensitivity and detection limit in electroanalytical determinations strongly depend on the current densities measured, there is a need to search for specific catalytic materials that would induce otherwise highly slow and irreversible redox processes of As(III) (oxidation) and, in particular, As(V) (reduction). In this regard, electrochemical determination of As(III) is better described in literature than that of As(V). It is reasonable to expect intense research in future aimed at the developing of new electroanalytical methods for direct selective determination of As(V). A possible route to the latter is suggested by the apparent need to bind or activate As(V) to a chemical compound as a route to its reduction. Designing effective electrocatalytic materials also is of importance to the development of more sensitive stripping methods and to monitoring of arsenic under chromatographic and flow conditions. Representative examples of catalytic systems have been provided and discussed here.

Instrumental approaches to enhanced sensitivity through increased Faradaic currents and reduced charging (background) currents cover various pulse methods including normal pulse voltammetry, differential pulse voltammetry or more complex square-wave voltammetric methods.⁹⁷ The detection limits are usually between 10⁻⁷ and 10⁻⁶ mol l⁻¹ but they can be lower. But even with no appreciable charging current contribution, practical utilization of pulse voltammetric approaches does require optimizing measurement parameters (e.g. sampling time, pulse width and height) because Faradaic background processes can contribute to, and often control, the detection limits. In applied analytical chemistry, classical cyclic voltammetry is more intuitively interpretable in chemical terms. Indeed, under some circumstances (complex samples), pulse techniques can produce distorted responses hiding true analytical information.

Having in mind recent progress in the fuel cell research, further enhancement of activity of catalytic noble metal sites will be achieved by exploring specific interactions and supporting them on nanostructured metal oxides⁹⁸ or various carbon materials including graphene nanocomposites.^99,100 Finally, progress is also expected on fabrication and characterization of highly sensitive and selective electrochemical biosensors for arsenic monitoring with use of certain microorganisms, enzymes (e.g. arsenate reductase)¹⁰¹ and immobilized biorecognition elements (e.g. L-leucine).¹⁰²