Diffusional protection of electrode surfaces using regular arrays of immobilised droplets: overcoming interferences in electroanalysis

Abstract

Regular arrays of ca. micron sized droplets on a gold electrode surface can block diffusion to the electrode surface of one metal ion (which binds with the material in the droplet) whilst having no significant effect on another (which does not), so allowing interference effects in electroanalysis to be eliminated.

---

Electrochemistry is a uniquely powerful tool when used to analyse trace amounts of heavy metals in aqueous (typically drinking water) solutions. An extremely wide ranging variety of metals are known to exhibit analytically useful signals at electrode surfaces,¹ making electrodes an extremely promising platform for producing cheap and portable yet highly sensitive trace analysis equipment.^2,3 It is however the large amounts of different metals that exhibit signals at electrode surfaces that can give rise to one of the biggest difficulties when using an electrochemical technique to measure the quantity of a specific trace metal in solution. The presence of other “interfering” metals can lead to unwanted signals;^2,3 this is of particular relevance in real sample analysis where a wide variety of metals may be present in significant concentrations. It is extremely difficult to preferentially electro-deposit one metal in favour of another where more than one species is present. Although the deposition potential, pH or electrode composition⁴ may be “tweaked” in order to preference the deposition of one metal species over another, significant “contamination” is regularly a problem leading to reduced sensitivity and difficulty in deconvoluting the target analyte peaks.

A new concept for the preferential removal of these contamination signals is “diffusional protection”.⁵ This has recently been introduced in this research group. This method utilises a ligand coating which partially covers the electrode surface. If the ligand in question strongly binds one metal in preference to another or the kinetics of absorption of one metal into the ligand are much faster for the contaminant metal than the analyte metal then the contaminant may be absorbed into the ligand coating rather than the electrode surface thereby removing the interference signal from the analysis conducted on the uncovered part of the electrode surface. To achieve this the complexing ligand must be sufficiently spread over the surface of the electrode such that the diffusion zones around the ligand patches covers the surface completely and overrides the ability of the contaminant metal to diffuse to the electrode surface. However bare patches of the electrode must be left between the ligand coatings so that the target analyte is still able to reach the electrode surface. The previous work by Wildgoose et al.⁵ utilised glassy carbon balls modified with L-cysteine methyl ester groups randomly spread over an electrode surface to remove contaminants. A different and entirely novel approach that uses a regular array of droplets rather than randomly spaced spheres is reported in this paper. A lithographically designed surface with specific hydrophobic (compared to the gold electrode surface) and inert silicon oxinitride blocks where a ligand dissolved in an organic solvent may simply be attached by evaporating droplets on the electrode surface which would preferentially form on the hydrophobic blocks and therefore ensure total and reproducible coverage of the blocks is used. It can be seen in the example given in Fig. 1 that using this method the interferent B would be preferentially absorbed by the regularly spaced droplets on the electrode surface before it can reach the electrode surface whereas the species of interest A is allowed to diffuse unhindered towards the bare electrode surface. Diffusion of B towards the electrode is prevented because the droplets act as sinks of this material, setting up concentration gradients which draw B to the droplet surface in a Fickian manner so protecting the electrode from B. This paper demonstrates proof of this concept where As(III) is preferentially absorbed into the droplets coating an electrode surface whilst Cu(II) is allowed to diffuse to the electrode surface. Regularly blocked gold electrodes 2.721 × 2.640 mm in size were obtained with 79404 five µm diameter circular silicon oxinitride blocks of 0.5 µm in height placed in a regular pattern on the gold surface, Fig. 2. The electrodes were made using standard photolithographic techniques, briefly in the first step of manufacturing a silicon wafer was oxidised (800 nm of SiO₂) at 1100 °C. The metallization was performed in a two step process, after deposition of a platinum layer, a titanium layer was deposited as an adhesion promoter (50 nm) followed by nickel (50 nm) and a gold (200 nm) layer. This was subsequently covered in a photoresist and subsequently patterned by a wet etching procedure. The surface was then covered by an insulating silicon oxinitride layer (500 nm) grown by plasma enhanced chemical vapour deposition (PECVD). This layer was patterned and the areas which act as working electrodes were then defined photolithographically and opened by dry etching leaving “blocks” of silicon oxinitride on the surface.^6,7

Fig. 1 Scheme showing the “diffusional protection” of an electrode by regularly placed droplets on the electrode surface. The interferent B is preferentially absorbed into the droplets while the target analyte A is left unhindered as it diffuses toward the bare electrode surface between the droplets. The dashed lines indicate the diffusional flux of B into the droplets.

Fig. 2 Diagram showing the layout and arrangement of blocks on the gold electrode surface, d = 5 µm and p = 10 µm. Image size 540 × 400 µm.

Before use the electrodes were electrochemically activated by repeatedly cycling the potential from 0 V to −1.5 V and back in 0.1 M KCl. The electrodes were then coated with a highly viscous industrial heavy metal complexing ligand Acorga P50 (5-nonyl-salicylaldoxime). The Acorga droplets were found to form only on the relatively hydrophobic silicon oxinitride blocks rather than the gold surface. The droplets were formed by dissolving 0.063 g of the ligand in 25 mL dichloroethane. A 10 µl droplet of this solution was then micropipetted onto the clean electrode surface and the DCE was allowed to evaporate. As the solvent evaporated 5 µm droplets of Acorga P50 were left on the electrode surface as dictated by the block dimensions and spacings. Fig. 3 contains a sequence of images depicting the droplet formation. Image 3(a) shows the bare uncoated electrode surface with the 5 µm blocks clearly visible. Image 3(f) shows the DCE droplet evaporating from right to left leaving behind droplets of Acorga P50 on the block surfaces. Fig. 3(i) shows that once the solvent has fully evaporated a perfect regular pattern of droplets of identical size is left on the blocks while the gold surface remained clear of droplets.

Fig. 3 A sequence of optical images depicting the Acorga P50 droplets formation over the hydrophobic blocks. Image (a) shows the bare electrode surface in which the blocks are clearly visible. From (b) to (i) the gradual evaporation of the 10 µl DCE droplet and the formation of individual 5 µm Acorga P50 droplets by image (i) is depicted.

The modified electrodes were then assessed for their performance with respect to simultaneous Cu(II) and As(III) detection. These metals are of particular interest as they regularly occur together in drinking water samples where the concentration of either As(III) or Cu(II) is required to be measured. Whilst the As(0) to As(III) and Cu(0) to Cu(II) stripping peaks occur within 100 mV of each other there is also a third peak based on the intermetallic compound Cu₃As₂.⁸ This makes the deconvolution of individual As or Cu peaks very difficult when the metals are both present in similar quantities. If the Acorga ligand preferentially absorbed one metal species over the other it could prevent a significant interference problem.

Fig. 4(A) demonstrates this problem at a gold electrode surface before coating with Acorga; cyclic voltammograms were performed in 0.1 M nitric acid at 100 mV s⁻¹ where the potential was swept from 0.4 V to −0.5 V (vs. SCE) and back. Four subsequent additions of 0.1 mM As(III) are shown to produce a characteristic As(III) stripping peak at ∼0.15 V. Subsequent additions of 0.1 mM Cu(II) were then performed and it can be seen that this results in a positive shift and amplification of the As(III) stripping peak as well as the emergence of the Cu₂As₃ intermetallic peak at ∼0.3 V. A very small peak due to Cu(II) stripping can be seen at ∼0 V.

Fig. 4 Cyclic voltammograms at (A) an uncoated and (B) an Acorga droplet coated electrode conducted in 0.1 M nitric acid and performed at 100 mV s⁻¹ beginning at 0.4 V (vs. SCE). A blank scan is shown followed by four successive additions of 0.1 mM As(III) and then five additions of 0.1 mM Cu(II).

Fig. 4(B) shows the same experiment repeated at an Acorga P50 droplet coated electrode, it can be seen that the initial five additions of 0.1 mM As(III) to the blank 0.1 M nitric acid solution produce no significant change in the voltammogram; the Acorga droplets appear to be acting to effectively block diffusion of As(III) towards the electrode surface. The following four additions of 0.1 mM Cu(II) are then seen to produce the reduction peak which can also be seen at ∼−0.4 V and the corresponding Cu(II) stripping peak at 0 V. The copper stripping peak now appears alone with no apparent significant interference from the arsenic present. Two tiny peaks are just visible representing the stripping of As(III) and Cu₂As₃ at 0.15 and 0.3 V respectively demonstrating that only an amount of As(III) an order of magnitude lower than that of Cu(II) is able to diffuse to the electrode surface.

In summary we have demonstrated the use of a diffusionally protected surface using a silicon oxinitride blocking structure that allows the size and spacing of viscous droplets on an electrode surface to be carefully and reproducibly positioned, ensuring the entire surface is protected. The modified gold electrode is then able to distinguish between two different metal ions as the droplet coatings prevents one of them from reaching the electrode surface and thereby removes the interference of one metal over another in a voltammetric experiment.

AOS thanks the EPRSC and Abington Partners Sensing for support via an Industrial CASE studentship.

Notes and references

1. J. Wang, Analytical Electrochemistry, 2nd edn, Wiley, New York, 2006.
2. A. O. Simm, C. E. Banks and R. G. Compton, Anal. Chem., 2004, 76, 5051.
3. H. L. Huang and P. K. Dasgupta, Anal. Chim. Acta, 1999, 380, 27.
4. X. Dai and R. G. Compton, Electroanalysis, 2005, 17, 1325.
5. G. G. Wildgoose, F. G. Chevallier, L. Xiao, C. A. Thorogood, S. J. Wilkins, A. Crossley, L. Jiang, T. T. G. Jones, J. H. Jones and R. G. Compton, 2006, submitted.
6. T. J. Davies, S. Ward-Jones, C. E. Banks, J. del Campo, F. X. Munoz and R. G. Compton, J. Electroanal. Chem., 2005, 585, 51.
7. F. G. Chevallier, N. Fietkau, J. del Campo, R. Mas, F. X. Munoz, L. Jiang, T. G. J. Jones and R. G. Compton, J. Electroanal. Chem., 2006 DOI:10.1016/j.jelechem.2006.06.015.
8. M. Kotoucek, J. Vascova and J. Ruzicka, Mikrochim. Acta, 1993, 111, 55.

---