Comparing plasmonic electrodes prepared by electron-beam lithography and electrochemical reduction of an Au (iii) salt: application in active plasmonic devices

All chemical products were purchased from Sigma-Aldrich and used without further purification. Indium tin oxide (ITO)-coated glass plates with a resistance of 30–50 Ω/square were purchased from Solems.

A CHI 440A electrochemical workstation (CH instruments) was employed in all electrochemical experiments, which were carried out with a conventional one-compartment three-electrode cell. A Zeiss Supra 40 scanning electron microscope was used to evaluate the shape and size of AuNPs. The LSPR of AuNPs was recorded using an Ocean Optics HR4000 UV-visible spectrophotometer.

AuNPs arrays were obtained by electron beam lithography (EBL) performed with a Zeiss Supra 40 scanning electronic microscope (SEM) equipped with a lithographic system (NPGS software). In a first step, a 180 nm thick polymethyl-methacrylate (PMMA in chlorobenzene, from AllResist) layer was spin coated onto freshly cleaned ITO (with acetone, ethanol and propan-2-ol) and baked at 180 °C to evaporate the solvent. Then, the substrates were exposed to the electron beam, with a typical dose of 370 μC cm⁻². After exposure, chemical development was performed using methylisobutylketone: propan-2-ol (1:3) and then pure propan-2-ol was used to stop the development. Then, a 50 nm thick layer of gold was deposited by an ultravacuum evaporation process (Plassys MEB550S) with an evaporation rate of approximately 0.05 nm s⁻¹. Finally, a lift-off process was performed using acetone in order to remove the remaining PMMA layer together with the gold layer on top whereas, in the areas exposed to the electron beam, the AuNPs resting directly on the ITO substrate remained. The total size of the gratings produced by EBL was limited to 100 μm squares.

ITO electrodes were cleaned by sonication for 20 min in dilute alkaline Extran®solution (Merck), then carefully rinsed with distilled water and then ethanol; they were stored in ultrapure milli-Q water (18.2 MΩ cm⁻¹) until required for AuNP electrodeposition. Electrodeposition experiments were carried out by the chronoamperometry technique in a stirred aqueous solution of 2.10⁻³ M KAuCl₄ and 0.25 M Na₂CO₃. The reference was a saturated calomel electrode (SCE) and the counter-electrode was a stainless steel grid. The solution was maintained under argon atmosphere during the whole experiment. AuNPs were deposited on ITO electrodes at a constant potential and a controlled total charge current density.

1-(2-Bisthienyl)-4-aminobenzene (BTAB) was synthesized as previously described [28]. Grafting onto substrates was performed using in situ diazonium salt formation: under argon 30 eq. of tert-butyl nitrite were added to a 0.5 mM solution of BTAB in acetonitrile containing 0.1 M of tetra-n-butylammonium tetrafluoroborate (TBABF₄). BTB was grafted by multi-scan cyclic voltammetry (CV) (15 successive cycles) in the cathodic range (between 0.3 and −0.5 V/SCE) at a scan rate of 100 mV s⁻¹. By this means, films less than 15 nm thick are generated. A full description of the measurement of the thickness of BTB on gold gratings generated by e-beam can be found in a previous publication [7].

To make LSPR measurements under electrochemical control, a home-made planar three-electrode cell with a stainless steel grid counter-electrode and a silver (Ag/AgCl) pseudo-reference electrode was used. The AuNP array-modified working electrode was placed horizontally on the bottom. The cell was then filled with the electrolyte solution. Absorption measurements were carried out with an Ocean Optics HR 4000 UV–visible spectrophotometer coupled with a fiber-optic system which made it possible to analyze an area of 80 μm × 80 μm. Modulation of LSPR signals was checked by switching the applied potential between a cathodic and an anodic value, either by chronoamperometry or by CV.

A classical route to AuNP substrates consists in using e-beam lithography. This method relies on the formation of a nanohole-based template in a resin, drawn by the electron beam of a scanning electron microscope (SEM), followed by evaporating gold into the holes. After a lift-off step allowing to remove the resin, an AuNP array appears on the substrate. The process is explained in more detail in the experimental part of this paper.

Figure 1(A) shows the SEM image of a typical AuNP array obtained by this technique and figure 1(B) depicts the corresponding spectrum. The extinction band is attributed to the localized surface plasmon resonance (LSPR) of the AuNPs. The peak is very sharp, with its full-width at half-maximum (FWHM) around 70 nm, which indicates the homogeneity in size of the AuNPs.

Zoom In Zoom Out Reset image size

The advantage of this technique is not only the possibility of obtaining very perfect AuNP arrays (every NP with the same size, a controlled distance between them, etc...), but also that of modulating the shape and size of the NPs, as well as the distance between them, at will. Nevertheless, this method is a multi-step process which requires not only a long time but also expensive apparatus (electron microscope and metal evaporation apparatus). Moreover, no large-area plasmonic substrates can be obtained by this technique (AuNP arrays are typically around 100 μm wide using e-beam lithography).

In this context, there is a strong need to develop low-cost methods to obtain large-area AuNP plasmonic substrates, with good LSPR characteristics.

We propose an alternative method to obtain AuNP substrates based on the electrochemical reduction of a gold Au (iii) salt. Electrodeposition of AuNP is performed using a chronoamperometric method from an aqueous KAuCl₄ solution (2 mM) containing Na₂CO₃ (0.25 M) as supporting electrolyte. After the electrochemical process, the ITO electrodes are red or blue, depending on the experimental conditions used, which is clear evidence for the deposition of gold at the nanoscale.

The influence of the potential value applied for gold deposition on the size and dispersity of the AuNPs has been investigated. Indeed, chronoamperometry was performed at various potentials, the total charge density used during the reduction process being fixed at 20 mC cm⁻². SEM images (figure 2) show that the size of the AuNPs decreases and their density increases when the electrodeposition potential becomes more negative. This is confirmed by the NP diameter distribution histograms constructed from these SEM images. Indeed, the average size of the AuNPs decreases from 150 nm to 90 nm when the reduction potential decreases from −0.6 V to −1.0 V/SCE. Similar analysis performed on various areas of the sample show no significant variation in the size of the AuNPs, which indicates that their deposition is quite uniform over the 2 cm² area of the ITO electrode.

Zoom In Zoom Out Reset image size

Figure 3 shows the optical extinction spectra of the five AuNP substrates corresponding to the SEM images of figure 2, i.e. obtained at different reduction potentials (between −0.6 and −1.0 V/SCE) in fixing the total charge density at 20 mC cm⁻². For all samples, the strong absorption in the visible range can be attributed to the LSPR of AuNPs. Spectra taken in various areas of the sample do not show significant variation of the LSPR, confirming that AuNPs are uniformly deposited on the ITO electrode. The LSPR maximum wavelength is blue-shifted from 680 nm to 580 nm when the applied potential goes from −0.6 to −1 V/SCE, which is consistent with the formation of smaller AuNPs when it becomes more negative. Moreover, the full width at half maximum (FWHM) of the LSPR decreases from 200 to 80 nm on going from −0.6 to −1 V/SCE, showing an improvement in the optical response of the substrates. In particular, the sharpness of the LSPR signals observed for reduction potentials between −0.8 and −1 V/SCE is characteristic of plasmonic substrates with high quality factor, the FWHM being comparable to that of substrates obtained by e-beam lithography.

Zoom In Zoom Out Reset image size

In these experiments, it appears that the potential used to reduce the gold salt has a marked influence on AuNP nucleation. Indeed, the higher the negative potential, the easier nucleation will be, leading to an increase in the number of nuclei and in the overall NP density. As the total charge density used during electroreduction was kept constant for each deposition, a high density manifests itself in the formation of smaller NPs (at high negative potentials) while a low density leads to larger NPs (at low negative potentials). A more detailed explanation concerning the processes involved during the electrosynthesis of AuNPs has been recently published [24].

3.3.1. BTB electrografting onto AuNP electrodes

These two kinds of AuNP substrates have been compared for use as plasmonic electrodes in electroactive plasmonic devices.

An ultrathin organic film of oligoBTB was deposited on each kind of electrode by electroreduction of the corresponding diazonium salt [28]. The resulting covalently bonded layer is less than 15 nm thick [7]. Such a molecular film is known to undergo a reversible conductance change with variation of the voltage [28, 29], switching between a totally blocking reduced state and a conducting oxidized state at close to 0.6 V/SCE. Such a conductance switch can induce a change in the dielectric constant of the BTB layer and affect the LSPR signal.

After BTB had been grafted onto the AuNP substrates, a red-shift of the LSPR was observed (figure 4), whatever the process used to perform AuNP substrates. This red-shift of 40–70 nm is attributed to the modification of the dielectric environment [30–32] induced by the deposition of the thin film. Such a red-shift is commonly observed when the dielectric constant of the medium surrounding the NPs increases [7, 30–32].

Zoom In Zoom Out Reset image size

3.3.2. Electrochemical polarization of BTB|AuNP|ITO electrodes

The effect of the BTB conductance switch on the LSPR of AuNPs was investigated by varying the polarization of the BTB|AuNP electrodes. Ferrocene (Fc) was added to the electrolyte solution in order to ensure the stability of the BTB layer, since BTB is oxidized to reactive radical-cations. Indeed, under these conditions, BTB acts as a molecular bridge to oxidize Fc in solution and, thus, the accumulation of charges in the BTB film and solid-state coupling between reactive BTB radical cations, leading to film degradation, are avoided.

Figures 5(A) and (B) display the optical extinction spectra of BTB|AuNP devices obtained by e-beam lithography and electrodeposition, respectively, during the electrochemical switch from the reduced state to the oxidized state of BTB. Similar effects are observed whatever the BTB|AuNP electrode used. Indeed, a broadening accompanied by a decrease in the intensity of the LSPR band is observed, clearly showing modulation of the LSPR by the switching of the thin layer of BTB. Nevertheless, no shift of the LSPR band is observed. These dampings are large compared to previously reported equivalent systems. Indeed, while marked dampings are observed when thick layers (for example, a layer of a conducting polymer such as a PANI film) are deposited on AuNPs [5], the effects are generally smaller for NP systems using ultrathin films or molecular monolayers [22]. The quite large damping observed in our case can be attributed to the covalent bonds between the BTB and the AuNPs [7].

Zoom In Zoom Out Reset image size

Such observations can be qualitatively explained by the Drude free-electron model [33–35], in which the dielectric function of a conductive material is complex (ε = ε_Re + iε_Im) while the dielectric function of an insulating material is described only by a real value. In our case, the observed damping of the LSPR band, when the BTB layer shifts from its reduced state to its oxidized state, indicates that the imaginary part of the dielectric constant (ε_Im) of the BTB increases sharply, which is in agreement with a switch of the BTB film between an insulating and a conducting state [7, 26, 27]. This damping can be attributed to charge transfer between plasmon-generated hot electrons and the BTB film in its conducting state, which acts as an efficient acceptor, the LSPR lifetime being decreased by the presence of an additional decay channel [36]. On the other hand, the absence of LSPR band shift when the BTB film is switched indicates that the real part of the dielectric constant (ε_Re) of the film does not change significantly [7, 26, 27].

A more detailed study can be performed by scanning the potential by CV. Figure 6 shows how the intensity at the LSPR maximum and its FWHM depend on the potential applied during the CV, in the case of a BTB|AuNP electrode obtained by electrodeposition. In the forward scan, as long as the BTB film remains in its blocking state, i.e. when the potential is below 0.6 V/SCE, no or very little LSPR modulation occurs. However, when charges are injected into the BTB film, leading to its conductive state, the LSPR signal is progressively damped. Note also that the initial LSPR signal was restored upon polarization of the film at 0 V.

Zoom In Zoom Out Reset image size