An ionogel composite including copolymer nanowires for disposable electrochemiluminescent sensor configurations

Abstract

Aniline derivatives such as luminol and benzidines can be electropolymerized for the preparation of electrochemiluminescent sensors. However, the development of these sensors requires formulations to improve the chemical stability and photoemissive and electrochemical properties. The use of ionic liquids confined in gels (ionogels) can provide a material for electrochemiluminescent sensor preparation and more specifically by means of hybrid organic–inorganic composites which provide flexibility, thermal stability, and ionic transport. By using this configuration, we prepared nanowires of luminescent polymers by electropolymerization in the pores of a hydrophilic silica ionogel with chitosan, offering a biocomposite that can be prepared in disposable screen-printed gold electrodes. The properties of the prepared biocomposite have been characterized by emissive luminescence, cyclic voltammetry, electrochemical quartz crystal microbalance and scanning electron microscopy. The results show that the electrochemiluminescent copolymer grows efficiently as nanowires in the pores of the biocomposite, forming sensing films that produce an intense luminescent blank signal due to the influence of the ionic liquid. Hydrogen peroxide was used as an analyte; it acts as a scavenger, reducing the emissive signal and resulting in a novel electrochemiluminescence configuration for sensing. This configuration broadens the field of electrochemiluminescence sensors due to the biocompatibility of the developed ionogel membrane.

---

1. Introduction

Electrochemiluminescence (ECL), the generation of light-emitting species by means of strong oxidizing and reducing agents,¹ has great potential as an analytical technique for detection in immunoassays, DNA probe assays, sensors and biosensors, mainly in flow and optical fibre formats, high throughput analysis, microfluidic devices and chromatographic and electrophoretic systems.^2–5 Proof of the growing interest in this technique in the scientific community is the exponential growth of ECL papers devoted to analytical chemistry (3056 papers in the last 20 years, doubling every 7 years).

The ECL can be carried out mainly by two ECL pathways: hot electron-induced and co-reactant. In hot electron-induced pathways, the ECL emission is initiated by hot energetic electrons from a thin insulator film covered in a conductor/insulator/electrolyte surface during cathodic pulse polarization.⁶ Sun et al.⁷ demonstrated that screen-printed carbon electrodes, instead of the common Al/Al₂O₃ electrode, can generate hot electrons; they proposed procedures for quercetin and dissolved oxygen determination.⁸

However, more common is the co-reactant ECL pathway where one-directional potential scanning is applied on the electrode in the presence of a luminophore and co-reactant, generating co-reactant intermediates that are either strong reducing agents (in oxidative-reduction ECL) or strong oxidizing agents (in reductive-oxidation ECL).⁹

When this co-reactant approach is used for disposable format sensors, immobilization is needed for the sensing chemistry, including luminophores; different immobilization strategies are used depending on the analytical scheme utilized and the type of luminophore.

In the case of ruthenium complexes as luminophores, mainly Ru(II)-bis(2,2′-bipyridyl), the common procedures for immobilization consist of a cationic exchange polymer, such as Nafion;¹⁰ inclusion in a self-assembled monolayer of ruthenium complex-aminopropyl imidazole on a gold-deposited screen printed cell containing carbon nanotubes;¹¹ or adsorption on electrochemically oxidized graphite in screen printed cells.¹²

However, the most popular chemistry is based on luminol, for which different forms of immobilization have been proposed including entrapment in membranes, ion-exchange, covalent bonding and electropolymerization.¹³ The ability of luminol to undergo electropolymerization by cyclic voltammetry is an interesting option to produce electrochemiluminescent polymers on the transducer surface. These have been prepared on graphite, glassy carbon, platinum, gold, screen-printed and indium tin oxide electrodes, typically in acidic conditions.^14–20

In previous papers, we studied the improvement in luminol immobilization as a copolymer with 3,3′,5,5′-tetramethylbenzidine¹³ or biotinylated pyrroles²¹ for the preparation of disposable ECL sensors by means of electrochemical copolymerization in terms of operational and temporal stability because the polymerization of luminol alone has some drawbacks such as undesirable mechanical properties, luminol adsorption in the polymer, short reusability and poor electroactivity in basic medium.^18,22

In order to improve the immobilization of reagents using the electropolymerization process, two approaches can be used: (a) hydrogels and (b) ionogels. In the case of hydrogels, i.e. gels based in water, the hybrid described by Montilla et al.²³ can be cited, where silica gel is used to produce polyaniline nanowires by electrodeposition, generating a polyaniline/silica composite material. With the aim of improving the ability of an ECL sensing film to host biorecognition elements and better hydration conditions, recently Loïc Blum et al.²⁴ have developed electrochemiluminescent biosensors based on hybrids of electrogenerated polyluminol nanowires and hydrogels such as silica and calcium alginate for hydrogen peroxide and choline detection.

The second approach for reagent immobilization consists of the preparation of ionogels, a new class of hybrid materials, characterized by the incorporation of ionic liquids (ILs) in the gel. ILs are salts of organic cations such as ammonium, phosphonium, imidazolium, or pyridinium.²⁵ These salts are characterized by their high ionic conductivity and their wide electrochemical potential window (up to 6.0 V). At this time they are widely considered as “green” materials due to their possible recyclability in most applications²⁶ and for their chemical stability. One of the reasons that ionogels can be suitable for ECL analysis is the improvement in both electrochemical and luminescent properties by the ready diffusion of analytes into the immobilized IL phase.²⁵

To date, ionogels consist of ILs hybridized with other components to improve the membrane properties, which may be the following: (1) organic, (2) inorganic or (3) hybrid organic–inorganic (e.g. polymer and inorganic fillers).²⁷

The first case includes two modalities: (a) organic ionogels where low molecular weight gelators are used such as L-glutamic or aspartame acids and (b) polymer gels, which combine the mechanical flexibility of a polymer and the characteristic conductivity, such as the described compatibility of imidazolium salts with poly(methylmethacrylate), among others.

The second case is based on inorganic ionogels called “bucky” gels, such as carbon nanotubes dispersed in ILs where a gelation effect occurs (crosslinking effect), and silica-based ionogels. Silica is a suitable matrix because its properties, such as pore size, orientation, and shape, can be easily tuned during synthesis.²⁸ In this case, two modalities can be distinguished: (a) dispersion of silica nanoparticles into ILs, where bare silica colloids were shown to be unstable in ILs, and (b) sol–gel processing, where the IL can be used as a template that is removed at the end of the gelation or as an enhancer of the conductivity, e.g. monolithic materials.

With respect to the possibilities of ECL sensor preparation using polymers, we focused on the third case, composite (hybrid organic–inorganic) ionogels that consist of the association of molecular organogelators (polymers) and inorganic nanoparticles providing flexibility, thermal stability, and ionic transport. As an example, Leroux et al.²⁹ prepared polymer nanocomposites by incorporating silica nanofillers of silicon tetraethyloxide covalently bonded to the polymer chains as poly(methylmethacrylate) with imidazolium salts as bis-(trifluoromethanesulfonyl)imide.

Due to the potential of this last type of ionogel, this paper explores the feasibility of preparing ionogels using different ILs on screen printed cells for sensing purposes, combining organic materials such as the copolymer luminol with 3,3′,5,5′-tetramethylbenzidine as luminophore and chitosan with inorganic silica by means of electrochemical growth of copolymer nanowires in the pores of the ionogel matrix. To date, this configuration has not been used and is proposed in this work.

2. Experimental

2.1 Chemicals

As monomers for electropolymerization, stock solutions of 1 mM of luminol (5-amino-2,3-dihydro-1,4-phthalazinedione), 97% and 3,3′,5,5′-tetramethylbenzidine, 97% (TMB), both supplied by Sigma (Sigma-Aldrich Química S.A., Madrid, Spain, https://www.sigmaaldrich.com/spain.html) were used. A stock buffer phosphate solution containing 0.5 M phosphate (Na₂HPO₄) and 0.25 M NaCl adjusted to the necessary pH by adding NaOH or HCl of appropriate concentrations was prepared. Different ionic liquids (ILs) were used, such as triethylsulfonium bis(trifluoromethylsulfonyl)imide ([TES][TFSI], 97%), 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIM][BF₄], 97%), 1-ethyl-3-methylimidazolium tetrafluoroborate ([EMIM][BF₄], 97%), 1-decyl-3-methyl-imidazolium chloride ([decyl-MIM][Cl], 96%), 1-hexyl-3-methyl-imidazolium hexafluorophosphate ([HMIM][PF₆], 97%), 1-dodecyl-3-methyl-imidazolium iodide ([dodecyl-MIM][I], 95%); 1-hexyl-3-methyl-imidazolium tetrafluoroborate ([HMIM][BF₄], 97%), 1-methyl-3-octyl-imidazolium chloride ([MOIM][Cl], 97%), and 1-butyl-3-methyl-imidazolium hexafluorophosphate ([BMIM][PF₆], 97%). Hydrogen peroxide solutions were prepared daily from H₂O₂ 33% w/v (110 vol) that was supplied from Panreac (Panreac Química S.L, Barcelona, Spain, http://www.panreac.es/) and was standardized iodimetrically. Other reagents and materials used were as follows: 1,1′-ferrocenedicarboxylic acid (97%) and Nafion 5% w/w solution in a mixture of aliphatic alcohols (Sigma-Aldrich). Chitosan was prepared by dissolving 1 g in 100 ml water containing 1 ml of acetic acid (95.5%, Panreac) and stirring for 4 hours; the prepared dilution of 1 g l⁻¹ stock solution of chitosan was prepared by dilution with 1% acetic acid. Tetraethyl orthosilicate (TEOS) stock solution (2.4 M) was prepared by mixing the following under vigorous stirring at room temperature, 4.46 ml of TEOS, reagent grade 98% (Sigma-Aldrich), 1.44 ml of water and 6 μl of 4.0 M HCl in a closed vial for 2 h. After that, 1 ml of this solution was mixed carefully with 1 ml of water and submitted to rotaevaporation with periodic control of the weight loss to 0.62 g due to the elimination of ethanol by an alkoxide hydrolysis reaction. The composite was prepared using the stock solutions indicated above (see Fig. 1), obtaining the ionogel solution. Reverse-osmosis type quality water (Milli-Q Plus185 from Millipore, Molsheim, France) was used throughout.

Fig. 1 Schematic illustration of the preparation steps of the composite on the disposable sensor.

2.2 Apparatus and software

The ECL emission from screen-printed cells was measured using an H8529 photomultiplier (PMT) interfaced to a C8855 USB photo counting unit, both from Hamamatsu (Hamamatsu Technologies K.K., Shizuoka, Japan), connected to a PC. The potentiostat used was an Autolab PGSTAT 128N with an F120 module for chrono-coulometric measurements (Metrohm Autolab B.V., Utrecht, Netherlands). The arrangement used for ECL studies has been described in previous works³⁰ and basically consists of a black box that contains two black methacrylate piece holders, where one piece fits on top of the other, one to insert the electrochemical cell in a fixed position and the other to hold the PMT.

An electrochemical quartz crystal microbalance (EQCM) (Metrohm Autolab B.V.) was used to study the polymerization process by recording the change in resonant frequency of a 6 MHz EQCM Crystal Au/TiO₂ quartz crystal oscillator. A Cary Eclipse fluorescence spectrometer (Varian Australia Pty Ltd.) was used for luminescence measurements. As the conventional Scanning Electron Microscopy technique induces an instantaneous dehydration of the sample due to the high vacuum needed, we used environmental scanning electron microscopy (ESEM), which makes it possible to examine any specimen, wet or dry, insulating or conducting, in situ and close to its natural state. An ESEM FEI model Quanta (FEI Co., Hillsboro, Oregon, USA) 3.5 nm was used for imaging with increases ranging from 7× to 1 000 000× and equipped with a ZEISS DSM 250 electronic microscope prepared to perform energy dispersive X-ray spectroscopy (EDX) microanalysis. Samples were mounted on Peltier cooling stages set at 2 °C with a working distance of about 10 mm and a humidity of 80%. A Crison digital pH-meter (Crison Instruments, Barcelona, Spain) with a combined glass-saturated calomel electrode was used for pH measurement.

Software programs used were: Statgraphics Centurion software package (Manugistics Inc. and Statistical Graphics Corporation, USA, 2007); Microsoft Office Suite 2013 (Microsoft Corp., Redmond, WA, USA); CSW32 v.1.3.3 (2001); CSWAIA v 1.7.3 (2001) (Dataapex software, Czech Republic); NOVA v.1.8.17 (2005–2012) (Metrohm Autolab B.V.) and GPES v.4.9 (2007) (Ecochemie, Netherlands).

2.3 SPE cells and pre-treatments

Low temperature (LT) gold screen-printed cells (SPE) were selected for sensor preparation based on our previous studies of a convenient electropolymerization of aniline derivatives.¹³ These cells were supplied by Dropsens S.A. (Oviedo, Spain, http://www.dropsens.com/). The cells consist of a round-shaped working electrode and a counter electrode, both of the same material, and a silver pseudo-reference electrode on a ceramic support. Before being used, the electrochemical cells were tested for uniform behaviour. In order to prepare a receptacle on the disposable electrochemical cell, we printed the round form on the working electrode by screen printing using a plastisol ink containing 30% Puff additive, which expands during heating (170–200 °C), giving a raised print used as a receptacle of approximately 50 μl capacity.

Initial pre-treatment of gold SPE cells was carried out by immersing them in 0.2 M H₂SO₄ and applying three voltammetric sweeps between −0.2 and 1.0 V. Then, the SPE cells were dried for 10 min at room temperature. After the addition of the ionogel solution to the SPE cell (as shown in the following section), a second pre-treatment was carried out by immersing the cell in the same 0.2 M H₂SO₄ solution and applying 40 voltammetric sweeps between −0.2 and 1.0 V. For the two pre-treatments, 0.1 V s⁻¹ was set as the scan speed.

2.4 Ionogel sensor preparation

An ionogel solution was prepared by mixing 20 μl TEOS stock solution (2.4 M), 20 μl of 1.0 g l⁻¹ chitosan solution and 30 μl of [TES][TFSI] stock solution in an ultrasonic bath. This ionogel solution can be used for a week if maintained at −20 °C. To prepare the composite and prompt the gelation, 70 μl of pH 7.4 PBS solution containing 0.2 M phosphate and 0.25 M NaCl was added finally to the inonogel solution. 5 μl of this mixture was dropped onto the working electrode of the gold LT SPE cell, waiting for two minutes at room temperature after use.

After the second pre-treatment, the luminol–TMB copolymer was grown in the pores of the prepared composite on the SPE cell by cyclic voltammetry working between −0.2 and 1.0 V, at 0.1 V s⁻¹, with 20 cycles from a solution of 1 mM luminol and TMB (1 : 1 ratio) in 0.2 M H₂SO₄ (standard conditions). Then, the sensor was washed with 0.2 M pH 7.4 phosphate buffer solution to eliminate adsorbed luminol or TMB traces. The prepared cells were stored in dark conditions at 4 °C until use (see Fig. 1).

2.5 ECL procedure and signal measurements

The experimental system for ECL measurement has been described elsewhere using the instrumentation shown in the apparatus and software section.³⁰ Aqueous standard solutions of different concentrations of H₂O₂ at pH 9.0 adjusted with a stock solution of saline phosphate buffer were prepared. 50 μl of the H₂O₂ solution was deposited into the receptacle of the SPE, which was then placed in a homemade holder and covered with a lid holding the PMT. The ECL signal obtained for analytical purposes was the integral signal over the course of 2 min coming from consecutives pulses of 0.6 V for 1 s each with a dwell time of 10 s.

3. Results and discussion

The goal of this study is the preparation and characterization of SPE cells with an ionogel composite³¹ with biocompatibility and biodegradability properties for electrochemiluminescent detection. This biocomposite is prepared by a ionogel solution that consists of the biopolymer chitosan, TEOS stock solution and an IL, inserting ECL nanowires into the pores of the solid ionogel prompted by 0.2 M PBS solution by in situ copolymerization of luminol and an aromatic amine: 3,3′,5,5′-tetramethylbenzidine. A study of biocomposite electro-preparation and ECL characterization on different SPE cells was performed.

3.1 SPE sensor composition

Sensing biocomposite layer. In previous works, our research group has studied the influence of reagents and the composition of the ECL transduction layer in sensor preparations using chitosan and polymers to enhance the mechanical and electrochemical properties.^13,32,33 However, while the use of sol–gel membranes can improve biocompatibility, they lack stability and conductivity and can undergo dehydration and cracking over time. A combination of different polymers such as the synthetic anionic polymer nafion and the cationic polysaccharide chitosan can improve the above characteristics.

In the first instance, a sol–gel configuration was used by immobilizing the electroformed poly(luminol–TMB) as nanowires, obtaining an intense ECL signal with H₂O₂ (see Table 1). However, the membrane was brittle and cracked some minutes after preparation due to the rapid dehydration of the membrane, resulting in an unstable sensor. To overcome the observed deficiencies it was necessary to include other components in the membrane in order to produce a hydrogel. Nafion, an anionic sulfonated perfluorinated vinyl ether polymer, widely used as an ion-exchange polymer, is described as a hydration agent in the preparation of sol–gel sensors by water adsorption at the hydrophilic ionic groups forming ionic clusters.³⁴ In this case, its use for composite preparation generated a reduction in the ECL signal of approximately 60% compared to the cell containing only sol–gel, and a retraction in the membrane with fast cracking in a few minutes. This behaviour can be attributed to the fact that nafion decelerates the electron transfer as well as the diffusion of luminol.³⁵

Table 1 ECL analytical signal of the different composites to obtain a better sensor configuration. In all cases, the analyte was hydrogen peroxide 5.0 μM and electropolymerized nanowires of poly(luminol–TMB) were included in all the composites

Composite	Layer composition	ECL analytical signal a.u. s	%
Sol–gel	TEOS	397	84.2
Hydrogel	TEOS/nafion	77	16.3
	TEOS/chitosan	311	66.1
	TEOS/chitosan/nafion	155	32.8
	TEOS/chitosan/nafion/ferrocene	11	2.3
	TEOS/chitosan/ferrocene	20	4.2
Ionogel	TEOS/chitosan/[TES][TFSI]	472	100.0

---

Moreover, chitosan, a positively charged natural polysaccharide, generates a sol–gel that is highly porous and structured, creating a hydrophilic environment.³⁶ The biocomposite is prepared by entrapping the chitosan in the porous silica matrix when the pH of the solution containing sol nanoparticles of silica and chitosan is neutralized, enabling crosslinking. In this biocomposite, the electron mobility is reduced as is, consequently, the ECL signal to some extent by approximately 21.5% with respect to the SPE cell containing only sol–gel, although the biocomposite membrane shows good mechanical properties.

The combined use of chitosan and nafion along with sol–gel in a hydrogel did not provide any differential characteristics, only an increase in the ECL signal with respect to that of nafion (50.4%) and a decrease with respect to that of chitosan (50.3%). With the goal of improving the ECL signal in the chitosan/sol–gel material, we studied the effect of the incorporation of electronic mediators and ionic conductors. The inclusion of a typical electronic mediator such as 1,1′-ferrocene dicarboxylic acid, generally used as an electronic transporter in sensing membranes,³⁷ dramatically reduced the ECL emission, centred at 440 nm, approximately 95% compared to the sol–gel layer, due to the absorption of ferrocene at 450 nm.

Contrastingly, the incorporation of ionic liquids (ILs) produces an ionogel that shows a large increase in the ECL signal, which makes these liquids useful for biocomposite preparation. ILs can be classified, according to their preparation, in three generations:³⁸ (1) those based on 1-alkyl-3-methylimidazolium salts with tetrachloroaluminates or halides (based on eutectics); (2) those obtained by replacing the anion by tetrafluoroborate or other anions (discrete anions); (3) covalent inclusion of a functional group in the cation, anion or both. In this study, we tested different ionogels with silica and chitosan prepared from ILs belonging to three generations (see Table 2). The results show that ILs from the first generation prevent the preparation of the biocomposite due to solubility problems, an aspect that is improved in second generation ILs due to their hydrophilic properties, which produce a water-stable composite. However, the produced ECL intensity reached only 21% of the maximum. The use of the third generation IL triethylsulfonium bis(trifluoromethylsulfonyl)imide ([TES][TFSI]) dramatically increased the ECL intensity.

Table 2 Preparation of ionogels using different ILs, grouped by generations. All composites prepared contained 1 mmol of IL and were tested against hydrogen peroxide (5.0 μM) with electropolymerized nanowires of poly(luminol–TMB). The ECL analytical signal was recovered

Composite	Generation	IL	ECL analytical signal a.u. s	%
Ionogel (TEOS/chitosan/IL)	First	[Decyl-MIM][Cl]	Solubility problems	—
		[Dodecyl-MIM][I]
		[MOIM][Cl]
	Second	[HMIM][PF6]	—	—
		[EMIM][BF4]	74	15.7
		[BMIM][BF4]	87	18.5
		[HMIM][BF4]	—	—
		[BMIM][ PF6]	98	20.9
	Third	[TES][TFSI]	472	100.0

---

The different behaviour of ionogels containing the studied ILs can be justified by their fluorescence properties. The ILs containing bulky imidazolium cations showed absorption bands around 350 nm with a fluorescence emission near 480 nm, although their confinement in the silica matrix modifies the ground and excited states involved in fluorescence, shifting their absorption and emission maxima, i.e., the emission peak, from 484 to 579 nm in the case of [BMIM][PF₆], consequently reducing the ECL emission due to absorption by imidazolium ILs.³⁹ In turn, the [TES][TFSI] showed an absorption band at 373 nm with the emission band at 431 nm, which apparently did not affect the ECL process. Thus, the [TES][TFSI] containing triflimide anion was selected. This IL is widely used for its relatively low viscosity and high ionic mobility and thus, for the electro-formation of conducting polymers,⁴⁰ such as polypyrrole⁴¹ and poly(3-methylthiophene).⁴²

A cyclic voltammogram on a gold SPE cell using 0.2 M H₂SO₄ solution (Fig. 2A(a)) showed an oxidation peak at 1.0 V corresponding to the oxidation of gold by water, and subsequently a reduction peak at 0.75 V due to the reduction of Au oxides formed in the anodic reaction on the working electrode. The placement of a sol–gel layer on the working electrode (Fig. 2A(b)) dramatically reduced the oxidation and reduction peaks due to the low conductivity of the material. The inclusion of chitosan in the sol–gel layer does not apparently improve the electrochemical properties (Fig. 2A(c)). The contribution of ionic liquid to the conductivity is highlighted in Fig. 2A(d), showing left-shifted gold oxidation and reduction peaks that result in a modified SPE electrode that improves the electrochemical characteristics of the bare gold SPE electrode.

Fig. 2 Preparation of poly(luminol–TMB) in the biocomposite using a gold SPE cell. (A) Cyclic voltammograms (last cycle of pre-treatment): black (a): bare cell; red (b): sol–gel; green (c): TEOS/chitosan; blue (d): TEOS/chitosan/IL. (B) Cyclic voltammogram (last cycle of copolymerization process): black (a): bare cell; red (b): sol–gel; green (c): TEOS/chitosan; blue (d): TEOS/chitosan/IL. (C) Cyclic voltammograms of luminol–TMB copolymer formation. Arrows in the figure indicate the direction of the successive cycles. (D) EQCM study of luminol–TMB copolymer formation. The monomer concentrations are 1 mM in 0.2 M H₂SO₄, and the pre-treatment was carried out at 0.2 M H₂SO₄.

Preparation and characterization of ECL nanowires. After the selection of different elements in the preparation of the sensor, we proceeded to study the preparation and characterization of the biocomposite containing the ECL copolymer nanowires that is initially evidenced by small dark blue dots that can be seen on the surface.

To study the electro-formation of the copolymer, first a bare Au SPE cell was used with luminol and TMB solutions working under the same electropolymerization conditions as previously studied by our group¹³ (between −0.2 and 1.0 V, at 0.1 V s⁻¹, from 1 mM solution of both luminol and TMB (1 : 1 ratio) in 0.2 M H₂SO₄), obtaining the results in Fig. 2B(a).

A study of luminol–TMB copolymer formation by cyclic voltammetry in the ionogel TEOS/chitosan/IL biocomposite is shown in Fig. 2C.

An increase in the intensity and left-shifted anodic and cathodic peaks was observed for copolymerization on a sol–gel layer on an Au SPE cell (Fig. 2B(b)). However, this membrane experienced a quick cracking effect due to poor hydration. The polymerization in the hydrogel TEOS/chitosan biocomposite produced a dramatic decrease in the intensity of the voltammetric peaks that appeared at the same voltage (Fig. 2B(c)). Finally, the inclusion of [TES][TFSI] IL in the biocomposite improved the formation of the copolymer (Fig. 2B(d)).

A study of luminol–TMB copolymer formation by cyclic voltammetry in the ionogel TEOS/chitosan/IL biocomposite is shown in Fig. 2C. The same previously described pattern can be observed:³³ TMB oxidation peak at 0.60 V; luminol oxidation peak at 0.89 V; copolymer oxidation at 0.63 V and reduction at 0.54 V. The EQCM study corroborates the electro-formation of the copolymer (Fig. 2D), showing a progressive increase in the mass by decreasing the resonance frequency by cycling. Using the Sauerbrey equation,⁴³ the calculated mass of the deposited copolymer after 20 cycles was 674.8 ng cm⁻².

The superficial topology of the porous biocomposite (Fig. 3B), obtained after the second pre-treatment, prepares the composite for the growth of copolymer strands in the porous structure (5 μm of pore size), giving rise to a massive growth of poly(luminol–TMB) nanowires (Fig. 3C). An interconnected filament shaped structure is observed when enlarging the image (Fig. 3D) in which the EDX spectrum confirms the presence of carbon from the copolymer and sulphur that can arise from either hydrogen sulfate or triflimide anions. In the latter case, this is the product of the ionic liquid that has been described as a dopant in polypyrroles.⁴¹

Fig. 3 ESEM images of biocomposite layer on gold SPE at different resolutions. (A) Bare biocomposite matrix; (B) pre-treated biocomposite matrix; (C) nanowires of poly(luminol–TMB) in the biocomposite; (D) filament shaped structure of poly(luminol–TMB) nanowires. EDX microanalysis spectra are included. The ECL polymer was prepared using the same monomer concentration of 1 mM in 0.2 M H₂SO₄.

SPE cell and ionogel cell pre-treatments. A first pre-treatment of a bare gold SPE cell generates a functional surface for polymerization, removing organic ink constituents and contaminants, stressing the surface roughness and preparing the gold particles for a redox response.¹⁹ In this case, the pre-treatment consisted of three voltammetric sweeps from −0.2 to 1.0 V at 0.1 V s⁻¹ in 0.2 M H₂SO₄.³³

Once the electrode with the biocomposite was prepared and before the electropolymerization process, a new pre-treatment step was necessary to improve the immobilization of the copolymer and the stability of the ionogel membrane. The morphology of the composite (Fig. 3A) is considerably modified after pre-treatment, increasing the number and distribution of the pores (Fig. 3B) and improving the nanowire formation on them.

After 40 cycles in the same previous conditions, a better electro-formation of the copolymer occurred after the second pre-treatment. This second pre-treatment can have three functions: (a) a polishing effect that eliminates traces or extra components of the composite; (b) a removing effect that permits the pore walls to strengthen on aging because some removed IL acts as a template,²⁷ as we can observe by an EQCM study where a constant loss of mass occurred after each new pre-treatment cycle, resulting in a mass loss of 858.9 ng cm⁻² (after 40 cycles); and (c) an interchange of the counter ion acetate present in chitosan by hydrogen sulphate. This is sustained by the EDX study that shows the presence of sulphur that was not present earlier in the composite structure after pre-treatment (see Fig. 3A and B). The increase in hydrogen sulphate in the membrane favours polymerization due to the fact that it acts as a dopant agent in copolymer formation.⁴⁴

Optimization of the sensing biocomposite layer. The composition of the biocomposite layer formed by TEOS, chitosan and the selected IL was optimized individually. The TEOS concentration was maintained constantly at 0.35 M as a starting point by using the TEOS stock solution previously described. Different concentrations of chitosan ranging from 0.01 to 1.10 g l⁻¹ along with 1 mmol [TES][TFSI] were used to prepare the sensing layer. The poly(luminol–TMB) was prepared in all cases by means of 40 cycles with the same electrodynamic conditions. The ECL signal increased with chitosan concentration up to 0.14 g l⁻¹, decreasing afterwards (Fig. 4A). This result is due to the increase in composite pore size described by El-Kadib and Bousmina⁴⁵ when the TEOS/chitosan ratio decreases, which involves an increase in the amount of copolymer formed as shown by cyclic voltammetry. A further increase in the amount of copolymer reduces the emitted ECL due to autoabsorption from the copolymer.³³

Fig. 4 Optimization of composite layer formation. (A) Influence of chitosan concentration of biocomposite layer on ECL signal. Conditions: 1 mM monomers (1 : 1 ratio), 0.7 M [TES][TFSI] and 0.35 M TEOS, ranging from 0.01 to 1.10 g l⁻¹ chitosan; (B) optimization of cycle number: blue (lower): without [TES][TFSI], red (upper): with [TES][TFSI]. Conditions: 1 mM monomers (1 : 1 ratio), 0.14 g l⁻¹ chitosan, 0.7 M [TES][TFSI] and 0.35 M TEOS. (C) In this bar chart, two sensors have been studied: Hydrogel (TEOS/chitosan) and ionogel (TEOS/chitosan/[TES][TFSI]). These sensors have been exposed to nitrogen atmosphere in order to observe the dissolved oxygen effect. Effectively, IL enhances the blank signal due to the stabilization of superoxide anion. The addition of hydrogen peroxide enhances the ECL signal in the hydrogel sensor but conversely decreases the ECL signal in the ionogel sensor. In this study, 5.0 μM of H₂O₂ was used.

The influence of the [TES][TFSI] concentration on the composite was studied between 0.2 and 1.2 M with 0.14 g l⁻¹ chitosan. The ECL values show low variation between 0.2 and 0.7 M in IL, decreasing afterwards due to the same reason discussed above: an increase in the autoabsorption due to an increase in the amount of copolymer formed and thus in the colour intensity. The final composition selected for the biocomposite was 0.14 g l⁻¹ chitosan, 0.7 M [TES][TFSI] and 0.35 M of TEOS.

Cycle number. Working at 1 mM (1 : 1 monomer ratio) in the optimized conditions discussed above, the analytical signal grew up to 20 cycles, decreasing afterwards (see Fig. 4B, red line). This is attributed to an increase in the mass of the formed nanowires along the number of cycles, which generates a self-absorption of the ECL emission (emission wavelength at 450 nm) produced by the increased formation of diimine TMB²⁺ cations (absorption wavelength at 450 nm) on the electrode surface.⁴⁶ Fig. 4B shows that in the absence of IL, the ECL signal decreased, probably due to a less efficient electro-formation process. An increase in the cycle number is necessary in cells prepared without IL to achieve similar signals (up to 60 cycles).

3.2 Sensor measurement conditions

As could be expected in a ECL sensor based on a luminol–TMB copolymer, the luminol reactivity against H₂O₂ increased with the basicity of the medium (pH > 8.0) due to the contribution of hydroxyl groups in catalysing the ECL reaction.⁴⁷ As in previous works, the blank signal increased with the pH but its standard deviation grew more strongly than the signal at a pH higher than 9.5.¹³ For this reason, pH 9.0 (phosphate buffer) was selected as the working pH. The volume of sample placed in the receptacle of the SPE cell has an effect on the ECL signal, with the ECL signal increasing with an increase in volume, up to 50 μl (receptacle maximum capacity).

To optimize the value of the electrode potential applied, a study using cyclic voltammetry was performed obtaining the maximum ECL emission at 0.6 V, a behaviour similar to that previously observed for poly(luminol–TMB) on bare gold SPE cells,¹³ suggesting that this new biocomposite maintains its electrochemical properties due to the use of chitosan and IL.

3.3 Mechanism of the ECL reaction

Luminol exhibits a blank ECL signal in the presence of oxygen due to the radical oxidation of luminol with the superoxide anion generated from the dissolved oxygen (eqn (1) and (2)).^48,49 The stabilization of the superoxide anion by the ionic liquid [TES][TFSI] due to the strong ionic interactions between cationic IL and anionic superoxide (O₂˙⁻), according to Mjalli et al.⁵⁰ increases the blank signal (see the expanded image of Fig. 5 and the bar chart of Fig. 4C). The consumption of the dissolved oxygen was observed when several electrochemical cycles were performed, with a decrease in the signal each time (Fig. 5).

O₂ + e → O₂˙⁻, E⁰ = −0.16 V (1)

Luminol + O₂˙⁻ → 3APA* → 3APA + hν (2)

Fig. 5 ECL response of the disposable sensor to hydrogen peroxide. The record shows the ECL signal observed for 2 min obtained from consecutive pulses of 0.6 V for 1 s each with a dwell time of 10 s. The ECL signal used for analytical purposes was the integral signal for this time, overshadowed into the red line that corresponds to the profile of a blank; blue line: 0.25 μM H₂O₂; pink line: 2.5 μM H₂O₂; and brown line: 6.0 μM H₂O₂. The insets show an enlargement of the first peak and the linear calibration function.

Moreover, imidazolium containing ionic liquids generate an unstable superoxide because the superoxide anion reacts with the imidazolium cation, giving the corresponding 2-imidazolones,⁵⁰ which reinforces the reason for the low ECL signal in imidazolium containing ionic liquids in the preliminary studies in Section 3.1.

In the presence of hydrogen peroxide, the usual increase of the ECL signal due to the chemical reaction of hydrogen peroxide and diazaquinone from luminol is observed in typical ECL sensors and less in hydrogel (TEOS/chitosan). This is not evidenced in the ionogel sensor (TEOS/chitosan/[TES][TFSI]) (see bar chart in Fig. 4C) because a decrease in the ECL intensity due to the addition of hydrogen peroxide is observed, which decreases with the concentration. One reason for this could be the conversion of the electroformed superoxide to oxygen with the addition of H₂O₂ following the Haber–Weiss reaction.⁵¹ In general, this reaction requires Fe(III) as a catalyst (Fenton reaction), but it can be electrochemically catalyzed at 0.48 V (pH 7.0), according to Koppenol,⁵¹ with the consequent elimination of the superoxide anion from the media (eqn (3)).

O₂˙⁻ + H⁺ + H₂O₂ → O₂ + HO˙ + H₂O, 0.48 V = ΔE⁰ (3)

For that reason, the sensor presented in this paper is sensitive to H₂O₂, acting as a scavenger of the increased blank signal generated by the biocomposite including the [TES][TFSI] ionic liquid as mediator.

3.4 ECL sensor analytical characterization

Analytical signal. As the analytical parameter, we studied the evolution of the ECL signal, working in chronoamperometric mode, by recording the height or peak area of the ECL intensity after each pulse. Fig. 5 shows the profile of the blank and hydrogen peroxide signals: an initial intense peak that decreases with subsequent pulses. The reproducibility of the ECL emitted from any excitation pulse (peak height) is around 20–40%, the same as the peak area (one pulse). However, the use of the peak area of a train of pulses produces a more reproducible value, typically less than 10%. The analytical signal used in this paper was the peak area of all the ECL emissions obtained from nine consecutive pulses lasting 1 s with a dwell time of 10 s. This analytical signal was denoted the ECL integrative signal (a.u. s).

Analytical parameters. The dependence of the ECL integrative signal on the H₂O₂ concentration was obtained. The linear dependence with negative slope observed between 0.0093 μM and 7.4 μM is the proposed analytical function. The ECL integrative signal presents square root dependence with H₂O₂ concentration. The limit of detection obtained by using the standard criteria⁵² was 0.8 nM and the limit of quantification was 9.3 nM. The repeatability of the biocomposite sensor was tested by using different screen-printed cells for each measure. Ten blanks produced a relative standard deviation (RSD) of 1.2% (high blank values) and working at 5.0 μM H₂O₂, the RSD was 8.5% (n = 5) (Table 3), acceptable values considering that commercial SPE cells used alone have a precision of around 5%, and the values are similar to those obtained for poly(luminol–TMB) gold SPE cells previously studied¹³.

Table 3 Analytical characteristics of biocomposite sensor for hydrogen peroxide determination

Parameter	Values
a LOD: limit of detection; LOQ: limit of quantification; RSD: relative standard deviation.
Range, μM	0.0093–7.4
Intercept, a.u. s	1301 ± 16
Slope, a.u. s M^−1/2	(−1.5 ± 0.06) × 10^4
R^2	0.996
LOD^a, nM	0.8
LOQ^a, nM	9.3
RSD^a (%) blank (n = 10)	1.2
RSD^a (%) H2O2, 5.0 μM (n = 5)	8.5

---

The lifetime of the disposable sensor was determined using a series of prepared sensors, regularly testing their response to a 5.0 μM H₂O₂ concentration. These sensors, when protected from light and kept at 4 °C, can be used for a month without substantial change. At room temperature and in the dark, the sensors can be used for a week, without the signal changing below 10% (254 ± 23 a.u. s). After that, the signal decreased some 20% on each subsequent day. The prepared sensors must be used in disposable format because each sensor cannot be reused more than 10 times, as we confirmed working at 5.0 μM H₂O₂, after which they showed a decrease in the signal of approximately 35% for each use.

A comparative study between different electrochemiluminescent and amperometric sensors from the bibliography (see Table 4) shows that these methods offer more than three orders of magnitude for H₂O₂ response, as observed in the sensor presented here (three orders). With respect to the limit of detection, the value found (0.8 nM) is in accordance with the results obtained by other researchers. In the case of the reproducibility, the obtained values from the bibliography using conventional electrodes are better than our results (8.5%), due to the fact that they do not work in disposable format.

Table 4 Analytical parameters of hydrogen peroxide detection in recent papers

System	Linear range μM	LOD nM	Linearity	Detection	Reproducibility
Hydrogel:^53 tetramethoxysilane/chitosan	250 to 3400	3000	0.998	Amperometry	4.0%
Dendrimer:^54 titanate nanotubes/poly(amidoamine) and luminol	0.001–0.9	1.0	0.998	ECL	—
Composite:^55 titanate nanotubes/nafion and luminol	0.001–0.5	0.88	0.998	ECL	2.1%
Composite:^56 poly(luminol–benzidine)	0.0002–0.1	0.060	0.997	ECL	5.0%
Composite:^13 poly(luminol–TMB)	0.0061–0.1	2.6	0.984	ECL	10.2%
Hydrogel and nanowires:^24 polyluminol/calcium alginate	0.4–5000	—	—	ECL	—
Ionogel and nanowires: (This work) TEOS/chitosan/IL and poly(luminol–TMB)	0.0092–7.4	0.8	0.998	ECL	8.5%

---

4. Conclusions

The use of a ionogel biocomposite prepared by the solution method from chitosan and ionic liquid in a silica matrix to host nanowires of poly(luminol–TMB) prepared by in situ electropolymerization in the pores of a ionogel make it possible to prepare a disposable reagentless electrochemiluminescent sensor with good mechanical, conductivity and luminescence properties, as well as good operational and temporal stability.

The preparation of ionogels with the inclusion of [TES][TFSI] as the ionic liquid containing triflimide anion produces: (a) an interesting increase in the conductivity of the biocomposite layer needed for ECL generation, (b) an increase in hydration along with chitosan needed for silica matrix stabilization, and (c) strong ionic interactions that stabilize the superoxide anion formed from the oxygen present, giving rise to an intense blank ECL emission.

The developed sensor is sensitive to hydrogen peroxide that acts as a scavenger of the stabilized superoxide anion, reducing the ECL signal with the increase in the H₂O₂ concentration, resulting in a reagentless disposable sensor with a linear range of three orders of magnitude but a better resolution than with usual disposable sensors because the acquisition of the luminescence signal is based on the measurement of the total luminescence area using a photomultiplier detector, such as an enhanced CCD camera. The modified SPE cell studied paves the way for the preparation of biomaterials such as oxidase-based, because binding ionogel with ECL polymers can offer new possibilities for disposable ECL biosensors.

Acknowledgements

We acknowledge financial support from the Junta de Andalucía (Proyecto de Excelencia P10-FQM-5974) and from the Ministerio de Economía y Competitividad (Spain) (CTQ2013-44545-R). These projects were partially supported by European Regional Development Funds (ERDF). We would like to acknowledge Manuel Agudo Acemel for their helpful and kind assistance.

Notes and references

1. A. Kapturkiewicz, Adv. Electrochem. Sci. Eng., 1997, 5, 1–60.
2. K. Muzyka, Biosens. Bioelectron., 2014, 54, 393–407.
3. M. Santhiago, E. W. Nery, G. P. Santos and L. T. Kubota, Bioanalysis, 2014, 6, 89–106.
4. S. Deng and H. Ju, Analyst, 2013, 138, 43–61.
5. L. Hu and G. Xu, Chem. Soc. Rev., 2010, 39, 3275–3304.
6. R. Pyati and M. M. Richter, Annu. Rep. Prog. Chem., Sect. C: Phys. Chem., 2007, 103, 12–78.
7. A. H. Wu, J. j. Sun, Y. M. Fang, R. J. Zheng and G. N. Chen, Electroanalysis, 2010, 22, 2702–2707.
8. R. J. Zheng, Y. M. Fang, S. F. Qin, J. Song, A. H. Wu and J. j. Sun, Sens. Actuators, B, 2011, 157, 488–493.
9. W. Miao, Chem. Rev., 2008, 108, 2506–2553.
10. Y. Xu, B. Lou, Z. Lv, Z. Zhou, L. Zhang and E. Wang, Anal. Chim. Acta, 2013, 763, 20–27.
11. C. H. Kang, Y. B. Choi, H. H. Kim, H. N. Choi and W. Y. Lee, Electroanalysis, 2011, 23, 2131–2138.
12. J. Ballesta-Claver, R. Rodriguez-Gomez and L. F. Capitán-Vallvey, Anal. Chim. Acta, 2013, 770, 153–160.
13. J. Ballesta-Claver, M. C. Valencia-Mirón and L. F. Capitán-Vallvey, Anal. Bioanal. Chem., 2011, 400, 3041–3051.
14. Y. T. Chang, K. C. Lin and S. M. Chen, Electrochim. Acta, 2005, 51, 450–461.
15. S. M. Chen and K. C. Lin, J. Electroanal. Chem., 2002, 523, 93–105.
16. C. H. Wang, S. M. Chen and C. M. Wang, Analyst, 2002, 127, 1507–1511.
17. G. F. Zhang and H. Y. Chen, Anal. Chim. Acta, 2000, 419, 25–31.
18. A. Sassolas, L. J. Blum and B. D. Leca-Bouvier, Anal. Bioanal. Chem., 2008, 390, 865–871.
19. A. Sassolas, L. J. Blum and B. D. Leca-Bouvier, Anal. Bioanal. Chem., 2009, 394, 971–980.
20. A. Sassolas, L. J. Blum and B. D. Leca-Bouvier, Sens. Actuators, B, 2009, 139, 214–221.
21. J. Ballesta-Claver, J. Ametis-Cabello, J. Morales-Sanfrutos, A. Megia-Fernandez, M. C. Valencia-Mirón, F. Santoyo-Gonzalez and L. F. Capitán-Vallvey, Anal. Chim. Acta, 2012, 754, 91–98.
22. G. Li, J. Lian, X. Zheng and J. Cao, Biosens. Bioelectron., 2010, 26, 643–648.
23. F. Montilla, M. A. Cotarelo and E. Morallon, J. Mater. Chem., 2009, 19, 305–310.
24. B. Leca-Bouvier, A. Sassolas and L. J. Blum, Anal. Bioanal. Chem., 2014, 1–11.
25. A. Vioux, L. Viau, S. Volland and J. Le Bideau, C. R. Chim., 2010, 13, 242–255.
26. A. I. Horowitz, Y. Wang and M. J. Panzer, Green Chem., 2013, 15, 3414–3420.
27. J. Le Bideau, L. Viau and A. Vioux, Chem. Soc. Rev., 2011, 40, 907–925.
28. R. Gobel, P. Hesemann, J. Weber, E. Moller, A. Friedrich, S. Beuermann and A. Taubert, Phys. Chem. Chem. Phys., 2009, 11, 3653–3662.
29. F. Gayet, L. Viau, F. Leroux, F. Mabille, S. Monge, J. J. Robin and A. Vioux, Chem. Mater., 2009, 21, 5575–5577.
30. J. Ballesta-Claver, M. C. Valencia-Miron and L. F. Capitán-Vallvey, Analyst, 2009, 134, 1423–1432.
31. Y. Shchipunov, Pure Appl. Chem., 2012, 84, 2579–2607.
32. J. Ballesta-Claver, V. Salinas, M. C. Valencia-Miron and L. F. Capitan-Vallvey, Talanta, 2011, 86, 78–185.
33. J. Ballesta-Claver, I. F. Díaz Ortega, M. C. Valencia-Mirón and L. F. Capitán-Vallvey, Anal. Chim. Acta, 2011, 702, 254–261.
34. C. D. Feng, S. L. Sun, H. Wang, C. U. Segre and J. R. Stetter, Sens. Actuators, B, 1997, 40, 217–222.
35. G. Xu, X. Zeng, S. Lu, H. Dai, L. Gong, Y. Lin, Q. Wang, Y. Tong and G. Chen, Luminescence, 2013, 28, 456–460.
36. M. A. Kim and W. Y. Lee, Anal. Chim. Acta, 2003, 479, 143–150.
37. L. Zhang, X. Gao, L. Yang, P. Yu and L. Mao, ACS Appl. Mater. Interfaces, 2013, 5, 8120–8124.
38. M. Hasanzadeh, N. Shadjou, M. Eskandani and M. d. l. Guardia, TrAC, Trends Anal. Chem., 2012, 41, 58–74.
39. M. P. Singh, R. K. Singh and S. Chandra, ChemPhysChem, 2010, 11, 2036–2043.
40. L. Viau, M. A. Neouze, C. Biolley, S. Volland, D. Brevet, P. Gaveau, P. Dieudonne, A. Galarneau and A. Vioux, Chem. Mater., 2012, 24, 3128–3134.
41. I. Villareal, E. Morales, T. F. Otero and J. L. Acosta, Synth. Met., 2000, 108, 57–65.
42. C. Fuvre, L. Abello and D. Delubouglise, Adv. Mater., 1997, 9, 722–725.
43. V. M. Mecea, Anal. Lett., 2005, 38, 753–767.
44. S. Cosnier and A. A. Karyakin in Electropolymerization: Concepts, Materials and Applications, Wiley-VCH, 1st edn, 2010.
45. A. E. Kadib and M. Bousmina, Chem.–Eur. J., 2012, 18, 8264–8277.
46. P. D. Josephy, T. Eling and R. P. Mason, J. Biol. Chem., 1982, 257, 3669–3675.
47. H. Chu, W. Guo, J. Di, Y. Wu and Y. Tu, Electroanalysis, 2009, 21, 1630–1635.
48. W. Wang, H. Cui, Z. X. Deng, Y. P. Dong and J. Z. Guo, J. Electroanal. Chem., 2008, 612, 277–287.
49. O. V. Reshetnyak, E. P. Koval'chuk and J. Blazejowski, Russ. J. Electrochem., 2011, 47, 1111–1118.
50. M. Hayyan, F. S. Mjalli, M. A. Hashim, I. M. AlNashef and X. M. Tan, J. Electroanal. Chem., 2011, 657, 150–157.
51. W. H. Koppenol, Redox Rep., 2001, 6, 229–234.
52. J. Mocak, A. M. Bond, S. Mitchell and G. Scollary, Pure Appl. Chem., 1997, 69, 297–328.
53. Y. Miao and S. N. Tan, Anal. Chim. Acta, 2001, 437, 87–93.
54. Y. Lin, H. Dai, G. Xu, T. Yang, C. Yang, Y. Tong, Y. Yang and G. Chen, Microchim. Acta, 2013, 180, 563–572.
55. G. Xu, X. Zeng, S. Lu, H. Dai, L. Gong, Y. Lin, Q. Wang, Y. Tong and G. Chen, Luminescence, 2013, 28, 456–460.
56. G. Li, X. Zheng and L. Song, Electroanalysis, 2009, 21, 845–852.

---