Redox-mediated electrochemiluminescence enhancement for bead-based immunoassay

Electrochemiluminescence (ECL) is a highly sensitive mode of detection utilised in commercialised bead-based immunoassays. Recently, the introduction of a freely diffusing water-soluble Ir(iii) complex was demonstrated to enhance the ECL emission of [Ru(bpy)3]2+ labels anchored to microbeads, but a comprehensive investigation of the proposed ‘redox-mediated’ mechanism was not carried out. In this work, we select three different water-soluble Ir(iii) complexes by virtue of their photophysical and electrochemical properties in comparison with those of the [Ru(bpy)3]2+ luminophore and the TPrA co-reactant. A systematic investigation of the influence of each Ir(iii) complex on the emission of the Ru(ii) labels on single beads by ECL microscopy revealed that the heterogeneous ECL can be finely tuned and either enhanced up to 107% or lowered by 75%. The variation of the [Ru(bpy)3]2+ ECL emission was correlated to the properties of each Ir(iii)-based mediator, which enabled us to decipher the mechanism of interaction and define guidelines for the future design of novel Ir(iii) complexes to further enhance the ECL emission of bead-based immunoassays. Ultimately, we showcase the potential of this technology for practical sample analysis in commercial instruments by assessing the enhancement of the collective ECL intensity from a bead-based system.


Introduction
Electrochemiluminescence (ECL) is the process of generating light via strongly exergonic electron transfer reactions between reactive electrogenerated species. 1,25][6][7] These benets have played a pivotal role in the establishment of ECL as a leading bioanalytical technique.To make ECL viable for clinical diagnostics, Roche Diagnostics produce fully automated analysers that exploit bead-based immunoassay technology.This system employs biotinylated antibodies and dye-functionalised antibodies that specically recognise a given antigen.When the analyte is present, the classical sandwich assay is formed.In this way, a 1 : 1 ratio between the antigen and the labelled antibody is achieved, making the ECL proportional to the analyte concentration. 8,9ypical ECL immunoassays exploit tris(2,2 ′ -bipyridine)ruthenium(II) ([Ru(bpy) 3 ] 2+ ) and tri-n-propylamine (TPrA) as luminophore and co-reactant, respectively.The latter is a sacricial molecular species that upon oxidation, undergoes an irreversible chemical step to generate strongly reducing radicals.The bead-based immunoassay follows the heterogeneous coreactant ECL mechanism (Fig. 1a), where the Ru(II) complexes on the beads cannot be directly oxidised because they are constrained further than the tunnelling distance of ∼1-2 nm from the electrode surface. 104][25][26] To date, there is a lack of straightforward methods to enhance the ECL signal of bead-based biosensors without further modifying the structure of the patented immunoassay.
However, a thorough investigation of the interaction mechanism between [Ru(bpy) 3 ] 2+ on the beads and [Ir(sppy) 3 ] 3− in solution has been lacking, and the pivotal features of the Ir(III) mediator that enable its participation in the enhancement of heterogeneous ECL remain unidentied.In this work, we introduced three different water-soluble iridium complexes in a model bead-based immunoassay system, namely [Ir(sppy) 3 ] 3− (Fig.  2c).Each Ir(III) complex displays unique photophysical and electrochemical properties, summarised in Fig. 2, such as different emission wavelengths, redox potentials, and positive or negative charges.In particular, their emission spectra exhibit a progressively increasing degree of overlap with the absorption spectrum of [Ru(bpy) 3 ] 2+ (Fig. 2d and Table S1 †).At the same time, each complex presents different electrochemical properties that modulate to different extent the homogeneous electron transfer with both the TPrA co-reactant and the a-aminoalkyl radical TPrAc (Fig. 2e and Table S1 †).
Herein, we utilise ECL microscopy (see Fig. S1 †) [29][30][31][32] to analyse the emission of Ru(II) labels attached to magnetic beads upon interaction with each Ir(III) mediator.The results were compared to the ECL signal of single beads generated at the same potential in a co-reactant solution without any Ir(III) complex.By systematically varying both the Ir(III) mediator and the experimental conditions (i.e., Ir(III) mediator concentration, TPrA concentration, working electrode potential), we were able to achieve an ECL enhancement up to 107%, or quenching by up to 75%.Notably, while observing a comparable ECL enhancement to that reported in our previous work under similar conditions, we also describe circumstances where the ECL signal is partially switched off.By correlating the properties of each freely diffusing Ir(III) complex to their inuence on the ECL emission of a model bead-based immunoassay, we present for the rst time a comprehensive study of this novel reaction mechanism.Our ndings enable the development of guidelines for the future design of novel redox mediators that maximise the effectiveness of such mechanism.Finally, we transitioned from the ECL microscopy to a collective beads experimental setup in which the entirety of the ECL generated by the beads is collected by a photomultiplier tube (PMT).The introduction of 50 mM [Ir(sppy) 3 ] 3− into the co-reactant solution resulted in a substantial 22.3% enhancement of the ECL signal from the [Ru(bpy) 3 ] 2+ labels on the beads, demonstrating the potential application of this technology for real sample analysis in commercial instruments.
The resulting spectra were corrected for the change in instrument sensitivity over the wavelength range and then normalised.

Beads preparation
Prior to use, magnetic beads covalently functionalised with [Ru(bpy) 3 ] 2+ were washed in a phosphate buffer (PB) solution and sonicated for 15 minutes.

Electrochemiluminescence
In the microscopy setup, the ECL and optical images were captured following the injection of a suspension of [Ru(bpy) 3 ] 2+ covalently functionalised beads in the electrochemical cell where the microspheres were collected on the working electrode surface using a magnet placed underneath.The ECL and optical imaging was performed using solutions of 0.3 M PB, variable TPrA concentrations (pH 6.8) and a redox mediator, in a homemade PTFE electrochemical cell comprising a Pt working (0.16 cm 2 ), Pt counter, and Ag/AgCl (3 M KCl) reference electrodes.The different solutions were inserted in the electrochemical cell with a pressure-driven ow controller (OB1 Mk3, Elveow) equipped with a ux sensor (Flow-04D working range from 0 to 1000 mL min −1 ) and exchanged, when necessary, with a 10-way bidirectional valve (MUX distributor).For microscopic imaging, an epiuorescence microscope from Nikon (Chiyoda, Tokyo, Japan) equipped with an ultrasensitive EMCCD camera (EM-CCD 9100-13 from Hamamatsu, Japan) was used with a resolution of 512 × 512 pixel and a size of 16 × 16 mm 2 .The microscope was enclosed in a homemade dark box to avoid interferences from external light.It was also equipped with a motorised microscope stage (Corvus, Märzhauser, Wetzlar, Germany) for sample positioning and with long-distance objective from Nikon (magnication 100×/numerical aperture 0.80/DL 4.5).Additionally, the integrated system included a SP-300 potentiostat (BioLogic Science Instrument, France) triggered with the camera (see Fig. S1 †).CV-ECL plots were collected by scanning the working electrode potential at 100 mV s −1 from open circuit potential (OCP) up to 2 V (vs.Ag/AgCl 3 M KCl), back to 0 V (vs.Ag/AgCl 3 M KCl) and, eventually, terminating the cycle at OCP.The beads emission during CV-ECL measurements was acquired every 200 ms to follow the temporal evolution of the signal.The integration time of the EM-CCD camera was set to 200 ms.ECL images were recorded by applying a double chronoamperometric pulse: OCP for 2 s and a suitable potential (vs.Ag/AgCl 3 M KCl) for the next 8 s.The total integration time of the EM-CCD camera was set to 10 s.Unless otherwise stated, gain and sensitivity parameters of the EM-CCD camera were set to 1 and 255, respectively.
In the collective beads conguration, 5.5 mL of a 0.72 mg mL −1 suspension of [Ru(bpy) 3 ] 2+ covalently functionalised beads were deposited on the working electrode surface, where they stick due to a magnet placed beneath.ECL measurements were carried out using 3 mL of solutions comprised of 0.3 M PB, 180 mM TPrA (pH 6.8) and a redox mediator (where required), in a PTFE homemade electrochemical cell comprising a Pt working (0.3 cm 2 ), Pt counter, and Ag/AgCl (3 M KCl) reference electrodes.The ECL signal was collected by a photomultiplier tube (PMT) positioned on top of the cell, whose voltage was set to 750 V.The recorded emission was amplied to a 000.0 mA level using a Keithley Model 6485 Picoammeter (Keithley Instruments Inc., Ohio, United States).Between the cell and the PMT we placed a longpass lter with a cut-on wavelength of 606 nm (Newport Corporation, Irvine, California, USA) to maximise the isolation of the ECL emission of [Ru(bpy) 3 ] 2+ labels from the Ir(III) complexes.The system was enclosed in a homemade dark box to avoid interferences from external light.The ECL emission is triggered by anodic potential sweep during cyclic voltammetry controlled by a SP-300 potentiostat.

Results and discussion
Generally, ECL sandwich immunoassay involves the capture of the antibody-antigen complex; therefore, as a proof of concept, we decided to employ magnetic beads decorated with Ru(II) labels.These covalently functionalised 2.8 mm beads (Ru@Beads) mimic the activity of the bead-based ECL immunoassay and generate an improved signal-to-noise ratio due to the higher surface concentration of Ru(II) labels.In this case, this feature is crucial to emphasise the beads signal over the background luminescence generated by homogeneous ECL from the Ir(III) complex.
Similarly, integrated ECL measurements were carried out on Ru@Beads/[Ir(dfppy) 2 (pt-TEG)] + using 100 mM of the Ir(III) complex.In contrast to [Ir(sppy) 3 ] 3− , which exhibits nearly identical oxidation potential to TPrA, [Ir(dfppy) 2 (pt-TEG)] + necessitates a signicantly more anodic potential (E ox = 1.44 V) to be oxidised.By modulating the working electrode potential, one can effectively control the presence of oxidised Ir(IV) species and investigate the ECL response in scenarios where only Ir(III) species are present in solution or where both Ir(III) and Ir(IV) species coexist.Essentially, the applied potential can work as a switch, enabling the manipulation of chemical conditions to tune the reaction mechanism.At rst, a potential of 1.5 V was imposed to achieve the oxidation of [Ir(dfppy) 2 (pt-TEG)] + .Quite surprisingly, instead of an enhancement, we observed a 24% drop in ECL intensity compared to Ru@Beads (see Fig. S20 †).Moreover, at 1.2 V, where only the co-reactant is oxidised, the ECL intensity suffers an even more pronounced attenuation (−75%) compared to the Ru@Beads reference (Fig. 4).Thus, contrary to [Ir(sppy) 3 ] 3− , this complex decreases the [Ru(bpy) 3 ] 2+ -emitted ECL either in presence or in absence of the oxidised Ir(IV) species.
All of the above supports electron transfer as the interaction mechanism leading either to heterogeneous ECL enhancement or quenching.We could actually exclude energy transfer as a major contributor to such an effect because of the poor overlap between the emission spectrum of [Ir(sppy) 3 ] 3− with the absorption spectrum of [Ru(bpy) 3 ] 2+ .
The opposite enhancement and quenching phenomena result from redox mediated mechanisms involving either the oxidised or reduced forms of the Ir(III) species, depending on their standard oxidation and reduction potentials relatively to those of the TPrA.The ECL of the Ru@Beads is kinetically controlled by the availability of short-lived TPrAc + and TPrAc species to react with the Ru(II)-anchored labels, and the alternative oxidants or reductants provided by the introduction of Ir(III) complexes thus results in considerable variation in the ECL emission intensity.Given the large excess of co-reactant in solution and the similarity between their oxidation potentials, 45,46 the oxidised [Ir(sppy) 3 ] 2− complex (as with the oxidised [Fe(bpy) 3 ] 3+ complex) homogeneously oxidise the TPrA to TPrAc + .This homogeneous generation of TPrAc + offers a kinetic advantage, especially at low overpotentials, compared to the sluggish electron transfer associated with the heterogeneous TPrA oxidation on a Pt electrode. 47This ultimately results in an increased production of TPrA radical.The catalytic route, altogether with the ability of the Ir(IV) species to oxidise the reduced [Ru(bpy) 3 ] + species, results in the large ECL enhancement in the potential region where the TPrA oxidation is under kinetic control.On the other hand, at 1.2 V (i.e., diffusion-controlled regime for TPrA oxidation), most of the current can be attributed to the direct electro-oxidation of TPrA and, in turn, most of the ECL signal arises from reaction with electrogenerated TPrAc + (Fig. S11 †).Therefore, we hypothesise that the homogeneous oxidation of the co-reactant may provide a signicant contribution at less anodic potentials (i.e., 0.9 V) where the oxidation of the TPrA at the electrode is notably sluggish, but at 1.2 V, the redox mediated pathway shown in Fig. 1b is the most effective.In this context, the negative charge on the [Ir(sppy) 3 ] 3− complex could promote its approach to the ECL label due to coulombic attraction, increasing the rate of interactions.In order to elucidate such 'redox mediated' ECL, we propose a simplied model (see ESI Section 13 †) combining a steadystate formalism to describe the surface concentration of the Ru-based species and a COMSOL simulation of the electrochemistry of TPrA and Ir species in solution.The ECL signal enhancement provided by the introduction of 100 mM of [Ir(sppy) 3 ] 3− upon application of 1.2 V can be reproduced (see modelled ECL prole in Fig. S28a †) without any increase in the emitting layer thickness by considering both the redox mediated formation of the emitter (Fig. 1b) and the redox mediated oxidation of the co-reactant: For the other two complexes (Fig. S28b †), the quenching of the ECL signal is consistent with TPrAc scavenging: [Ir(C^N) 2 (pt-TEG)] + + TPrAc / [Ir(C^N) 2 (pt-TEG)] 0 + P where P is the TPrA oxidation product.It is noteworthy that these complexes also allow the connement of the TPrAc radical in the vicinity of the electrode surface, preventing radical chemistry path (or reactive radical species formation) in the solution bulk.This may be adventitious for further use of TPrA as co-reactant for ECL imaging of real biological systems.The extent of the ECL quenching can be modulated by considering the contribution of the inter-species electron transfer (see Reactions (S3.11) and (S4.11), † see modelled ECL prole in Fig. S28b †), where the faster this step, the weaker the quenching: In both situations, the simplied computed ECL model could indeed reproduce the observed trends at 1.2 V. Finally, the involvement of all redox states from a redox mediator, as for the [Fe(bpy) 3 ] 2+ case, could also be reproduced (see modelled ECL prole in Fig. S28c †).
Leveraging our deepened understanding of the impact of Ir(III) redox mediators on the ECL emission of a bead model system, we have extended our investigation to showcase a tangible application of the addition of [Ir(sppy) 3 ] 3− in the coreactant solution by emulating the experimental conditions commonly employed in clinical bioanalysis.The transition from the original microscopy conguration to a collective conguration, utilising a photomultiplier tube (PMT) instead of an EM-CCD camera, enabled the comprehensive capture of emitted light from all the beads immobilised on the working electrode surface (Fig. 5a).Given the absence of spatial resolution, distinguishing the [Ru(bpy) 3 ] 2+ labels emission from the homogeneous ECL of [Ir(sppy) 3 ] 3− would, in principle, seem challenging.To address this, we introduced a longpass optical lter (l cut-on = 606 nm) to minimise the background emission from the Ir(III) species while still capturing the peak of the [Ru(bpy) 3 ] 2+ emission (see Fig. S29 †).Any residual background signal attributed to [Ir(sppy) 3 ] 3− homogeneous ECL was subsequently subtracted during the data processing stage (see Fig. S30 †).
Upon integrating the ECL signal generated during the CV-ECL cycle, we determined that the addition of 50 mM of [Ir(sppy) 3 ] 3− yields a 22.3% enhancement in beads emission (Fig. 5b).However, it is noteworthy that the signal gain observed in the two experimental setups, under the same chemical conditions, displays marked difference.The enhancement achieved in collective beads conguration is much smaller, and this divergence can be ascribed to the spatial arrangement of the beads.While in microscopy experiments, the beads were injected into the cell, ensuring a uniform spatial distribution across the electrode surface, in the collective setup they were directly deposited on top of the working electrode where they tend to cluster together.This clustering hinders the ideal diffusion of TPrA radicals and electro-oxidised [Ir(sppy) 3 ] 2− to the core, thereby resulting in a less pronounced ECL enhancement.

Conclusions
We have demonstrated that the ECL reactions between [Ru(bpy) 3 ] 2+ and three Ir(III) complexes involve an electron transfer mechanism in which the quenching or enhancing nature of the redox mediators is dependent on their redox potentials.More facile reduction yields scavenging of the critical radical species derived from the co-reactant and therefore quenching of the ECL of [Ru(bpy) 3 ] 2+ .On the other hand, the production of the oxidised Ir(IV) species may boost the ECL, particularly when the reduction route is not thermodynamically feasible (such as for [Ir(sppy) 3 ] 3− ).These ndings provide a new framework to design mediators to further enhance the ECL signal-to-noise ratio in bead-based assays.
Ultimately, the proposed shi in experimental design provides a concrete example of the relevance of this strategy in a setting that mirrors industrial bioanalysis practices, bridging the gap between fundamental understanding and real-sensor application.

Fig. 3
Fig. 3 (a) CV-ECL measurement performed on Ru@Beads/[Ir(sppy) 3 ] 3− in a 0.3 M PB solution at pH 6.8 with 180 mM TPrA and 100 mM [Ir(sppy) 3 ] 3− .The working electrode potential was scanned at 100 mV s −1 from OCP up to 2 V (vs.Ag/AgCl), back to 0 V (vs.Ag/AgCl) and, eventually, terminating the cycle at OCP.The beads ECL emission (red line) was acquired each 200 ms and, for each frame, the maximum value of the beads ECL profile is plotted versus the applied potential.The background signal is eliminated by subtracting, for each frame, the average ECL intensity value of the background retrieved over a 50 × 50 pixel square centred in a region of the image where no beads are present.(b) Comparison between ECL intensities of Ru@Beads (red bar) and of Ru@Beads/[Ir(sppy) 3 ] 3− in a 0.3 M PB solution at pH 6.8 with 180 mM TPrA and different [Ir(sppy) 3 ] 3− concentrations: 5 mM (pink bar), 10 mM (yellow bar), 20 mM (green bar), 50 mM (blue bar) and 100 mM (black bar).The ECL intensities were obtained from ECL images captured using an EM-CCD camera during a two-step chronoamperometry measurement: 2 s at OCP and 8 s at 1.2 V vs. Ag/AgCl.Magnification, 100×; objective numerical aperture, 0.8; gain, 1; sensitivity, 255.Data are averaged over a minimum of six beads (n $ 6).Each bar represents the maximum value of the respective ECL profile, and the error bars show the standard errors.(c and d) ECL images of a single 2.8 mm bead covalently labelled with [Ru(bpy) 3 ] 2+ in a 0.3 M PB solution at pH 6.8 with 180 mM TPrA (c) without (Ru@Beads) and (d) with 100 mM [Ir(sppy) 3 ] 3− (Ru@Beads/[Ir(sppy) 3 ] 3− ).The background signal in (d) is generated by [Ir(sppy) 3 ] 3− * following the conventional homogeneous ECL pathways.The images were obtained with an EM-CCD camera by recording the ECL signal for 10 s during a two-step chronoamperometry measurement: 2 s at open circuit potential (OCP) and 8 s at 1.2 V vs. Ag/AgCl.Magnification, 100×; objective numerical aperture, 0.8; gain, 1; sensitivity, 255; contrast intensity scale: 2000 to 11 000; scale bar: 3 mm.(e) Comparison between the single-bead ECL intensity profiles of Ru@Beads (grey line) and Ru@Beads/[Ir(sppy) 3 ] 3− (red line).Inset: histogram of the comparison between the respective averaged maximum values of ECL intensity where the error bars show the standard error.Data are averaged over a minimum of six beads (n $ 6).

Fig. 5
Fig. 5 (a) Experimental setup employed to record the ECL signal in collective beads configuration.The beads were deposited on the surface of a Pt working electrode where they remain because of a magnet placed underneath.To discern the emission of [Ru(bpy) 3 ] 2+ labels from the homogeneous ECL of [Ir(sppy) 3 ] 3− , the ECL generated during the anodic potential sweep first passed through an optical filter that cuts off the all the light below 606 nm and, eventually, strikes the PMT that capture all the light without spatially resolving the signal.Yet, a small but nonnegligible background signal due to [Ir(sppy) 3 ] 3− homogeneous ECL could still be detected, thus the ECL intensity generated by Beads/ [Ir(sppy) 3 ] 3− (i.e., ECL signal of the same amount of non-labelled streptavidin-coated beads in the co-reactant solution with 50 mM of [Ir(sppy) 3 ] 3− ) was subtracted to Ru@Beads/[Ir(sppy) 3 ] 3− during data processing.(b) CV-ECL measurement performed on Ru@Beads (grey line) and of Ru@Beads/[Ir(sppy) 3 ] 3− at 50 mM concentration of [Ir(sppy) 3 ] 3− (red line), both in a 0.3 M PB solution at pH 6.8 with 180 mM TPrA.The working electrode potential was scanned at 100 mV s −1 from OCP up to 2.5 V (vs.Ag/AgCl), back to 0 V (vs.Ag/AgCl) and, eventually, terminating the cycle at OCP.The inset represents a comparison between ECL intensities of Ru@Beads (grey bar) and of Ru@Beads/[Ir(sppy) 3 ] 3− at 50 mM concentration of [Ir(sppy) 3 ] 3− (red bar) in a 0.3 M PB solution at pH 6.8 with 180 mM TPrA.Each bar represents the ECL intensity obtained by integrating the whole CV-ECL cycle and the error bars show the standard errors.Data are averaged over two different measurements.