Insights into the mechanism of coreactant electrochemiluminescence facilitating enhanced bioanalytical performance

Electrochemiluminescence (ECL) is a powerful transduction technique with a leading role in the biosensing field due to its high sensitivity and low background signal. Although the intrinsic analytical strength of ECL depends critically on the overall efficiency of the mechanisms of its generation, studies aimed at enhancing the ECL signal have mostly focused on the investigation of materials, either luminophores or coreactants, while fundamental mechanistic studies are relatively scarce. Here, we discover an unexpected but highly efficient mechanistic path for ECL generation close to the electrode surface (signal enhancement, 128%) using an innovative combination of ECL imaging techniques and electrochemical mapping of radical generation. Our findings, which are also supported by quantum chemical calculations and spin trapping methods, led to the identification of a family of alternative branched amine coreactants, which raises the analytical strength of ECL well beyond that of present state-of-the-art immunoassays, thus creating potential ECL applications in ultrasensitive bioanalysis.


Functionalization of beads with Ru (Ru@beads):
For each bead dimension, the total surface area was constant of 7×10 9 μm 2 . Indeed, the volume of bead solution was 133.9 (300 nm), 134.5 (500 nm), 265.6 (1 µm) and 6000.0 (2.8 µm) µl, respectively. Each bead solution was poured in a 20 ml vial, collected with a magnet for 2 minutes, and afterwards the supernatant was discharged. Then, washing with 10 ml bead buffer was repeated 2 times for 5 minutes each. Beads were incubated with 18 ml of R2 (2 h at 37°C) by using a tube rotator. Separation was carried out with a magnet for 2 minutes and the supernatant was discharged. This whole procedure was repeated 5 times. Finally, beads were stored in the bead buffers from Roche Diagnostics SAP assay bottle M (2.8-1 μm) or Ademtech buffer (500-300 nm) in a total volume of 800 ml.

Ru quantification:
The amount of Ru conjugated to beads (Ru@beads) was measured by inductively coupled plasma-mass spectrometry (ICP-MS).
Briefly, 500 μl of beads were dissolved in 358 ml of nitric acid (70%) and double distilled water for a final volume of 5 ml, and kept overnight at 80°C. After dissolution a clear solution was obtained. The total amount of Ru, as ppb concentration, was normalized for the total surface area of each bead type, to obtain the density Ru μm -2 , as shown in the following table. Supplementary

Turnover number (TON) and Turnover frequency (TOF)
TON and TOF are defined as: ECL emission (ECLRu@bead) was quantified by integration of ECL images (.tiff), as obtained with the CCD camera in the surface generation-beads emission for a time (t) of 0.5 s. Software ImageJ was used to integrate the signals from ECL images with squares of 50×50px, 30×30px, 20×20px, and 20×20px for 2.  taken at 1.4 V clearly shows a not homogeneous ECL emission generated with the large Pt electrode, but not observed with the small Pt electrode. This is a well-known behavior associated to hindrance of TPrA diffusion exerted by the large tip. 5

Supplementary Methods 3: Spin Trapping experiments and analysis of radical intermediates
In order to unequivocally identify the radicals generated following the oxidation of TPrA we performed prolonged electrolysis and we used spin trapping for stabilization of the electrogenerated radicals (as it is schematized in Supplementary Figure 12). Spin trapping experiments allow to transform short-living radicals into long-living systems with (normally) paramagnetic character, therefore suitable to be characterized by Electron Paramagnetic Resonance (EPR) and mass spectrometry. Spin-trapping experiments were conducted by adding the spin traps 5,5-dimethyl-pyrroline N-oxide (DMPO) (≥98%, for ESR, Sigma, 0.05 M before the electrolysis) to a solution of TPrA 180 mM, formic acid pH 6.9. The electrolysis was performed at 1.4 V (vs Ag/AgCl) for 1 h under Ar saturated atmosphere to reduce oxygen content. After electrolysis, the solution was analyzed by mass spectrometry and Electron Paramagnetic Resonance (EPR). According to literature 6,7 oxidation of TPrA can lead to the following radical products. In addition, the nitroxide group can lately reduce to give the analogous nitroxylamine (DMPOH) adduct. The results of the mass spectra evidence the formation of species DMPO-DPrA(C), therefore, as reported for TPrA, the nitrogen radical cation undergoes a transposition to form a C-centered radical before to be trapped by DMPO.
For the primary radical of TPrA, instead, we observe a peak relative to ions with Z=1 of either species DMPO-TPrA(N) + and/or DMPO-TPrA(C)-H +, with an extra proton. This is not surprising, and often nitroxides in acid solution can undergo disproportion to hydroxylamine and oxammonium cation. 11 From the mass spectra, it is not possible to distinguish between the formation of adduct DMPOH-TPrA(C) and DMPOH-TPrA(N) + , as the masses of the relative charged ions are the same Therefore, mass spectra account for the formation of DMPO-DPrA(C), and likely of the precursor DMPO-DPrA(N), and for the formation of a C-centered or a N-centered radical adduct of TPrA with DMPO; for this adduct we have identified the reduced form, where the nitroxide turned into hydroxylamine. These last species have a larger signal that might be related to a larger concentration in the solution.

Supplementary Methods 4: Computational results
Starting geometry of TPrA and DPIBA: due to the flexible substituents of the TPrA molecule, as a first point the most stable conformer has been assessed as a function of the dihedral angle described by Nefgof each alkyl chain, in neutral state (Supplementary Figure S14). A scan of molecular energy for varying dihedral angles (from -180 to 180°, scan step=15°) allowed to identify and optimize the most stable conformer, employed in subsequent calculations. The characteristics of the most interesting conformers identified from the scan are summarized in Supplementary Table 4. All of them have an N atom with sp3 hybridization and a pyramidal arrangement of C-N bonds. The most stable structure is labeled T1 (Supplementary Table 4, left), having all the alkyl chains similarly arranged with a dihedral angle of 60°. Other stable conformers have an appreciably higher relative energy with respect to the T1 structure, and consequently a significantly lower relative population, estimated by means of the Boltzmann distribution (< 0.3). Moreover, the different chain orientation does not determine significant differences regarding the C-N bond length, which is 1.46 Å for all conformers. Similar considerations hold true for the highest occupied orbital (HOMO, Supplementary Figure 15) energy, that is very close for all conformers, with differences below 0.05 eV. Hence in the subsequent discussion, only the most stable conformer was considered, namely T1. With regards to DPIBA, it has been assumed that the methylation does not significantly influence the alkyl chains orientation. Hence, only conformer D1, Supplementary Figure 16 . This dissociation pathway is not competitive in the absence of an electric field owing to the large barrier. Therefore assuming an inner sphere electron transfer, we reevaluated the potential energy surface in the presence of an external electric field (E) applied along the C-N bond to be cleaved (directed towards the N atom) and with 0.025 a.u (~10 8 V/cm) of magnitude (Supplementary Figure  13, green line). For both TPrA and DPIBA, the resulting PESs flattens, with a marked tendency towards stabilizing the products in solvent, after overcoming a single barrier of ca. 1 eV for TPrA and ca. 0.6 eV for DPIBA.

Supplementary
To characterize the electronic structure evolution of the molecular fragments upon dissociation, the Mulliken spin densities are employed to identify unpaired radicals (Supplementary Table 3), along with the partial charges based on NBO (Natural Bond Orbitals) analysis; 12 in particular the total charge of the propyl (for TPrA) or isobutyl (for DPIBA) cleaved fragment is evaluated, as a sum of the NBO atomic partial charges of all the atoms belonging to the cleaved fragment (Supplementary Table 6). In the absence of the electric field (E=0), the C-N dissociation is covalent and leaves a localized unpaired electron on C(2) atom of the cleaved bond (Supplementary Table 5) without the formation of a neat charge on the fragment (Supplementary Table  6). On the contrary, in the presence of the field (E = 0.025 a.u), the C-N dissociation acquires a ionic character with the formation of a carbocation (Supplementary Table 6); the ionic character is further confirmed by the absence of appreciable spin density on C(2) atom (Supplementary Table 5).
Supplementary Table 5. Spin densities on C atoms (see Supplementary Figure 13 for numbering) of cleaved propyl fragment for TPrA (left) and DPIBA (right) from calculations in H2O without (E =0) and with (E = 0.025a.u) external electric field. Table 6. Cleaved propyl fragment final charge, computed from NBO partial charges, from calculations in H2O without (E =0) and with (E = 0.025a.u) external electric field. All determinations were performed on a cobas® 8000 / cobas e 801 analyzer using the implemented assay protocols. Commercially available system reagents CleanCell M and PreClean II M, AssayTip/AssayCup trays, Elecsys® assay reagent kits and Elecsys calibrator sets were used (see Supplementary table 7). All calibrator sets were reconstituted as indicated in the product insert. ProCell II M was replaced by 0.2 M PB (pH 6.9) with 180 mM TPrA and 0.1% polidocanol, which were supplemented with 0 mM (reference), 30 mM, 50 mM and 100 mM DPIBA. Conductivity (25 mS cm -1 ) and pH (6.9) were similar among these buffers.