Exploiting a New Strategy to Prepare Water‐Soluble Heteroleptic Iridium(III) Complexes to Control Electrochemiluminescence Reaction Pathways in Aqueous Solution

Abstract

The direct derivatization of heteroleptic iridium(III) complexes with sulfonate groups removes the limitations of prior strategies for the preparation of water‐soluble analogues, in which complexes were prepared from a narrow range of ligands with suitably polar or charged functional groups. Phenylpyridine, phenylpyrazole, and phenylbenzothiazole ligands were selectively sulfonated opposite to their cyclometalated carbon within iridium(III) complexes containing bipyridine or N‐heterocyclic carbene ancillary ligands. Altering reaction conditions enabled additional sulfonation of the benzothiazole fragments. Informed by the electrochemical and photophysical properties of the parent complexes in organic solvents, we adopted this strategy to design six novel luminophores that proceeded through three different sets of coreactant electrochemiluminescence (ECL) reaction pathways under aqueous conditions. The intensity of the indirect coreactant ECL of one of the iridium(III) luminophores was enhanced by over an order of magnitude by introducing a redox mediator, extending this promising analytical approach beyond the conventional [Ru(bpy)₃]²⁺ electrochemiluminophore.

Water‐soluble luminophores prepared by sulfonation of heteroleptic iridium(III) complexes enable judicious tailoring of properties for unprecedented control of electrochemiluminescence (ECL) reaction pathways in aqueous solution. This approach removes a longstanding barrier to the translation of high‐efficiency and multicolored ECL systems developed in organic solvents to the aqueous conditions desired for many analytical applications.

1. Introduction

Cyclometalated iridium(III) complexes have been studied extensively over the past few decades for light emitting (e.g., organic light‐emitting diodes,^{[ 1} , ^{2 ]} light‐emitting electrochemical cells,^{[ 3 ]} electrochemiluminescence (ECL),^{[ 4} , ^{5 ]} chemical sensing,^{[ 6 ]} and cellular imaging^{[ 7} , ^{8 ]}) and light harvesting (e.g., solar energy conversion,^{[ 9 ]} photoredox catalysis,^{[ 10} , ^{11 ]} and photodynamic therapeutics^{[ 8 ]}) applications. High phosphorescence quantum yields and excited state lifetimes amenable to electron and energy transfer can be attained in iridium(III) complexes exhibiting a wide range of triplet state energies and redox properties, tuneable through ligand structure.^{[ 12} , ^{13 ]}

Additional considerations for ligand design arise from the disparate chemical environments of the target applications; this can be particularly challenging with those involving aqueous conditions.^{[ 14 ]} Homoleptic cyclometalated iridium(III) complexes such as the archetypal fac‐Ir(ppy)₃ (Figure 1a) are typically charge‐neutral and insoluble in water.^{[ 15 ]} The widely used monocationic heteroleptic iridium(III) complexes incorporating ligands with and without cyclometalation, exemplified by [Ir(ppy)₂(bpy)]⁺ (Figure 1b), are more hydrophilic, but often still insufficient for applications in solely aqueous solution. Approaches to impart greater solubility in water have predominantly involved developing ligands with charged/polar functional groups, such as N‐alkyl pyridiniums,^{[ 16} , ¹⁷ , ¹⁸ , ^{19 ]} quaternary ammoniums,^{[ 20} , ^{21 ]} polyethylene glycols,^{[ 22} , ²³ , ²⁴ , ^{25 ]} saccharides,^{[ 26} , ^{27 ]} and sulfonates.^{[ 23} , ²⁸ , ²⁹ , ³⁰ , ^{31 ]} The need to incorporate one or more of this limited selection of ligand structures to assemble water‐soluble iridium(III) complexes, however, complicates the tuning of photophysical properties.

Figure 1.

Archetypal (a) homoleptic and (b) heteroleptic iridium(III) complexes (left) and their water‐soluble, sulfonated derivatives (right), including the novel [Ir(sppy)₂(bpy)]⁻ prepared in this study. Reagents: (i) sulfuric acid and trifluoroacetic acid anhydride, (ii) saturated aqueous sodium hydrogen carbonate.

In an alternative strategy introduced by Wenger and coworkers,^{[ 14} , ^{32 ]} the facial isomers of Ir(ppy)₃, Ir(f‐ppy)_3, and Ir(df‐ppy)₃ were decorated with sulfonate groups (e.g., Figure 1a, right structure) to create highly water‐soluble analogues exhibiting similar electrochemical and optical spectroscopic properties to their parent complexes. These compounds have been exploited for innovative aqueous photochemistry including enzyme and photoredox catalysis,^{[ 32 ]} visible‐light generation of hydrated electrons,^{[ 33} , ^{34 ]} and photochemical upconversion.^{[ 35 ]} They also enabled a breakthrough in coreactant ECL detection,^{[ 36} , ³⁷ , ³⁸ , ^{39 ]} in which the water‐soluble iridium(III) complex serves as both an electrocatalyst of coreactant oxidation and a precursor for more efficient chemi‐excitation of the [Ru(bpy)₃]²⁺ luminophore, in some cases increasing the ECL intensity by over an order of magnitude.

Nevertheless, despite their effectiveness as redox‐mediator enhancers of ECL, these homoleptic complexes are relatively poor electrochemiluminophores.^{[ 37 ]} Realizing the promise of iridium(III) complexes for ECL detection^{[ 4} , ^{5 ]} necessitates the preparation of water‐soluble analogues of heteroleptic species for which both chemi‐excitation and emission are highly efficient.

Herein, we prepare a series of highly water‐soluble heteroleptic iridium(III) complexes (Figure 1b), adapting the synthetic approach outlined by Wenger and coworkers.^{[ 32} , ^{35 ]} The scope is illustrated through the sulfonation of various cyclometalated ligands, including phenylpyridine, phenylpyrazole, and phenylbenzothiazole ligands, in iridium(III) complexes containing bipyridine or N‐heterocyclic carbene (pyridyl‐idazolylidene and ‐benzimidazolylidene) ligands (Scheme S3). The novel complexes are examined as luminophores for coreactant ECL in aqueous solution, in which the electrochemical and photophysical properties introduced by different ligand combinations are used to control the chemi‐excitation pathways to the emitting species.

2. Results and Discussion

The chloride salts of five cationic heteroleptic complexes (Scheme S3) served as precursors to the novel sulfonated species. The Na[Ir(ppy)₂(bpy)], Na[Ir(ppz)₂(bpy)], and Na[Ir(pbt)₂(bpy)] complexes (where ppy, ppz, and pbt = cyclometalated phenylpyridine, phenylpyrazole, or phenylbenzothiazole ligands respectively; Scheme S3), were prepared by heating the selected iridium dimeric precursor complex ([Ir(μ‐Cl)(ppy)₂]₂, [Ir(μ‐Cl)(ppz)₂]₂, and [Ir(μ‐Cl)(pbt)₂]₂; Scheme S2) with 2,2′‐bipyridine in a 1:1 mixture of dichloroethane and ethanol at 80 °C for 1 hour using microwave irradiation. The N‐heterocyclic carbene complexes [Ir(ppy)₂(pyim)]Cl and [Ir(ppy)₂(pybm)]Cl (where pyim = pyridylidazolylidene, pybm = pyridylbenzimidazolylidene; Scheme S3) were synthesized using a Ag(I) transmetalation protocol,^{[ 40} , ^{41 ]} where the azolium salts 4 or 5 (Scheme S1) were heated with Ag₂O and [Ir(μ‐Cl)(ppy)₂]₂ in a 1:1 mixture of dichloroethane and ethanol using microwave irradiation.

The sulfonated heteroleptic iridium(III) complexes (Figure 2) were prepared from the five parents using a modification of a procedure for the synthesis of sulfonated homoleptic iridium(III) complexes.^{[ 32} , ^{35 ]} In the present study, the chosen complex was stirred in a mixture of sulfuric acid (2.2 eq.) and trifluoroacetic anhydride (TFAA) for 15 minutes at room temperature. After neutralizing the reaction mixture with NaHCO₃, the crude products were purified using column chromatography on silica.

Figure 2.

Structures of sulfonated heteroleptic iridium(III) complexes.

Analysis of the products formed in these reactions showed that for each complex, the phenyl rings of the cyclometalating ligands had been sulfonated at the position opposite to the cyclometalated carbon. The reaction time used here was significantly shortened compared to those described previously (12 hours) for the synthesis of sulfonated homoleptic complexes.^{[ 32} , ^{35 ]} The choice of a shorter reaction time was the result of a systematic series of synthetic studies where different reaction times were examined in addition to heating the reaction mixture using either conventional or microwave heating methods. These studies showed that long reaction times or heating generally led to higher levels of complex decomposition in the acidic media and difficulties associated with the purification of the desired sulfonated products. This work showed that the sulfonation reaction is rapid and good yields were obtained with relatively short reaction times.

Interestingly, in the synthesis of Na[Ir(spbt)₂(bpy)], analysis of the crude reaction mixture by TLC showed that two new luminescent products were formed. These products were separated using column chromatography, and the structures of each compound were determined using a combination of NMR, XRD, and mass spectrometric analysis. The first product was found to be the target [Ir(spbt)₂(bpy)]⁻, where the phenyl rings of the 2‐phenylbenzothiazole cyclometalating ligands were sulfonated at the expected positions. Careful analysis of the second product showed that it was tetrasulfonated (i.e., [Ir(dspbt)₂(bpy)]³⁻). For this compound, the NMR results indicated that sulfonation had also occurred on the phenyl rings of the benzothiazole fragments of the 2‐phenylbenzothiazole ligands, as depicted in Figure 2. Subsequently, it was found that the tetrasulfonated product could be formed exclusively if the quantity of added sulfuric acid was doubled and the reaction time was extended to 1 hour.

2.1. Structural Studies

Crystals suitable for X‐ray diffraction were obtained for Na[Ir(sppy)₂(bpy)], Na[Ir(spbt)₂(bpy)], and Na[Ir(sppy)₂(pyim)]. The crystallographic data are presented in Table S1 and selected bond lengths (Å) are presented in Table 1. The crystal structures (Figure 3) show that for each complex, the iridium(III) centre is in an octahedral coordination environment, consisting of two cyclometalated ligands, which are sulfonated at the position opposite to the cyclometalated carbon, and an ancillary ligand. The octahedral iridium(III) complexes are bridged by sodium counterions bound to the sulfonate oxygen atoms, generating polymeric Ir(III) – Na(I) solvate chains (solvate = methanol for Na[Ir(sppy)₂(bpy)] and Na[Ir(spbt)₂(bpy)], and methanol + water for Na[Ir(sppy)₂(pyim)]). A detailed description of the extended structures for these complexes is presented in the Supporting Information. In the case of Na[Ir(sppy)₂(bpy)] and Na[Ir(spbt)₂(bpy)], the structures contain crystallographically distinct octahedral iridium(III) complexes (denoted A and B in Table 1), which are present in equal amounts.

Table 1.

Selected bond lengths (Å) for complexes present in the crystal structures of Na[Ir(sppy)₂(bpy)], Na[Ir(spbt)₂(bpy)], and Na[Ir(sppy)₂(pyim)].

Bond	Na[Ir(sppy)2(bpy)] (A)	Na[Ir(spbt)2(bpy)] (A)	Na[Ir(sppy)2(pyim)]	Bond	Na[Ir(sppy)2(bpy)] (B)	Na[Ir(spbt)2(bpy)] (B)
Ir1‐N1	2.053(4)	2.082(9)	2.076(7)	Ir2‐N5	2.047(4)	2.090(8)
Ir1‐N2	2.048(4)	2.094(8)	2.067(7)	Ir2‐N6	2.050(4)	2.089(9)
Ir1‐N3	2.152(4)	2.152(8)	2.180(7)	Ir2‐N7	2.148(4)	2.161(8)
Ir1‐N4	2.144(4)	2.167(8)		Ir2‐N8	2.140(4)	2.143(9)
Ir1‐C1	2.011(5)	2.027(10)	2.018(9)	Ir2‐C3	2.018(5)	2.036(11)
Ir1‐C2	2.013(4)	2.046(10)	2.060(8)	Ir2‐C4	2.013(5)	2.028(10)
Ir1‐C3			2.101(8)

Figure 3.

Representations of the X‐ray crystal structures of complexes (a) [Ir(sppy)₂(bpy)]⁻, (b) [Ir(spbt)₂(bpy)]⁻, and (c) [Ir(sppy)₂(pyim)]⁻. Sodium cations along with solvent molecules, MeOH for [Ir(sppy)₂(bpy)]⁻ and [Ir(spbt)₂(bpy)]⁻, and MeOH + H₂O for [Ir(sppy)₂(pyim)]⁻ have been omitted for clarity. Thermal ellipsoids are shown at 40% probability.

2.2. Spectroscopy

The UV‐Vis absorption spectra for the six sulfonated heteroleptic iridium(III) complexes in water were similar to those of their parent complexes in acetonitrile (Figure S24) except that the bands below ∼340 nm attributable to ligand centered (LC) π→π* transitions were generally more intense. The bands above ∼340 nm can be assigned predominantly to spin‐allowed and spin‐forbidden charge transfer transitions, as previously discussed in detail.^{[ 42} , ⁴³ , ⁴⁴ , ^{45 ]} Photoluminescence emission spectra of the parent complexes in acetonitrile were in good agreement with published data (Table S2).^{[ 24} , ^{42 − 47 ]}

The [Ir(ppy)₂(bpy)]⁺ and [Ir(ppz)₂(bpy)]⁺ complexes exhibited broad, featureless emissions (Figures S25a, S25b and Table 2) with large Stokes shifts from their lowest energy absorption bands, representative of triplet metal‐ligand‐to‐ligand charge‐transfer (³MLLCT) character.^{[ 42} , ^{43 ]} The selective sulfonation of their cyclometalated ligands resulted in large hypsochromic shifts of 41 nm. At low temperature, the emission of [Ir(ppy)₂(bpy)]⁺ (Figure S26a) comprises multiple electronic transitions, which is more evident in derivatives containing alkyl groups on the bpy ligand (Figure S27) or sulfonate groups on the ppy ligands (Figure S28a), which hinders assessment of the energy gap between the zeroth vibrational levels of the ground and lowest triplet excited state (E _0‐0). The spectral distribution of the phenylpyrazole analogues indicates a similar occurrence. For estimations of E _0‐0, we therefore considered previous assignments of electronic transitions,^{[ 48} , ⁴⁹ , ^{50 ]} and the energy difference in the electrochemical oxidation and reduction of each complex (Table 2).

Table 2.

Selected spectroscopic, electrochemical, and ECL data.

Complex	PL [r.t.] λmax [nm][ [a] , [b] ]	PL [l.t.] λmax [nm][ [b] , [c] ]	E 0‐0 [eV][ d ]	E 0′[M/M‐] vs Fc+/0 [V][ [e] , [f] ]	E 0′[M+/M] vs Fc+/0 [V][ e ]	ΔE [V]	E 0′[M*/M‐] vs Fc+/0 [V][ g ]	E 0′[M+/M*] vs Fc+/0 [V][ g ]	E 0′[M+/M] vs Ag/AgCl [V] Aq.[ h ]	ECL Int. [PMT][ i ]	ECL Int. [CCD][ i ]
[Ru(bpy)3]2+	620	581, 629	2.13	−1.73	0.89	2.62	0.40	−1.24	1.08	100	100
[Ir(sppy)3]3−	518	481, 516	2.58	−2.56[ j ]	0.31	2.87	0.02	−2.27	0.79	0.4	0.5
[Ir(df‐sppy)3]3−	484	449, 480	2.76	−2.59[ j ]	0.68	3.27	0.17	−2.08	1.08	1.7	1.3
[Ir(ppy)2(bpy)]+	608	520, 537	2.38[ k ]	−1.78	0.88	2.66	0.51	−1.41			200[ l ]
[Ir(sppy)2(bpy)]−	567	469, 503	2.46[ k ]	−1.78	0.82	2.60	0.68	−1.64	1.32	91	31
[Ir(ppz)2(bpy)]+	590	500(sh), 517	2.47[ k ]	−1.80	0.96	2.76	0.67	−1.51
[Ir(sppz)2(bpy)]−	549	465(sh), 489	2.54[ k ]	−1.80	0.87	2.67	0.74	−1.67	1.36	56	28
[Ir(pbt)2(bpy)]+	525, 565	514, 555	2.41	−1.73	1.03	2.76	0.68	−1.38
[Ir(spbt)2(bpy)]−	525, 562	513, 554	2.42	−1.72	1.00[ j ]	2.72	0.70	−1.42	1.45[ j ]	44	22
[Ir(dspbt)2(bpy)]3−	524, 562	515, 556	2.41	−1.73	1.00[ [j] , [m] ]	2.73	0.68	−1.41	1.51[ j ]	61	24
[Ir(ppy)2(pyim)]+	474, 503	466, 500	2.66	−2.34[ j ]	0.80	3.14	0.32	−1.86			1.5[ n ]
[Ir(sppy)2(pyim)]−	466, 496	461, 494	2.69	−2.36[ j ]	0.76[ j ]	3.12	0.33	−1.93	1.25	17	10
[Ir(ppy)2(pybm)]+	470, 501	463, 497	2.68	−2.30[ j ]	0.85	3.15	0.38	−1.83			3.2[ n ]
[Ir(sppy)2(pybm)]−	464, 494	460, 493	2.70	−2.31[ j ]	0.83[ j ]	3.14	0.39	−1.87	1.31	19	8.3

^[a] Photoluminescence emission maxima of metal complexes at 10 µM in acetonitrile ([Ru(bpy)₃]²⁺ and parent iridium(III) complexes) or deionized water (sulfonated complexes) at room temperature.

^[b] Corrected for the change in instrument sensitivity over the wavelength range.

^[c] Photoluminescence emission maxima of metal complexes at 5 µM in ethanol:methanol (4:1) at 85 K (up to two most prominent bands listed, with most intense shown in bold; sh = shoulder).

^[d] Energy gap between the zeroth vibrational levels of the ground and triplet excited states, estimated from the highest energy peak of the low‐temperature emission spectrum, unless otherwise indicated.

^[e] Metal complex at 0.5 mM in acetonitrile ([Ru(bpy)₃]²⁺ and parent iridium(III) complexes) or dimethylformamide (sulfonated complexes) with 0.1 M TBAPF₆ supporting electrolyte; scan rate: 0.1 V s⁻¹.

^[f] Only first reduction potentials listed.

^[g] Excited state potentials estimated from the ground state potentials and E _0‐0.^{[ 52 ]}

^[h] Metal complex at 0.5 mM in aqueous phosphate buffer (0.1 M, pH 7.5).

^[i] Comparison of ECL intensities in ProCell solution using 80 cycles of 0.1 s at [E ⁰′(M⁺/M) + 0.1 V] and 0.2 s at 0 V, or in acetonitrile with TPrA coreactant under conditions as stated in cited papers.

^[j] Peak potential of an irreversible electron transfer.

^[k] Based on other considerations than the shortest wavelength band of low‐temperature emission (see main text).

^[l] From ref. ^{[ 53 ]}

^[m] Estimated from potential for [Ir(spbt)₂(bpy)]⁺.

^[n] From ref. ^{[ 47 ]}

In contrast, the emission spectra of complexes with the phenylbenzothiazole or N‐heterocyclic carbene ligands showed vibronic fine structure even at ambient temperature (Figures S25c^–S25f; Table 2), indicative of more pronounced ³LC contributions,^{[ 51 ]} and like the homoleptic cyclometalated iridium(III) complexes,^{[ 37 ]} the parent and sulfonated analogues exhibited similar spectral distributions (Δλ _max ≤ 10 nm). The spectra for these complexes at low temperature (Figures S26c^–S26e and S28c^–S28f) showed greater vibronic detail, enabling more accurate estimations of E _0‐0 (Table 2).

2.3. Electrochemistry

Reduction potentials for the parent complexes in acetonitrile determined by cyclic voltammetry (Figure S29) were in agreement with previous reports (Table S2).^{[ 24} , ⁴² , ⁴³ , ⁴⁵ , ⁴⁶ , ^{47 ]} Complexes [Ir(ppy)₂(bpy)]⁺, [Ir(ppz)₂(bpy)]⁺, and [Ir(pbt)₂(bpy)]⁺ showed reversible or quasi‐reversible oxidations between 0.88 V and 1.03 V versus Fc^+/0 and reductions from −1.73 V to −1.80 V versus Fc^+/0. The smaller difference in the reduction potentials reflects the localization of the LUMO on the 2,2ʹ‐bipyridine ligand.^{[ 42} , ^{43 ]} Replacement of the bipyridine with the N‐heterocyclic carbene ligands (to give [Ir(ppy)₂(pyim)]⁺ and [Ir(ppy)₂(pybm)]⁺) resulted in a negative shift in the potentials required for oxidation (0.80 and 0.85 V versus Fc^+/0) and to a much greater extent, reduction (−2.34 and −2.31 V versus Fc^+/0). The reduction of these two complexes was chemically irreversible even when increasing the scan rate to 0.5 V s⁻¹.

The sulfonated complexes in N,N‐dimethylformamide showed similar potentials to those of the parent complexes in acetonitrile (Figure S29; Table 2), albeit with a loss of the chemical reversibility in the oxidation of the complexes with phenylbenzothiazole or N‐heterocyclic‐carbene ligands. In the case of [Ir(sppy)₂(pyim)]⁻ and [Ir(sppy)₂(pybm)]⁻, the anodic peak potential was within the electrochemical window of the matrix. For [Ir(spbt)₂(bpy)]⁻, however, the oxidation was obscured but could be elucidated by square wave voltammetry.

For the exploration of the novel sulfonated iridium(III) complexes as electrochemiluminophores in aqueous solution, the potentials required for their oxidation in 0.1 M phosphate buffer (pH 7.5) were also examined (Figure S30; Table 2). The chemical reversibility of the oxidation generally increased with scan rate from 0.1 to 1.0 V s⁻¹, with the exception of [Ir(dspbt)₂(bpy)]³⁻, exhibiting the highest potential for oxidation, which remained irreversible over the entire scan‐rate range. Square wave voltammetry was used to confirm the position of each potential.

2.4. ECL

In ECL, the electronically excited state responsible for the emission is attained through chemical reactions between electrochemically generated intermediates.^{[ 54 ]} The majority of analytical applications of ECL utilize tris(2,2′‐bipyridine)ruthenium(II) ([Ru(bpy)₃]²⁺) as the luminophore and tri‐n‐propylamine (TPrA) as a coreactant that enables the reaction to be initiated in aqueous solution by a single applied potential (Reactions 1–9, where M = [Ru(bpy)₃]²⁺, TPrA is the α‐amino alkyl radical reductant Pr₂NC^•HCH₂CH₃, and P⁺ is an iminium ion, Pr₂N⁺C═CH₂CH₃, which hydrolyses in aqueous solution).^{[ 55 ]} If both reactants can be oxidized, the emitting species can be generated by reactions 1–5, referred to as the “direct” pathway. The [Ru(bpy)₃]²⁺ luminophore can also be excited via reduction by TPrA (reaction 6) and then oxidation by TPrA (reaction 8), labelled the “indirect” pathway, which is dominant if the luminophore is not oxidized due to insufficient applied potential, or when the complex is immobilized away from the electrode surface. When the luminophore is free in solution at concentrations much lower than the coreactant, ECL from the indirect and direct pathways can be seen as two distinct “waves” (Figure 4a) at potentials sufficient to oxidize the TPrA and [Ru(bpy)₃]²⁺, respectively.^{[ 55 ]}

$$\mathrm{TPrA}\to {\mathrm{TPrA}}^{•+}+e^{-}$$ (1)

$$\mathrm{M}\to {\mathrm{M}}^{+}+e^{-}$$ (2)

$${\mathrm{M}}^{+}+\mathrm{TPrA}\to \mathrm{M}+{\mathrm{TPrA}}^{•+}$$ (3)

$${\mathrm{TPrA}}^{•+}\to {\mathrm{TPrA}}^{•}+{\mathrm{H}}^{+}$$ (4)

$${\mathrm{M}}^{+}+{\mathrm{TPrA}}^{•}\to {\mathrm{M}}^{\ast }+{\mathrm{P}}^{+}$$ (5)

$$\mathrm{M}+{\mathrm{TPrA}}^{•}\to {\mathrm{M}}^{-}+{\mathrm{P}}^{+}$$ (6)

$${\mathrm{M}}^{+}+{\mathrm{M}}^{-}\to {\mathrm{M}}^{\ast }+\mathrm{M}$$ (7)

$${\mathrm{M}}^{-}+{\mathrm{TPrA}}^{•+}\to {\mathrm{M}}^{\ast }+\mathrm{TPrA}$$ (8)

$${\mathrm{M}}^{\ast }\to \mathrm{M}+h\nu$$ (9)

Figure 4.

ECL intensities (normalized) obtained when cycling the applied potential between 0 V and 1.6 V versus Ag/AgCl for (a) [Ru(bpy)₃]²⁺ (b) [Ir(sppy)₂(bpy)]⁻ (blue plot) and [Ir(sppz)₂(bpy)]⁻ (red plot), (c) [Ir(spbt)₂(bpy)]⁻ (blue plot) and [Ir(dspbt)₂(bpy)]³⁻ (red plot), (d) [Ir(sppy)₂(pyim)]⁻ (blue plot) and [Ir(sppy)₂(pybm)]⁻ (red plot), obtained using 10 µM metal complex in ProCell solution. The dashed colored lines show the E ⁰′(M⁺/M) (or E _p) for the metal complexes from Table 2. The dashed black line shows the estimated E ⁰′ for the oxidation of the TPrA coreactant in aqueous solution.^{[ 55 ].}

To extend these well‐established excitation pathways to iridium(III) complexes, however, their redox potentials and triplet excited state energies (Table 2) must be considered.^{[ 24} , ^{56 ]} For all six of the sulfonated heteroleptic iridium(III) complexes, reduction of the oxidized metal complex by the TPrA radical (reaction 5; −2.1 V versus Fc^+/0)^{[ 57 ]} is sufficiently energetic to attain the electronically excited state, where the difference in the electrochemical potentials of the reactants (E ⁰′(M⁺/M) – E ⁰′(TPrA)) is greater than the excited state energy (E _0‐0). This can be seen in Table 2, where (E ⁰′(M⁺/M*) for each sulfonated heteroleptic iridium(III) complex is less negative than E ⁰′(TPrA). While these estimates indicate the feasibility of the reaction, the ECL intensity (photons per time unit) and ECL quantum yield (photons emitted per electron transfer) are determined by both the excitation efficiency (the rate of the light‐producing pathway versus various deleterious reactions), and emission efficiency (luminescence quantum yield). A large increase in ECL intensity was observed when scanning potentials nearing the E ⁰′(M⁺/M) of each sulfonated heteroleptic iridium(III) complex when using TPrA as a coreactant (Figure 4b−d), corresponding to the direct ECL pathway.

The emission from each reaction was confirmed to arise from the luminophore by comparing the ECL spectra (Figure 5) with the corresponding photoluminescence of the iridium(III) complex (Figure S31). The only significant difference between these spectra was the diminished vibrational structure, particularly for the complexes exhibiting the sharpest and highest energy peaks ([Ir(sppy)₂(pyim)]⁻ and [Ir(sppy)₂(pybm)]⁻), which can be attributed to the lower resolution of the charge coupled device (CCD) spectrometer used to collect the ECL spectra. The ECL of the six novel iridium(III) complexes was higher in energy than that of the conventional [Ru(bpy)₃]²⁺ luminophore. Photographs of the emissions from the working electrode surface showed colors from teal to yellow‐green (Figure 5), distinct from the orange of [Ru(bpy)₃]²⁺.

Figure 5.

ECL spectra for [Ir(sppy)₂(pybm)]⁻ (dashed blue plot), [Ir(sppy)₂(pyim)]⁻ (blue plot), [Ir(dspbt)₂(bpy)]³⁻ (dashed green plot), [Ir(spbt)₂(bpy)]⁻ (green plot), [Ir(sppz)₂(bpy)]⁻ (dashed black plot), [Ir(sppy)₂(bpy)]⁻ (black plot), and [Ru(bpy)₃]²⁺ (orange plot), in ProCell solution upon application of 0.1 V above the oxidation potential of the metal complex. The photographs show the ECL from the working electrode surface for the complexes in the same order to that listed above.

The ECL intensities (integrated area under the emission spectrum) of the six sulfonated heteroleptic iridium(III) complexes with TPrA as a coreactant upon application of 0.1 V beyond the E ⁰′(M⁺/M) of the metal complex were between 8% and 31% of that for the conventional [Ru(bpy)₃]²⁺ luminophore (Figure 6; green columns, complexes 3–8), using the CCD spectrometer as the photodetector. ECL intensities, however, are more commonly measured as the integrated emission over time, using photomultiplier tubes that are considerably more sensitive toward the blue‐green than the red region of the visible region. When using a trialkali photomultiplier tube, the relative coreactant ECL intensities of the sulfonated heteroleptic iridium(III) complexes were increased to between 17% and 91% to that of [Ru(bpy)₃]²⁺ (Figure 6; blue columns, complexes 3–8), due to the shorter wavelengths of the iridium(III)‐complex emissions.

Figure 6.

ECL intensity (relative to [Ru(bpy)₃]²⁺ = 100) of iridium(III) complexes: (1) [Ir(sppy)₃]³⁻, (2) [Ir(df‐sppy)₃]³⁻, (3) [Ir(sppy)₂(bpy)]⁻, (4) [Ir(sppz)₂(bpy)]⁻, (5) [Ir(spbt)₂(bpy)]⁻, (6) [Ir(dspbt)₂(bpy)]³⁻, (7) [Ir(sppy)₂(pyim)]⁻, and (8) [Ir(sppy)₂(pybm)]⁻, at 1 µM in ProCell solution upon application of 0.1 V above the E ⁰′(M⁺/M) of the metal complex, where the light was measured using a photomultiplier tube (blue columns) or CCD spectrometer (green columns). The error bars show ± 1 standard deviation.

The heteroleptic iridium(III) complexes examined in this study gave much greater coreactant ECL intensities than the previously reported^{[ 35} , ^{37 ]} homoleptic species [Ir(sppy)₃]³⁻ and [Ir(df‐sppy)₃]³⁻ (Figure 6) that were prepared through a similar sulfonation procedure (Figure 1). Notably, the ECL intensity of [Ir(sppy)₂(bpy)]⁻ was over 200‐fold greater than [Ir(sppy)₃]³⁻, and comparable to the archetypal [Ru(bpy)₃]²⁺ luminophore. Similarly, no ECL is observed with the homoleptic [Ir(sppz)₃]³⁻ complex with TPrA coreactant,^{[ 39 ]} but the ECL intensity of the heteroleptic [Ir(sppz)₂(bpy)]⁻ is 56% that of [Ru(bpy)₃]²⁺. In general, the stabilized frontier orbitals provided by the bipyridine ligand shifts the electrochemical potentials in the positive direction, increasing the efficiency of the ECL excitation pathways with TPrA as a coreactant. Luminophores incorporating this ligand can also be readily converted into ECL labels by replacement with the butanoic acid derivative, which will have minimal effect on luminophore properties.

The iridium(III) complexes described in this study were targeted in part because they were predicted to proceed through three distinct mechanisms at potentials at which TPrA but not the luminophore is oxidized, where the chemi‐excitation pathway depends on reactions with TPrA and TPrA (i.e., the indirect route of coreactant ECL).^{[ 24} , ^{47 ]} For the complexes containing N‐heterocyclic carbene ancillary ligands, [Ir(sppy)₂(pyim)]⁻ and [Ir(sppy)₂(pybm)]⁻, the potentials required for their reduction (Table 2) are beyond that of the TPrA radical (E ⁰ʹ(M/M⁻) < −2.1 V versus Fc^+/0)^{[ 57 ]} and therefore no reaction was anticipated without oxidation of the metal complex, which was the observed outcome (Figure 4d).

The other four sulfonated complexes should be reduced by TPrA. Estimates of E ⁰(M*/M⁻) shown in Table 2 suggest that subsequent oxidation by TPrA (E ⁰′ = 0.48 V versus Fc^+/0) is not sufficiently energetic to generate their electronically excited state. However, there are numerous sources of error in these predictions, and these should be considered borderline cases. When increasing the applied potential, the onset of coreactant ECL for [Ir(sppy)₂(bpy)]⁻ and [Ir(sppz)₂(bpy)]⁻ was not observed until the applied potential neared that required to oxidize the metal complex (Figure 4b), but for [Ir(spbt)₂(bpy)]⁻ and [Ir(dspbt)₂(bpy)] ³⁻, two distinct ECL waves were seen, the first of which occurred when only TPrA was oxidized (Figure 4c), albeit at lower relative intensities than the ECL of [Ru(bpy)₃]²⁺ generated through the analogous indirect pathway (Figure 4a).

2.5. Redox‐Mediator‐Enhanced ECL

Having established the excitation pathways available to the novel water‐soluble luminophores, we selected two complexes ([Ir(sppy)₂(bpy)]⁻ and [Ir(spbt)₂(bpy)]⁻) to gain new insight into the redox‐mediated enhancement of coreactant ECL, which previously had only been applied to [Ru(bpy)₃]²⁺.^{[ 36} , ³⁷ , ³⁸ , ³⁹ , ^{58 ]} For the redox mediator in these experiments, we used the homoleptic sulfonated iridium(III) complex [Ir(sppy)₃]³⁻ (Figure 1a), previously found effective for [Ru(bpy)₃]²⁺.^{[ 36} , ^{37 ]}

The effect of the mediator was evaluated using an electrochemical pulse sequence to collect ECL spectra over a series of applied potentials, which can be presented as contour plots (Figures S32–S34).The emission spectrum obtained at each potential was then deconvoluted into contributions from the luminophore and redox mediator (Figure S35), so that the ECL intensity of the luminophore with and without the mediator can be compared (Figure 7). As in our previous work,^{[ 36 ]} the coreactant ECL intensity of [Ru(bpy)₃]²⁺ was enhanced by [Ir(sppy)₃]³⁻ most prominently at potentials close to E ⁰′(M⁺/M) of the mediator (Figure 7a), where the emission arises from the indirect ECL pathway (reactions 1, 4, 6, 8, and 9). The mechanisms of this enhancement have been elucidated as electrocatalysis of coreactant oxidation (reactions 10 and 11) and an alternative chemi‐excitation step (reaction 12).^{[ 37} , ³⁸ , ^{58 ]}

$${{\left[\mathrm{Ir}{\left(\mathrm{sppy}\right)}_3\right]}^3}^{-}\to {{\left[\mathrm{Ir}{\left(\mathrm{sppy}\right)}_3\right]}^2}^{-}+e^{-}$$ (10)

$${{\left[\mathrm{Ir}{\left(\mathrm{sppy}\right)}_3\right]}^2}^{-}+\mathrm{TPrA}\to {{\left[\mathrm{Ir}{\left(\mathrm{sppy}\right)}_3\right]}^3}^{-}+{\mathrm{TPrA}}^{•+}$$ (11)

$${{\left[\mathrm{Ir}{\left(\mathrm{sppy}\right)}_3\right]}^2}^{-}+{\mathrm{M}}^{-}\to {{\left[\mathrm{Ir}{\left(\mathrm{sppy}\right)}_3\right]}^3}^{-}+{\mathrm{M}}^{\ast }$$ (12)

Figure 7.

ECL intensity (integrated spectrum) of 1 µM (a) [Ru(bpy)₃]²⁺, (b) [Ir(sppy)₂(bpy)]⁻, or (c) [Ir(spbt)₂(bpy)]⁻, in ProCell solution without (dashed plots) or with (solid plots) 100 µM [Ir(sppy)₃]³⁻, at a series of different applied potentials. The blue and green boxes depict the regions at which the indirect and direct ECL mechanisms were anticipated to occur, based on the potentials required to oxidize the coreactant and the luminophore, respectively. The contribution of [Ir(sppy)₃]³⁻ to the ECL intensity has been removed. Experimental and fitted ECL intensities from all species are presented in Figures S32–S36.

In contrast, the coreactant ECL of [Ir(sppy)₂(bpy)]⁻ was partially quenched by the addition of [Ir(sppy)₃]³⁻ (Figure 7b). Although the novel luminophore is reduced by TPrA (reaction 6), neither reaction 8 of the indirect pathway nor reaction 12 with the mediator is exergonic, and therefore no ECL is observed at potentials insufficient to oxidize the luminophore. At higher potentials, electrocatalytic oxidation of the coreactant is less competitive with its electrochemical oxidation, and reaction 5 of the direct ECL pathway is diminished by the consumption of TPrA by the oxidized redox mediator (reaction 13).

$${{\left[\mathrm{Ir}{\left(\mathrm{sppy}\right)}_3\right]}^2}^{-}+{\mathrm{TPrA}}^{•}\to {\left({{\left[\mathrm{Ir}{\left(\mathrm{sppy}\right)}_3\right]}^3}^{-}\right)}^{\ast }+\mathrm{P}$$ (13)

Under these experimental conditions, even in the absence of the mediator, the first wave of the coreactant ECL of [Ir(spbt)₂(bpy)]⁻ (Figure 7c, dashed plot) was more prominent than when scanning the applied potential (Figure 4c, blue plot). Introduction of the redox mediator increased the ECL intensity by 12‐fold at 1.0 V versus Ag/AgCl, similar to that for [Ru(bpy)₃]²⁺ (13‐fold at 0.95 V versus Ag/AgCl), which can be confidently ascribed to reactions 10–12.

3. Conclusion

This study demonstrates a synthetic approach to impart water‐solubility to iridium(III) complexes with different ligand combinations, including the capacity to introduce different numbers of sulfonate groups to cyclometalated ligands with additional aromatic rings. Considering the extensive range of commercially available heteroleptic iridium(III) complexes, and the much wider range reported in the literature, this approach offers an unprecedented “toolbox” of water‐soluble analogues with diverse electrochemical and photophysical properties that can be estimated from those of the parent compounds, for the development of new photocatalysts and luminophores. For ECL detection, this approach removes a longstanding barrier to the translation of high‐efficiency, multicolored, and potential‐resolved iridium(III)‐based systems, predominantly developed in organic solvents, to the aqueous conditions desired for many analytical applications.

4. Experimental Section

General synthesis

All reagents were purchased from Sigma Aldrich, Alfa Aesar, or Precious Metals Online and were used without further purification. NMR spectra were recorded on either a Bruker Avance ARX‐400 (400.13 MHz for ¹H, 100.61 MHz for ¹³C) or a Bruker Avance ARX‐500 (500.13 MHz for ¹H, 125.77 MHz for ¹³C) spectrometer and were internally referenced to solvent resonances. High‐resolution mass spectra were obtained using an Agilent 6530 QTOF LC/MS mass spectrometer fitted with an Agilent electrospray ion (ESI) source.

Sulfonated heteroleptic iridium(III) complexes

Na[Ir(sppy)₂(bpy)]. Sulfuric acid (19 mg, 10.2 µL, 0.19 mmol) was added to TFAA (3 mL, 21.60 mmol) and the mixture was stirred at RT for 15 minutes. To the mixture was added a solution of [Ir(ppy)₂(bpy)]Cl (0.060 g, 0.087 mmol) in dichloroethane (12 mL) and the reaction was stirred at RT for 15 minutes. The reaction was then neutralized by adding a concentrated solution of NaHCO₃ and the solvent was removed on a rotatory evaporator. The solid residue was then dissolved in methanol (15 mL) and filtered and the filtrate was concentrated to dryness on a rotatory evaporator. The solid residue was redissolved in ethyl acetate (15 mL), gently heated to 50 °C for 10 minutes and filtered through Celite and the filtrate was purified by column chromatography on silica using MeOH and EtOAc (7:3 v/v) as eluent to afford the pure product as a yellow solid. (Yield = 0.037 g, 50%). ¹H NMR (500 MHz, DMSO‐d₆ ): 8.88 (d, J = 8.20 Hz, 2H), 8.28 (t, J = 7.80 Hz, 2H), 8.24 (d, J = 8.10 Hz, 2H), 8.07 (s, 2H), 7.95 (t, J = 7.80 Hz, 2H), 7.89 (d, J = 5.3 Hz, 2H), 7.75 (t, J = 6.70 Hz, 2H), 7.63 (d, J = 5.70 Hz, 2H), 7.19 (d, J = 7.05 Hz, 2H), 7.17 (t, J = 8.05 Hz, 2H), 6.17 (d, J = 7.85 Hz, 2H). ¹³C NMR (125 MHz, DMSO‐d₆ ): δ(ppm) = 166.9, 155.8, 151.2, 150.3, 149.4, 143.8, 143.2, 140.2, 139.5, 130.7, 129.2, 128.0, 125.5, 124.5, 122.5, 120.5. ESI‐MS: m/z = 815.0597 [C₃₂H₂₂IrN₄O₆S₂]⁻, (calcd = 815.0605).

Na[Ir(sppz)₂(bpy)]. This compound was prepared as described for Na[Ir(sppy)₂(bpy)] from [Ir(ppz)₂(bpy)]Cl (0.080 g, 0.12 mmol). The pure product was obtained as a yellow solid. (Yield = 0.028 g, 29%). ¹H NMR (500 MHz, DMSO‐d₆ ): 9.02 (d, J = 2.8 Hz, 2H), 8.85 (d, J = 8.40 Hz, 2H), 8.28 (t, J = 7.70 Hz, 2H), 8.04 (d, J = 5.6 Hz, 2H), 7.89 (s, 2H), 7.72 (t, J = 5.18 Hz, 2H), 7.17 (d, 2H), 7.10 (d, J = 7.85 Hz, 2H), 6.66 (t, 2H), 6.14 (d, J = 7.50 Hz, 2H). ESI‐MS: m/z = 793.0530 [C₂₈H₂₀IrN₆O₆S₂]⁻, (calcd = 793.0526).

Na[Ir(spbt)₂(bpy)]. This compound was prepared as described for Na[Ir(sppy)₂(bpy)] from [Ir(pbt)₂(bpy)]Cl (0.060 g, 0.072 mmol). The pure product was obtained as a yellow solid. (Yield = 0.030 g, 44%). ¹H NMR (500 MHz, DMSO‐d₆ ): 8.79 (d, J = 7.85 Hz, 2H), 8.29 (t, J = 7.30 Hz, 2H), 8.32 (d, J = 8.00 Hz, 2H), 8.10 (s, 2H), 8.07 (d, J = 5.30 Hz, 2H), 7.81 (t, J = 7.00 Hz, 2H), 7.43 (t, J = 7.90 Hz, 2H), 7.18 (d, J = 8.30 Hz, 2H), 7.14 (t, J = 8.75 Hz, 2H), 6.30 (d, J = 8.05 Hz, 2H), 6.09 (d, J = 8.15 Hz, 2H). ¹³C NMR (125 MHz, DMSO‐d₆ ): δ(ppm) = 181.3, 156.3, 151.6, 150.7, 148.9, 143.9, 140.9, 139.6, 132.6, 131.7, 129.6, 129.5, 128.6, 126.6, 125.2, 125.1, 123.9, 117.1. ESI‐MS: m/z = 927.0097 [C₃₆H₂₂IrN₄O₆S₄]⁻, (calcd = 927.0047).

Na₃[Ir(dspbt)₂(bpy)]. Concentrated sulfuric acid (0.031 g, 16.9 µL, 0.32 mmol) was added to TFAA (6 mL, 43.20 mmol) and the mixture was stirred at RT for 15 minutes. To the mixture was added a solution of [Ir(pbt)₂(bpy)]Cl (0.060 g, 0.072 mmol) in dichloroethane (12 mL) and the reaction was stirred at RT for 1 hour. The reaction was then neutralized by adding a concentrated solution of NaHCO₃ and the solvent was removed on a rotatory evaporator. The solid residue was then dissolved in methanol (15 mL) and filtered and the filtrate was concentrated to dryness on a rotatory evaporator. The solid residue was redissolved in ethyl acetate (15 mL), gently heated to 50 °C for 10 minutes and filtered through Celite. The filtrate was purified by column chromatography on silica gel using MeOH and EtOAc (7:3 v/v) as eluent to afford the pure product as a yellow solid. (Yield = 0.024 g, 29%). ¹H NMR (500 MHz, DMSO‐d₆ ): 8.79 (d, J = 8.55 Hz, 2H), 8.43 (s, 2H), 8.26 (t, J = 7.25 Hz, 2H), 8.06 (s, 2H), 8.04 (d, J = 5.3 Hz, 2H), 7.79 (t, J = 6.90 Hz, 2H), 7.31 (d, J = 8.95 Hz, 2H), 7.15 (d, J = 7.30 Hz, 2H), 6.35 (d, J = 7.70 Hz, 2H), 5.99 (d, J = 8.75 Hz, 2H). ¹³C NMR (125 MHz, DMSO‐d₆ ): δ(ppm) = 182.3, 156.3, 151.6, 150.6, 148.9, 146.5, 144.3, 140.9, 139.5, 132.8, 131.4, 129.8, 129.5, 126.4, 125.3, 123.9, 122.3, 116.6. ESI‐MS: m/z = 553.9471 [C₃₆H₂₀IrN₄O₁₂S₆Na]⁻, (calcd = 553.9472).

Na[Ir(sppy)₂(pyim)]. This compound was prepared as described for Na[Ir(sppy)₂(bpy)] from [Ir(ppy)₂(pyim)]Cl (0.060 g, 0.086 mmol). The pure product was obtained as a yellow solid. (Yield = 0.025 g, 34%). ¹H NMR (500 MHz, DMSO‐d₆ ): 8.52 (s, 1H), 8.36 (d, J = 8.25 Hz, 1H), 8.26 (d, J = 7.75 Hz, 1H), 8.23 (t, J = 9.05 Hz, 1H), 8.18 (d, J = 7.85 Hz, 1H), 8.05 (d, J = 4.95 Hz, 2H), 7.99 (d, J = 5.60 Hz, 1H), 7.96 (t, J = 7.70 Hz, 2H), 7.63 (d, J = 4.66 Hz, 2H), 7.57 (s, 1H), 7.49 (t, J = 6.50 Hz, 1H), 7.20–7.16 (m, 3H), 7.08 (d, J = 7.90 Hz, 1H), 6.25 (d, J = 7.95 Hz, 1H), 6.10 (d, J = 7.60 Hz, 1H). ¹³C NMR (125 MHz, DMSO‐d₆ ): δ(ppm) = 206.9, 177.8, 167.9, 166.7, 164.9, 154.1, 153.6, 150.3, 149.5, 144.1, 143.9, 142.6, 142.3, 142.0, 139.2, 138.4, 130.1, 129.9, 128.0, 127.5, 125.9, 125.1, 124.8, 124.1, 122.3, 122.2, 120.6, 120.3, 118.3, 113.3, 36.9. ESI‐MS: m/z = 818.0743 [C₃₁H₂₄IrN₅O₆S₂]⁻, (calcd = 818.0724).

Na[Ir(sppy)₂(pybm)]. This compound was prepared as described for Na[Ir(sppy)₂(bpy)] from [Ir(ppy)₂(pybm)]Cl (0.060 g, 0.080 mmol) in dichloroethane (12 mL) and the reaction was stirred at RT for 15 minutes. The pure product was obtained as a yellow solid. (Yield = 0.021 g, 29%). ¹H NMR (500 MHz, DMSO‐d₆ ): 8.75 (d, J = 8.45 Hz, 1H), 8.55 (d, J = 7.60 Hz, 1H), 8.29 (t, J = 8.05 Hz, 1H), 8.24–8.19 (m, 3H), 8.08 (s, 1H), 7.96–7.93 (m, 2H), 7.76 (d, J = 3.45 Hz, 2H), 7.68 (d, J = 5.40 Hz, 1H), 7.60–7.51 (m, 3H), 7.22 (d, J = 7.70 Hz, 1H), 7.16–7.08 (m, 3H), 6.27 (d, J = 7.85 Hz, 1H), 6.06 (d, J = 7.60 Hz, 1H), 3.39 (s, 3H). ¹³C NMR (125 MHz, DMSO‐d₆ ): δ(ppm) = 187.6, 167.6, 166.6, 165.6, 154.8, 154.2, 150.8, 149.9, 149.7, 144.2, 143.8, 142.8, 139.4, 138.6, 136.8, 131.6, 130.1, 129.9, 128.1, 127.7, 125.6, 125.3, 125.2, 124.4, 124.2, 122.4, 120.7, 120.5, 114.4, 113.4, 113.0, 31.2. ESI‐MS: m/z = 868.0734 [C₃₅H₂₆IrN₅O₆S₂]⁻, (calcd = 868.0881).

X‐ray crystallography

Single crystals of complexes Na[Ir(sppy)₂(bpy)], Na[Ir(spbt)₂(bpy)], and Na[Ir(sppy)₂(pyim)] suitable for X‐ray diffraction studies were grown by the diffusion of vapours between diethyl ether and methanol or acetonitrile solutions of these compounds. Crystallographic data for all structures determined are given in Table S1. For all samples, crystals were removed from the crystallization vial and immediately coated with Paratone oil. A suitable crystal was mounted in Paratone oil on a mylar loop and cooled in a stream of cold nitrogen using an Oxford low‐temperature device. Diffraction data were measured using an Oxford Supernova diffractometer mounted with Mo‐Kα λ = 0.71 073 Å and Cu‐Kα λ = 1.54 184 Å. Data were reduced and corrected for absorption using the CrysAlis Pro program.^{[ 59 ]}

Structures were solved using SHELXT^{[ 60 ]} (intrinsic phasing) and refined with the SHELXL^{[ 61 ]} program by the full‐matrix least‐squares procedure based on F ². The nonhydrogen atoms were refined with anisotropic displacement parameters, except for a disordered methanol carbon atom in Na[Ir(sppy)₂(pyim)], which when refined with anisotropic displacement parameters resulted in unstable refinement. Hydrogen atoms were placed in geometrically estimated positions and refined using the riding model. Coordinates and anisotropic thermal parameters of all nonhydrogen atoms were refined. All calculations were carried out using the program Olex2.^{[ 62 ]} CCDC 2 395 239 (Na[Ir(sppy)₂(pyim)]), 2 395 240 (Na[Ir(spbt)₂(bpy)]), and 2 395 241 (Na[Ir(sppy)₂(bpy)]) contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via https://www.ccdc.cam.ac.uk/structures/.

UV‐visible spectroscopy

UV‐visible absorption spectra were collected using a Cary 60 or Cary 300 UV/Vis spectrophotometer (Agilent, Australia). Ambient temperature photoluminescence spectra of the sulfonated complexes (10 µM) in ultrapure water, and parent complexes (10 µM) in acetonitrile, were obtained using a Cary Eclipse fluorescence spectrophotometer (Agilent). Low‐temperature photoluminescence spectra of the metal complexes (5 µM) in 4:1 ethanol:methanol were obtained using an OptistatDN Variable Temperature Liquid Nitrogen Cryostat (Oxford Instruments, UK) with a custom‐made quartz sample holder.^{[ 63 ]} A temperature of 85 K was used to avoid damage to the sample holder as the solution approached 77 K.^{[ 64 ]} Under our experimental conditions, no difference in the wavelengths of maximum emission were previously observed between these two temperatures.^{[ 63 ]} All photoluminescence emission spectra were corrected using factors established with a quartz halogen tungsten lamp to account for variation in instrumental sensitivity over the wavelength range.

Electrochemistry and ECL

The electrochemical cell comprised a cylindrical quartz cell with a flat base and Teflon lid with four holes to accommodate the electrodes and gas line (when required). The cell was housed in a light‐tight Faraday cage. The three‐electrode configuration consisted of a glassy carbon working electrode (CH Instruments), a platinum wire counter electrode and a leak‐free Ag/AgCl reference electrode (model ET069, eDAQ, Australia). Potentials were applied using an Autolab PGSTAT204 or PGSTA128N potentiostat (Metrohm Autolab, B.V., Netherlands). Data were recorded using Nova software (version 1.11). ECL was measured by interfacing the base of the electrochemical cell with: (i) a photomultiplier tube (PMT, extended‐range trialkali S20, ET Enterprises model 9828B) for intensity; (ii) a charge coupled device (CCD, QEPro, Ocean Optics, via collimating lens and optical fibre) for the spectral distribution; or (iii) a digital camera (Canon EOS 6D DSLR camera with Tonika AT‐X PRO MACRO 100 mm f/2.8 D lens) for images. The CCD and digital camera were synchronized with the potentiostat using a HR4000 break‐out box (Ocean Optics).

For experiments in nonaqueous solvents, the metal complexes were prepared in freshly distilled acetonitrile or dimethylformamide with 0.1 M tetrabutylammonium hexafluorophosphate (TBAPF₆) supporting electrolyte. Prior to each experiment, the working electrode was polished with 0.05 µm alumina powder on a felt pad with deionized water, sonicated in and rinsed with appropriate solvents and dried with a lint‐free wipe. The counter electrode was rinsed with ethanol, then cleansed by flame until incandescent. The reference electrode was rinsed with appropriate solvents, then dried with a lint‐free wipe. Where required, the reference electrode was sonicated twice in fresh solvent for 5 minutes, to avoid contamination. Solutions were degassed with grade 5 argon for 10 minutes. The ferrocenium/ferrocene (Fc^+/0) couple was used as an in situ reference potential.

For experiments in aqueous solvents, the metal complexes were prepared in phosphate buffer (0.1 M, pH 7.5) or ProCell solution (containing 180 mM TPrA, 0.1% surfactant, and a preservative in 0.3 M phosphate buffer at pH 6.8).^{[ 65 ]} The working electrode was rinsed with ethanol, then water, and polished with 0.05 µm alumina powder, rinsed with water again, and dried with a lint‐free wipe. The counter electrode was rinsed with ethanol, then ultrapure water after every experiment, then cleansed and dried by flame until incandescent. The reference electrode was rinsed with water, ethanol, then water again, then dried with a sterile, lint‐free wipe. The working and reference electrode were regularly sonicated in water or phosphate buffer solution, to avoid contamination.

Enhanced ECL

“3D ECL” data (intensity versus emission wavelength and applied potential), also referred to as “ECL spooling”,^{[ 66} , ^{67 ]} were obtained using an automated procedure created in Nova software.^{[ 37 ]} A series of 10 s pulses of increasing potential were interspersed with 10 s at 0 V versus Ag/AgCl. Each potential pulse was synchronized with the QEPro CCD detector via an HR4000 break‐out box (Ocean Optics) to acquire ECL spectra. The presented data are the average of three replicate experiments, for which the relative standard deviation of the spectrum (integrated area) at the potential that elicited the greatest ECL intensity was generally less than 6%. Contour plots (Figure S32–S34) were created using OriginPro (OriginLab, USA). ECL spectra were deconvoluted (Figure S35) using Microsoft Excel with the Solver add‐in, by minimizing the sum of the magnitude difference between the model (I _total = c _Ru I _Ru + c _Ir I _Ir) and measured intensity (I _meas) at each wavelength, where c represents the metal complex concentration. ECL intensity versus applied potential plots (Figures 7 and S36) were obtained from the experimental and modelled data by integrating the ECL spectral distribution at each applied potential.

Synthesis of precursor compounds, further structural studies, and additional spectroscopic, electrochemical, and ECL data are included as supporting information.

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting information

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.