of Birmingham High performance optical oxygen sensors based on iridium complexes exhibiting interchromophore energy shuttling

A doubly pyrene-grafted bis-cyclometallated iridium complex with engineered electronically excited states demonstrates reversible electronic energy transfer between adjacent chromophores giving rise to extremely long-lived red luminescence in solution ( τ = 480 μ s). Time-resolved spectroscopic studies a ﬀ orded determination of pertinent photophysical parameters including rates of energy transfer and energy distribution between constituent chromophores in the equilibrated excited molecule ( ca . 98% on the organic chromophores). Incorporation into a nanostructured metal – oxide matrix (AP200/19) gave highly sensitive O 2 sensing ﬁ lms, as the detection sensitivity was 200 – 300% higher than with the commonly used PtTFPP and approaches the sensitivity of the best O 2 -sensing dyes reported to date.


Introduction
From food packaging to electronics and from beer fermentation to blood analysis, monitoring concentrations of molecular oxygen is crucial in life and environmental sciences. 1-9 The use of optical sensors for determining oxygen concentration has gained particular interest in recent years [10][11][12][13][14][15][16][17][18][19] due to their advantageous characteristics: they are simple, fast, sensitive, accurate, non-invasive, non-destructive, and they are much simpler in terms of instrumental implementation and sample preparation than traditional electrochemical methods. 7 Additionally, optical sensors can be prepared in several formats ( planar, nanoparticles, paints, fibre-optic, etc.) and sizes, and they allow remote monitoring as well as wireless readout. 7,20 The oxygen-sensitive luminescent dye is a central component to the optical sensor. As the excited dye has to be quenched by the triplet ground state of molecular oxygen, phosphorescent complexes based on heavy metals such as Ir(III), Pt(II), Pd(II), Re(I) and Ru(II) complexes have been particularly studied, which correspond to kinetic and energetic prerequisites. 6,7,10,[21][22][23][24][25][26][27][28][29][30][31] The lifetime of the emissive excited state, τ 0 , and hence likelihood to interact with diffusing O 2 , is a key parameter influencing the performance of the sensor, which is evaluated by measuring the Stern-Volmer constant K SV = k q ·τ 0 with k q being the quenching rate constant. Consequently, the most commonly used highly sensitive dyes to date are Pd(II)/Pt(II) porphyrins such as Pt/PdTFPP (TFPP = mesotetra( pentafluorophenyl)porphin) that have excited triplet state lifetime >50 μs, possibly reaching milliseconds for Pd porphyrins. 21,22,32 Increasing the lifetime of the excited triplet state of dyes is therefore an appealing strategy to improve the performance of the oxygen sensor. [33][34][35][36][37] To this end, the triplet state of organic and metalloid-containing organic molecules has been successfully harnessed for highly sensitive O 2 -sensing systems due to lifetime of excited states easily reaching values >milliseconds. For example, C 70 (τ 0 > 20 ms) reached K SV ≈ 230 kPa −1 using ethyl cellulose as the matrix. 34 Aluminum complexes of 6-hydroxy-7H-benzo[de]anthracen-7-one (τ 0 = 350 ms) reach K SV ≈ 330 kPa −1 when highly O 2 -permeable Teflon AF is used as the supporting matrix. 36 However, in these cases, k q is low (k q (C 70 ) ≈ 11.5 Pa −1 s −1 and k q (Al complex) ≈ 0.94 Pa −1 s −1 ) compared to iridium complexes. For example, N969 (K SV1 (0-10% O 2 ) = 4.79 kPa −1 and τ 0 = 3.8 μs) 32 and EB146 (K SV1 (0-10% O 2 ) = 1.70 kPa −1 and τ 0 = 1.9 μs) 32 have k q values around 1000 Pa −1 s −1 . Such compounds are therefore very promising as highly sensitive O 2sensing dyes for ultra-low O 2 concentrations upon greatly extending the lifetime of the excited state (i.e. >100 μs).
Much effort has been devoted to increasing the excited state lifetime of emissive transition metal complexes. The strategies used rely upon modification of the ligands to modify the energetics of the 3 MLCT and 3 MC states [38][39][40][41] or to modify the geometry around the metal centre. 42 Another strategy is to design systems containing additional weakly interacting chromophores resulting in reversible electronic energy transfer (REET) between the core complex and the appended organic chromophore(s). [43][44][45][46][47] The latter strategy is very effective at significantly increasing the excited state lifetime without changing other parameters such as energetics (emission wavelength, redox potentials) and photoluminescence quantum yields and has been successfully applied to cyclometallated iridium complexes. [47][48][49] Surprisingly, REET has not been widely adopted for oxygen sensing beyond gaining insights into the photophysical processes occurring in the systems 50,51 or simply measuring the emission intensity ratio between air-equilibrated and degassed solutions. We have found only one report of a bichromophoric ruthenium-pyrene system exhibiting REET actually used for O 2 -sensing related application, a pressure sensitive paint. 52 Yet this system has very low sensitivity to oxygen, 0.11 kPa −1 for the main component, which is unsuitable for measuring low concentrations of oxygen.
In this work we demonstrate with a series of dyes, Ir0, Ir1 and Ir2, that REET can be used to improve the sensitivity of low concentration oxygen sensing dyes far beyond commonly used platinum(II) porphyrins and approaches the best sensitivities reported to date. The dyes are based on a bis-cyclometallated iridium core complex of the type Ir(C^N) 2 (bpy) + with C^N = 2-(naphthalen-1-yl)pyridine and bpy = 2,2′-bipyridine substituted with 0, 1 and 2 pyrene moieties (Scheme 1). It was anticipated that the two auxiliary pyrene groups would substantially increase the luminescence lifetime of Ir2 compared to that of complex Ir1 with a single appended pyrene (due to entropy-induced modification of the effective driving force for energy transfer, which favours location of excitation energy on organic chromophores in the former case exalting the energy reservoir effect), and hence greatly benefit the oxygen sensing performance, without compromising the emission profile and quantum yield of emission. 44

Synthesis and characterisation
The syntheses of Ir0 and Ir1 have been reported previously. 48 Notably, while Ir0 was obtained following standard procedures reacting the naphthylpyridine-based chloro-bridged dinuclear iridium complex directly with t Bubpy ( t Bubpy = 4,4'-di-tert-butyl-2,2′-bipyridine), Ir1 necessitated a synthetic strategy involving an intermediate bis-solvato iridium complex to be obtained pure. Ir2 was therefore synthesized directly with the latter strategy and obtained as an orange solid with a yield of 58%. The complex was characterized by 1 H NMR, high-resolution mass spectrometry (HRMS), and its purity assessed by elemental analysis. Slow diffusion of hexane into CH 2 Cl 2 solutions of Ir1 and Ir2 afforded single crystals suitable for X-ray diffraction analysis. The ORTEP diagrams of both iridium complexes and crystallographic parameters are given in ESI. †

Electrochemistry
Redox potentials were measured for the three complexes in acetonitrile (Table 1). Ir0 shows only quasi-reversible oxidation of the iridium ion and reduction of the bipyridine ancillary ligand as expected for such bis-heteroleptic iridium complexes. Upon addition of a pyrene group in Ir1, two irreversible signals appear at about 0.9 and −1.7 V. As the intensity of these signals increase upon addition of the second pyrene moiety in Ir2, they are attributed to the presence of the pyrene groups. 53 Nevertheless, the redox potentials of the core complex are not significantly impacted by the presence of the pyrene groups, demonstrating the absence of electronic coupling between the two chromophores, as a result of the saturated hydrocarbon spacer.

Electronic absorption spectroscopy
The photophysical properties of Ir0, Ir1 and Ir2 are summarised in Table 1. Electronic absorption spectra in acetonitrile at room temperature ( Fig. 1) are dominated by ligand-centred absorption bands in the UV region and similar MLCT absorption in the visible for all studied samples. Additional pyrene absorption bands are observed in Ir1 and Ir2 compared to Ir0, which retain their vibronic fine structure implying only weak ground-state coupling between the chromophores. As such, Scheme 1 Series of iridium complexes studied in this work. each chromophore is anticipated to retain its own specific properties in the ensemble.

Emission spectroscopy
Steady-state luminescence spectra show MLCT-based red emission for the complexes Ir0, Ir1 and Ir2 (λ em. max = 625 nm, Fig. 2), which have similar emission quantum yields in degassed acetonitrile (Φ degas = 0.1). However, higher oxygen sensitivity in dilute solution is observed for Ir2 compared to Ir1 and parent Ir0. While the latter molecule shows a Φ degas /Φ air ratio of 26, a much higher value of 190 is obtained with Ir1 and 950 with Ir2, which is consistent with much longer excited-state lifetimes on successive introduction of pyrene chromophores. Indeed, incorporation of two pyrene chromophores is clearly advantageous in terms of prolonged emission with respect to a single pyrene. Low temperature phosphorescence measurements (77 K) showed MLCT emission (λ em. max = 572 nm) with Ir0, while emission emanating from dyad Ir1 and triad Ir2 was located further towards the red spectral region (λ em. max = 595 nm) and is ascribed to pyrene emission (see ESI †). 55,56 A small energy gap value can be estimated at ∼700 cm −1 between high energy emission features (see ESI †), corresponding to the energy gap between lowest lying electronically excited states on adjacent chromophores assuming similar 3 MLCT levels in all complexes (at room temperature similar absorption and emission spectra are observed for the emitting moiety). Note: More accurate measurements of the energy gap can be obtained through time-resolved spectroscopies, vide infra.
A small energy gap value is conducive with reversible interchromophore electronic energy transfer taking place at room temperature. This would result in an excited state equilibrium being reached, which would consequently result in prolonged luminescence lifetime and oxygen sensitivity. 44

Time resolved spectroscopy
Time-resolved spectroscopic studies on the sub-ps to ms regimes give supplementary information on the nature of the excited molecule and a clear insight into excited-state processes in the supermolecule. An emission lifetime of 8.3 μs was obtained for Ir0 in degassed acetonitrile solution (Fig. 3), while a luminescence lifetime, which is over 25-times longer (225 ± 15 μs) was measured for a micromolar solution of Ir1, which doubled (480 ± 15 μs) for Ir2. With Ir2 proving more sensitive to O 2 and having a longer luminescence lifetime, the focus will be principally directed towards this molecule for the rest of the discussion. Additional information on Ir0 and Ir1 can be found in ref. 48.   Complementary transient absorption spectroscopy studies allow elucidation of the management of energy by the excited molecule prior to emission, and transient absorption signatures for both pyrene and iridium complexes have been described. 48 Excitation into the MLCT absorption band of Ir2 at 430 nm rapidly led to the population of the pyrene triplet, denoted by a characteristic T n ← T 1 absorption at 410 nm, see Fig. 4. The kinetics of deexcitation of this absorption band exactly parallels those of the emission, clearly showing that while energy is principally located on this triplet (see ESI †), it is quantitatively transferred to the metal centre where it is subsequently emitted. This observation implies the presence of quasi-isoenergetic excited states on the adjacent chromophores permitting rapid and reversible electronic energy transfer, leading to a dynamic excited-state equilibrium.
The establishment of a dynamic excited-state equilibrium from an initial non-equilibrated excited state can be observed in real time by transient absorption spectroscopy, and a 5 ns and 1.8 ns rise time of the pyrene triplet absorption signature was measured for Ir1 and Ir2, respectively. This value gives the rate of establishment of equilibrium (k eq = 2 × 10 8 s −1 and k eq = 5.6 × 10 8 s −1 ), and is equal to the sum of forward (k f ) and back (k b ) energy transfer processes. 44 The increased rate of attaining equilibrium Ir2 compared to Ir1 can be ascribed to the presence of two degenerate pyrene triplet manifolds in the former case and one in the latter complex. Due to an overlap between Ir-centre ground state bleaching signal with the transient MLCT absorption signal, the relative k f and k b values cannot be determined by direct observation of transient absorption signatures.

Oxygen sensing
The three Ir(III) dyes were immobilised in polystyrene (PS) and in a porous aluminium oxide solid support (AP200/19) 23,24,58 chosen for its excellent performance as supporting matrix for dyes for sensing traces of oxygen in gas. 32,59 They have been characterized by intensity measurements and the Ir(III) dye showing the best analytical performances (Ir2) was also characterised by multifrequency phase-resolved luminescence using a short duty cycle rectangular-wave as the excitation signal. 60 The variations of the luminescence intensity with the oxygen concentration as well as the Stern-Volmer plots for Ir2-AP200/19 and Ir2-PS are shown in Fig. 5, which also demonstrate the reversibility of the sensing process. The fitting parameters for all five studied sensing films are reported in Table 2. Ir0 immobilised in AP200/19 provided very irreproducible sensing films and therefore was not studied further.
The increasing number of pyrene moieties on the complex clearly improves the oxygen sensitivity; the sensitivity of Ir2-PS is 1.5 and 5 times higher than the sensitivities of Ir1-PS and Ir0-PS, respectively, following the increase of excited state lifetime. The same effect is observed with the AP200/19 matrix, as the sensitivity of Ir2-AP200/19 is 2.4 times higher than with Ir1-AP200/19. The latter also highlights the effectiveness of the nanostructured support AP200/19 to increase the oxygen sensi-  tivity, as previously observed for other oxygen sensitive dyes. [23][24][25]32 Based on intensity measurements, the most sensitive sensing film is Ir2-AP200/19 (K SV of 26.96 kPa −1 ). Comparing these results with classical Pt(II), Ru(II), and Ir(III) sensing films using the same nanostructured matrix, it is clear that this sensing film is very promising for O 2 -sensing applications, showing a higher Stern-Volmer constant than one of the most   32 In order to determine if Ir2 can be used for analysing low oxygen concentration, lifetime measurements were carried out in the frequency domain ( phase-resolved method) 32,59,60 because intensity measurements are not sensitive enough to characterize the sensing films in the ranges 0-1 and 0-0.21 kPa pO 2 . The description of the methodology as well as the protocol for carrying out the measurements are described by Medina-Rodríguez et al. 59,60 The modulation frequencies used for the measurement in these ranges (see ESI; Fig. S11 †) were 1130 Hz for the Ir2-PS sensing film (average phase difference of 51.90 degrees) and 2260 Hz for the Ir2-AP200/19 sensing film (average phase difference of 32.61 degrees). The variation of the apparent luminescence lifetime of Ir2 incorporated into PS and into AP200/19 when exposed to different oxygen concentrations between 0-1 kPa pO 2 are shown in Fig. 6 and the results are summarized in Table 2, demonstrating the high sensitivity of the film (K SV1 of Ir2-AP200/19 is 94.81 kPa −1 and Δτ 0.05 of Ir2-AP200/19 = 68.51 ± 0.21%); the Δτ 0.05 of PtTFPP-AP200/19 measured in the same conditions is 62.53 ± 3.66%. 32 Finally, the measurement capability of Ir2-PS and Ir2-AP200/  19 was also determined by comparing the measured oxygen  concentrations with the real O 2 concentrations (see ESI,  Table S2 †), which shows the reliability of the sensing films.

Conclusions
In summary, the benefit of reversible intramolecular energy transfer as a strategy to improve the oxygen sensitivity of phosphorescent iridium complexes by dramatically increasing the excited state lifetime is shown herein using a series of complexes, Ir0, Ir1 and Ir2, bearing 0, 1 and 2 pyrene moieties, respectively. Ir2 presents an excited state lifetime in solution of 480 μs, almost twice the excited state lifetime of Ir1 containing only one pyrene group. Incorporation into AP200/19, a nanostructured inorganic matrix, Ir2 leads to an oxygen sensing performance (K SV = 26.96 kPa −1 ) similar to the best reported film using the same matrix, PtTFPP-AP200/19 with K sv = 25.68 kPa −1 . This is the first time that a non-porphyrinbased transition metal complex reaches such high performance. Furthermore, using a lifetime measurement method for very low oxygen concentration 0-1 kPa, Ir2-AP200/19 provides excellent performance with Ksv = 94.81 kPa −1 to be compared to 31.02 kPa −1 for PtTFPP in the same conditions. 32 Such performance is only bested by some organic and metalloidcontaining organic dyes that, however, have low k q values (k q < 10 Pa s −1 ) compared to the system reported here (k q up to ≈ 2000 Pa s −1 ). These results clearly demonstrate the potential of reversible intramolecular energy transfer as a strategy to develop high performance oxygen sensing dyes.
A key advantage of this strategy is that it provides a rational approach to further enhancing the performance of the dyes: simply increasing the number of energy reservoir units will increase the excited state lifetime linearly, as previously demonstrated with ruthenium dyes, 61 which will result in an increase of sensitivity. Equally, reversible energy transfer could a priori be instilled in Pd/Pt porphyrins in an analogous fashion. Clearly ultrasensitive oxygen sensing dyes are now accessible.