Heteroleptic Re(CO)₂⁺ and Re(CO)₃⁺ complexes with α-diimines: similarities and differences in their luminescence properties

Abstract

The photophysical properties of two series of phosphorescent rhenium(I) complexes, [Re(CO)₂(N^N)(tpp)₂]⁺ and [Re(CO)₃(N^N)(tpp)]⁺ with carbon monoxide (CO), triphenylphosphine (tpp) and α-diimine (N^N) ligands have been investigated in deoxygenated acetonitrile solution at room temperature and in solid methanol/ethanol 1 : 1 matrices at 77 K. The complexes display moderate to strong phosphorescence which is related to the N^N ligand modulated metal-to-ligand charge-transfer S₀ ← ³*MLCT or intraligand S₀ ← ³*LC transitions. Luminescence properties of the investigated series have been found to be very similar but some intrinsic differences between them are clearly seen. Whereas the [Re(CO)₂(N^N)(tpp)₂]⁺ series shows MLCT emission in both temperature regimes studied, the [Re(CO)₃(N^N)(tpp)]⁺ series exhibits intrinsic changes in its emission character when the measurement temperature is lowered from 298 to 77 K. In both investigated series, their emission characteristics are strongly affected by the nature of coordinated α-diimine N^N ligands. The observed trends, changes in the radiative k_r and non-radiative k_nr deactivation rate constants, have been compared with those found for the previously investigated [Re(CO)₃(N^N)(Cl)], [Re(CO)₃(N^N)(CH₃CN)]⁺, and [Re(CO)₂(N^N)(dppv)]⁺ series (dppv = cis-1,2-bis(diphenylphosphino)-ethene). Similarities and differences in the spectroscopic and photophysical properties of five series of the Re(CO)₃⁺ and Re(CO)₂⁺ complexes have been analyzed in the view of results from DFT and TD-DFT computation and the emission band-shape analyses performed according to the Marcus–Jortner formalism.

---

Introduction

Since the pioneering work of Wrighton and Morse on the luminescent [Re(CO)₃(1,10-phenathroline)(Cl)] molecule,¹ rhenium(I) complexes have occupied a prominent position in organometallic luminophores with a d⁶ central metal ion.^2–4 Among these, the [Re(CO)₃(N^N)(L)]^0/+ chelates have attracted special attention due to their rich excited state behaviour that can be widely tuned by modification of the main N^N or ancillary L ligands and the medium or temperature.^5–12

The photophysical properties of [Re(CO)₃(N^N)(L)]^0/+ species are governed by the relative energetic position and the interplay of the closely lying excited states of different characters. Particularly, the energy gaps between the excited ³*LC (ligand centred) and ³*MLCT (metal-to-ligand charge-transfer) triplet states are relatively small and allow efficient electronic interaction between them. The excited states of the fac-Re(CO)₃⁺ complexes are mixed and their “real” excited states can be regarded as a superposition of the initial “pure” ³*MLCT and ³*LC components. This leads to monotonic changes in the nature of these emitters from the excited ³*MLCT to ³*LC character when appropriate changes of N^N and/or L ligands are involved. Consequently, the spectroscopic and photophysical properties of [Re(CO)₃(N^N)(L)]^0/+ complexes may be quite different, even for pretty similar N^N and/or L ligands present in their structures. This is because energies of the “pure” excited ³*LC and ³*MLCT states are affected in different ways.

The above-described behaviour is generally characteristic for many other transition metal complexes consisting d⁶ ions and N^N ligands^13,14 including much less elaborated cis-Re(CO)₂⁺ complexes^15–20 as well. The latter, containing monodentate PR₃ or bidentate P^P phosphines as ancillary ligands in their [Re(CO)₂(N^N)(PR₃)₂]⁺ or [Re(CO)₂(N^N)(P^P)]⁺ structures, are emissive in the spectral range consistent with the emission range observed for analogous [Re(CO)₃(N^N)(Cl)] complexes. Their emissive properties, however, are noticeably better (higher quantum yields ϕ_em and longer lifetimes τ_em of emission) as compared to those found for their [Re(CO)₃(N^N)(Cl)] analogues. Thus, the Re(CO)₂⁺ based luminophores can be considered as very promising alternative in a wide range of applications in which the Re(CO)₃⁺ chelates have been already applied. Among them, the most noticeable examples include photocatalysis,^21–23 luminescent sensors,^24–26 organic light emitting devices²⁷ or dye-sensitized solar cell.²⁸

Luminescence properties of the [Re(CO)₂(N^N)(P^P)]⁺ series²⁰ bearing cis-1,2-bis(diphenylphosphino)-ethene – dppv and different α-diimines as P^P and N^N ligands, have been recently studied in more detail. It has been found that the radiative as well as nonradiative deactivation processes of the excited ³*[Re(CO)₂(N^N)(dppv)]⁺ states are considerably suppressed as compared to the ³*[Re(CO)₃(N^N)(Cl)] ones.¹² This results in longer τ_em values and, due to still more pronounced suppression of the nonradiative deactivation, in definitely higher quantum yields ϕ_em. The observed changes in the τ_em and ϕ_em values were found to be connected with lowering of the rate constants describing the radiative k_r = ϕ_em/τ_em and nonradiative k_nr = (1 − ϕ_em)/τ_em deactivation of the excited ³*MLCT states. This can be attributed to the presence of P^P ligand or/and smaller number of CO group (two instead of three) in the previously investigated [Re(CO)₂(N^N)(dppv)]⁺ complexes. To clarify the occurring issue we have decided to perform a more systematic, comparative study of luminescent Re(CO)₃⁺ and Re(CO)₂⁺ complexes bearing different α-diimine N^N (cf. Fig. 1) and triphenylphosphine – tpp ligands. For the studies reported here, two series of Re(I) chelates, [Re(CO)₃(N^N)(tpp)]⁺ and [Re(CO)₂(N^N)(tpp)₂]⁺ were selected because their emission spectral ranges are expected matching these characteristic for their [Re(CO)₂(N^N)(dppv)]⁺, [Re(CO)₃(N^N)(Cl)] or [Re(CO)₃(N^N)(CH₃CN)]⁺ analogues.^12,16,20

Fig. 1 Structures of N^N ligands investigated and their acronyms used in the text. 2,2′-Bipyridine – bpy, and 4,4′-di-tert-butyl-2,2′-bipyridyne – dtbbpy, 1,10-phenanthroline – phen, 2,9-dimethyl-1,10-phenanthroline – 29dmphen, 4,7-dimethyl-1,10-phenanthroline – 47dmphen, 3,4,7,8-tetramethyl-1,10-phenanthroline – tmphen, 5,6-dimethyl-1,10-phenanthroline – 56dmphen, and 4,7-diphenyl-1,10-phenathroline – dpphen, respectively.

Investigations reported in this paper are also devoted to the relationships between the nature of the given MLCT emitter and the k_r or k_nr rate constants characterizing the radiative and non-radiative S₀ ← ³*MLCT transitions in the α-diimine complexes. The experimentally observed huge variety in the k_r and k_nr values^{16,20,29–31} are explainable by the N^N ligand induced changes in the electronic structure of these emitters from the ³*MLCT to more pronounced ³*LC character reflecting different mixing between the “pure” excited LC and MLCT configurations. Within this approximation, one can discuss the anticipated mixing taking into account the states energetically closest. In the simplest case, one can assume that the triplet ³*MLCT configuration interacts with the lowest excited triplet state ³*LC, typically localized within the N^N ligand. Due to the mixing between the “pure” ³*MLCT and ³*LC states, one can describe the resulting “real” emissive state as their superposition with the mixing coefficients c_MLCT and c_LCT

where the V₃₃ and E_LCT − E₀₀ are the electronic coupling element responsible for the interactions and the energy differences between the 0–0 transitions, the E_LCT and E₀₀ values characterizing phosphorescence of the isolated N^N ligand and the complex, respectively. The latter quantity is affordable from the emission band-shape analysis.³¹

Despite of all its approximation, the LC/MLCT mixing approach seems to be applicable in any more quantitative discussion of the luminescent ³*MLCT states including interpretation of the N^N ligand induced changes of the k_r and k_nr values. In the presented work, this has been tested for the [Re(CO)₂(N^N)(tpp)₂]⁺ and [Re(CO)₃(N^N)(tpp)]⁺ complexes. The obtained results have been compared with those previously obtained for the [Re(CO)₃(N^N)(Cl)], [Re(CO)₃(N^N)(CH₃CN)]⁺, and [Re(CO)₂(N^N)(dppv)]⁺ complexes. Spectroscopic and photophysical properties of five series of the Re(CO)₃⁺ and Re(CO)₂⁺ complexes are discussed in the view of their emission band-shape analysis. This work presents also the results from DT and TD-DFT computation performed to clarify the observed similarities and differences.

Results and discussion

S₀ → ¹*MLCT absorption and S₀ ← ³*MLCT/³LC emissions

Fig. 2 and 3 present examples of the room temperature absorption spectra of the studied complexes recorded in acetonitrile solutions. The spectra show superposition of the overlapping bands as characteristic for the transition metal complexes exhibiting typical MLCT features. Whereas, the high-energy bands can be ascribed to the π → π* transitions localized within ligands attached to the central Re(I) core, the lowest energy bands with relatively low intensities can be attributed to the MLCT transitions. Generally, the spectra are very similar to those characterizing other Re(I) complexes with chelating α-diimine ligands. More specifically, the UV-vis spectra characterizing the studied [Re(CO)₂(N^N)(tpp)₂]⁺ chelates correspond well to their [Re(CO)₂(N^N)(dppv)]⁺ analogues,²⁰ whereas the [Re(CO)₃(N^N)(tpp)]⁺ complexes exhibit features closer to that found for their [Re(CO)₃(N^N)(CH₃CN)]⁺ counterparts.¹² This is reasonable because the spectrochemical parameters of tpp and CH₃CN ligands are very close one to another.³³ The same explanation holds for tpp and dppv ligands that explains resemblances between [Re(CO)₃(N^N)(tpp)]⁺ and [Re(CO)₂(N^N)(dppv)]⁺ series. Additionally, the positions of the CO and phosphine ligands in the spectrochemical series explain the observed bathochromic shift of the MLCT absorption and emission bands between the Re(CO)₃⁺ and Re(CO)₂⁺ complexes reported in this work (cf. data in Table 1).

Fig. 2 Band profiles of the UV-vis absorption and emission spectra recorded for the [Re(CO)₂(47dmphen)(tpp)₂]⁺ (top) and [Re(CO)₃(47dmphen)(tpp)]⁺ (bottom) complexes. Room temperature absorption (black lines) in CH₃CN solutions. Room temperature (red lines) and 77 K (blue lines) emission in CH₃CN solutions and 1 : 1 CH₃OH/C₂H₅OH matrices, respectively. Dashed lines present expanded the low energy part of the UV-vis absorption.

Fig. 3 Band profiles of the UV-vis absorption and emission spectra recorded for the [Re(CO)₂(bpy)(tpp)₂]⁺ (top) and [Re(CO)₃(bpy)(tpp)]⁺ (bottom) complexes. Room temperature absorption (black lines) in CH₃CN solutions. Room temperature (red lines) and 77 K (blue lines) emission in CH₃CN solutions and 1 : 1 CH₃OH/C₂H₅OH matrices, respectively. Dashed lines present expanded the low energy part of the UV-vis absorption.

Table 1 Spectroscopic and photophysical properties of [Re(CO)₂(N^N)(tpp)₂]⁺ and [Re(CO)₃(N^N)(tpp)]⁺ complexes (data in CH₃CN solutions at room temperature and methanol/ethanol 1 : 1 glasses at 77 K). Absorption maxima [small nu, Greek, tilde]^max_abs and molar extinction coefficients ε_M of MLCT bands. Emission maxima [small nu, Greek, tilde]^max_em, emission quantum yields ϕ_em, and emission lifetimes τ_em of S₀ ← T₁ transitions

Ligand N^N	Absorption at 298 K		Emission at 298 K			Emission at 77 K
	[small nu, Greek, tilde]^maxabs/cm^−1	εM/M^−1 cm^−1	[small nu, Greek, tilde]^maxem/cm^−1	ϕem	τem /μs	[small nu, Greek, tilde]^maxem/cm^−1	τem/μs
a For [Re(CO)3(phen)(tpp)]^+ and [Re(CO)3(dpphen)(tpp)]^+ complexes bi-exponential emission decays were observed. Values given in parentheses are the normalized amplitudes of the short-lived and long-lived components of emission decays.
[Re(CO)2(tmphen)(tpp)2]^+	27 050	5.6 × 10^3	17 100	0.52	21.6	18 450	38
[Re(CO)2(47dmphen)(tpp)2]^+	26 300	5.0 × 10^3	17 000	0.33	10.4	18 050	31
[Re(CO)2(29dmphen)(tpp)2]^+	25 000	2.7 × 10^3	15 950	0.045	0.43	17 600	7.5
[Re(CO)2(phen)(tpp)2]^+	27 050	3.8 × 10^3	16 150	0.20	4.6	17 800	22
[Re(CO)2(56dmphen)(tpp)2]^+	24 250	3.9 × 10^3	16 200	0.19	3.7	18 050	18
[Re(CO)2(dpphen)(tpp)2]^+	23 250	7.1 × 10^3	15 700	0.24	6.8	16 950	25
[Re(CO)2(dtbbpy)(tpp)2]^+	24 950	3.5 × 10^3	16 100	0.080	0.95	18 200	15
[Re(CO)2(bpy)(tpp)2]^+	24 250	2.8 × 10^3	15 950	0.042	0.63	17 950	15
[Re(CO)3(tmphen)(tpp)]^+	27 050	2.8 × 10^3	20 550, 19 350	0.070	19.9	21 200, 19 750, 18 400, 16 950	480
[Re(CO)3(47dmphen)(tpp)]^+	27 050	4.7 × 10^3	19 300	0.083	9.8	21 200, 19 850, 18 450, 17 100	430
[Re(CO)3(29dmphen)(tpp)]^+	26 600	2.4 × 10^3	19 100	0.071	2.6	21 450, 20 250	41
[Re(CO)3(phen)(tpp)]^+	27 050	3.3 × 10^3	18 800	0.088	2.3	21 600, 20 150, 18 800	42 (0.63), 156 (0.37)^a
[Re(CO)3(56dmphen)(tpp)]^+	25 850	3.1 × 10^3	19 800, 18 800	0.12	56	20 550, 19 200, 17 800	890
[Re(CO)3(dpphen)(tpp)]^+	26 050	6.0 × 10^3	17 650	0.36	45	19 750, 18 550	70 (0.70), 160 (0.30)^a
[Re(CO)3(dtbbpy)(tpp)]^+	29 050	4.0 × 10^3	18 800	0.040	0.17	22 400, 21 150, 19 990	13
[Re(CO)3(bpy)(tpp)]^+	28 750	3.5 × 10^3	18 400	0.040	0.29	22 200, 20 700, 19 500	8.6

---

Both series compared above are similar one to another in quantitative way as well. Although, due to the presence of the overlapping of many intra-ligand bands the high-energy parts of the recorded spectra are barely informative, one can draw some decisive conclusions comparing MLCT regions. Particularly, the overall MLCT intensities are very similar for the given N^N ligand and nearly independent of the remaining ligands attached to the central Re(I) ion (cf. Fig. 4). Fig. 4 presents data for the [Re(CO)₂(N^N)(tpp)₂]⁺ and [Re(CO)₃(N^N)(tpp)]⁺ series, but nearly the same relationship between the molar extinction coefficients ε_M is characteristic for [Re(CO)₃(N^N)(Cl)] and [Re(CO)₂(N^N)(dppv)]⁺ pair. Although the above examples concern complexes with sufficient separation of the MLCT and the intra-ligand absorption bands, similar situation seems to take place also for these α-diimine Re(I) complexes, [Re(CO)₃(N^N)(CH₃CN)]⁺, where their MLCT bands are partly obscured by the intra-ligand absorption. Thus, one can emphasize very similar intensities of the MLCT absorption band as the characteristic feature of the Re(CO)₃⁺ as well as Re(CO)₂⁺ complexes.

Fig. 4 Relation between the values of the molar extinction coefficients characterizing the MLCT band in the UV-vis absorption spectra of the [Re(CO)₃(N^N)(tpp)]⁺) and [Re(CO)₂(N^N)(tpp)₂]⁺ complexes in CH₃CN solutions. Experimental points are labelled with the N^N ligand abbreviations. The slope and intercept of the dashed line are equal to 1 and 0, respectively.

Although, due to overlapping MLCT and LC bands, a quantitative characterization of the MLCT bands intensities is rather difficult, some estimates are possible taking into account the spectral positions [small nu, Greek, tilde]^max_abs and spectral width Δ[small nu, Greek, tilde]_1/2 of these bands. Using so-called the “lazy man's” approximation^34,35 one can approximately evaluate “effective” values of the transition dipole moments M_abs describing these bands from the respective extinction coefficient ε_M data. In nice accordance with the observed similar MLCT bands intensities, a quite small variation of the M_abs values is characteristic for the Re(I) complexes with α-diimine ligands. In most cases, the obtained M_abs values fall into the range of 1.5–2.5 D, somewhat larger M_abs ∼ 3.0 D are characteristic only for the complexes with dpphen ligand. This finding implies nearly constant sum of oscillator strengths f of all S₀ → MLCT transitions potentially contributing to the MLCT bands of the discussed Re(I) complexes. One can rationalize the ∑f = const claim assuming that the sum of the oscillator strength of the individual d → *π transitions (d_xy → *π, d_xz → *π, etc.) is constant as well. Then, independently how the given d orbital(s) will contribute to the occupied molecular orbitals involved in the transitions consisting the MLCT band (HOMO, HOMO−1, etc.), one can expect very similar values of its overall intensity. This expectation is also valid when the configuration interactions would be required for the proper description of the individual MLCT transitions. In the case of discussed Re(I) complexes it seems to be obvious that their UV-vis spectroscopic properties are connected with the presence of low energetically lying ¹*MLCT states. Results from the DFT and TD-DFT computations, performed for the ground state optimized geometries (vide infra), confirm this anticipation without any doubts.

All of the Re(I) complexes under study are luminescent both at room temperature and at 77 K. The luminescence spectra recorded for deaerated solutions in CH₃CN at 298 K are typically broad and, in most cases, show no vibronic structure. The spectral positions of the emission bands depend on the nature of the N^N ligand present in the structures of the studied [Re(CO)₂(N^N)(tpp)₂]⁺ as well as [Re(CO)₃(N^N)(tpp)]⁺ series (cf. data in Table 1). The observed trends in the emission maxima follow that expected for the S₀ ← ³*MLCT transitions taking into accounts changes in the electron withdrawing properties of the coordinated N^N ligand. These changes, caused by the presence of the methyl (electron donor) or the phenyl (electron acceptor) substituents attached to the parent phen or bpy ligands, lead to hypsochromic or bathochromic shifts of the emissions, respectively.

Intrinsic differences are, however, characteristic for the emission recorded at 77 K in the CH₃OH/C₂H₅OH matrices. The investigated [Re(CO)₂(N^N)(tpp)₂]⁺ complexes exhibit broad and structureless bands. The observed hypsochromic shift allows concluding destabilization of the emissive ³*MLCT states caused by hindered solvent/solute relaxation caused by the extreme viscosity of the low temperature glasses.^36,37 The studied [Re(CO)₃(N^N)(tpp)]⁺ complexes demonstrate, however, distinctly different 77 K behaviour exhibiting nicely structured emission bands. Their positions and shapes resemble emissions from the isolated N^N ligands.^38–40 Thus, the observed rigidochromism can be related to the temperature induced changes in the nature of the emissive ³*[Re(CO)₃(N^N)(tpp)]⁺ species, from the excited ³*MLCT to the excited intra-ligand ³*LC.

Similarly, as it can be seen in the UV-vis absorption spectra, the emissive properties of the [Re(CO)₂(N^N)(tpp)₂]⁺ and [Re(CO)₃(N^N)(tpp)]⁺ complexes depend on the nature of the N^N ligand present in their structure. This is particularly true for the kinetic parameters associated with the radiative and non-radiative deactivations of the excited ³*[Re(CO)₂(N^N)(tpp)₂]⁺ and ³*[Re(CO)₃(N^N)(tpp)]⁺ species (cf. data in Table 2). The rate constants for these processes, determined from k_r = ϕ_em/τ_em and k_nr = (1 − ϕ_em)/τ_em relationships, exhibit large diversity, much larger as compared to the variations of the ε_M values. This is also true for the transition dipole moments M_em of the S₀ ← ³*MLCT emissions as determined using the following relationship

k_r = (16π³/3hε_o)(n[small nu, Greek, tilde]^max_em)³|M_em|² (2)

where n and ε_o denote the solvent refractive index and the vacuum permittivity. When the M_abs values depend marginally on the N^N ligand, the M_em values differ by almost one order of magnitude.

Table 2 Kinetic and energetic parameters characterizing S₀ ← ³*[Re(CO)₂(N^N)(tpp)₂]⁺ and S₀ ← ³*[Re(CO)₃(N^N)(tpp)]⁺ emissions (data in CH₃CN solutions at room temperature). Rate constants of nonradiative k_nr and radiative k_r deactivation processes, transition dipole moments M_em of S₀ ← T₁ transitions, and electronic coupling elements V₃₀. Fitted values of 0–0 transitions energies E₀₀, reorganization energies λ_LM and λ_H, and vibrational quanta hν_H

Ligand N^N	Kinetic parameters				Energetic parameters (from band-shape analysis)
	knr/s^−1	kr/s^−1	Mem/D	V30/eV	E00/eV	λLM/eV	λH/eV	hνH/eV	knr/s^−1 (calc)^a
a knr values calculated using E00, λLM, λH, and hνH parameters from performed emission band-shape analyses and V30 values as estimated using the Hush–Mulliken relationship V30 = hc[small nu, Greek, tilde]^maxemMem/Δμ.
[Re(CO)2(tmphen)(tpp)2]^+	2.2 × 10^4	2.4 × 10^4	0.08	0.011	2.23	0.28	0.22	0.15	3.1 × 10^3
[Re(CO)2(47dmphen)(tpp)2]^+	6.4 × 10^4	3.2 × 10^4	0.09	0.013	2.20	0.31	0.20	0.17	7.1 × 10^4
[Re(CO)2(29dmphen)(tpp)2]^+	2.2 × 10^6	1.0 × 10^5	0.18	0.024	2.04	0.38	0.18	0.20	2.7 × 10^6
[Re(CO)2(phen)(tpp)2]^+	1.7 × 10^5	4.3 × 10^4	0.12	0.015	2.04	0.40	0.16	0.19	2.2 × 10^5
[Re(CO)2(56dmphen)(tpp)2]^+	2.2 × 10^5	5.1 × 10^4	0.13	0.017	2.07	0.37	0.17	0.18	1.3 × 10^5
[Re(CO)2(dpphen)(tpp)2]^+	1.1 × 10^5	3.5 × 10^4	0.11	0.014	2.00	0.35	0.18	0.17	1.6 × 10^5
[Re(CO)2(dtbbpy)(tpp)2]^+	9.7 × 10^5	8.4 × 10^4	0.16	0.022	2.08	0.41	0.20	0.18	1.0 × 10^6
[Re(CO)2(bpy)(tpp)2]^+	1.5 × 10^6	6.7 × 10^4	0.15	0.019	2.02	0.43	0.18	0.19	1.1 × 10^6
[Re(CO)3(tmphen)(tpp)]^+	4.7 × 10^4	3.5 × 10^3	0.03	0.004	2.55	0.11	0.28	0.17	3.6 × 10^2
[Re(CO)3(47dmphen)(tpp)]^+	9.4 × 10^4	8.5 × 10^3	0.04	0.006	2.53	0.12	0.30	0.16	9.5 × 10^2
[Re(CO)3(29dmphen)(tpp)]^+	3.6 × 10^5	2.7 × 10^4	0.07	0.011	2.41	0.38	0.20	0.22	7.1 × 10^4
[Re(CO)3(phen)(tpp)]^+	4.0 × 10^5	3.8 × 10^4	0.09	0.014	2.37	0.38	0.20	0.23	3.3 × 10^5
[Re(CO)3(56dmphen)(tpp)]^+	1.6 × 10^4	2.1 × 10^3	0.02	0.003	2.47	0.12	0.25	0.17	1.6 × 10^2
[Re(CO)3(dpphen)(tpp)]^+	1.4 × 10^4	8.0 × 10^3	0.04	0.006	2.33	0.19	0.28	0.17	1.9 × 10^4
[Re(CO)3(dtbbpy)(tpp)]^+	5.6 × 10^6	2.4 × 10^5	0.22	0.036	2.39	0.42	0.24	0.22	5.3 × 10^6
[Re(CO)3(bpy)(tpp)]^+	3.3 × 10^6	1.4 × 10^5	0.17	0.025	2.35	0.43	0.23	0.22	2.8 × 10^6

---

Analysing more deeply the room temperature data for the [Re(CO)₂(N^N)(tpp)₂]⁺ and [Re(CO)₃(N^N)(tpp)]⁺ complexes, one can see intrinsic correlation between the determined k_r and k_nr rate constants. Monotonic relationships between these rate constants (cf. Fig. 5) are characteristic for the excited triplet ³*[Re(CO)₂(N^N)(tpp)₂]⁺ and ³*[Re(CO)₃(N^N)(tpp)]⁺ states. Very likely, such behaviour, reported previously for other Re(I) complexes,^12,20 seems to be a general rule for the α-diimine chelates of Re(CO)₃⁺ and Re(CO)₂⁺ ions. Although the experimental points are somewhat scattered, the found coincidences between k_nr and k_r values allow concluding that all three discussed series of the Re(CO)₃⁺ complexes (with tpp, CH₃CN, and Cl⁻ ancillary ligands) are very similar to each other. In an analogous way, one can also postulate close analogy between two [Re(CO)₂(N^N)(tpp)₂]⁺ and [Re(CO)₂(N^N)(dppv)]⁺ series.

Fig. 5 Relations between the values of k_r and k_nr rate constants for the Re(CO)₂⁺ (top) and Re(CO)₃⁺ (bottom) complexes. Data for [Re(CO)₂(N^N)(tpp)₂]⁺ (yellow symbols), [Re(CO)₂(N^N)(dppv)]⁺ (grey symbols),²⁰, [Re(CO)₃(N^N)(tpp)]⁺ (cyan symbols), [Re(CO)₃(N^N)(CH₃CN)]⁺ (green symbols),¹² and [Re(CO)₃(N^N)(Cl)] (red symbols)¹² complexes in CH₃CN solutions at room temperature.

Band-shape analysis of the S₀ ← ³*MLCT emission spectra

The observed k_r vs. k_nr relations point to noticeable association between the radiative and non-radiative S₀ ← ³*MLCT processes as specific for the discussed Re(I) complexes. This finding remains consistent with that expected from close relation between the thermal and optical charge-transfers occurring in so-called inverted Marcus region where these processes are bound one to another. Pursuing the close analogy between them, one can relate their kinetic description to the same set of the energetic parameters associated with the radiative and nonradiative S₀ ← ³*MLCT transition.

For more quantitative description of these processes one can use a commonly accepted formalism introduced by Marcus^41–43 and developed further by many other authors.^44,45 A moderately simple approach, based on the separation of low frequency λ_L, medium frequency λ_M and high frequency λ_H reorganization energies allows the description of the charge-transfer emission profile, i.e., the emission intensity I([small nu, Greek, tilde]_em) vs. the emitted photon energy hc[small nu, Greek, tilde]_em. In the case of the transition metal complexes exhibiting the MLCT emission, the λ_M and λ_H energies are mainly connected with changes of the intra-ligand and the ligand–metal bonds, whereas the λ_L energy is mostly associated with the solvent shell reorganization. When the semi-classical treatment of the medium-frequency modes together with the classical and quantum treatment of the low frequency and the high frequency modes are applied,^32,46 the following expression can be obtained

where E₀₀, hν_H, k_B and T are the energy of the 0–0 transitions, the average spacing of the quantized high frequency intra-molecular modes undergoing reorganization upon charge-transfer, the Boltzmann constant, and absolute temperature, respectively. Parameter S in eqn (3) corresponds to the electron-vibrational coupling constant defined as S = λ_H/hν_H.

Within the same framework, the values of k_nr rate constants are predictable from the following expression^29,47,48

where V₃₀ is related to an effective electronic coupling matrix element describing electronic interactions between the S₀ and ³*MLCT states participating in the non-radiative S₀ ← ³*MLCT thermal charge-transfer.

The λ_LM term present in eqn (3) and (4), contain contributions from the low frequency (treated classically) and medium frequency (treated semi-classically) reorganization energies. Within these assumptions, one can approximate the resulting effective λ_LM value as follows^32,48

λ_LM = λ_L + λ_M(hν_M/2k_BT)coth(hν_M/2k_BT) (5)

where hν_M corresponds to the average spacing of the quantized medium frequency intra-molecular vibrations participating in the S₀ ← ³*MLCT transitions.

Representative examples of the numerical fits, presented in Fig. 6, show that, despite all approximations of the single frequency model applied, one can reproduce quite well the experimental emission profiles of the studied complexes. Emission spectra of the complexes under investigations were fitted by the application of a one-mode Franck–Condon analysis according to eqn (3) with the quantities relevant for their radiative charge-transfer (i.e., E₀₀, λ_LM, λ_H, and hν_H summarized in Table 2) varied as free fit parameters. It should be noted, however, that the fitted quantities turn out to be somewhat correlated. This leads to numerical uncertainty (±0.02 eV) of their fitted values. Due to approximate character of the applied model, the real uncertainty of the fitted parameters can be slightly larger. Despite that, the obtained energetic quantities give somewhat deeper insight into the nature of the emissive ³*MLCT species.

Fig. 6 Band profiles of the emission spectra recorded for the [Re(CO)₂(tmphen)(tpp)₂]⁺ (yellow line), [Re(CO)₂(dtbbpy)(tpp)₂]⁺ (orange line), [Re(CO)₃(tmphen)(tpp)]⁺ (cyan lines), and [Re(CO)₃(dtbbpy)(tpp)₂]⁺ (green line), complexes in CH₃CN solutions at room temperature. Dashed black lines correspond to the numerical fits according to eqn (5).

In the fitting procedure as used in this work, the applied classical, semi-classical and quantum-mechanical treatment of the low, medium and high frequency modes seems to be justified in the cases of 298 K emissions. One can additionally check this anticipation because the same set of parameters is, according to eqn (4), related to the rate constants of the non-radiative S₀ ← ³*MLCT transitions. Thus, one can additionally test the data from band-shape analyses by comparing the experimentally found rate constants k_nr with those calculated using eqn (4). Such calculations are possible because one can deduce the required V₃₀ values from the experimentally available values of the MLCT emission maxima [small nu, Greek, tilde]^max_em and the transition dipole moments M_em. Without going into details of the interaction between the coupled S₀ and ³*MLCT states, one can estimate the V₃₀ values using the Mulliken–Hush relationship^49–52

In the case of the discussed Re(I) complexes, one can approximate required Δμ value assuming the whole electron transfer over the distance between the central Re(I) ion and the centre of N^N ligands. Within this approach, one can obtain Δμ = 15 D. Similarly, as it was done in our previous works, this quite reasonable value^53,54 was used for estimation of the V₃₀ terms. The obtained V₃₀ values together with the resulted k_nr rate constants are collected in Table 2. The agreement between the experimentally found and calculated rate constants k_nr is more than satisfactory for nearly all Re(I) complexes discussed in this work (cf. Fig. 7). In these cases, the discrepancies between the experimentally found and computed values do not exceed a factor of 3–4.

Fig. 7 Relationship between the calculated and experimentally found k_nr values for the Re(CO)₂⁺ and Re(CO)₃⁺ complexes. Data for [Re(CO)₂(N^N)(tpp)₂]⁺ (yellow symbols), [Re(CO)₂(N^N)(dppv)]⁺ (grey symbols),²⁰, [Re(CO)₃(N^N)(tpp)]⁺ (cyan symbols), [Re(CO)₃(N^N)(CH₃CN)]⁺ (green symbols),¹² and [Re(CO)₃(N^N)(Cl)] (red symbols)¹² complexes in CH₃CN solutions at room temperature.

For some of the [Re(CO)₃(N^N)(tpp)]⁺ complexes (with 47dmphen, 56dmphen, and tmphen ligands) one can see, however, significant discrepancy between the calculated and experimentally found k_nr values. In the case of these complexes, their largest E₀₀ values may suggest the presence of an additional, thermally activated non-radiative deactivation channel. More specifically, the presence of the excited metal-centred ³*MC state, often contributing to the non-radiative deactivation,^55,56 might be operative. However, one can eliminate this option because the expected energy splitting of the Re(I) d and *d orbitals as high as ca. 4 eV can be estimated using the spectrochemical parameters of the ligand present in the [Re(CO)₃(N^N)(tpp)]⁺ complexes. This precludes possible ³*MLCT → ³*MC → S₀ deactivation path due to enough large energy gap between the ³*MLCT and ³*MC states. Thus, further work seems to be necessary for any convincing explanation of the appearing issue. Until yet the most of the MLCT emitters analysed as descried above belong to the c_MLCT > c_LCT class, only for very few examples reported in this work c_MLCT ≈ c_LCT could be anticipated (vide infra). Therefore, any more systematic data for the MLCT emitters with c_MLCT < c_LCT would be particularly interesting.

For the complexes under study, the E₀₀ energies found, as expected, depend on the nature of N^N ligand. Smaller E₀₀ values are characteristic for the N^N ligands with stronger electron withdrawing properties. Generally, as it could be expected, the E₀₀ values for the [Re(CO)₃(N^N)(tpp)]⁺ complexes are larger (by ca. 0.30 eV) than those characterizing their [Re(CO)₂(N^N)(tpp)₂]⁺ analogues. The fitted hν_H values in the range of 0.15–0.23 eV correspond well to averaged contributions from the vibrational modes of the N=C, C=C, and C≡O bonds stretching in the N^N diimine (1200–1600 cm⁻¹) and CO (1850–2050 cm⁻¹) ligands, correspondingly. For the [Re(CO)₂(N^N)(tpp)₂]⁺ complexes, the hν_H values (0.15–0.20 eV) are somewhat smaller as compared with their [Re(CO)₃(N^N)(tpp)]⁺ analogues (0.16–0.23 eV). In a similar way, the λ_H values (0.20–0.30 eV) found for the [Re(CO)₃(N^N)(tpp)]⁺ complexes are somewhat larger than those characterizing their [Re(CO)₂(N^N)(tpp)₂]⁺ counterparts (0.16–0.22 eV). One can tentatively attribute the observed differences in the fitted λ_H and hν_H values to reorganization of two or three C≡O bonds. The found differences in the λ_H and hν_H values explain the experimentally observed differences in the τ_em values for the Re(CO)₂⁺ and Re(CO)₃⁺ series. The longer τ_em values characteristic for the Re(CO)₂⁺ complexes are evidently connected with suppression of the k_nr rate constants caused, in the way characteristic for the inverted Marcus region, by smaller λ_H and hν_H values.

For both investigated series, the fitted λ_LM energies depend on the E_LCT − E₀₀ energy gap. The larger the E_LCT − E₀₀ term, the larger is the λ_LM value. The observed trends can be rationalized taking into account the electronic structures of the ³*[Re(CO)₂(N^N)(tpp)₂]⁺ and ³*[Re(CO)₃(N^N)(tpp)]⁺ emitters. The smaller the E_LCT − E₀₀ term, the larger contribution of the “pure” ³*LC to the wave function of the given emitter is expected. Assuming that the λ_LM energy is mostly connected with the amount of the “pure” ³*MLCT contribution to the observed “real” ³*MLCT state, one can obtain the following expression

where λ⁰_LM is the reorganization energy characterizing the “pure” ³*MLCT state. The variability in the fitted values of the λ_LM energies, up to 0.15 eV for the [Re(CO)₂(N^N)(tpp)₂]⁺ and up to 0.32 eV for the [Re(CO)₃(N^N)(tpp)]⁺ series, is significantly more pronounced.

For the [Re(CO)₂(N^N)(tpp)₂]⁺ and [Re(CO)₃(N^N)(tpp)]⁺ series of the Re(I) complexes, the λ_LM values were fitted according to eqn (7) using λ⁰_LM and V₃₃ as free fit parameters with obtaining good agreement between the experimental and fitted values (cf. Fig. 8). Whereas similar λ⁰_LM values, 0.52 and 0.50 eV are characteristic for the [Re(CO)₃(N^N)(tpp)]⁺ and [Re(CO)₂(N^N)(tpp)₂]⁺ series, the fitted V₃₃ values (0.16 and 0.25 eV, respectively) are distinctly different. Such difference in the V₃₃ values seems to be a general rule for the Re(CO)₃⁺ and Re(CO)₂⁺ complexes (cf. data in Table 3). The fits performed according to eqn (7) were done assuming V₃₃ = const within the given complex series. Thus, one should treat the obtained parameters fit parameters as an averaged V₃₃ and λ⁰_LM values.

Fig. 8 Relationship between the λ_LM and E_LCT − E₀₀ terms for [Re(CO)₂(N^N)(tpp)₂]⁺ (yellow symbols) and [Re(CO)₃(N^N)(tpp)]⁺ (cyan symbols) complexes in CH₃CN solutions at room temperature. Dashed curves present fits according to eqn (7).

Table 3 λ⁰_LM and V₃₃ values obtained from the fits performed according to eqn (7)^a

Complex series	λ^0LM	V33
a V33 and λ^0LM values for the [Re(CO)3(N^N)(Cl)], [Re(CO)3(N^N)(CH3CN)]^+, and [Re(CO)2(N^N)(dppv)]^+ series obtained analysing previously published data.^12,20
[Re(CO)2(N^N)(tpp)2]^+	0.50	0.25
[Re(CO)2(N^N)(dppv)]^+	0.48	0.26
[Re(CO)3(N^N)(Cl)]	0.52	0.13
[Re(CO)3(N^N)(CH3CN)]^+	0.39	0.11
[Re(CO)3(N^N)(tpp)]^+	0.52	0.16

---

Similarly to the observed variety of the λ_LM energies, one could expect the N^N ligand induced changes in the λ_H and hν_H terms. Their values, however, depend rather weakly on the N^N ligand present in the emitters under study. One can rationalize the lack of the significant λ_H and hν_H variation considering possible values of these parameters associated with the S₀ ← ³*LC emissions. For these emissions, with well-structured spectra, one can assume very small λ_LM and relatively large λ_H values. Then, independently of the c_LCT and c_MLCT coefficients attributed to the given MLCT emitter, the observed λ_H values should remain nearly constant. In an analogous way, this explains relatively small changes in the averaged high frequency intra-molecular vibrations.

The fitted λ_LM values contain contributions from the λ_L and λ_M reorganization energies. In a similar way, the λ_L and λ_M terms contributing to the overall λ_LM values are expected to follow analogous relationships. Separation of the both contributions is principally possible but requires some additional data and/or assumptions. Nominally, one can fit the emission band using more advanced model with two quantized modes corresponding to the medium ν_M and high ν_H frequency vibrations. However, since the one-mode approximation with expression (5) already gives a satisfactory agreement with experimental spectra the additional parameters do not follow from a free fit. Therefore, some of the fitting parameters (e.g., medium ν_M and high ν_H frequencies) should be somewhat arbitrary fixed. Optionally, one can compare the λ_LM values at distinctly different temperatures (e.g., 77 and 298 K) but this opportunity is not available for the investigated [Re(CO)₃(N^N)(tpp)]⁺ complexes due to changes of their emission character, from the MLCT at 298 K to LC at 77 K. Such comparison, however, is possible for the [Re(CO)₂(N^N)(tpp)₂]⁺ complexes exhibiting MLCT emission at both temperatures. This was performed in the manner as described previously for the [Re(CO)₂(N^N)(dppv)]⁺ complexes²⁰ with finding the close analogy between both series of the Re(CO)₂⁺ emitters.

The fitted V₃₃ values together with the E₀₀ energies provides some additional information about the nature of the given MLCT emitter allowing estimation of the c_LCT and c_MLCT coefficients. The c_MLCT coefficients calculated according to eqn (1) with the obtained V₃₃ values comprise the ranges of 0.88–0.96 and 0.66–0.98 for the Re(CO)₂⁺ and Re(CO)₃⁺ series, respectively. This allows classifying nearly all the discussed complexes as the emitters with dominant MLCT nature. Only in some cases, namely the [Re(CO)₃(N^N)(CH₃CN)]⁺ or [Re(CO)₃(N^N)(tpp)]⁺ complexes bearing 47dmphen, 56dmphen or tmphen as N^N ligand, comparable contributions from the “pure” ³*LC and ³*MLCT excitations can be anticipated from the estimated c_LCT values. This remains in agreement with traces of the vibronic structures observed in the 298 K emission spectra of these complexes.

Nature of the lowest excited triplet state

DFT and TD-DFT computations were performed to obtain a deeper insight into the nature of the emissive ³*[Re(CO)₂(N^N)(tpp)₂]⁺ and ³*[Re(CO)₃(N^N)(tpp)]⁺ species. To attain more comparable results, additional computation performed for the all discussed series of the Re(I) complexes were done at the same level of theory. Geometries of the investigated complexes were optimized at the B3LYP level⁵⁷ for each stationary structure in the ground S₀ and the lowest triplet T₁ electronic states. Electronic transitions were calculated for the optimized structures using the TD-DFT method.⁵⁸ Calculation performed in CH₃CN solutions were done by means of the polarizable continuum solvation model.⁵⁹ Combination of Lanl2DZ basis set⁶⁰ (Re element) and the 3-21G* or 6-31G* basis sets (light elements) were employed in the present computations. Relatively small 3-21G* basis set was applied in the optimizations because, due to flexibility of tpp ligand, optimizations required much more computing time than it could be expected taking into account only number of atoms consisting the given tpp complex. The electronic transitions, however, were afforded using the Lanl2DZ and 6-31G* combination. Such approach was applied successfully for the heteroleptic Ru(II) and Os(II) complexes bearing α-diimine and phosphine ligands.⁶¹ Our testing computations performed for selected [Re(CO)₃(N^N)(Cl)] complexes provided results compatible with those presented in the literature.^62–65 This is particularly true for the shapes of the molecular orbitals and orbital assignment of the lowest excited states. In a similar way, agreement with the literature data for the [Re(CO)₃(bpy)(tpp)]⁺ complex⁶⁶ was obtain.

The performed DFT and TD-DFT computations confirm the MLCT nature of the lowest triplet state anticipated for the discussed complexes. For the optimized T₁ geometry, the S₀ → T₁ transitions are described mainly (with CI coefficients 0.66–0.69) between the HOMO and LUMO orbitals, localized on the Re(I) core and N^N ligand, respectively (cf. data in Table 3). All other computed quantities associated with the S₀ → T₁ excitation confirm the MLCT nature of their lowest excited triplet state as well. Comparing the bond lengths computed for the optimized S₀ and T₁ states, one can see expected changes in the metal–ligands and intra-ligands bonds, e.g., shortening of the Re–N bonds and elongation of the Re–P and Re–C bonds. This remains in accordance with increased positive charge on the central Re(I) ion. In a similar way, the computed changes in the C=N and C=C bond lengths are consistent with the introduction of an additional electron on the π* orbital of the N^N ligand. Congruently, the S₀ → T₁ transitions lead to changes of the dipole moments characterizing the ground S₀ and the excited T₁ states. The observed lowering reflects well the charge redistribution associated with the MLCT excitation.

Analysis of the distribution of atomic charges and spin densities in the excited T₁ state of the discussed complexes supports intrinsic MLCT nature of these states. Upon inspecting the calculated spin densities for the investigated complexes, one can see that one of the unpaired electrons is localized mainly on the N^N ligand, whereas the second one on the remaining parts of the complexes. The observed ca. 1 : 1 spin redistribution symmetry is distinctly different from the 2 : 0 asymmetry expected for the ³*LC excitation localized solely within the N^N ligand. For the discussed complexes, the 1 : 1 symmetry predictable for “pure” MLCT excitation is, however, to some extent broken. Whereas, the effect is relatively small for the investigated Re(CO)₂⁺ complexes (e.g., 1.13 : 0.87 for [Re(CO)₂(bpy)(dppv)]⁺ chelate), the spin redistribution is distinctly more dissymmetric for [Re(CO)₃(bpy)(tpp)]⁺ (1.22 : 0.78) or [Re(CO)₃(bpy)(CH₃CN)]⁺ (1.31 : 0.69) complexes. The computation performed on the same level of theory gave for the prototype [Re(CO)₃(bpy)(Cl)] complex nearly theoretical 1.04 : 0.96 ratio. Reasonably, larger spin redistribution dissymmetry arises from the DFT computation performed for other complexes under investigations. For example, the 1.49 : 0.51 and 1.31 : 0.69 ratios were found for the [Re(CO)₃(47dmphen)(CH₃CN)]⁺ and [Re(CO)₃(47dmphen)(tpp)]⁺ chelates. Noteworthy, the found differences in the spin redistribution extracted from the DFT computation follow adequately the c_LCT coefficients values that can be estimated using data the emission band-shape analysis. Using eqn (1) with affordable V₃₃ and E_LCT − E₀₀ one can obtain the c_LCT values ranging from 0.15 to 0.71–0.75 for the [Re(CO)₃(bpy)(Cl)] and [Re(CO)₃(47dmphen)(CH₃CN)]⁺ or [Re(CO)₃(47dmphen)(tpp)]⁺ complexes, respectively. For c_LCT = 0.15 one can expect the 1.02 : 0.98 spin redistribution ratio, whereas for c_LCT = 0.75 the 1.56 : 0.44 dissymmetry is anticipated. Thus, one can conclude nice congruence between the DFT and the band-shape analysis outcomes.

The obtained shapes of the molecular orbitals as well as the resulted orbital assignments of the low energy transitions remain in agreement with the MLCT nature of the low energy bands of the complexes under study. Shapes of their molecular orbitals (cf. Fig. 9) confirm metallic character of their HOMO, HOMO-1 and HOMO-2 levels involved in the low energy electronic transitions. The lowest LUMO orbitals participating in these transitions are essentially π* orbitals localized mainly on the N^N ligand with minor contributions from the Re(I) ion and other components forming the discussed complexes. However, although the HOMO, HOMO−1, and HOMO−2 possess mainly the metallic d orbital character, some contributions from the ligand attached to the Re(I) core are clearly seen. Thus, one can consider the low energy singlet and triplet states of the discussed complexes as the MLCT states with more or less pronounced admixture of the intra-ligand excitations. This is characteristic for the transitions within the singlet as well as the triplet manifolds. Moreover, both these excitation types should be treated as the charge-transfer between the whole Re(CO)₃(tpp)⁺, Re(CO)₃(CH₃CN)⁺, Re(CO)₃(Cl), Re(CO)₂(tpp)₂⁺ or Re(CO)₂(dppv)⁺ fragments to the N^N ligand instead of somewhat oversimplified “pure” d_Re(I) → N^N or d_Re(II) ← N^N⁻ assignments.

Fig. 9 Shapes of the frontier molecular orbitals for the bpy and 47dmphen complexes. Data (for CH₃CN solutions) from the TD-DFT computations performed for the lowest triplet states structures optimized in vacuum. Plots views (along z-axis) of the isodensity surfaces with contour value Z = 0.02.

Despite all found similarities, there are some intrinsic differences in properties of the discussed complexes. These include essential variety in the oscillator strengths of the computed vertical electronic transitions from the minima of the S₀ potential curves (cf. Fig. 10) as well from the S₀ states at the T₁ geometries, i.e., at geometry of the Franck–Condon states reached in the S₀ ← ³*MLCT emissions (cf. data in Table 4). Such behaviour seems to be associated with different contributions of the individual Re(I) d orbitals to the molecular HOMO, HOMO−1, and HOMO−2 orbitals, respectively. Considering the shapes of the molecular orbitals involved in the individual transition, one can conclude that their symmetries are mainly responsible for the observed changes in the f patterns Nevertheless, the found sums of oscillator strengths f of the transitions with MLCT character remains (for the given N^N ligand) nearly constant over these chelates. The found small variety of the ∑f values (e.g., 0.08–0.12 for the bpy complexes) corresponds well with the observed similar values of the ε_m coefficients.

Fig. 10 MLCT parts of the UV-vis absorption spectra of the [Re(CO)₂(bpy)(tpp)₂]⁺, [Re(CO)₂(bpy)(dppv)]⁺, [Re(CO)₃(bpy)(Cl)], [Re(CO)₃(bpy)(CH₃CN)]⁺, and [Re(CO)₃(bpy)(tpp)]⁺ complexes (from left to right) in CH₃CN solutions. Blue vertical bars correspond to the positions and relative intensities of the electronic transitions as obtained from TD-DFT computations.

Table 4 Energies E (in eV) and oscillator strengths f for the vertical electronic transitions with their orbital assignments.^a Data (in CH₃CN solutions) from the TD-DFT computation performed for the S₀ states at geometries corresponding to the vacuum optimized structures of the lowest triplet state

Complex	S0 → T1	S0 → S1	S0 → S2	S0 → S3	S0 → S4
a H/L, H−1/L, H−2/L, H−3/L, H−4/L and H/L+1 denote HOMO → LUMO, HOMO−1 → LUMO, HOMO−2 → LUMO, HOMO−3 → LUMO, HOMO−4 → LUMO and HOMO → LUMO+1 transitions, respectively.
[Re(CO)2(bpy)(tpp)2]^+	E = 1.80	E = 2.27, f = 0.103	E = 2.63, f = 0.004	E = 2.99, f = 0.011	E = 3.44, f = 0.009
	98% H/L	99% H/L	99% H−1/L	99% H−2/L	82% H/L+1 + 13% H−4/L
[Re(CO)2(bpy)(dppv)]^+	E = 1.85	E = 2.29, f = 0.044	E = 2.38, f = 0.010	E = 2.84, f = 0.076	E = 3.38, f = 0.004
	94% H/L	85% H/L+ 10% H−1/L	89% H−1/L + 9% H/L	93% H−2/L	98% H/L+1
[Re(CO)3(bpy)(Cl)]	E = 1.92	E = 2.10, f = 0.001	E = 2.48, f = 0.077	E = 2.77, f = 0.004	E = 3.61, f = 0.001
	92% H/L	99% H/L	98% H−1/L	99% H−2/L	87% H/L+1 + 11% H−3/L+1
[Re(CO)3(bpy)(CH3CN)]^+	E = 2.09	E = 2.49, f = 0.000	E = 2.80, f = 0.004	E = 2.95, f = 0.120	E = 3.76, f = 0.250
	78% H/L + 17% H−3/L	99% H/L	87% H−2/L + 12% H−1/L	86% H−1/L + 12% H−2/L	84% H−3/L + 11% H/L
[Re(CO)3(bpy)(tpp)]^+	E = 2.12	E = 2.60, f = 0.049	E = 2.93, f = 0.035	E = 3.02, f = 0.033	E = 3.36, f = 0.020
	82% H/L	93% H/L	77% H−1/L + 18% H−2/L	79% H−2/L + 16% H−1/L	95% H−3/L
[Re(CO)2(47dmphen)(tpp)2]^+	E = 1.81	E = 2.45, f = 0.148	E = 2.71, f = 0.032	E = 2.75, f = 0.017	E = 3.10, f = 0.000
	92% H/L	95% H/L	81% H/L+1 + 13% H−1/L	82% H−1/L + 16% H/L+1	97% H−2/L
[Re(CO)2(47dmphen)(dppv)]^+	E = 1.84	E = 2.58, f = 0.149	E = 2.67, f = 0.003	E = 2.80, f = 0.038	E = 3.05, f = 0.064
	90% H/L	86% H/L + 7% H/L+1	95% H−1/L	88% H/L+1	92% H−2/L
[Re(CO)3(47dmphen)(Cl)]	E = 1.98	E = 2.45, f = 0.054	E = 2.67, f = 0.069	E = 2.82, f = 0.048	E = 3.01, f = 0.009
	87% H/L	82% H/L	81% H−1/L	87% H/L+1	60% H−2/L + 32% H−3/L
[Re(CO)3(47dmphen)(CH3CN)]^+	E = 2.10	E = 2.92, f = 0.072	E = 3.07, f = 0.092	E = 3.20, f = 0.005	E = 3.30, f = 0.076
	80% H/L	69% H/L + 26% H−1/L	69% H−1/L + 18% H/L	91% H−2/L	77% H/L+1, 9% H/L
[Re(CO)3(47dmphen)(tpp)]^+	E = 2.06	E = 2.90, f = 0.181	E = 3.14, f = 0.010	E = 3.23, f = 0.046	E = 3.28, f = 0.008
	81% H/L	92% H/L	88% H−1/L + 11 H/L+1	75% H/L+1 + 9% H−1/L	60% H−2/L + 32% H−3/L

---

For all discussed Re(I) complexes the lowest S₀ → T₁ transitions involve the HOMO and LUMO orbitals in accordance with their MLCT character. Energies of these transitions, computed at the lowest triplet state geometries, correspond quite well with the experimental hc[small nu, Greek, tilde]^max_em values. Although the found agreement (cf. Fig. 11) could be somewhat better, one can regard the differences between experimental and computed values (up to ca. 0.3 eV) as acceptable taking into account low level of the theory. In similar way, the energies of the S₀ → ¹*MLCT transitions seem to be underestimated with comparable errors (cf. Fig. 10). Thus, one can conclude that the energy gaps between the lowest ³*MLCT and ¹*MLCT states (ΔE_ST) available from the performed TD-DFT computations are, due to cancellation of possible errors, quite sensible values. One can also hypothesise that the TD-DFT energy gaps between T₁ and S₂ or T₁ and S₃ states are reasonable as well.

Fig. 11 TD-DFT energies of the S₀ → T₁ (red bars) and S₀ → S_n transitions (green and blue bars) for the [Re(CO)₂(N^N)(tpp)₂]⁺, [Re(CO)₂(N^N)(dppv)]⁺, [Re(CO)₃(N^N)(Cl)], [Re(CO)₃(N^N)(CH₃CN)]⁺, [Re(CO)₃(N^N)(tpp)]⁺ complexes (from the left to right). Data for the bpy (left column) and 47dmphen (right column) ligands. Blue bars indicate transitions from these HOMO−1 or HOMO−2 where their metallic character is connected with the Re(I) d_xy orbitals whereas green bars correspond to transitions from the HOMO, HOMO−1 or HOMO−2 originated from the d_xz and/or d_yz orbitals. Yellow star symbols denote the energies of the S₀ ← ³*MLCT emissions. Insert presents relation of the TD-DFT energy gaps ΔE_ST between the lowest excited ³*MLCT and ¹*MLCT states to the ΔE_ST values calculated according to eqn (12). Data for the [Re(CO)₂(N^N)(tpp)₂]⁺ (yellow symbols), [Re(CO)₂(N^N)(dppv)]⁺ (grey symbols), [Re(CO)₃(N^N)(Cl)] (red symbols), [Re(CO)₃(N^N)(CH₃CN)]⁺ (green symbols), [Re(CO)₃(N^N)(tpp)]⁺ (cyan symbols) complexes with 47dmphen (marked with * character) and bpy ligands.

According to the data presented in Table 4, the computed ΔE_ST values, spanning the range from 0.18 to 0.84 eV, are much larger than one can expect taking into account only electronic interactions between the excited ³*MLCT and ³*LC states. In such a case, due to the electronic interactions within the ¹*LC/¹*MLCT and ³LC/³*MLCT manifolds, the initial degeneracy of the “pure” ¹*MLCT and ³*MLCT states is removed. Thus, one can approximate the appearing energy splitting ΔE_ST between the “real” ¹*MLCT and ³*MLCT states as follows:²⁰

where the V₁₁ denotes the electronic coupling element responsible for the interactions between the initial ¹*MLCT and ¹*LC states, the latter characterized by the 0–0 transition energy E_LCS. Correspondingly, the electronic coupling element V₁₀ describes the electronic interaction between the initial ¹*MLCT and the ground S₀ states with energy gap between them close to E₀₀. Due to expected similar values of the V₁₁ and V₁₀ terms as well as comparable values of the E₀₀ and E_LCS − E₀₀ energy gaps, one can postulate cancelling of the contributions from the second and third terms in eqn (8) because they are working in the opposite directions. Thus, the first term on the right side of eqn (8) is dominantly contributing to the ΔE_ST:

ΔE_ST = V₃₃²/(E_LCT − E₀₀) (9)

With the V₃₃ and E₀₀ quantities available from the emission band-shape analysis, one can estimate the ΔE_ST terms according to eqn (9). The resulting values, ranging from 0.02 to 0.25 eV, are up to nearly one order of magnitude smaller than the TD-DFT ones. Since such large discrepancy seems only hardly attributable to the possible errors in the V₃₃ and E₀₀ values, some other factors must affect the energy splitting between the lowest ³*MLCT and ¹*MLCT states.

Comparing the ΔE_ST values estimated according to eqn (9) with those from the TD-DFT computations one can find significant correlation. The larger the c_LCT coefficient characterizing the given MLCT emitter, the larger is the observed discrepancy. This finding suggests important contributions from the exchange interactions between unpaired electrons of the excited ³*MLCT and ¹*MLCT states.^67,68 Thus, one can rationalize the observed ΔE_ST inconsistency taking into account the nature of the ψ_S and ψ_T wave functions of the lowest excited S₁ and T₁ states. If the S₁ state is nearly “pure” MLCT state, one can assume localization of the unpaired electrons mainly on the d and π* orbitals. In such a case, with neglecting smaller contributions from other ligand or metal orbitals, one can express the ψ_S wave function as follows:

ψ_S = ψ_d + ψ_L (10)

where ψ_d and ψ_L are the wave functions of the unpaired electrons localized on the d and π* orbitals. The latter is the LUMO orbital of the N^N ligand. In the excited ³*MLCT state the unpaired electrons are localized, due to mixing between “pure” ³*LC and ³*MLCT states, on the d, π and π* orbitals. Thus, one can express the ψ_T wave function as follows:

ψ_T = c_MLCTψ_d + c_LCTψ_H + ψ_L (11)

where ψ_H is the wave function describing the unpaired electrons localized on the π orbital (HOMO of the N^N ligand). Since the nature of the ψ_S and ψ_T wave functions determine the exchange energy, one can expect larger ΔE_ST values for the MLCT emitters with larger c_LCT coefficients. Searching of any connection between the TD-DFT results and the c_LCT and ΔE_ST parameters from the emission band-shape analysis, we have found following empirical relationshipwhere J_ex is the energy exchange characterizing the given N^N ligand. Taking into account the respective J_ex values, 1.00 eV for 47dmphen⁵³ and 1.05 eV for bpy^55,69 ligands as determined from the fluorescence and phosphorescence spectra of their Zn(II) complexes, one can calculate the “theoretical” ΔE_ST values using eqn (12). Agreement between the TD-DFT values of the ΔE_ST energy gap with those calculated according to eqn (12) (cf. Fig. 11) can be regarded as more than satisfactory. Some deviations are explainable by errors in the analysed ΔE_ST, c_LCT and V₃₃ values. Although Fig. 11 and Table 4 present data only for bpy and 47dmphen complexes, the found relationship seems to have a more general meaning. This conclusion arises from preliminary TD-DFT results obtained for some other [Re(CO)₃(N^N)(CH₃CN)]⁺ and [Re(CO)₃(N^N)(Cl)] complexes where the ΔE_ST vs. c_LCT relationship is fulfilled as well.

Transition dipole moments of S₀ ← ³*MLCT emission

The close analogy between complexes within the Re(CO)₃⁺ and Re(CO)₂⁺ series is also seen in the N^N ligand induced changes in the M_em values (cf. Fig. 12) where the M_em values are plotted against the E_LCT − E₀₀ difference. The applied correlation is based on our previous works,^{12,20,29–31} where following relationship between the M_em and (E_LCT − E₀₀)/hc[small nu, Greek, tilde]^max_em was introducedwhere V_SOC is the spin–orbit coupling element between the lowest ³*MLCT and ¹*MLCT states and Δμ is the difference in the dipole moments of the S₀ and “pure” ¹*MLCT states. Eqn (12) predicts linear correlation (with slopes equal to the χ_M = V₁₀V_SOCΔμ/V₃₃² terms) between the experimentally available M_em and (E_LCT − E₀₀)/hc[small nu, Greek, tilde]^max_em values when the remaining parameters are constant (or nearly constant) within analysed series of MLCT emitters. This may be expected when the coordinated N^N ligands are varied whereas the central metal ion and other coordinated ligands remain the same. The trend in the M_em values observed for the [Re(CO)₂(N^N)(tpp)₂]⁺ series is similar to that found for the [Re(CO)₂(N^N)(dppv)]⁺ one whereas the [Re(CO)₃(N^N)(tpp)]⁺ complexes are more resembling their [Re(CO)₃(N^N)(CH₃CN)]⁺ or [Re(CO)₃(N^N)(Cl)] counterparts. The found χ_M slopes for the Re(CO)₂⁺ series are distinctly smaller than those found for the Re(CO)₃⁺ ones.

Fig. 12 Relationships between the M_em and E_LCT − E₀₀ quantities for the Re(CO)₃⁺ (top) and Re(CO)₂⁺ (bottom) complexes. Data for [Re(CO)₂(N^N)(tpp)₂]⁺ (yellow symbols), [Re(CO)₂(N^N)(dppv)]⁺ (grey symbols),²⁰, [Re(CO)₃(N^N)(tpp)]⁺ (blue symbols), [Re(CO)₃(N^N)(CH₃CN)]⁺ (green symbols),¹² and [Re(CO)₃(N^N)(Cl)] (red symbols)¹² complexes in CH₃CN solutions at room temperature.

The above presented findings might suggest applicability of the eqn (12) in the quantitative interpretation of the experimentally found M_em values. Some fundamental reservations, however, allow regarding the observed coincidence as somewhat accidental. These arises because eqn (13) was derived assuming that the intensity borrowing from the lowest excited ¹*MLCT state is responsible for the radiative S₀ ← ³*MLCT deactivation. Then the transition dipole moment M_em is related to the ΔE_ST energy gap according to following relationship⁷⁰

M_em = M₀₁V_SOC/ΔE_ST (14)

where M₀₁ is the transition dipole moment of the S₀ ← ¹*MLCT transition. Combining eqn (14) with eqn (9) and assuming that the M₀₁ values follow the Hush-Mulliken formalism

M₀₁ = ΔμV₁₀/hc[small nu, Greek, tilde]^max_em (15)

one can straightforwardly obtain eqn (13) which can be treated as the limiting c_MLCT ≈ 1 and c_LCT ≈ 0 case. Since in any more realistic situation c_MLCT < 1 and c_LCT > 0, one can correct eqn (13) taking into account eqn (12) with additional assumption that the effective spin–orbit coupling constants V_SOC depend on the c_MLCT coefficients as follows

V_SOC = c_MLCTV⁰_SOC (16)

where V⁰_SOC characterize the spin–orbit coupling between the “pure” ³*MLCT and ¹*MLCT states. Then one can simply obtain following expression

Since the c_LCT and c_MLCT coefficients are affordable from eqn (1) and typically c_LCTJ_ex ≫ V₃₃²/(E_LCT − E₀₀) one can further simplify eqn (17) to

The obtained expression, similarly to eqn (13), predicts again linear relationship between the experimentally determinable M_em and (E_LCT − E₀₀)hc[small nu, Greek, tilde]^max_em quantities. Potentially one could apply eqn (18) in interpretation of the found similarities and variation of the experimentally found M_em values but some doubts appear concerning the spin–orbit coupling induced mixing of the lowest excited ³*MLCT and ¹*MLCT states. This is because a ³*MLCT state may couple effectively with a ¹*MLCT state if their electronic configurations involve the same π* ligand orbital, but different d metal orbitals.^71,72 Thus, the required spin orbit coupling may be only weakly operative. Moreover, in the case of some α-diimine fac-Re(CO)₃⁺ complexes, the oscillator strengths of their lowest S₀ → ¹*MLCT transitions (cf. data in Table 4) are very low, too low for an effective intensity borrowing responsible for the S₀ ← ³*MLCT emission.

Possible mixing between the lowest ³*MLCT state with other singlet states may be treated as any plausible option explaining the experimentally observed M_em values. Then the resulting transition dipole moment M_em of the S₀ ← ³*MLCT emission can be expressed as follows⁷⁰

where the M and V_SOC are the dipole moment and the spin–orbit coupling operators, respectively. The V_SOC operator mixes the lowest triplet state T₁ with energy E(T₁) with the excited singlet states S_n with energies E(S_n) whereas the M operator is responsible for the transition dipole moments of the spin allowed S₀ ← S_n transitions. In the case of organometallic luminophore its ¹*MLCT states close in energy to the ³*MLCT one are expected contributing mostly to the overall M_em value. This conclusion arises from relatively small energy gaps E(³*MLCT) − E(¹*MLCT) and anticipated large values of the spin–orbit coupling constants. The latter is caused by the metallic character of the interacting ³*MLCT and ¹*MLCT states. With this assumption, one can consider three excited ¹*MLCT state as potentially operative in the required intensity borrowing. Because one can exclude the electronic HOMO → LUMO transition from considerations, only the HOMO−1 → LUMO and HOMO−2 → LUMO transitions need more attention.

The perceived congruity between the band-shape analysis data and results from the performed TD-DFT computation allows considering the TD-DFT data as quite trustworthy. Thus, despite of low level of our computations, one can apply them to discuss the experimental M_em values in more details. Analysing the data collected in Table 4, however, it is rather difficult to find any simple correlation between the M_em and 〈S_n|M|S₀〉/[E(S_n) − E(T₁)] quantities. According to eqn (17) one can also expect (in some specific cases) the M_em values independent off the oscillator strengths redistribution over tree considered S₀ ← ¹*MLCT transitions. However, this could be only possible when the numerators 〈¹*MLCT_n|V_SOC|³*MLCT〉 in eqn (19) would be proportional to the denominators E(¹*MLCT) − E(³*MLCT). Then, for all three conceivable contributions, the V_SOC/ΔE terms remain constant and the resulting M_em values is simply proportional to the sum of 〈¹*MLCT_n|M|S₀〉. Perhaps one can consider this special case as imaginable explanation of the experimental findings, but at the present stage of investigations, it remains as possible but very unlikely opportunity.

The second option is the intensity borrowing from the spin allowed T₁ → T_n transitions. In such a case, one can relate the M_em values to the spin–orbit induced coupling between the ground S₀ state and T_n states 〈T_n|V_SOC|S₀〉 and the transition dipole moments of the transitions 〈T_n|M|T₁〉 occurring within the triplet manifold⁷⁰

This option is, however, still less credible because the energy gaps E(T_n) − E(S₀) are relatively large that should result in distinctly smaller V_SOC values. Because values of the 〈T_n|V_SOC|S₀〉 and 〈S_n|V_SOC|T₁〉 terms are anticipated to be similar, one could consider this option only for the 〈T_n|M|T₁〉 values much larger than their 〈S_n|M|S₀〉 complements. This is, however, very unlikely. Thus, one can expect potentially possible intensity borrowing from the T₁ → T_n transitions as rather small, most probably too small to explain the experimentally observed M_em values.

Finally, one can also consider a permanent dipole difference contribution arising from the direct spin–orbit coupling induced interactions between the ground S₀ and the excited ³*MLCT states. Both involved states are metallic in their character that precludes possibly effective spin–orbit induced interactions between them. In this particular case the M_em values will depend on the differences between the dipole moments, μ(T₁) and μ(S₀), of the states involved in the S₀ ← ³*MLCT transition and the spin–orbit coupling 〈T₁|V_SOC|S₀〉 element responsible for interaction between them⁷⁰

One can attribute the energy gap E(T₁) − E(S₀) in eqn (21) to the emission maxima hc[small nu, Greek, tilde]^max_em of the given S₀ ← ³*MLCT transition. For both remaining terms, μ(T₁) − μ(S₀) and 〈T₁|V_SOC|S₀〉, one can expect their functional dependence on amount of the metallic character in the ³*MLCT emitter. Both considered quantities depend on the value of the c_MLCT coefficient. Thus, the μ(T₁) − μ(S₀) difference is approximately equal to c_MLCT²Δμ. Correspondingly, one can approximate the 〈T₁|V_SOC|S₀〉 value as c_MLCTV⁰_SOC. With the above remarks, one can simply obtain the following expressions

According to eqn (22), one can anticipate the experimental M_em values related to the c_MLCT³ factor. In fact a monotonic relationship between M_em and c_MLCT seems to be evident (cf. Fig. 13) but, instead of the expected M_em ∼ c³_MLCT association, the M_em values can be linearized against c_MLCT⁹. Similarly, as one can see in Fig. 12, the trends in the M_em values reflect the MLCT character of the given emitter. The experimental points are similarly scattered but the picture is more coherent, all five discussed Re(I) series follow the same layout. The larger the c_MLCT values in the ³*MLCT state, the larger are the transition dipole moments attributed to the S₀ ← ³*MLCT emission. Evidently, the c_MLCT coefficient is the most important factor affecting the M_em values. Most probably, the higher MLCT characteristic involved in the singlet and triplet excited states, the spin–orbit coupling is more pronounced. Any exact origin of the observed correlation remains, however, an open question. Thus, one should treat the observed M_em ∼ c_MLCT⁹ relation as an empirical rule. On the other hand, similar correlations, e.g., between the M_em and (E_LCT − E₀₀)/hc[small nu, Greek, tilde]^max_em values, have been found for in our previous works concerning some Os(II) or Ru(II) complexes.^29–31 Preliminary results from our investigations of the cyclometalated Ir(III) complexes [Ir(C^N)₂(N^N)]⁺ suggests that these chelates follow the same trend as well. Thus, the correlations described in this work may have a general meaning and the problem is worthy for further investigations. To obtain an adequate solution, however, one can speculate that it will be necessary to go beyond simple perturbation theory. Very likely, a proper account of the spin–orbit coupling between the ground S₀ and the lowest ³*MLCT states will require considerations of the spin-vibronic coupling.⁷³

Fig. 13 Relationships between the M_em and c_MLCT quantities for the Re(CO)₂⁺ and Re(CO)₃⁺ complexes. Data for [Re(CO)₂(N^N)(tpp)₂]⁺ (yellow symbols), [Re(CO)₂(N^N)(dppv)]⁺ (grey symbols),²⁰, [Re(CO)₃(N^N)(tpp)]⁺ (blue symbols), [Re(CO)₃(N^N)(CH₃CN)]⁺ (green symbols),¹² and [Re(CO)₃(N^N)(Cl)] (red symbols)¹² complexes in CH₃CN solutions at room temperature.

Concluding remarks

Comparative studies of the luminescence properties of Re(CO)₂⁺ and Re(CO)₃⁺ α-diimine chelates point to very similar nature of their lowest excited states. In both types of the complexes, their emissions in solutions at room temperature take place from the lowest excited T₁ states possessing distinct MLCT character. For most of the complexes discussed in this work, the 77 K emissions in solid matrices exhibit MLCT character as well. In some cases, however, the character of the 77 K emission change from MLCT to LC. Particularly well pronounced temperature effect can be seen for the [Re(CO)₃(N^N)(tpp)]⁺ series where, independently of the coordinated N^N ligand, structured emissions have been recorded.

Despite all similarities, the investigated Re(CO)₂⁺ and Re(CO)₃⁺ complexes exhibit some important differences, particularly in the reorganization energies accompanying the electron transfer between the metallic center and the coordinated N^N ligand. Typically, the reorganization energies associated with high and low/medium frequency modes are distinctly smaller for the Re(CO)₂⁺ complexes as compared to the Re(CO)₃⁺ ones. This affects strongly the non-radiative deactivation of the excited ³*MLCT states making the Re(CO)₂⁺ chelates better emissive that is reflected in the higher emission quantum yields and longer emission life-times.

One can regard the reported complexes as well suited for the fundamental studies of the structure/properties relations for the MLCT emitters. In this work, two different approaches, analysis of emission band shapes and TD/TD-DFT computations, have been applied to clarify the observed changes in their luminescence properties as caused by the nature of the main and ancillary ligands. Well congruent results obtained from both applied methodologies shown that the data from band shape analyses are applicable in testing results from the quantum-mechanical computations, and vice versa. Particularly, both applied approaches point to crucial role of the lowest excited state ³*LC localized within the coordinated N^N ligands. Mixing of the “pure” ³*LC and ³*MLCT configurations determines the nature of the given MLCT emitter affecting all quantities associated with the radiative as well as non-radiative S₀ ← ³*MLCT transitions.

Results from the performed band shape analyses of the S₀ ← ³*MLCT emissions have been applied in the interpretation of the experimentally determined rates of the non-radiative deactivations of the excited ³*MLCT species. In most cases, the theoretically predicted k_nr values remain in nice agreement with those found experimentally. This confirms a close connection between the radiative and non-radiative deactivation of the excited states with the MLCT character. Intrinsic deviations between calculated and experimental k_nr values were, however, found for some of the MLCT emitters with large contributions of the ³*LC excitation. Any more detailed explanation of this finding requires further systematic investigations.

Another unclear issue, origin of the electronic coupling between the lowest excited ³*MLCT and the ground S₀ state requires further systematic investigations as well. The S₀ ← ³*MLCT transitions are, due to spin conservation rule, forbidden processes. They might become possible when the intensity borrowing through the spin–orbit coupling effects cause the mixing between the singlet and triplet states. Analysing the obtained TD-DFT data we are, however, not able to find any S₀ → S_n or T₁ → T_n transition responsible individually for the required intensity borrowing. At least not one which could be common for all five discussed series of Re(I) complexes with adequate explanation of the N^N ligand induced changes in the M_em values. Perhaps, one should treat each of the analysed MLCT emitters independently but this option seems to be improbable. The latter conclusion is based on the monotonic relationship between the M_em values and c_MLCT coefficients characterizing the MLCT character of the given MLCT emitter. Although the found M_em ∼ c_MLCT⁹ correlation looks somewhat amazing, this finding can have a more general meaning because similar behaviour is characteristic for other α-diimine complexes. At the present stage of investigations, however, it is rather difficult to provide any reliable explanation of these findings. Further work in such direction seems to be required for any decisive answer.

Experimental

Materials

Solvent used in UV-vis absorption and emission studies, acetonitrile – ACN, methanol, and ethanol, were of spectroscopic grade purchased from Aldrich. Reagents and analytical grade solvents, used without further purification in performed syntheses and purification of the investigated complexes, were purchased from Trimen, Acros Organics, Sigma-Aldrich or Alfa Aesar companies.

Both investigated series, [Re(CO)₂(N^N)(tpp)₂]⁺ and [Re(CO)₃(N^N)(tpp)]⁺, have been synthesized in the form of PF₆⁻ salts from [Re(CO)₃(N^N)(Cl)] precursors^1,10,11 prepared according to well-known procedure reacting equimolar mixtures of Re(CO)₅Cl and appropriate N^N ligand in refluxing toluene under argon for 4–5 h. The obtained precursors were further converted into [Re(CO)₂(N^N)(tpp)₂]⁺⋯PF₆⁻ salts by refluxing their deoxygenated o-dichlorobenzene solutions containing TlPF₆ as dehalogenation agent and seven fold excess of tpp ligand for 2–4 h in dark.¹⁶ After filtering of precipitated TlCl, the reaction mixture was cooled to room temperature and an excess of diethyl ether/hexane mixture was added to precipitate [Re(CO)₂(N^N)(tpp)₂]⁺⋯PF₆⁻ products. The synthesized complexes were further purified by means of the column chromatography on activated acidic alumina with CH₂Cl₂/acetone 3 : 1 v/v mixture as eluent. The [Re(CO)₃(N^N)(tpp)]⁺⋯PF₆⁻ salts⁷⁴ were prepared in a similar way using chlorobenzene as reaction medium. The reaction mixture was heated to 105–110 °C under argon in dark for 3 h. Due to the lower reaction temperature, refluxing of the deoxygenated chlorobenzene solutions containing equimolar amount of [Re(CO)₃(N^N)(Cl)] and tpp with slight excess of TlPF₆ leads to [Re(CO)₃(N^N)(tpp)]⁺⋯PF₆⁻ as main products. Column chromatography, performed on silica gel with CHCl₃/CH₃OH 100 : 1 v/v as an eluent, resulted in final separation and purification of these complexes. Identification of all synthesized complexes was done by means of FT-IR, ³¹P NMR and ¹H NMR spectroscopy. Acquired spectroscopic data confirm the expected structures of the synthesized complexes without any doubts. The recorded FT-IR spectra exhibit the presence of two (1919–1931 and 1830–1854 cm⁻¹) or three ((2027–2037, 1935–1954 and 1902–1935 cm⁻¹) sharp and intense absorption band in the ν_C≡O stretching region in accordance with the presence of two or three CO ligands in the [Re(CO)₂(N^N)(tpp)₂]⁺ or [Re(CO)₃(N^N)(tpp)]⁺ series, respectively. All the investigated complexes exhibit common features in their ³¹P NMR spectra, exhibiting the presence of characteristic septet signals arising from the PF₆⁻ counterions (δ = −144.7 ppm, J ∼ 706 Hz). Additional singlet signals (with singlet-to-septet 2 : 1 or 1 : 1 intensity ratio) from the coordinated tpp ligands, in the range of 17–23 ppm for the [Re(CO)₂(N^N)(tpp)₂]⁺ or 15–19 ppm for [Re(CO)₃(N^N)(tpp)]⁺ series, correspond to the presence of two or one tpp ligand(s) in the analysed species. The presence of N^N and tpp ligands in these complexes is also nicely reflected in their ¹H NMR spectra where the integrated intensities of signals and their positions reproduce nicely the expected numbers of protons. Here the performed syntheses and identification of the synthesized complexes are only briefly mentioned. Any more detailed description will be reported elsewhere.

Instrumentation and procedures

FT-IR, ³¹P and ¹H NMR spectra were acquired using Shimadzu IRAffinity-1 and VARIAN 400-MR spectrometers, respectively. UV-vis absorption spectra were measured using Shimadzu UV 3100 spectrophotometer, whereas corrected steady-state luminescence spectra and emission decays by means of Gilden Photonics FluoroSense and FluoroSense-P fluorimeters. In the case of emission studies, the investigated CH₃CN solutions were carefully deaerated by the prolonged saturation with preliminary purified and dried argon.

As a quantum yield standard, a solution of quinine sulphate in 0.1 N H₂SO₄ (ϕ_ref = 0.51)⁷⁵ was used. The obtained emission quantum yields ϕ_em were measured with the estimated 10% accuracy. Emission spectra were fitted by means of a least-square method using OriginPro 9.0 software (Origin Lab Corp.) with user-defined functions. The experimental decay curves were analysed by the single-curve method using the reference convolution based on the Marquardt algorithm.⁷⁶ with the χ² test and the distribution of residuals serving as the main criteria in the evaluation of fit quality. Emission lifetime τ_em values, characterizing the recorded decays were measured with the temporal resolution of ca. 0.01 μs. DFT and TD-DFT results presented in this work were obtained with the Gaussian software supported by GaussView 5.0.⁷⁷

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work is a part of the research project no. 24/20/B supported by the Siedlce University of Natural Sciences and Humanities.

References

1. M. Wrighton and D. L. Morse, J. Am. Chem. Soc., 1974, 96, 998.
2. R. A. Kirgan, B. P. Sullivan and D. P. Rillema, Top. Curr. Chem., 2007, 281, 45.
3. A. Kumar, S.-S. Sun and A. J. Lees, Top. Organomet. Chem., 2010, 29, 1.
4. H. Takeda, K. Koike, T. Morimoto, H. Inumaru and O. Ishitani, Adv. Inorg. Chem., 2011, 63, 137.
5. D. J. Stufkens and A. Vlcek Jr, Coord. Chem. Rev., 1998, 177, 127.
6. B. D. Rossenaar, D. J. Stufkens and A. Vlcek Jr, Inorg. Chem., 1996, 35, 2902.
7. L. Wallace, D. C. Jackman, D. P. Rillema and J. W. Merkert, Inorg. Chem., 1995, 34, 5210.
8. L. Sacksteder, M. Lee, J. N. Demas and B. A. DeGraf, J. Am. Chem. Soc., 1993, 115, 8230.
9. J. K. Hiño, L. Della Ciana, W. J. Dressick and P. Sullivan, Inorg. Chem., 1992, 31, 1072.
10. L. A. Worl, R. Duesing, P. Chen, L. Della Ciana and T. J. Meyer, J. Chem. Soc., Dalton Trans., 1991, 849.
11. J. V. Caspar and T. J. Meyer, J. Phys. Chem., 1983, 87(6), 952.
12. A. Woźna and A. Kapturkiewicz, Phys. Chem. Chem. Phys., 2015, 17, 30468.
13. Top. Curr. Chem., ed. V. Balzani and S. Campagna, 2007, vol. 280.
14. Top. Curr. Chem., ed. V. Balzani and S. Campagna, 2007, vol. 281.
15. E. Schutte, J. B. Helms, S. M. Woessner, J. Bowen and B. P. Sullivan, Inorg. Chem., 1998, 37, 2618.
16. J. L. Smithback, J. B. Helms, E. Schutte, S. M. Woessner and B. P. Sullivan, Inorg. Chem., 2006, 45, 2163.
17. Y. Shen, B. Maliwal and J. Lakowicz, J. Fluoresc., 2001, 11, 315.
18. Y. Reece and D. G. Nocera, J. Am. Chem. Soc., 2005, 127, 9448.
19. H. Tsubaki, A. Sekine, Y. Ohashi, K. Koike, H. Takeda and O. Ishitani, J. Am. Chem. Soc., 2005, 127, 15544.
20. A. Kamecka, K. Prachnio and A. Kapturkiewicz, J. Lumin., 2018, 203, 409.
21. S. Sato and O. Ishitani, Coord. Chem. Rev., 2015, 282–283, 50.
22. C. D. Windle and R. N. Perutz, Coord. Chem. Rev., 2012, 256, 2562.
23. Q. Zhao, F. Li and C. Huang, Chem. Soc. Rev., 2010, 39, 3007.
24. K. K.-W. Lo, K. Y. Zhang and S. P.-Y. Li, Eur. J. Inorg. Chem., 2011, 3551.
25. K. K.-W. Lo, Acc. Chem. Res., 2015, 48, 2985.
26. L. C.-C. Lee, K.-K. Leung and K. K.-W. Lo, Dalton Trans., 2017, 46, 16357.
27. G.-W. Zhao, J.-H. Zhao, Y.-X. Hu, D.-Y. Zhang and X. Li, Synth. Met., 2016, 212, 131.
28. Y. Yamazaki, H. Takeda and O. Ishitani, J. Photochem. Photobiol., C, 2015, 25, 106.
29. A. Kamecka and A. Kapturkiewicz, Phys. Chem. Chem. Phys., 2015, 17, 23332.
30. A. Kamecka, K. Suwińska and A. Kapturkiewicz, Phys. Chem. Chem. Phys., 2016, 18, 28982.
31. A. Kamecka, V. Muszyńska and A. Kapturkiewicz, J. Lumin., 2017, 192, 842.
32. A. Ito and T. J. Meyer, Phys. Chem. Chem. Phys., 2012, 14, 13731.
33. Y. Shimura, Bull. Chem. Soc. Jpn., 1988, 61, 693.
34. P. Chen and T. J. Meyer, Inorg. Chem., 1996, 35, 5520.
35. R. A. Marcus, J. Phys. Chem., 1990, 94(12), 4963.
36. M. Bixon, J. Jortner and J. W. Verhoeven, J. Am. Chem. Soc., 1994, 116, 7349.
37. J. W. Verhoeven, T. Scherer, B. Wegewijs, R. M. Hermant, J. Jortner, M. Bixon, S. Depaemelaere and F. C. De Schryver, Recl. Trav. Chim. Pays-Bas, 1995, 114, 443.
38. M. Senoue, T. Iwaki, K. Seki and M. Yagi, J. Photochem. Photobiol., A, 1996, 101, 257.
39. M. A. A. Rashid, B. Saad, E.-S. M. Negim and M. I. Saleh, World Appl. Sci. J., 2012, 17, 958.
40. J. Higuchi, K. Suzuki, H. Arai, A. Saitoh and M. Yagi, J. Phys. Chem., 1986, 90, 1270.
41. R. A. Marcus, J. Chem. Phys., 1956, 24, 1261.
42. R. A. Marcus, Annu. Rev. Phys. Chem., 1964, 15, 155.
43. R. A. Marcus, J. Phys. Chem., 1989, 93, 3078.
44. Electron Transfers – From Isolated Molecules to Biomolecules: Advances in Chemical Physics, ed. J. Jortner and M. Bixon, vol. 106–107, Wiley, New York, 1999.
45. Electron Transfer in Chemistry, ed. V. Balzani, vol. 1, Wiley-VCH, Weinheim, 2001.
46. J. Cortes, H. Heitele and J. Jortner, J. Phys. Chem., 1994, 98, 2527.
47. J. Jortner, J. Chem. Phys., 1976, 64, 4860.
48. M. Bixon and J. Jortner, J. Phys. Chem., 1991, 95, 1941.
49. R. S. Mulliken, J. Am. Chem. Soc., 1952, 74, 811.
50. R. S. Mulliken, J. Phys. Chem., 1952, 56, 801.
51. N. S. Hush, Prog. Inorg. Chem., 1967, 8, 391.
52. N. S. Hush, Electrochim. Acta, 1968, 13, 1005.
53. K. A. Walters, Y.-J. Kim and J. T. Hupp, Inorg. Chem., 2002, 41, 2909.
54. A. B. Maurer, E. J. Piechota and G. J. Meyer, J. Phys. Chem., 2019, 123, 8745.
55. J. V. Caspar and T. J. Meyer, J. Am. Chem. Soc., 1983, 105, 5583.
56. S. E. Angell, Y. Zhang, C. W. Rogers, M. O. Wolf and W. E. Jones Jr, Inorg. Chem., 2005, 44, 7377.
57. J.-F. Guillemole, V. Barone, L. Joubert and C. Adamo, J. Phys. Chem. A, 2002, 106, 11354.
58. M. E. Casida, C. Jamorski, K. C. Casida and D. R. Salahub, J. Chem. Phys., 1989, 108, 4439.
59. S. Miertus, S. Scrocco and T. Tomasi, Chem. Phys., 1981, 55, 117.
60. P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 299.
61. Z.-Z. Xie and W.-H. Fang, J. Mol. Struct.: THEOCHEM, 2005, 717, 179.
62. L. Yang, A.-M. Ren, J.-K. Feng, X.-J. Liu, Y.-G. Ma, M. Zhang, X.-D. Liu, J.-C. Shen and H.-X. Zhang, J. Phys. Chem. A, 2004, 108, 6797.
63. R. Heydová, E. Gindensperger, R. Romano, J. Sýkora, A. Vlček Jr, S. Záliš and C. Daniel, J. Phys. Chem. A, 2012, 116, 11319.
64. A. El Nahhas, C. Consani, A. M. Blanco-Rodríguez, K. M. Lancaster, O. Braem, A. Cannizzo, M. Towrie, I. P. Clark, S. Záliš, M. Chergui and A. Vlček Jr, Inorg. Chem., 2011, 50, 2932.
65. B. Machura, M. Wolff, M. Jaworska, P. Lodowski, E. Benoist, C. Carrayon, N. Saffon, R. Kruszynski and Z. Mazurak, J. Organomet. Chem., 2011, 696, 3068.
66. F. Zhao, J.-X. Wang, W.-Q. Liu and Y.-B. Wang, Comput. Theor. Chem., 2012, 985, 90.
67. B. J. Powell, Sci. Rep., 2015, 5, 10815.
68. A. C. Jacko, B. J. Powell and R. H. McKenzie, J. Chem. Phys., 2010, 133, 124314.
69. J. Kotlicka and Z. R. Grabowski, J. Photochem., 1979, 11, 413.
70. G. Baryshnikov, B. Minaev and H. Ågren, Chem. Rev., 2017, 117, 6500.
71. H. Yersin, Highly Efficient OLEDs with Phosphorescent Materials, Wiley-VCH, Weinheim, 2008.
72. E. Y.-T. Li, T.-Y. Jiang, Y. Chib and P.-T. Chou, Phys. Chem. Chem. Phys., 2014, 16, 26184.
73. T. J. Penfold, E. Gindensperger, C. Daniel and C. M. Marian, Chem. Rev., 2018, 118, 6975.
74. M. R. Gonçalves and K. P. M. Frin, Polyhedron, 2017, 132, 20.
75. R. A. Velapoldi, in National Bureau of Standards Special Publication 378, Accuracy in Spectrophotometry and Luminescence Measurements. Proc. Conf. NBS, Gaithersburg, MD, 1972, p. 231.
76. D. W. Marquardt, J. Soc. Ind. Appl. Math., 1963, 11, 431.
77. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09, Revision D.01, Gaussian, Inc., Wallingford CT, 2013.

---