Charge-Transfer and Spin-Flip States: Thriving as Complements

2.1 Charge-Transfer States

CT transitions involve the redistribution of charge within a molecule. The Coulomb attraction between the separated charges, the depopulation of (weakly) bonding orbitals, and the occupation of antibonding orbitals lead to a geometric distortion of the CT state. In a simplified single configurational coordinate diagram, this can be understood as a horizontal displacement of the potential energy well (Figure 3a). The result is a strong vibronic coupling between the GS and the CT state as well as broad absorption and emission bands (strong coupling limit).⁴⁸ CT transitions are not restricted by parity rules and, thus, feature large transition dipole moments as well as high oscillator strengths, thereby resulting in large molar absorption coefficients ϵ in the range of 10² to 10⁶ M⁻¹ cm⁻¹ and high radiative rates k_r.^{80, 81}

For comparison with SF emitters, it is interesting to note that CT luminescence can occur as fluorescence or phosphorescence. Although ISC to long-lived triplet excited states after initial spin-allowed excitation is very common, often ultrafast, and highly efficient in organometallic CT emitters, there are complexes that show CT fluorescence or thermally activated delayed fluorescence (TADF, see Section 3.2).^{72, 82-84} ISC is facilitated by the heavy atom effect by spin-orbit coupling (SOC) or vibronic coupling,⁸⁵ which also affects the radiative rates k_r of phosphorescence of typically 10⁴ to 10⁶ s⁻¹.⁸⁶

For a detailed discussion, we take a look at [Ru(bpy)₃]²⁺ and its photophysics as the prototypical CT chromophore with CT dynamics (Figures 1 and 4a). [Ru(bpy)₃]²⁺ has a low-spin 4d⁶ electron configuration (t_2g)⁶ with a singlet ground state (¹GS). Blue and green light excites the complex to its ¹MLCT states. Quantitative ISC to the ³MLCT states occurs within 15±10 fs thanks to strong SOC.^{88, 91} During and after vibrational cooling (VC) within 300 fs,^{88, 92, 93} the complex shows a broad phosphorescence, which reaches a maximum at 620 nm with a lifetime of 806 ns and an emission quantum yield of 9.5 % in acetonitrile under deaerated conditions (Figure 5a, Table 1).^{26, 28, 94}

Apart from the MLCT states, inter-configurational ³MC and ⁵MC states with electrons occupying M−L antibonding e_g* orbitals can play an important role. Population of these ³MC/⁵MC states leads to large geometric distortions or even ligand dissociation and facilitates nonradiative deactivation of the ³MLCT states.^{57, 89, 95, 96} The MC energies are determined by the ligand-field splitting and the geometric distortion. Strong bpy ligands in combination with the large Ru^II ion place the energy of the relaxed ³MC states about 3600 cm⁻¹ above the energy of the lowest ³MLCT state.⁸⁷ In this case, thermally activated nonradiative relaxation via these ³MC states is a relevant deactivation pathway for the ³MLCT states at room temperature.⁸⁷ The ⁵MC states in [Ru(bpy)₃]²⁺ are at very high energy due to the large ligand-field spitting imposed by the 4d ion and consequently play no further role.³²

As the 3d transition metal ions have a much lower intrinsic ligand-field splitting, CT states can efficiently relax nonradiatively via the low-energy ³MC and ⁵MC states.³² For example, [Fe(bpy)₃]²⁺ (Figure 2) is nonluminescent, as its ³MLCT state evolves to the ^3/5MC states within 50 fs and decays back to the ¹GS via the ⁵MC state within 650 ps.^{99, 100} Only recently, ³MLCT emissions with 3d⁶ metals such as Cr⁰, Mn^I, and Fe^II were achieved by using very strong ligands that raised the energy of the ^3/5MC states sufficiently to prevent efficient depopulation of the ³MLCT state, thereby allowing competitive phosphorescence and photocatalysis.^101-106

A different strategy to avoid this problem entirely is using metal centers with a d⁰ or d¹⁰ electronic configurations such as Ti^IV, Zr^IV, or Cu^I.^{10, 15, 107-115} As a consequence of their full or empty d-shells, no excited MC states exist. For linear Cu^I complexes, nonradiative deactivation of the LL′CT states could be shut down completely, thereby resulting in emission quantum yields of >99.9 % in solution.¹¹⁶

2.2 Spin-Flip States

In contrast to CT states, SF transitions effect the smallest possible change in the electronic structure of a molecule: the flipping of a single electron spin. Since the electron configuration of the GS is preserved, no bonding orbitals are depopulated and no antibonding orbitals are occupied, SF states are only weakly distorted, and their potential wells are nested with the GS potential (weak coupling limit, Figure 3b). Therefore, SF emission bands are generally sharp (Figure 5).⁴⁹ Exceptions may occur when several individual SF transitions overlap¹¹⁸ or when unsymmetric (ungerade) modes are activated in centrosymmetric complexes to enable the emission.^{66, 119}

SF transitions might bear some similarity to transitions observed in electron paramagnetic resonance (EPR) spectroscopy, as in both cases the relative orientation of unpaired electron spins is important. However, EPR probes transitions between states that mainly arise from differences in the orientation of the electron spins relative to an external magnetic field (different spin quantum number m_S),¹²⁰ whereas SF transitions occur between electronic states with a different orientation of the electron spins with respect to each other (different total spin S).

Transition dipole moments of MC transitions are generally small. Moreover, SF transitions are spin and Laporte forbidden, which is especially pronounced in centrosymmetric complexes. Consequently, radiative rates k_r are low (20–200 s⁻¹), leading to very long emission lifetimes of up to ms provided that the nonradiative rates are also low.^{49, 53, 61, 66}

The nature of SF states requires a spin change by ΔS=−1 with respect to the GS (which obeys Hund's rule of maximum multiplicity) and thus an ISC process after the initial spin-allowed excitation to the Franck–Condon state.⁴⁹ Alternatively, SF states can be directly populated by energy-transfer (EnT) processes (see Sections 4.3 and 5).^{121, 122} The prime example is ¹O₂, which is commonly sensitized by EnT^{123, 124} and shows a weak long-lived SF phosphorescence at 1275 nm from its lowest excited singlet state (¹Δ_g) with two electrons paired in a π orbital.^{125, 126}

As a prototypical SF emitter, we selected [Cr(ddpd)₂]³⁺, which has the nickname “molecular ruby”. This chromium(III) complex possesses a d³ electron configuration, which gives a ⁴A₂ GS with three unpaired electrons occupying the t_2g orbitals (Figure 4b). The complex can be excited to the inter-configurational metal-centered ⁴T₂ state by irradiation at 435 nm.⁵⁹ Ultrafast ISC (<200 fs) from vibrationally hot ⁴T₂ states to the intra-configurational doublet excited states is facilitated by a large density of ²T₂ states in this energy region. Vibrational cooling (VC) and internal conversion (IC) within 3.2 ps yield the two long-lived doublet SF states of ²T₁ and ²E character.⁶¹ Whereas the three electrons are evenly distributed in the t_2g orbitals in the ²E state but with one spin flipped, two of the electrons are paired in the ²T₁ state (Figure 4b), so these states differ in orbital angular momentum. Please note that the orbitals and terms are given in the O notation, but in fact the D₂ symmetry of the complex splits the E and T terms into two and three states, respectively.⁶¹ Interestingly, the electronic situation of the lowest doublet states of [Cr(ddpd)₂]³⁺ is similar to that in singlet oxygen. The two singlet states in ¹O₂, ¹Δ_g and ¹Σ_g⁺, feature two electrons in π orbitals that are spin-paired in the same orbital or occupy different orbitals with antiparallel spin, respectively. However, the spin-paired ¹Δ_g state is markedly stabilized by its favorable orbital angular momentum.¹²⁶

The molecular ruby shows a sharp dual phosphorescence reaching maxima at 738 and 775 nm with a common lifetime of 1.12 ms and a quantum yield of 13.7 % from the thermally equilibrated lowest ²T₁ and ²E derived states in deaerated acetonitrile (Figure 5b, Table 1).^{59, 60} Residual fluorescence from the distorted ⁴T₂ states is very weak (Φ=0.01 %), in agreement with the fast ISC process.⁶¹

For a 3dⁿ metal complex (n<10), the emission lifetime, quantum yield, as well as the photostability of [Cr(ddpd)₂]³⁺ are exceptionally high and achieved by a large ligand-field splitting, which raises the energy of the ⁴T₂ states above the ²T₁/²E states in the Franck–Condon geometry. This large energy gap prevents thermally activated back-ISC occurring to the ⁴T₂ states followed by nonradiative deactivation, (delayed) fluorescence, or dissociation.^{59, 61, 62}

3.1 Tuning of the Excited-State Energy

The simplest way to modulate the energy of CT states is by changing the (solvent) environment. As a result of their charge-separated nature, CT states are usually quite sensitive to the surrounding matrix (solvatochromism, see Section 4.3 for more details). Furthermore, the emission energy of CT emitters can be tuned by modifications of the complex changing the energy gap between the donating and accepting orbitals (Figure 6a).^{31, 127} In MLCT transitions, an electron is formally transferred from a d orbital to a ligand's π* orbital, which results in an oxidized metal center and a ligand radical anion. Consequently, the excited state energy can be tuned by using a different metal center or by modifying the ligands.^{31, 128-130}

Changing the metal center of a complex results in different energies for the donating d orbitals. For example, using Os^II instead of Ru^II yields lower MLCT energies because of the higher energy of the 5d orbitals compared to the 4d orbitals (Figure 6a).¹²⁸

In principle, it is expected that the energy of an MLCT state will decrease on introducing electron-withdrawing substituents that lower the accepting π* orbital energy and will increase with electron-donating substituents that raise the π* orbital energy. However, ligand modifications also affect the energy of the d orbitals. For example, an electron-donating substituent also increases the energy of the metal's d orbitals in addition to destabilizing the ligands’ π* orbitals.¹²⁸ Introducing two diethylamino substituents onto the bpy ligands red-shifts the emission band to 700 nm for the homoleptic ruthenium(II) complex because the destabilizing effect on the d orbitals is stronger than on the π* orbitals, thereby leading to a lower energy gap between the two levels.¹³¹

As a consequence of this intricate interplay between the ligand and metal center outlined above, careful control of the donating and accepting orbital energies is needed to achieve MLCT emission in the near-infrared (NIR) spectral region. For example, a combination of electron-rich ligands to destabilize the d orbitals and an electron-poor ligand to provide a low-energy π* accepting orbital enabled NIR phosphorescence to be achieved.¹²⁸ Other designs include extended aromatic systems on the ligands and polynuclear complexes, amongst others.^{129, 130, 132-134} In any case, the low-energy excited states required for NIR emission suffer from enhanced rates of nonradiative deactivation according to the energy gap law, which negatively impacts lifetimes and quantum yields.¹³⁵ In contrast, efficient green or even blue emission from CT states is readily achieved, for example from complexes derived from Ir(ppy)₃ (Figure 6a).^{25, 136, 137}

The typical spectral range for SF emission is deep-red to NIR-I (780–1000 nm) or even NIR-II (1000–1700 nm).⁴⁹ No charge separation and no (or only a slight) change in orbital occupation occurs in SF transitions. Instead, the energy of SF states results from additional repulsive interactions between the electrons in nearly degenerate orbitals caused by a lower exchange energy, which are often quantified with the Racah parameters B and C derived from ligand-field theory.^{64, 81, 141, 142} Hence, strategies for tuning emission energy are less obvious than in the CT case.⁴⁹

Successful strategies are changing the covalency of the metal–ligand bonds and varying the size of the d orbitals. More covalent M−L bonds lower the energies of the SF state because the interelectronic repulsion is reduced by delocalization of the electrons onto the ligands (nephelauxetic effect). This principle was demonstrated with Cr^III complexes containing carbanionic ligands such as ppy⁻ (Figure 6b) or amido ligands leading to NIR emission band maxima up to 1067 nm.^{139, 142, 143}

More ionic M−L bonds result in higher emission energies, but this is difficult to achieve in molecular systems. Recently, [Cr(bpmp)₂]³⁺ (Figure 2, bpmp=2,6-bis(2-pyridylmethyl)pyridine) was reported, which emits at 709 nm (1.75 eV) with a high quantum yield of 19.6 % and a millisecond lifetime. The ligand design was guided by high-level quantum mechanical calculations.⁶³

The central metal ion profoundly affects the energy of the SF state. Metal centers with larger d orbitals (lower effective charge, higher period) reduce the interelectronic repulsion and hence the SF state energy, while those with small d orbitals (higher effective charge, lower period) increase the SF energy.⁴⁹ For example, the Racah parameter B of the free ions qualitatively predicts that V^III [B(V³⁺)=861 cm⁻¹] and Mo^III complexes [B(Mo³⁺)=610 cm⁻¹] should possess lower energy SF states than Cr^III complexes [B(Cr³⁺)=918 cm⁻¹] in similar environments and assuming a similar ratio C/B.^{49, 81} In fact, [V(ddpd)₂]³⁺ (Figure 2) and MoCl₃(urea)₃ emit between 982–1109 nm⁷² and at 1095 nm,¹⁴⁴ respectively, while the emission from Cr^III complexes is usually located between 720 and 800 nm.^{48, 49, 53, 58} We would like to point out that the Racah parameters B and C should be used with caution for quantitative comparisons, as the assumptions in the underlying ligand-field theory (ligands treated as point charges, ideal coordination geometries)⁸¹ might not be valid, especially for complexes with low symmetry and highly anisotropic bonding situations.

In contrast to CT emitters, it is challenging to blue-shift SF emission by varying the metal center. One problem is that ions in high oxidation states such as Mn^IV with contracted d orbitals can form highly covalent M−L bonds, eventually leading to low-energy SF states.⁷⁸ Additionally, very high or low oxidation states of the metal center can also introduce low-energy CT states, which can facilitate nonradiative deactivation, for example through mixing with the luminescent SF states.^{49, 78, 145}

3.2 Tuning the Emission Lifetime

An efficient way of prolonging τ is rigidification of the ligand scaffold (Figure 7). This can decrease k_nr through the restriction of intramolecular motions in the excited state. This strategy succeeded for CT, SF, and organic emitters alike.^{62, 148-150} For example, the phen complexes of Ru^II and Cr^III (Figure 2) have emission lifetimes that are longer by factors of about 1.5 and 4, respectively, compared to their bpy analogues.^{31, 55} Furthermore, Cr^III complexes with tris(bidentate) chelation such as [Cr(bpy)₃]³⁺ show a trigonal distortion in the excited state that facilitates nonradiative deactivation of the SF states.^{151, 152} This mode can be shut down by employing tridentate ligands, such as in [Cr(ddpd)₂]³⁺.⁶²

In CT states, the geometric distortion can be reduced by stronger delocalization of the excited state on an extended π-system of the ligand (Figure 7).^{102, 148} This diminishes the impact of the charge separation on individual bonds and, thus, coupling to the GS. In SF emitters, a delocalization of the excited state, for example by mixing with CT states, has the contrary effect: it increases the geometric distortion and facilitates nonradiative deactivation.^{78, 142, 145}

Lifetimes of Ru^II complexes were prolonged by a larger ³MLCT-³MC energy gap limiting nonradiative decay via the detrimental ³MC states (Figure 7). For example, a larger bite-angle of the chelating ligand raises the ligand-field splitting and, thus, the ³MC energy, while electron-withdrawing substituents on the ligands lower the ³MLCT states.¹⁵³ Similarly, increasing the energy of the ⁴T₂ MC states in [Cr(ddpd)₂]³⁺ proved to be key to its remarkable optical properties, as it prevents deactivation of the SF states via back-ISC.^{59, 61, 62}

Deuteration of the ligands or solvents is a particularly useful strategy that lowers k_nr for low-energy luminescence of any kind (Figure 7).^{154, 155} Above an emission wavelength of about 700 nm (<14 300 cm⁻¹), the energy of the electronically excited state is in the region of X−H (X=C, N, O, 3000–3500 cm⁻¹) overtone vibrations of the ligand or solvent, and deactivation through energy transfer becomes relevant. This process requires a spectral overlap between the emission band of the complex and the overtone absorption band. Deuteration mitigates the negative effect of these oscillators, as X−D vibrations are significantly lower in energy than their X−H counterparts [$$ (C−H)≈3000 cm⁻¹, $$ (C−D)≈2200 cm⁻¹], thus requiring a higher overtone to match the emission energy. The lower molar absorption coefficient of the higher overtone leads to a smaller spectral overlap and reduced deactivation. For the molecular ruby, statistical deuteration of the ddpd ligand increases the emission lifetime from 1.12 to 2.3 ms and the quantum yield from 13.7 to 30 %.¹⁵⁶ Since the rate of energy transfer strongly decreases with the M⋅⋅⋅X−H distance (≈r⁻⁶), the α−C−H oscillators are the most important ones. Hence, deuteration of the α−C−H groups of the bpmp ligands in [Cr(bpmp)₂]³⁺ (Figure 2) boosts the lifetime from 1.8 to 2.5 ms and the quantum yield from 19.6 to 24.6 %.⁶³ Similarly, but to a lesser extent, perdeuteration of [Ru(bpy)₃]²⁺ increases the emission lifetime by about 20 % in aqueous solution.¹⁵⁷ The smaller effect arises from the higher emission energy of [Ru(bpy)₃]²⁺.

The lifetime of CT emitters such as [Ru(bpy)₃]²⁺ can be increased by exploiting the so-called reservoir effect (Figure 7). The ligands are decorated with aryl substituents, such as pyrene, that possess similar triplet energies as the parent complex.^{146, 158-162} After photoexcitation and ISC, a thermal equilibrium between the ³MLCT state and the pyrene-based ³(π-π*) state is established through an intramolecular energy transfer. The very long-lived pyrene triplet state (τ=9.4 ms in ethanol)¹⁶³ serves as a reservoir that slowly repopulates the emissive ³MLCT state and leads to microsecond emission lifetimes.¹⁶¹ Adding 9,10-diphenylanthracene (DPA, Figure 8) to a solution of [Cr(bpmp)₂]³⁺ (Figure 2) also prolongs the lifetime through a doublet-triplet energy-transfer equilibrium between the SF states and triplet state of DPA. However, unlike the CT emitters, the lifetime was not increased beyond the value of 1.8 ms found for the complex without additives.¹⁶⁴ As excited-state lifetimes are already very long, this also seems unnecessary as a molecular design strategy for SF emitters.

As a consequence of the spin-forbidden nature of phosphorescence transitions, the possibilities to increase their radiative rates include increasing the metal character of the involved wavefunctions (e.g. by changing the character of the lowest state from LL′CT to MLCT),¹⁶⁵ installing heavy atoms to increase spin-orbit coupling (SOC)^{166, 167} and to increase the density of states which can interact with the emissive state through SOC, for example, by a multi-metal approach.¹⁶⁸ Another strategy that was thoroughly explored for CT emitters is thermally activated delayed fluorescence (TADF, Figure 7).^{84, 109, 169-173} Here, an emissive triplet state is in thermal equilibrium with an energetically higher lying singlet state (ΔE_ST<1600 cm⁻¹)^{174, 175} through back-ISC. The repopulated singlet state decays through fluorescence to the ¹GS with a high radiative rate, as no change in multiplicity is involved.

For Cr^III complexes as prototypical SF emitters, TADF occurs in cases where back-ISC from the ²E/²T₁ states to the ⁴T₂ state is thermodynamically possible.⁵⁶ However, apart from fluorescence, other deactivation pathways of the ⁴T₂ state, such as ligand dissociation and surface crossing, are available due to the strong Jahn–Teller distortion of the ⁴T₂ state.^{176, 177} Hence, enabling back-ISC in Cr^III complexes comes at the cost of poor quantum yields, as discussed above. Therefore, TADF cannot be considered a viable strategy to tune SF emitters, at least with the chromium(III) complexes developed so far.

The metal-confined nature of the SF states enables tuning of the radiative rate constant k_r by exploiting Laporte's rule (Figure 7).¹¹⁹ By introducing an inversion center in [Cr(tpe)₂]³⁺ using the tripodal ligand tpe (=1,1,1-tris(pyrid-2-yl)ethane, Figure 2), the radiative rate of the spin- and Laporte-forbidden ²E_g/²T_1g→⁴A_2g transitions was lowered to merely k_r=18 s⁻¹, thereby resulting in a record luminescence lifetime of 4.5 ms.⁶⁶ This strategy is not available for CT emitters, since the charge separation does not occur in orbitals of the same parity and always leads to symmetry breaking.

4.1 Temperature

Some of the effects of temperature on the ³CT and SF luminescence, such as thermal deactivation via MC states and TADF, have been discussed in the previous sections.

Equilibria can also exist between phosphorescent states of different character. For example, temperature-dependent dual emission arises from an excited-state equilibrium between a ³MLCT and a ligand-centered ³LC state in an [Ir(ppy)₂(N^N)]⁺ complex featuring a pyrimidyltriazole ligand (N^N).¹⁶⁵

In SF emitters such as the molecular ruby, temperature-dependent changes in the number of emission bands and their shape are largely governed by the population of close-lying SF states that arise from the splitting of SF states degenerate in an ideal octahedral symmetry (²E, ²T₂, Figure 4b). As the energy gaps amount to around 200–800 cm⁻¹ with small barriers, the population follows a Boltzmann distribution.^{48, 49, 63, 64, 178} Hence, the molecular ruby has been used as a self-referencing ratiometric optical thermometer.^{178, 179} The three ³MLCT states in derivatives of [Ru(phen)₃]²⁺ (Figure 2) are also split in D₃ symmetry, but only by about 10 and 60 cm⁻¹.^{31, 180}

Although the population of the close-lying SF states varies with temperature, the energies of the individual states generally remain unaffected in most cases^{64, 66, 178} with rare exceptions.⁷³ In contrast, drastic changes in the emission energy and band shape can be observed for CT emitters upon freezing of the solution. This phenomenon, known as rigidochromism, will be discussed in more detail in Section 4.3 as it is not primarily a temperature effect.

4.2 Pressure

Pressure can increase the excited-state lifetimes and photoluminescence quantum yields of some CT and SF emitters. This can be explained by a stronger interaction between the metal center and the ligands under pressure, which increases the ligand-field splitting and thus the energy of deactivating MC states.^{95, 181, 182} For [NH₄]₃[CrF₆] in the solid state, this effect even causes a change in the emission type: at low pressures the complex fluoresces weakly from the inter-configurational ⁴T₂ MC state, while SF phosphorescence occurs at pressures above 71 kbar.¹⁸³ In [Ru(bpy)₃]²⁺, the pressure response of the luminescence intensity is strongly dependent on the environment. While the emission quantum yield increases in acetonitrile solution and doped solids, a decrease in Φ was found in single crystals.^{95, 181, 184, 185} High pressures can restrict vibrational modes and limit nonradiative deactivation, as shown for [Ru(bpy)₂(py)₂]²⁺ with the more flexible pyridine ligands (py).^{186, 187}

Hydrostatic pressure can also affect the energy of the SF emission. The most well-known example is the shift of about −0.8 cm⁻¹ kbar⁻¹ of the SF emission from the gemstone ruby (Al₂O₃ : Cr³⁺), which is exploited for optical measurements of high pressures in diamond anvil cells.^{188, 189} Much stronger shifts of up to −14.8 cm⁻¹ kbar⁻¹ are found in the molecular ruby [Cr(ddpd)₂]³⁺ in solution and the solid state.⁹⁰ This pressure effect on the SF emission is caused by subtle changes in the coordination geometry and the π(M−L) overlap affecting the M−L covalency and the nephelauxetic effect.⁹⁰

For CT emitters, the effect is usually less pronounced. The emission energy of [Ru(bpy)₃]²⁺ shows red-shifts of −2 to −6 cm⁻¹ kbar⁻¹ depending on the matrix and pressure range.^{95, 184, 187} For some CT emitters, the shift can be traced back to pressure-induced changes of solvent parameters (see Section 4.3).¹⁹⁰ A strong shift of −13 cm⁻¹ kbar⁻¹ was found in single crystals of a heteroleptic ruthenium(II) complex, which was ascribed to a pressure-dependent hydrogen bonding of the complex to a salt bridge.¹⁹¹

4.3 Environment

The influence of the solvent on the emission properties is very much tied to the nature of the excited states involved. CT transitions are more sensitive to changes in the environment because of their large change in the dipole moment.¹⁹² SF states only show a minimal electron redistribution with respect to the GS. Here, effects of the matrix are limited to changes in the emission quantum yields and lifetimes,⁴⁹ whereas for CT states the energy can strongly vary with solvent parameters.^193-197 If the excited state has a lower dipole moment than the GS, the emissive state is destabilized in more polar solvents, which leads to a higher emission energy (negative solvatochromism). Excited states that are more polar than the GS show a red-shift in more polar solvents (positive solvatochromism).^{198, 199} For example, the positive solvatochromism of [Ru(bpy)₂(CN)₂]⁺ enables tuning of the emission energy in the range of 623 to 712 nm.²⁰⁰

Apart from solvent, charged complexes can also be affected by interactions with their counterions. For example, variations in ion pairing can open different reaction pathways.^{201, 202} With regards to luminescence, strong effects have been observed for CT emitters.^203-205 The emission quantum yield of an iridium(III) ³MLCT emitter in the solid state could be increased 12-fold by employing bulky counterions. The effect was rationalized by a larger average distance between the complex molecules leading to reduced self-quenching.²⁰⁶

In the molecular ruby, counterions and the solvent influence the emission lifetime and quantum yield, but not the ultrafast ISC and internal conversion (IC) processes after excitation.^{60, 156} For [Cr(bpy)₃]³⁺, it was shown that the counterions and co-solutes help to rigidify the ligand scaffold and reduce nonradiative decay.²⁰⁷

In general, phosphorescence can be quenched by ³O₂ through a Dexter energy transfer (EnT) which leads to the formation of ¹O₂ and significantly reduced excited-state lifetimes (Table 1).^{123, 208, 209 1}O₂ can be a useful reagent, for example, for the α-cyanation of amines^{210, 211} or photodynamic therapy,^{212, 213} but in some cases quenching by O₂ is undesired.²¹⁴ In principle, steric shielding can help to limit the EnT process, as it requires orbital overlap between the energy donor and energy acceptor. For CT emitters with spatially extended excited states, however, such a steric protection is challenging due to large delocalization of the ³MLCT wavefunction onto the ligands.²¹⁴ The SF states in the molecular ruby, on the other hand, can be efficiently protected from external influence by shielding the metal center.²¹⁵

A phenomenon that can be found in CT emitters is rigidochromism. It describes the blue-shift of the emission band upon rigidification of the environment.^{192, 216} Although it can occur during freezing of a solution, rigidochromism is not a temperature effect, as similar shifts can be observed when comparing the emission spectra recorded in a frozen solution and in the solid state at room temperature. For example, the emission of [Ru(bpy)₃]²⁺ shifts from 601 nm at room temperature²¹⁷ to 580 nm in a frozen diethyl ether : isopentane : ethanol solution at 77 K²¹⁸ as well as in the solid state at room temperature.²¹⁹