Introduction

Phosphorescent transition-metal complexes have recently caught significant attention because of their unique photophysical properties, which are useful in applications, such as dopants for OLEDs (organic light-emitting diode) [1,2,3,4,5], light-emitting electrochemical cells (LECs) [6,7,8,9], dye-sensitized solar cells [3, 10], water splitting [11, 12], and biological phosphorescent labels and sensors [13, 14]. Those complexes exhibit phosphorescence due to very strong spin–orbit coupling, which causes mutual isoenergetic transition from the singlet to the triplet state and conversely, known as intersystem crossing (ISC) [15]. Many families of heavy metal complexes, such as Os(II) [16], Ru(II) [17], Pt (II) [17], and Ir(III) [18,19,20], have been extensively investigated with the aim of better understanding their photophysical properties. In particular, cyclometalated iridium(III) complexes are very promising for a large range of luminescence-based applications because of their high photoluminescence quantum yields, relatively short excited-state lifetime, and general thermal and electrochemical stability [21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59]. Another interesting feature is the possibility of tuning the emission energy of Ir(III) complexes from blue to red light over the entire visible range, which is a key step for realizing the full-color displays and large-area solid-state lighting in OLED fields [4, 5, 19]. Emission color tuning is possible by varying emitter’s HOMO–LUMO gap [60], either by using various ligand cores or playing with acceptor/donor character of the substituents on the main or the ancillary ligands.

In recent years, a number of intensive studies have been carried out to design and synthesize phosphorescent materials for highly efficient OLEDs. Nevertheless, most of the studies concerning this subject have been focused on the synthesis and photophysical properties of neutral and ionic iridium(III) compounds, those stabilized with various types of cyclometalated ligands built on the basis of 2-phenylpyridine (ppy) cores [61]. However, there has been little attention paid to iridium(III) compounds containing cyclometalating ligands, such as benzo[h]quinoline (bzq), which is an analogue of ppy [62]. This fact has prompted us to focus on homoleptic iridium(III) compounds with bzq ligands. They are termed homoleptic when three identical chelating ligands are present in the complex structure as illustrated below.

Furthermore, there are two possible isomers, facial and meridional (fac and mer), for such complexes having three identical but unsymmetrical bidentate ligands [63,64,65,66] (Fig. 1).

Fig. 1
figure 1

Chemical formulas of Ir(III) complexes examined

Full size image

In this work, we present a series of iridium(III) complexes with three cyclometalating ligands based on bzq. We report herein an extensive theoretical investigation of the structural, electrochemical, and photophysical properties of these compounds. A systematic comparison of fac and mer isomer pairs aimed at identification of the differences in their stability and photophysical properties is presented. In addition, as HOMO and LUMO energy levels are likely to be affected by the substitution effects, we were interested in red or blue shifts in the emissions of these fac and mer Ir(III) complexes.

Computational details and theory

The ground-state geometries were fully optimized using the density functional theory (DFT) [67] with B3LYP [68,69,70], M06 [71], and WB97XD [72] functionals. These methods were selected on the basis of the results from extensive comparative studies, they are also recommended for robust and fast calculations for large organometallic compounds [68,69,70,71,72]. The 6-31G(d,p) basis set was used for H, C, N, O, and F atoms [73] and LANL2DZ basis set was adopted for the Ir atom [74]. A relativistic effective core potential (ECP) on Ir replaced the inner core electrons leaving the outer core (5s25p6) electrons and the valence electrons (5d6) of Ir(III). There were no symmetry constraints on these Ir(III) complexes during the geometry optimizations. Vibrational analyses for the optimized structures of isolated molecules were performed to verify if a given structure corresponded to potential energy minima and to calculate zero-point vibrational energies, entropies, and thermal corrections for Gibbs free energies. Solvent–solute interactions were taken into account with the aid of the polarizable continuum model (PCM) [75,76,77] and acetonitrile or dichloromethane as the solvent usually used in experimental studies [78, 79]. All calculations were carried out with Gaussian09 software package [80] in PL-Grid infrastructure.

In this paper, we focused our attention on several R-substituted (R = H, F, OMe, OPh, NMe2, C6F5, and p-C6H4-NPh2) Ir(5-R-bzq)3 complexes in both fac and mer isomeric forms. These tris–cyclometalated Ir(III) complexes were studied to explore preferences toward fac or mer isomers of Ir(III) complexes.

Results and discussion

Geometries in the ground state (S0)

The representative optimized structures of fac-[Ir(bzq)3] and mer-[Ir(bzq)3] in the ground state (S0) along with the numbering of some key atoms are shown in Fig. 2. In order to investigate the solvent effect, the ground-state geometry optimization was also carried out within the PCM method [75,76,77]. The selected optimized geometry parameters for [Ir(bzq)3] in the gas phase and the acetonitrile environment are summarized in Table 1S (Supplementary materials).

Fig. 2
figure 2

Optimized geometries of fac-[Ir(bzq)3] and mer-[Ir(bzq)3] in the ground state

Full size image

Table 1S (Supplementary materials) illustrates the parameters of Ir–ligand bond lengths and bond angles in the gas phase and CH3CN media for [Ir(bzq)3]. Calculated Ir–N, Ir–C, and Ir–O bond lengths are slightly larger in the gas phase than in the acetonitrile environment. The maximum deviation in bond distances between coordinating atoms and Ir(III) is 0.008 Å, while changes in bond angles are less than 1.0°. However, the geometries of all complexes present similar features, indicating that the solvent environment has little effect on the geometry of the complexes.

The optimized structures showed the expected pseudo-octahedral coordination geometry around the iridium center (the corresponding parameters provided as Supplementary material). According to Fig. 2 and Table 1S, two coordinated atoms (N(2) and C(3)) in the fac isomer are in trans position and the valence angle of N(2)–Ir–C(3) is nearly 180°, while the coordinating atoms (N(3) and C(2)) are in cis position and the valence angle value N(3)–Ir–C(2) is close to 90°. Furthermore, two valence angles between the coordinating atoms from the same ligand and the central iridium atom N(2)–Ir–C(1) and N(1)–Ir–C(3) are nearly identical, ca. 90° (Table 1S and Fig. 1). Likewise, the valence angles N(2)–Ir–N(1) and C(1)–Ir–C(3) in the mer isomer are also ca. 90°. For fac-[Ir(bzq)3], Ir–N bonds lengths (Ir–N1, Ir–N2, Ir–N3) are in the range of 2.195 – 2.199 Å, while Ir–C bonds (Ir–C1, Ir–C2, Ir–C3) vary from 2.033 to 2.035 Å (Table 1). It was found that Ir–C bond lengths are significantly shorter by ∼0.15 Å compared to Ir–N bonds. A similar electronic environment of Ir–C/Ir–N bonds leads to their similar bond distance (0.002 Å and 0.004 Å). It is also notable that the change of substituent (R = H, F, OMe, OPh, NMe2, C6F5, and p-C6H4-NPh2) in position 5 in bzq ligand very slightly affects Ir–C and Ir–N bond lengths of fac complexes containing these functionalized ligands.

Table 1 Selected calculated bond distances (Å) in the ground states (S0) for the studied complexes
Full size table

In the case of complex mer-[Ir(bzq)3], Ir–C and Ir–N bond lengths are different from those in the fac isomer. In mer configuration, two donor carbon atoms (C1 vs. C2) and two donor nitrogen atoms (N2 vs. N3) adopt a trans orientation, leaving the remaining C3 and N1 atoms trans to each other. The different local environments result in different bond lengths for these Ir–C and Ir–N bonds. For instance, owing to the strong trans influence of carbon atom C3, Ir–N bond (Ir–N1) trans to it is much longer (about 0.13 Å) than the other two Ir–N ones (Ir–N2, Ir–N3, see Table 1 and Fig. 2). It was also found that Ir–C and Ir–N bond distances follow the trend: Ir–C1 > Ir–C2 > Ir–C3 and Ir–N1 > Ir–N3 > Ir–N2. Moreover, hardly any variation in Ir–C and Ir–N bond lengths (the differences are 0.005 Å and 0.004 Å) was observed upon the substituent change in position 5 of these mer complexes.

On the basis of the obtained Gibbs free energies for iridium(III) complexes, populations of conformers were calculated using a standard Boltzmann formalism. The percentage of a conformer X is given as:

$$ \%X=\frac{\exp \left(-\frac{{\Delta G}_x^0}{RT}\right)}{\sum_i\exp \left(-\frac{{\Delta G}_i^0}{RT}\right)}\times 100\% $$
(1)

where ∆Gx0 is the relative energy of a conformer X.

The relative free energies of the optimized conformations in the ground state along with their distributions are summarized in Table 2. In all cases, the most energetically preferred structure was the fac configuration. The relative energy of fac-[Ir(5-R-bzq)3] in comparison to the mer-[Ir(5-R-bzq)3] was more favorable from about 8 kcal mol–1 to 9.2 kcal mol–1. As we can see, the fac isomer is more stable than mer, which is consistent with the experimental observations for similar ppy-based complexes [81]. The calculated populations of conformers suggest that virtually only the facial isomers are likely to be observed in the equilibrium.

Table 2 Comparison of fac/mer relative energies in the ground state obtained in B3LYP calculations and population of obtained conformers
Full size table

NMR chemical shifts

The 1H NMR calculations were carried out for the two isomers of complex [Ir(bzq)3] in CH2Cl2 solvent using PCM model and compared to experimental 1H chemical shifts reported in literature [81]. It must be underlined here that experimental data relate to a mixture of isomers [82] because the isolation of pure facial or meridional isomer was not achieved; as reported by Lamansky et al. [81]: “The Ir(bzq)3 product from this reaction is a mixture of fac- and mer-isomers. Several wash cycles (acetone and dichloromethane) cause significant enrichment of the mixture in fac-product but still does not allow isolation of a pure facial complex.”. Therefore, the experimental data was compared with computational data obtained for both isomers (Table 3). The spectral data of the complexes fac-[Ir(bzq)3] and mer-[Ir(bzq)3] have been summarized in Table 2S and Table 3S (Supplementary materials).

Table 3 Calculated 1H NMR chemical shifts (ppm) together with the experimental data for fac-[Ir(bzq)3] and mer-[Ir(bzq)3]
Full size table

As shown in Fig. 1S (Supplementary materials), calculated chemical shifts of 1H NMR spectra for fac-[Ir(bzq)3] and mer-[Ir(bzq)3] are in good agreement with the experimental data, the obtained correlation equaled 0.95. The comparison of calculated and experimental data let us assume that signals in experimental spectrum come from both isomers. In the 1H NMR spectrum of fac-[Ir(bzq)3], eight protons of a single bzq ligand are displayed because the three ligands surrounding the iridium atom are magnetically equivalent. The signals appearing at 8.03 ppm, 7.53 ppm, and 7.39 ppm originate from fac-[Ir(bzq)3] and the other signals originate from mer-[Ir(bzq)3].

Frontier molecular orbitals analysis

Results from numerous literature reports indicated that frontier molecular orbitals analysis constitutes a useful proxy for experimentally recorded photophysical properties of iridium(III) complexes [78, 79, 82,83,84]. It is known that frontier molecular orbitals (FMO) of complex ground state S0 are related to its spectral properties [86]. Emission color of iridium(III) complexes can be adjusted by changing their HOMO–LUMO bandgap, which can be achieved on the course of ligand functionalization with electron-donating and electron-withdrawing substituents [86], and values of HOMO–LUMO gaps predicted for Ir(III) complexes by DFT methods showed surprisingly good correlation with the experimentally recorded values of energies of emitted photons even in the case of phosphorescence, see for example [21, 22, 83,84,85, 87,88,89,90]. Contour plots of frontier orbitals of both [Ir(bzq)3] isomers are depicted in Fig. 3, while the visualizations of complexes 2–7 are collected in Table 4S (Supplementary materials).

Fig. 3
figure 3

Contour plots of HOMOs (bottom) and LUMOs (top) of fac-[Ir(bzq)3] (left) and mer-[Ir(bzq)3] (right)

Full size image

As shown in Fig. 3, HOMO of fac-[Ir(bzq)3] is predominantly localized on the iridium atom and over benzo moieties of the three bzq ligands. Similarly, LUMO is localized mostly on pyrido fragments of the bzq ligands. It is also noteworthy that change of the substituent in position 5 (R = H, F, OMe, OPh, NMe2, C6F5, and p-C6H4–NPh2) causes minor effects on the electron densities of the HOMOs and LUMOs of complexes 1–7 (Supplementary material).

In the case of mer isomers, HOMO dominants residues on Ir metal center and benzo moieties of two bzq ligands containing C1 and C3 donor atoms. LUMO of mer-[Ir(bzq)3] is mainly spread over pyrido moieties of the bzq ligands containing N1 and N2 donor atoms. Similarly, as in the case of fac-[Ir(bzq)3], HOMO and LUMO distributions seem not to be very sensitive to substituents influence.

Energy levels of frontier orbitals and HOMO–LUMO energy gap for studied complexes are plotted in Fig. 4. It can be seen that the impact of R substituent variation on HOMO and LUMO energy levels for fac-[Ir(5-R-bzq)3] series follows the same trend as in the case of mer-[Ir(5-R-bzq)3] series. In general, electron-withdrawing substituents (R = F, C6F5) are responsible for lowering energy levels of HOMO and LUMO, but does not extend the HOMO–LUMO energy gap in comparison to unmodified [Ir(bzq)3]. According to that, electron-donating groups (R = OMe, OPh, NMe2, and p-C6H4–NPh2) cause a higher increase of HOMO and LUMO levels in reference to unmodified [Ir(bzq)3]. In addition, such substituents induce more destabilization of HOMO than LUMO, resulting in smaller HOMO–LUMO energy gaps (Eg).

Fig. 4
figure 4

Energy levels and HOMO–LUMO energy gaps of the studied complexes (in eV) calculated at the B3LYP/LANL2DZ/6-311++G(d,p) level of theory

Full size image

Moreover, it should also be pointed out that LUMO levels of complexes in fac configuration are similar to their mer analogues, while HOMO energy is decreased in comparison to mer isomers, thereby leading to larger energy gaps of fac-[Ir(5-R-bzq)3] than the corresponding mer-[Ir(5-R-bzq)3] (Fig. 4). The energies of FMOs calculated with the use of different methods were listed in Tables 5S–7S in Supplementary materials.

Conclusions

Electronic structures and photophysical properties of facial and meridional Ir(III) complexes (fac-[Ir(5-R-bzq)3] and mer-[Ir(5-R-bzq)3]) series were examined with the use of density functional theory. For all of the studied iridium complexes, a greater thermodynamic stability of fac isomers in reference to mer isomers were observed. DFT calculation with different functionals (B3LYP, M06, and WB97XD) showed approximately 9 kcal mol–1 stabilization of total energy in favor of fac isomers. The differences in the ligand orientation around the metal center causes the HOMO energies of meridional isomers to be higher than those of facial forms, while LUMO energies are roughly the same. Consequently, HOMO–LUMO energy gaps are wider for fac-[Ir(5-R-bzq)3] than the corresponding mer-[Ir(5-R-bzq)3]. The obtained results show that geometrical structures of the complexes are hardly affected by change of substituent (R = H, F, OPh, NMe2, C6F5, and p-C6H4–NPh2) in position 5 of cyclometalated ligand, while HOMO and LUMO energy levels are significantly influenced by the substitution effects. In particular, the incorporation of electron-donating substituent p-C6H4–NPh2 leads to decreased HOMO–LUMO gap, thus suggesting red shift on the absorption/emissions spectra lines. In addition, the predicted change of HOMO and LUMO energy levels for both fac and mer forms are similar. Moreover, we suggested which signals in 1H NMR spectra are likely to correspond to fac or mer isomers, which should help in interpretation of experimental NMR spectra.