Triplet Emitting C^N^C Cyclometalated Dibenzo[c,h]Acridine Pt(II) Complexes

In a series of Pt(II) complexes [Pt(dba)(L)] containing the very rigid, dianionic, bis-cyclometalating, tridentate C^N^C2− heterocyclic ligand dba2– (H2dba = dibenzo[c,h]acridine), the coligand (ancillary ligand) L = dmso, PPh3, CNtBu and Me2Imd (N,N’-dimethylimidazolydene) was varied in order to improve its luminescence properties. Beginning with the previously reported dmso complex, we synthesized the PPh3, CNtBu and Me2Imd derivatives and characterized them by elemental analysis, 1H (and 31P) NMR spectroscopy and MS. Cyclic voltammetry showed partially reversible reduction waves ranging between −1.89 and −2.10 V and increasing along the series Me2Imd < dmso ≈ PPh3 < CNtBu. With irreversible oxidation waves ranging between 0.55 (L = Me2Imd) and 1.00 V (dmso), the electrochemical gaps range between 2.65 and 2.91 eV while increasing along the series Me2Imd < CNtBu < PPh3 < dmso. All four complexes show in part vibrationally structured long-wavelength absorption bands peaking at around 530 nm. TD-DFT calculated spectra agree quite well with the experimental spectra, with only a slight redshift. The photoluminescence spectra of all four compounds are very similar. In fluid solution at 298 K, they show broad, only partially structured bands, with maxima at around 590 nm, while in frozen glassy matrices at 77 K, slightly blue-shifted (~580 nm) bands with clear vibronic progressions were found. The photoluminescence quantum yields ΦL ranged between 0.04 and 0.24, at 298 K, and between 0.80 and 0.90 at 77 K. The lifetimes τ at 298 K ranged between 60 and 14040 ns in Ar-purged solutions and increased from 17 to 43 µs at 77 K. The TD-DFT calculated emission spectra are in excellent agreement with the experimental findings. In terms of high ΦL and long τ, the dmso and PPh3 complexes outperform the CNtBu and Me2Imd derivatives. This is remarkable in view of the higher ligand strength of Me2Imd, compared with all other coligands, as concluded from the electrochemical data.

In contrast to the dppy system, in which the C-C single bonds between the phenyl and the pyridyl groups provides some flexibility for deformations (Scheme 1, left), the condensation of the three aromatic rings in the dba 2-(H2dba = dibenzo[c,h]acridine) ligand (Scheme 1, right) led to a far more rigid coordination environment. In our previous study, In contrast to the dppy system, in which the C-C single bonds between the phenyl and the pyridyl groups provides some flexibility for deformations (Scheme 1, left), the condensation of the three aromatic rings in the dba 2− (H 2 dba = dibenzo[c,h]acridine) ligand (Scheme 1, right) led to a far more rigid coordination environment. In our previous study, we found markedly increased PL quantum yields Φ L for the complex [Pt(dba)(dmso)] (0.23 at 298 K in CH 2 Cl 2 solution), compared with derivatives of [Pt(dppy)(dmso)] [33] (no emission under these conditions), which is in line with the idea that molecular rigidity helps to reduce or circumvent radiationless decay paths. This makes these Pt dba complexes interesting candidates for various applications.
Beginning with the previously reported complex [Pt(dba)(dmso)] (1) [33], we sought to improve the photoluminescence properties by exchanging the dmso coligand for PPh 3 , CNtBu (tert-butyl-isocyanide), or Me 2 Imd (N,N'-dimethylimidazolydene) (complexes 2, 3, and 4; Scheme 2). All three of the new coligands are expected to produce a stronger ligand field splitting (or ligand field stabilization energy) due to their superior σ-donating and πaccepting capabilities. This should destabilize the dark d-d* excited states further, and thus lead to superior luminescence quantum yields for the corresponding complexes. In addition to UV-vis absorption and time-resolved photoluminescence spectroscopy, we also studied the electrochemistry of the complexes, using cyclic voltammetry and spectroelectrochemical UV-vis spectroscopy, as well as modelling the excited states and electronic transitions employing (TD)-DFT calculations. The 7-phenyl substituted derivatives [Pt(db(ph)a)(L)] (H 2 db(ph)a = 7-phenyldibenzo[c,h]acridine) were also included in this study for comparison, as the experimental data for L = dmso (complex 5) was already available [33], while further derivatives with L = PPh 3 , CNtBu, and Me 2 Imd (6, 7, and 8) were calculated to probe if the 7-phenyl derivative of dba is potentially more interesting for phosphorescent Pt(II) complexes. All four compounds gave single crystals suitable for XRD structure analysis, and the crystallographic data is shown in Table S1 of the Supplementary Material (SM). All of the structures, with the exception of [Pt(dba)(Me2Imd)] (4), are characterized by pronounced head-to-tail π-stacks of two molecules each (Figures S1 to S3, SM), which is typical for square planar complexes with planar tridentate or tetradentate ligands [19,21,[25][26][27][28][29][33][34][35][36]39]. For 4, the carbene ligand interrupts this interaction and a complex packing motive is observed (Figure 1 and Figure S4). All four compounds gave single crystals suitable for XRD structure analysis, and the crystallographic data is shown in Table S1 of the Supplementary Material (SM). All of the structures, with the exception of [Pt(dba)(Me 2 Imd)] (4), are characterized by pronounced head-to-tail π-stacks of two molecules each ( Figure S1 to Figure S3, SM), which is typical for square planar complexes with planar tridentate or tetradentate ligands [19,21,[25][26][27][28][29][33][34][35][36]39]. For 4, the carbene ligand interrupts this interaction and a complex packing motive is observed ( Figure 1 and Figure S4). the molecular structure in the crystal (B) with thermal displacement elipsoids at 50% probability and including the numbering; H atoms were omitted for clarity. DFT-optimized S0 structure of 4 (C,D) displaying the specific angles listed in Table 1.  1. Crystal structure of [Pt(dba)(Me 2 Imd)] (4) viewed along the crystallographic a axis (A) and the molecular structure in the crystal (B) with thermal displacement ellipsoids at 50% probability and including the numbering; H atoms were omitted for clarity. DFT-optimized S 0 structure of 4 (C,D) displaying the specific angles listed in Table 1. A closer inspection of the π-stacks showed that the centroid ... centroid distance of the central pyridyl ring of [Pt(dba)(dmso)] constitutes approximately 3.62 Å, with an offset of approximately 1.25 Å, which lies in the range of typical π-interactions [45,55,56,58,59]. In contrast, for the PPh 3 and the CNtBu derivatives 2 and 3, the bulky coligands enlarge these centroid ... centroid distances, to 4.53 Å (0.93 Å offset) for 2 and to 5.59 Å (1.12 Å offset) for 3. Therefore, dimerization through stacking in solution might have an impact on the photophysical properties only for the dmso complex (1). On the other hand, our previous study did not give evidence for this phenomenon [33].
The molecular structures show the expected square planar coordination of the d 8 configured Pt(II) (Figure 1). The dba ligand is also completely planar, and the S, P, or C donor atoms of the ancillary ligands complete the planar coordination with N-Pt-X angles (α angle, Figure 1) of around 180 • , while the residual atoms of the coligands fill the space at this fourth coordination site. The planar Me 2 Imd ligand in 4 shows a tilt angle towards the Pt,C,C,N coordination plane of about 58 • (δ angle). The Pt-C-N unit (γ angle) in the CNtBu complex 3 is almost linear ( Table 1).
The optimized geometries of the singlet ground state S 0 ( Figure 1 and Figure S5) and the lowest excited triplet state T 1 ( Figure S6) were calculated by DFT methods. The S 0 geometries are in good agreement with the experimental data from the X-ray diffractometry ( Table 1). The S 0 and T 1 geometries hardly differ between all of the complexes; the largest differences occurred for the dmso and PPh 3 derivatives, where the N-Pt-X and C-Pt-X angles change by a few degrees upon relaxation in the T 1 state. In the case of PPh 3 , the planarity of the complex is broken in the T 1 state, with the Pt(II) center clearly protruding out of the coordination plane. The Pt1-C1-C6-C10 dihedral angle ( Figure 1) measures 7 • .

Electrochemistry and DFT-Calculated Orbitals
On the first view, the complexes show irreversible oxidation processes ranging between around 0.55 and 1 V (Figure 2 and Figure S7, data in Table 2); based on the DFT calculations, they can be assigned to an essentially Pt-centered Pt(II)/Pt(III) couple. The DFT-calculated highest occupied molecular orbitals (HOMO) are energetically ordered in growing order as Me 2 Imd > PPh 3 > CNtBu > dmso ( Figure S8). The electrochemical potentials increase along the same series for the dba complexes, which roughly match the expected increasing σ-donating strength of the coligand. out of the coordination plane. The Pt1-C1-C6-C10 dihedral angle (Figure 1) measures 7°.

Electrochemistry and DFT-Calculated Orbitals
On the first view, the complexes show irreversible oxidation processes ranging between around 0.55 and 1 V (Figures 2 and S7, data in Table 2); based on the DFT calculations, they can be assigned to an essentially Pt-centered Pt(II)/Pt(III) couple. The DFT-calculated highest occupied molecular orbitals (HOMO) are energetically ordered in growing order as Me2Imd > PPh3 > CNtBu > dmso ( Figure S8). The electrochemical potentials increase along the same series for the dba complexes, which roughly match the expected increasing σ-donating strength of the coligand.    A first reduction is observed for all of the complexes at around −2 V. This process is partly reversible for the complexes 2 to 4, but irreversible for the dmso complex 1 ( Figure S7). Regardless of the slightly different reversibility, the very similar potentials point to a largely ligand π*(dba) centered reduction processes. This is in line with the DFT-calculated lowest unoccupied molecular orbitals (LUMO), which show the growing energetic sequence Me 2 Imd > PPh 3 > CNtBu > dmso ( Figure S8); this correlates with the expected higher π-backbonding ability of the newly introduced coligands compared to dmso.
The irreversibility of both the oxidation and reduction processes were explained by rapid chemical reactions (C) following the electrochemical processes (E). If the two steps are timely correlated, this is called an EC mechanism or EC process [31,33,36,60,61]. Thus, secondary yet small, reduction waves found at −2.8 to −3.0 V represent the EC products. The introduction of the pending phenyl group led to markedly higher reduction and oxidation potentials for [Pt(db(ph)a)(dmso)], which is in line with the contribution of the dba core to the HOMO.
The electrochemical gaps increase along the series of dba complexes Me 2 Imd < CNtBu < PPh 3 < dmso following the trend of decreasing ligand strength (Table 2).

UV-Vis Absorption Spectroscopy
The experimental UV-vis absorption spectra of the four [Pt(dba)(L)] complexes are very similar (Figure 3 and Figure S9). Four series of absorption maxima are discernible with very intense bands between 250 and 300 nm, along with two vibrationally structured medium-intense groups in the range between 300 and 420 nm, as well as a fourth structured band system between 450 and 550 nm that tails down to cut-offs (zero absorption) between 570 and 580 nm ( Table 3).  The long-wavelength absorption and cut-off energies are only slightly smaller (by 0.03 to 0.05 eV) for the complexes with the new coligands, 2, 3, and 4, compared with the parent dmso complex 1 (Table 3) and vary only slightly for these three complexes, in contrast to the marked differences in the electrochemical potentials and gaps ( Table 2). The corresponding intensities are very different and the CNtBu complex reaches not even half the intensities of the PPh3 derivative. Nevertheless, the similar band shapes and energies point to comparable excited states; we assume transitions into π-π* configurations for the two high-energy band systems (200 to 350 nm) and transitions into mixed π-π*(LC)/metal(d)-to-ligand(π*)(MLCT) states for the two long-wavelength absorption processes (350 to 550 nm), in line with previous reports [10,11,[16][17][18][19][20][21][45][46][47][49][50][51][52][53][54][55][56]62]. We infer a relatively small metal participation from the rather low impact of the coligand exchange.

TD-DFT Calculated Absorption Spectra
The most intense calculated transitions are found in the range between 225 and 325 nm (Figure 3, left). The maxima for the complexes with L = dmso and Me2Imd show very good agreement with the experimental data, while deviating by only 4 to 9 nm, whereas The λ cut-off is defined as the wavelength of zero absorption and is obtained through extrapolation of the low-energy slope of the lowest energy absorption band.

TD-DFT Calculated Absorption Spectra
The most intense calculated transitions are found in the range between 225 and 325 nm ( Figure 3, left). The maxima for the complexes with L = dmso and Me 2 Imd show very good agreement with the experimental data, while deviating by only 4 to 9 nm, whereas for PPh 3 and CNtBu, the deviation is somewhat more pronounced (ca. 25 nm). The

Time-Resolved Photoluminescence Spectroscopy
At room temperature and in diluted CH 2 Cl 2 solutions, the four dba Pt(II) complexes 1 to 4 show a red luminescence with a broad band centered at around 600 nm ( Figure  S10), most likely arising from the mixed 3 LC/ 3 MLCT excited states; this is in line with previous reports [10,11,[16][17][18][19][20][21][45][46][47][49][50][51][52][53][54][55][56]62], as well as with our TD-DFT calculations (vide infra). The change of the monodentate coligand does not substantially influence the emission profile of the complexes; only the vibrational progression is marginally affected. The absence of 3 O 2 in the Ar-purged solutions generally increased the lifetime of the complexes, compared to the air-equilibrated solutions, in line with a triplet character of the luminescence ( Table 4). The most remarkable finding was that the replacement of L = dmso by the stronger ligands PPh 3 , CNtBu, and Me 2 Imd had almost no effect on the emission energies and a rather detrimental effect on Φ L , which decreases along the series of coligands dmso >> CNtBu >> Me 2 Imd > PPh 3 and is generally lower for the complexes with stronger coligands L than for the parent complex [Pt(dba)(dmso)]. In the more concentrated solutions, we observed red-shifted emissions which were likely a result of the aggregation of the molecules. However, in this report we wanted to focus on the luminescence properties of the isolated molecules and avoided aggregation by using low concentrations (c = 10 −5 M). In future studies, we will investigate the aggregation of these complexes in solution and in solid materials.
.279 ± 0.008 Me 2 Imd (4) Uncertainties of τ are given by the decay fitting. n.d. = not determined, due to bad signal-to-noise ratios.
The short-lived component of the adjusted biexponential decays for τ in air-equilibrated CH 2 Cl 2 solutions are due to traces of the free H 2 dba ligand (τ dba ≈ 5 ns) (Table 4). However, in most cases, this effect was not seen in Ar-purged solution due to the vast increment of Φ L for the complexes under these conditions. The fluorescence quantum yield of dba is not negligible in CH 2 Cl 2 at 298 K (Φ L = 0.44); thus, even traces of the ligand are detected in case of the poorly phosphorescent complexes in the presence of oxygen. Therefore, a trace amount of the ligand below 0.01% might escape the analysis by NMR, elemental analysis and other methods, while still being detectable through intrinsic fluorescence. Importantly, the studied dba Pt(II) complexes were purified by column chromatography, and NMR spectra show no signals for the H 2 dba precursor (with the exception of the PPh 3 complex, see Figures S31-S35 in the Supporting Material). We assume that, in addition to the complexes as molecular species, small stacks (see crystal structures) of complexes with the ligand precursor H 2 dba are also eluted. Our findings also emphasize that special care has to be taken when measuring weakly phosphorescent complexes stemming from a brightly fluorescent ligand precursor, in order to avoid potential confusion with dual emission [55].
In frozen glassy matrices at 77 K, the charge-transfer states are destabilized, which lowers the MLCT character of the emissive T 1 state, while causing a blue-shifted photoluminescence (576 nm for the dmso complex). Therefore, the observed emission at 77 K (Figure 4) predominantly originates from ligand-centered triplet states ( 3 LC), as reflected by the enhanced vibrational progression, compared with 298 K ( Figure S10). Under these conditions, all of the complexes show prolonged excited state lifetimes and quantum yields between 0.8 and 1.0 due to a diminished k nr (Table 4). In addition, it is worth noting that at 77 K, diverse comparably stable conformations can be locked in the glassy matrix, leading to biexponential decays, as was previously observed for other Pt(II) complexes [22,31,32,55,56,62]. ure 4) predominantly originates from ligand-centered triplet states ( 3 LC), as reflected by the enhanced vibrational progression, compared with 298 K ( Figure S10). Under these conditions, all of the complexes show prolonged excited state lifetimes and quantum yields between 0.8 and 1.0 due to a diminished knr (Table 4). In addition, it is worth noting that at 77 K, diverse comparably stable conformations can be locked in the glassy matrix, leading to biexponential decays, as was previously observed for other Pt(II) complexes [22,31,32,55,56,62].  In addition, Φ L and τ at 77 K were used to determine the radiative (k r ) and radiationless (k nr ) deactivation rate constants (Table 4), according to equations 1 to 4, and assuming a unitary intersystem crossing efficiency for S 1 →T n →T 1 , the following relations are valid where k IC is the internal conversion rate constant and k ISC is the intersystem crossing rate constant (T 1 →S 0 ). For biexponential decays, amplitude-weighted average lifetimes (τ av_amp ) were used to estimate the rate constants [64].
Depending on the nature of the coligand, minor changes in both k r and k nr are found (Table 4). In a frozen glassy matrix, at 77 K, the complexes are trapped in their most stable conformation(s) within the solid and the differences in solvation environments and rotovibrational relaxation effects are diminished; therefore, only minor differences in k r and k nr can be found.
For the complexes [Pt(naphen)(X)] (X = Cl or C≡CPh) with the rigid CˆNˆN-coordinating naphen − ligand, or the slightly more flexible tetrahydro-derivative [Pt(thnaphen)(X)], similar broad PL spectra were found at 298 K in the CH 2 Cl 2 solution, peaking at 630 nm for the naphen and 570 nm for the thnaphen derivatives [55], i.e., slightly red-shifted with respect to the dba complexes. The NˆCˆN-coordinated complex [Pt(bdq)(Cl)] containing the rigid bdq − ligand shows a slightly more pronounced vibronic structure and a maximum at 568 nm [55]. The quantum yields Φ L ranged between 0.08 and 0.32, and are similar to the dba complexes.
At low T (77 K, frozen glassy matrices), vibrationally structured emission profiles were found for these complexes, and they appear similar to those observed for the dba complexes, but with slightly higher emission energies for the naphen (600 nm) and thnaphen (530 nm) complexes, yet with almost no shift for the bdq complex. The Φ L values were close to 1 for these complexes [55]. Taking into account the photoluminescence lifetime of the [Pt(naphen)(Cl)] complex, we assign the phosphorescence to mixed 3 LC/ 3 MLCT excited states with only a small metal contribution and dominant 3 LC character [55].

DFT-Calculated Emission Spectra
The calculated vibrationally resolved emission spectra (red lines) and the zero-point corrected 0-0 emission energies (green dashed lines) are in very good agreement with the experimental data (blue lines). For all of the complexes ( Figure 5, left), the main peak and the most intense shoulder deviate by less than 10 nm (except for PPh 3 ). Nonetheless, for the calculated 0-0 energy of the PPh 3 , the agreement is still reasonable. A vibrationally resolved emission spectrum could not be computed for PPh 3 due to numerical issues (see Computational Details). The calculated orbitals contributing to the description of the emissive T1 state show that the HOMO and LUMO are partially localized on the metal ion, as well as on the ligand scaffold ( Figure 6). For Me2Imd, partial contributions from the coligand can also be identified. In general, the occupied orbitals have higher metal contributions than their unoccupied counterparts. For the dmso complex, exclusively, the T1 state can be described as a predominant HOMO→LUMO excitation, which involves a significant amount of charge transfer between the metal and the ligand ( Figure 6).
However, for all of the complexes, the T1 state is composed of a number of monoelectronic excitations. For the dmso complex, there are two main contributions to the T1 state, namely, HOMO→LUMO (44.1%) and HOMO-1→LUMO (37.6%). In addition, there is a small contribution from the HOMO→LUMO+1 excitation (5.3%). The PPh3 derivative shows a major contribution (65.9%) from the HOMO-1→LUMO excitation. Other smaller contributions come from HOMO-3→LUMO (7.6%) and HOMO→LUMO+1 (5.7%). The Me2Imd complex exhibits two approximately equal contributions from HOMO-2→LUMO (32.8%) and HOMO-1→LUMO (29.5%), as well as a significant contribution from the HOMO→LUMO+1 (14.1%) excitation. In the case of CNtBu, the main contributions stem from the HOMO-1→LUMO (47.1%), HOMO→LUMO (33.3%), and HOMO→LUMO+1 (8.5%) excitations. Generally speaking, for all of the complexes, the character of the T1 state can be described as a mixture of intraligand LC and MLCT character; this is in line with our initial assumptions. When comparing the complexes [Pt(dba)(L)] and [Pt(db(ph)a)(L)] for the different coligands ( Figure 5, right), the main peak shifts to shorter wavelengths (for both cases) when the dmso ligand is replaced by Me 2 Imd. In the case of [Pt(db(ph)a)(L)], the main emission peak for PPh 3 lies inbetween the other two. Interestingly, the vibronic coupling pattern is very different for the two CˆNˆC chelate ligands and the change from dba to db(ph)a causes a general, yet slight, red-shift.
The calculated orbitals contributing to the description of the emissive T 1 state show that the HOMO and LUMO are partially localized on the metal ion, as well as on the ligand scaffold ( Figure 6). For Me 2 Imd, partial contributions from the coligand can also be identified. In general, the occupied orbitals have higher metal contributions than their unoccupied counterparts. For the dmso complex, exclusively, the T 1 state can be described as a predominant HOMO→LUMO excitation, which involves a significant amount of charge transfer between the metal and the ligand ( Figure 6).
However, for all of the complexes, the T 1 state is composed of a number of monoelectronic excitations. For the dmso complex, there are two main contributions to the T 1 state, namely, HOMO→LUMO (44.1%) and HOMO−1→LUMO (37.6%). In addition, there is a small contribution from the HOMO→LUMO+1 excitation (5.3%). The PPh 3 derivative shows a major contribution (65.9%) from the HOMO−1→LUMO excitation. Other smaller contributions come from HOMO−3→LUMO (7.6%) and HOMO→LUMO+1 (5.7%). The Me 2 Imd complex exhibits two approximately equal contributions from HOMO−2→LUMO (32.8%) and HOMO−1→LUMO (29.5%), as well as a significant contribution from the HOMO→LUMO+1 (14.1%) excitation. In the case of CNtBu, the main contributions stem from the HOMO−1→LUMO (47.1%), HOMO→LUMO (33.3%), and HOMO→LUMO+1 (8.5%) excitations. Generally speaking, for all of the complexes, the character of the T 1 state can be described as a mixture of intraligand LC and MLCT character; this is in line with our initial assumptions. A quantitative description of the T1 character was obtained through a correlated electron-hole pair analysis [65]. The resulting decomposition into LC (π−π*), MLCT (d−π*), coligand-to-dba charge transfer (L'LCT), coligand-to-metal charge transfer (L'MCT), and MC (d−d*) contributions (Figure 7) is based on a partitioning of the complexes into the dba ligand, the metal and the coligand (as shown for 3 in Figure 7).
Overall, the analysis shows that all of the complexes exhibit a similar character for the T1 state. The dmso, PPh3, Me2Imd, and CNtBu complexes all predominantly show a LC character (86.8%, 76.3%, 76.6%, and 91.1%, respectively), and only a small degree of CT character. The proportion of the MLCT character increases from dmso (4.8%) via CNtBu (5.4%) and PPh3 (6.8%) to Me2Imd (7.9%). A quantitative description of the T 1 character was obtained through a correlated electron-hole pair analysis [65]. The resulting decomposition into LC (π−π*), MLCT (d−π*), coligand-to-dba charge transfer (L'LCT), coligand-to-metal charge transfer (L'MCT), and MC (d−d*) contributions (Figure 7) is based on a partitioning of the complexes into the dba ligand, the metal and the coligand (as shown for 3 in Figure 7).
Overall, the analysis shows that all of the complexes exhibit a similar character for the T 1 state. The dmso, PPh 3 , Me 2 Imd, and CNtBu complexes all predominantly show a LC character (86.8%, 76.3%, 76.6%, and 91.1%, respectively), and only a small degree of CT character. The proportion of the MLCT character increases from dmso (4.8%) via CNtBu (5.4%) and PPh 3 (6.8%) to Me 2 Imd (7.9%). The high LC character of the excited states explains our spectroscopic findings and helps us to understand the marginal influence of the coligands L on the PL properties. However, it cannot explain why the change from dmso to the stronger σ-donating and πaccepting ligands PPh3, CNtBu and Me2Imd drastically reduces the ΦL at 298 K, namely, from a remarkable 0.23 (dmso) to almost zero for PPh3. At the same time, at 77 K, the introduction of the new ligands led to increased ΦL values, with the highest performance observed for the PPh3 complex. In future work, we will seek insights into the dynamics of the excited states by dedicated calculations. In fact, the results depicted in Figure 7 only show the composition concerning the initial excited states [65], which might change over time, as recent work on related luminescent complexes has revealed [66,67].

Spectroelectrochemistry (SEC)
The UV-vis spectroscopic response upon cathodic reduction and anodic oxidation (SEC) was studied for all of the for complexes in the THF/n-Bu4NPF6 solution (Figure 8, more figures and data in the SM) First of all, the SEC experiments show that the initial reductions and oxidations of all four compounds occur, irreversibly, on the timescale of minutes. In the CV experiments, the first reduction showed some degree of reversibility, but the CV experiment is on a timescale of seconds. Nevertheless, the spectral responses of the compounds upon reduction or oxidation leave the impression that defined EC (electrochemical + subsequent chemical reaction) [61,[68][69][70] reactions occur. This assumption is derived from the almost isosbestic character of the produced spectral traces (Figure 8).
Upon reduction, a pronounced band, at around 435 nm, was observed for the set of reduced species, while long-wavelength bands at around 650 nm were only observed for the dmso (1) and Me2Imd (4) complexes. Assuming a ligand π*-centered reduction, we were surprised to see no long-wavelength transitions into π*-π* states, which are usually observed for complexes with reduced heteroaromatic ligands [27,36,[68][69][70][71][72]. They most likely occur for the dba system at wavelengths below 1100 nm, beyond the detection limit of our spectrometer.
The oxidation of the four complexes leaves UV-vis absorption traces that are very dissimilar, thus confirming the mixed metal/coligand character of the oxidation process. Importantly, no evidence for stable or transient Pt(III) species was observed, in line with The high LC character of the excited states explains our spectroscopic findings and helps us to understand the marginal influence of the coligands L on the PL properties. However, it cannot explain why the change from dmso to the stronger σ-donating and π-accepting ligands PPh 3 , CNtBu and Me 2 Imd drastically reduces the Φ L at 298 K, namely, from a remarkable 0.23 (dmso) to almost zero for PPh 3 . At the same time, at 77 K, the introduction of the new ligands led to increased Φ L values, with the highest performance observed for the PPh 3 complex. In future work, we will seek insights into the dynamics of the excited states by dedicated calculations. In fact, the results depicted in Figure 7 only show the composition concerning the initial excited states [65], which might change over time, as recent work on related luminescent complexes has revealed [66,67].

Spectroelectrochemistry (SEC)
The UV-vis spectroscopic response upon cathodic reduction and anodic oxidation (SEC) was studied for all of the for complexes in the THF/n-Bu 4 NPF 6 solution (Figure 8, more figures and data in the SM). a rapid chemical reaction after oxidation (EC). Bands for Pt(III) species are expected in the range 500 to 1800 nm [72].

General Information
Commercially available chemicals were purchased from Sigma-Aldrich, Acros, First of all, the SEC experiments show that the initial reductions and oxidations of all four compounds occur, irreversibly, on the timescale of minutes. In the CV experiments, the first reduction showed some degree of reversibility, but the CV experiment is on a timescale of seconds. Nevertheless, the spectral responses of the compounds upon reduction or oxidation leave the impression that defined EC (electrochemical + subsequent chemical reaction) [61,[68][69][70] reactions occur. This assumption is derived from the almost isosbestic character of the produced spectral traces (Figure 8).
Upon reduction, a pronounced band, at around 435 nm, was observed for the set of reduced species, while long-wavelength bands at around 650 nm were only observed for the dmso (1) and Me 2 Imd (4) complexes. Assuming a ligand π*-centered reduction, we were surprised to see no long-wavelength transitions into π*−π* states, which are usually observed for complexes with reduced heteroaromatic ligands [27,36,[68][69][70][71][72]. They most likely occur for the dba system at wavelengths below 1100 nm, beyond the detection limit of our spectrometer.
The oxidation of the four complexes leaves UV-vis absorption traces that are very dissimilar, thus confirming the mixed metal/coligand character of the oxidation process. Importantly, no evidence for stable or transient Pt(III) species was observed, in line with a rapid chemical reaction after oxidation (EC). Bands for Pt(III) species are expected in the range 500 to 1800 nm [72].

General Information
Commercially available chemicals were purchased from Sigma-Aldrich, Acros, ABCR or Fisher-Scientific and were used without further purification. Dry THF was obtained from distillation over sodium/potassium alloy. All reactions and measurements involving metal complexes were conducted under argon, using standard Schlenk techniques. The complex [Pt(dba)(dmso)] (1) was synthesized and characterized as previously described [33]. Finally, 1,3-dimethylimidazolium iodide was prepared from commercially purchased methylimidazole and methyl iodide as described in the literature [73].

Instrumentation
1 H and 31 P NMR spectra were recorded on a Bruker Avance II 600 MHz spectrometer ( 1 H: 600.13 MHz, 31 P: 242.88 MHz) with a triple resonance (TBI) 5 mm inverse probehead with z-gradient coil using a triple resonance. Chemical shifts are relative to TMS, respectively. UV-vis absorption spectra were recorded with Varian Cary 05E or Cary 50 scan spectrophotometers. Photoluminescence spectra at room temperature and at 77 K were recorded with a PicoQuant FluoTime 300. The same equipment was used for lifetime measurements. Lifetime analysis was performed using the commercial EasyTau 2.2 software. The quality of the fit was assessed by minimising the reduced chi-squared function. Photoluminescence quantum yields were determined with a Hamamatsu Photonics absolute PL quantum yield measurement system (C9920-02), equipped with a L9799-01 CW Xenon light source, monochromator, photonic multichannel analyser and integrating sphere (error of maximum ± 2% for Φ L is estimated). All of the solvents were of spectroscopic grade and were degassed prior to use. Elemental analyses were obtained using a HEKAtech CHNS EuroEA 3000 analyzer. EI-MS spectra were measured with a Finnigan MAT 95, and HR-ESI-MS using a THERMO Scientific LTQ Orbitrap XL. MS Simulations were performed using ISOPRO 3.0. Electrochemical measurements were carried out in 0.1 M n-Bu 4 NPF 6 /THF solution using a three-electrode configuration (glassy carbon working electrode, Pt counter electrode, Ag/AgCl reference electrode) and a Metrohm Autolab PGSTAT30 potentiostat and function generator. The ferrocene/ferrocenium couple served as internal reference.

X-ray Diffractometric Analysis of Single Crystals for Structure Determination
The crystals of 1 to 4 were obtained via isothermic evaporation at 0 • C from acetone/Et 2 O mixtures (v:v = 2:1). The measurement of [Pt(dba)(Me 2 Imd)] was performed using graphitemonochromated Mo-K α radiation (λ = 0.71073 Å) with an IPDS 2T diffractometer (STOE and Cie.) at 100 K. [Pt(dba)(dmso)] and [Pt(dba)(CNtBu)] were measured on a D8 Advance diffractometer (Bruker), with the same wavelength, at 100 K. Diffraction data of [Pt(dba)(PPh 3 )] were collected at the beamline P24 of PETRA III at the German Electron Synchrotron facility DESY using a wavelength of λ = 0.56076 Å and at a temperature of 100 K. The data were recorded on a Pilatus 1M CdTe detector. All of the structures were solved by dual space methods (SHELXT-2015) [74] and refined by full-matrix least-squares techniques against F 2 (SHELXL-2017/1) [75,76] The non-hydrogen atoms were refined with anisotropic displacement parameters, without any constraints. The hydrogen atoms were included by using appropriate riding models. The numerical absorption correction of the data of [Pt(dba)(Me 2 Imd)] (X-RED V1.31; STOE and Cie, 2005, Darmstadt, Germany) were performed after optimising the crystal shapes using the X-SHAPE V1.06 (STOE & Cie, 1999, Darmstadt, Germany) [77,78]. For the data set acquired with synchrotron radiation, no absorption correction was omitted. Residual electron density in the pores alongside the crystallographic b-axis of [Pt(dba)(PPh 3 )] was removed. The crystallographic details are summarized in Table S1

Vibrationally Resolved Emission Spectra
The quantum-chemical calculations of the vibrational Franck-Condon spectra and optimized geometries were carried out with the quantum chemistry package, Gaussian 09 Rev. D.01 [79]. The CAM-B3LYP functional [80] was used in the density functional theory (DFT) calculations and the SDD basis was set, which applies an effective core potential for the Pt atoms [81] and the D95 basis set for H, C, N and O atoms [82]. The vibrational Franck-Condon spectra were calculated according to the method of Barone et al. [83][84][85], with Kohn-Sham DFT-based geometry optimizations in the S 0 and T 1 states, followed by frequency analysis calculations.
The energies were corrected by zero-point vibrational energies and thermal free energy contributions. In order to calculate the overlap integrals for the vibronic spectra, the transitions are divided into classes C n , where n is the number of the excited normal modes in the final electronic state. The maximum number of quanta per mode was set to 100 and the maximum number of quanta for combinations of two modes to 65. The number of integrals calculated was limited to 1.5 × 10 8 . A maximum of 20 classes were computed. The line spectrum was broadened using Gaussian functions with a half-width at half-maximum of 500 cm −1 . The solvent CH 2 Cl 2 was taken into account by the polarizable continuum model (PCM) in an integral equation formalism framework [86] with atomic radii from the universal force field model (UFF) [87].

Character of the T 1 State
The character of the emissive T 1 state is determined by TD-DFT calculations, including 20 excited singlet and triplet states at the T 1 geometry, optimized with Kohn-Sham DFT with multiplicity 3.

Absorption Spectra
To obtain the UV-vis absorption spectrum of the complexes, TD-DFT calculations of the 40 lowest excited singlet states were performed with the PBE0 functional [88] and the SDD basis set. A Lorentzian broadening with a half-width at a half-maximum (HWHM) of 10 nm was used for each transition.
Cyclic voltammetry showed partially reversible reduction waves ranging between −1.89 and −2.10 V and irreversible oxidation processes between 0.55 (L = Me 2 Imd) and 1.00 V (dmso), depending on the ancillary ligand L. The strongly variable oxidation potentials raised the electrochemical gaps from 2.65 to 2.91 V along the series Me 2 Imd < CNtBu < PPh 3 < dmso along the trend of decreasing ligand strength.
All four dba complexes partially show vibrationally structured long-wavelength absorption bands peaking at around 530 nm. TD-DFT calculated absorption spectra are in good agreement with the experimental findings, showing that they are slightly red-shifted, but perfectly reproduce the observed trends. The stronger coligands only cause a minor red-shift of the long-wavelength absorption bands (0.03 to 0.05 eV). The photoluminescence spectra of all four compounds are very similar and strongly depend on the presence or absence of 3 O 2 , in line with the triplet character of the emissive state. In a solution at 298, they show broad bands with a minor vibronic structure and maxima at around 590 nm; in frozen glassy matrices, slightly blue-shifted (~580 nm) bands with vibrational progression were found. In contrast to the comparable emission energies, the photoluminescence quantum yields Φ L differ between the four complexes and the lifetimes increase along the series Me 2 Imd < CNtBu < PPh 3 < dmso from 0.06 to 14 µs; surprisingly, the dmso complex constitutes the most efficient emitter, at 298 K (Φ L(Ar) = 0.24). This is highly counterintuitive, as the ligand strength is considered to decrease from Me 2 Imd to dmso, as supported by the electrochemical results. In frozen glassy matrices at 77 K, the lifetimes are comparable and increase from 17 µs for the Me 2 Imd complex, to approximately 30 µs for the other three derivatives. The TD-DFT calculated emission spectra agree well with the experimental data for all of the complexes. Decomposition of the excited T 1 states showed mostly ligand-centered (LC) contributions (80 to 90%) to the description of the emissive states, which is in line with the vibrationally structured emission bands, a minor metal-to-ligand charge transfer (MLCT) character (5 to 8%), marginal coligand-to-metal (L'MCT) or coligand-to-chelate ligand charge transfer (L'LCT) contributions and virtually no metal-centered (MC) character.
The exclusively theoretical study on the dba-7-phenyl substituted derivatives [Pt(db(ph)a)(L)] (H 2 db(ph)a = 7-phenyldibenzo[c,h]acridine) with L = PPh 3 , CNtBu, and Me 2 Imd showed that these complexes were not superior to the dba congeners, in line with the previous experimental comparison of [Pt(db(ph)a)(dmso)] with the dba derivative.
The far lower Φ L for the complexes containing the stronger coligands Me 2 Imd, CNtBu, and PPh 3 (in comparison with the parent dmso complex [Pt(dba)(dmso)] in solution at 298 K) is an unexpected result as most of the previous studies showed that the exchange of the coligand by a stronger σ-donor and/or π-acceptor improved the photoluminescence (i.e., higher quantum yields). However, at 77 K in frozen glassy matrix, the "ligand effect" was found as expected (higher Φ L for stronger ligands). The elucidation of the cause for these surprising observation at 298 K is the object of ongoing research. At the same time, the assessment of their suitability for optoelectronic devices will require in-depth photophysical studies in the solid state, e.g., in polymeric matrices, which will be carried out in the near future. This will also include a study of the aggregation of these complexes in solution and in solid materials.