ILCT lifetimes using ancillary ligands in - American Chemical Society

Article Cite This: J. Am. Chem. Soc. 2018, 140, 4534−4542

pubs.acs.org/JACS

Dramatic Alteration of 3ILCT Lifetimes Using Ancillary Ligands in [Re(L)(CO)3(phen-TPA)]n+ Complexes: An Integrated Spectroscopic and Theoretical Study Georgina E. Shillito,† Thomas B. J. Hall,† Dan Preston,† Philipp Traber,‡ Lingjun Wu,§ Katherine E. A. Reynolds,§ Raphael Horvath,§,⊥ Xue Z. Sun,§ Nigel T. Lucas,† James D. Crowley,† Michael W. George,*,§,∥ Stephan Kupfer,*,‡ and Keith C. Gordon*,† †

Department of Chemistry, University of Otago, P.O. Box 56, Dunedin, New Zealand Institute for Physical Chemistry, Friedrich Schiller University Jena, Helmholtzweg 4, 07743 Jena, Germany § School of Chemistry, University of Nottingham, Nottingham NG7 2RD, United Kingdom ∥ Department of Chemical and Environmental Engineering, University of Nottingham Ningbo China, 199 Taikang East Road, Ningbo 315100, China ‡

S Supporting Information *

ABSTRACT: The ground and excited state photophysical properties of a series of fac-[Re(L)(CO)3(α-diimine)]n+ complexes, where L = Br−, Cl−, 4-dimethylaminopyridine (dmap) and pyridine (py) have been extensively studied utilizing numerous electronic and vibrational spectroscopic techniques in conjunction with a suite of quantum chemical methods. The α-diimine ligand consists of 1,10-phenanthroline with the highly electron donating triphenylamine (TPA) appended in the 5 position. This gives rise to intraligand charge transfer (ILCT) states lying lower in energy than the conventional metal-to-ligand charge transfer (MLCT) state, the energies of which are red and blueshifted, respectively, as the ancillary ligand, L becomes more electron withdrawing. The emitting state is 3ILCT in nature for all complexes studied, characterized through transient absorption and emission, transient resonance Raman (TR2), time-resolved infrared (TRIR) spectroscopy and TDDFT calculations. Systematic modulation of the ancillary ligand causes unanticipated variation in the 3ILCT lifetime by 2 orders of magnitude, ranging from 6.0 μs for L = Br− to 27 ns for L = py, without altering the nature of the excited state formed or the relative order of the other CT states present. Temperature dependent lifetime measurements and quantum chemical calculations provide no clear indication of close lying deactivating states, MO switching, contributions from a halide-to-ligand charge transfer (XLCT) state or dramatic changes in spin−orbit coupling. It appears that the influence of the ancillary ligand on the excited state lifetime could be explained in terms of energy gap law, in which there is a correlation between ln(knr) and Eem with a slope of −21.4 eV−1 for the 3ILCT emission.

■

INTRODUCTION The photophysical behavior of Re(I) complexes belonging to the class, fac-[Re(L)(CO)3(α-diimine)]n+, have been the subject of extensive study due to their dynamic and manipulatable excited state properties.1,2 The lowest energy electronic absorbing and emitting state in compounds of this class is most commonly a dπ(Re)→π*(α-diimine), metal-toligand charge transfer (MLCT) state. This was first characterized by Wrighton and Morse for the benchmark Re(I) carbonyl complex, [Re(Cl)(CO)3(phen)] and related derivatives.2 The energetics of the electronic states can be tuned through modulation of both the α-diimine ligand and the ancillary monodentate ligand, L which predominantly affect the energies of the π*(diimine) and dπ(Re) orbitals, respectively.3 Casper and Meyer demonstrated how increasing the electron © 2018 American Chemical Society

withdrawing ability of the ancillary ligand could be utilized to raise the energy of the MLCT transition.1 The emission energies and nonradiative decay constants reported were consistent with the energy gap law. Though MLCT states are predominantly the lowest lying, ancillary ligand variation can also alter the nature of the lowest energy absorbing state to a ligand-to-ligand charge transfer (LLCT) state.4−7 Likewise, modulation of the diimine chromophore can also tune the energies of the excited states present.2,7−15 Re(I) tricarbonyl diimine complexes are often emissive, possessing relatively long-lived excited states, usually attributed Received: December 5, 2017 Published: March 14, 2018 4534

DOI: 10.1021/jacs.7b12868 J. Am. Chem. Soc. 2018, 140, 4534−4542

Article

Journal of the American Chemical Society to 3MLCT phosphorescence.1,2,7,12,16 However, mixing and even inversion of 3MLCT and diimine based 3(π→π*) states has been reported. Pomestchenko et al. investigated the influence of Au(I)-acetylide substituted 1,10-phenanthroline in a [Re(Cl)(CO)3(phen)] complex, where the lowest excited state was characterized as a mixture of MLCT, involving the Re center and a triplet intraligand (3IL) state localized on the Au(I) acetylide substituted diimine.17 For Re(I) complexes possessing dipyrido[3,2-a:2′,3′-c]phenazine (dppz) as the diimine ligand, the lowest lying emissive state can be a dppz centered 3IL state.9,18−21 These results were shown convincingly through the use of time-resolved infrared (TRIR) spectroscopy. The CO bands act as spectroscopic handles, as their relative shifts reflect the change in electron density over the Re(CO)3 unit and can be used to elucidate the electronic excited state being accessed.22−25 An additional level of complexity can be added to these already photophysically rich systems by the addition of an electron donating substituent to the diimine ligand. This introduces the possibility of another electronic state known as an intraligand charge transfer (ILCT) state.26−33 This work aims to build upon our previous research, whereby introduction of an electron donating triphenylamine (TPA) group to a dppz ligand facilitated population of low lying, long-lived ILCT states that were lower in energy than the MLCT states.30,31,34,35 Herein we report a series of complexes of the type fac[Re(L)(CO)3(phen-TPA)]n+ where a TPA donor is substituted in the 5-position of the 1,10-phenanthroline ligand (Figure 1).

order to gain insight into the nature of the electronic states within the Franck−Condon region as well as into the proceeding excited state relaxation processes. Time-dependent density functional theory (TDDFT) was applied to elucidate the excited states contributing to the initial photoactivation, resonance Raman (RR) spectra as well as TR2, TRIR and emission spectra. Relativistic effects were incorporated by means of the scalar relativistic zeroth-order regular approximation (SR-ZORA). SOC was incorporated by the twocomponent perturbative approach.

■

RESULTS AND DISCUSSION Ground State Electronic Absorption. The ground state electronic absorption spectra of the Re(I) complexes and the phen-TPA ligand are presented in Figure 2. Complexation of

Figure 2. Ground state electronic absorption spectra in CH2Cl2 of the [Re(L)(CO)3(phen-TPA)]n+ complexes and phen-TPA ligand. Bright singlet excitations contributing to the absorption are indicated by vertical bars.

the ligand to the Re(L)(CO)3 core results in the formation of a low energy absorption band around 400 nm. Studies of related [Re(L)(CO)3(α-diimine)]n+ complexes lacking an α-diimine appended donor group, commonly have a lowest energy MLCT state.1,2 Increasing the electron withdrawing ability of the ancillary ligand, L results in a stabilization of the dπ(Re) orbitals and consequently a blue-shift in the MLCT band (λmax) is observed.1 However, for the reported complexes where the αdiimine consists of 1,10-phenanthroline with an appended electron donating TPA moiety, the opposite trend is observed, with λmax displaying a red-shift as the electron withdrawing ability of L increases, with the nonhalide complexes red-shifted relative to that of the halide species. This suggests that the lowest energy electronic state is not MLCT in nature. In accordance with previous studies of donor−acceptor complexes involving a TPA substituent, the lowest energy band is instead dominated by absorption to an ILCT state, which is consistent with that observed in the resonance Raman and TDDFT calculations (vide inf ra). In the ILCT transition the donor MO lies on the TPA and the acceptor as the π*phen. As the ancillary ligand becomes more electron withdrawing, stabilization of the π*phen acceptor MOs relative to that to the TPA donor is apparent,30 hence the redshift in the spectra is observed. TDDFT simulations were performed to unravel the nature of the electronic transitions underlying the electronic absorption

Figure 1. Structure of the [Re(L)(CO)3(phen-TPA)]n+ complexes studied, where L = Br−, Cl−, dmap and py; with ILCT and MLCT transitions indicated.

These systems were designed to possess ILCT states akin to those of the formerly investigated dppz complexes, but at a higher energy such as to allow for the possibility of mixing with the MLCT states which has been observed previously.29 Alongside the phen-TPA ligand, four complexes are reported: [Re(Br)(CO) 3 (phen-TPA)], [Re(Cl)(CO) 3(phen-TPA)], [Re(dmap)(CO)3(phen-TPA)]+ and [Re(py)(CO)3(phenTPA)]+, subsequently denoted as Re(Br), Re(Cl), Re(dmap) and Re(py) respectively. Electronic tuning of the CT states is achieved by variation of the dπ(Re) orbital energy through modulation of the electron withdrawing ability of the ancillary ligand such that pyridine (py) > dimethylaminopyridine (dmap) > Cl− > Br−. The photophysical properties of the ground and excited states have been extensively characterized, utilizing a suite of electronic and vibrational spectroscopic techniques capable of observing ground and excited state species; these include transient resonance Raman (TR2) and time-resolved infrared (TRIR). Quantum chemical simulations were performed in 4535


Article

Journal of the American Chemical Society spectra of the Re(I) complexes. These calculations clearly reveal that the broad absorption feature between approximately 500 and 350 nm is mainly associated with the excitation into two ILCT states, denoted S1 and S2 as well as into a MLCT state, S4. Both ILCT states are associated with a CT from the TPA-moiety (πTPA, HOMO) to π*phen-orbitals (LUMO and LUMO+1), whereas the oscillator strength of S2 is considerably higher than for S1. The smaller oscillator strength of S1 with respect to S2 is reasoned by the poor overlap of the MOs involved in the transition. The UV flank of the absorption band is correlated to the excitation into the broad MLCT state (S4) populating the LUMO localized on the phen-fragment. In agreement with the experimentally observed red-shift with respect to the electron donating ability of the ancillary ligand, the quantum chemical calculations reveal a stabilization of S1 and S2 from Re(Cl) (401 and 385 nm/3.09 and 3.22 eV) to Re(Br) (406 and 387 nm/3.05 and 3.20 eV), Re(dmap) (437 and 409 nm/2.84 and 3.03 eV) and Re(py) (445 and 414 nm/ 2.79 and 3.00 eV). The red-shift of the ILCT states, S1 and S2, is mainly rationalized by the pronounced stabilization of the LUMO as well as of the LUMO+1 of Re(Cl), Re(dmap) and Re(py) with respect to Re(Br), see Figure S9. Simultaneously, the MLCT state (S4) is destabilized from 356 to 352, 334 and 329 nm (3.49, 3.53, 3.72 and 3.77 eV) from Re(Br) to Re(Cl), Re(dmap) and Re(py), which is in agreement with the stabilization of the dπ(Re) orbitals illustrated in Figure S9. Interestingly, the MLCT state of the halides, Re(Cl) and Re(Br), shows, opposed to the ILCT states, a pronounced mixing with triplet states as indicated by SR-ZORA-TDDFT. More details with respect to the computational results on the absorption spectra of the four complexes are presented in Tables S1−S10 and Figures S5−S9 in the Supporting Information. Resonance Raman Spectroscopy. To explore the nature of the electronic transitions RR spectra were recorded at a series of excitation wavelengths (λex) across the lowest energy absorption band. The RR spectra of Re(Cl) are presented in Figure 3. When λex is 351 nm, the CO band at 2025 cm−1 is enhanced compared to that of the 1064 nm nonresonant spectrum and the dominant TPA modes at 1586 and 1622 cm−1 are comparatively weak, indicting the presence of a MLCT state in this region. This finding is in line with the RR spectrum simulated using the short-time approximation (STA),36 which allows identification of the characteristic intensity pattern of each electronic state contributing to the RR signal. Thus, in contrast to the signals of the ILCT states, S1 and S2 shown in Figure S10d,e, the STA-RR spectrum of the MLCT state (S4), illustrated in Figure S10f, features the highest contribution of the normal mode associated with the symmetric CO-stretch normal mode (mode 159 at 2024 cm−1), whereas no intense TPA modes were calculated for the band structure at approximately 1600 cm−1. As the excitation wavelength changes to 375 nm, the TPA modes become more predominant; however, the CO band intensity is still significant, such that it is likely that there is overlap of both MLCT and ILCT states in this region. This has been observed previously in Re(I) hexaazanaphthalene complexes.26 Upon further red-shifting λex to 407 nm, the pattern of relative band intensity changes significantly, with the TPA modes from 1586 to 1622 cm−1 increasing in intensity, dominating the spectra and phen-based vibrations such as those at 1376, 1422 (coincident with a CH2Cl2 band) and 1522 cm−1 are also enhanced, supporting the presence of an ILCT state.

Figure 3. Resonance Raman spectra of Re(Cl) in CH2Cl2 (solvent bands marked with *). The band intensities have been scaled relative to the 702 cm−1 CH2Cl2 band. See Figure S4 for comparison with simulated spectra.

This state likely corresponds to the higher energy, strongly absorbing S2, ILCT state predicted by TDDFT. This assertion is supported by means of the simulated RR spectra for the three individual states (STA for S1, S2 and S4) and the intensity pattern obtained at an excitation wavelength of 407 nm (at IMDHOM level of theory), Figure S10d−f, Figure S11b and Table S11. With the excitation wavelengths further to the red tail of the absorption band, the TPA modes are less enhanced but most notably, their pattern of relative enhancement changes, with the 1586 and 1622 cm−1 bands decreased relative to that of the 1601 cm−1 signal, indicating that a different state is being probed in this region. This weakly absorbing state likely corresponds to the lower energy ILCT, S1 state predicted by TDDFT calculations. The assumption that the observed changes in the RR spectra between the 407−413 nm and the 448−491 nm regions are a result of resonance with two different ILCT states is supported by the simulated RR spectra, which show differences between the RR enhancement profiles of the S1 and S2 states. The largely phen-based mode predicted at 1599 cm−1 (mode 151, Tables S11 and 12), shows an increased contribution in the S1 state. This likely corresponds to the relative increase of the experimentally observed 1601 cm−1 band in the 448−491 nm region. The resonance Raman spectra of Re(Br), Re(py) and Re(dmap) complexes are similar to that of Re(Cl), (see Figures S1−3 and Figures S10 and S11) with a higher energy MLCT state and two lower ILCT states contained within the absorption band. The key difference is that the enhancement patterns indicate that the ILCT states occur at slightly longer λex values for Re(dmap) and Re(py), in accordance with the red-shift in their electronic absorption spectra, relative to that of Re(Cl) and Re(Br). More details with respect to the simulated 4536


Article

Journal of the American Chemical Society

Table 1. Experimental Electronic Absorption and Emission Data for the [Re(Cl)(CO)3(phen-TPA)]n+ Complexes in CH2Cl2 at 293 Ka λabs (nm) Re(Br) Re(Cl) Re(dmap) Re(py) a

389 386 410 401

ε (103 M−1 cm−1) 10.1 11.1 13.7 11.5

λem (nm)

Eem (eV)

605 607 678 689

2.05 2.04 1.83 1.80

τ (ns) 6000 3900 44 27

φem

kr (s−1)

0.016 0.005 0.002 0.001

× × × ×

2.7 1.3 3.5 3.7

knr (s−1) 3

10 103 104 104

1.6 2.5 2.3 3.8

× × × ×

105 105 107 107

Estimated error of ±10% on lifetimes and an absolute error of ±5 × 10−4 for the quantum yields.

RR spectra as well as the contributions of the individual ILCT and MLCT states are collected for all complexes in the Supporting Information. Steady State Emission Spectroscopy. Similarly to related Re(I) carbonyl compounds, the complexes studied are emissive and the emission energies (Eem) are tuned systematically with respect to the electronic nature of the ancillary ligand. In accordance with the electronic absorption spectra, the more electron withdrawing the ancillary ligand the more red-shifted λem, occurring at 605, 607, 678 and 689 nm (2.05, 2.04, 1.83 and 1.80 eV) for Re(Br), Re(Cl), Re(py) and Re(dmap), respectively; as presented in Table 1. This trend suggests that the emitting state is also ILCT in nature, which is confirmed by the quantum chemical calculations. The calculations predict emission from the T1 state of ILCT character at 578, 567, 640 and 658 nm (2.14, 2.19, 1.94 and 1.89 eV) for Re(Br), Re(Cl) and Re(dmap) and Re(py), respectively; differing from the experimental measurements by only 0.1 eV. The complexes possess similar Stokes shifts between 10,400 and 9,200 cm−1, consistent with a charge transfer transition. The calculations indicate that structural variations predicted to occur upon relaxation in the triplet ground state from the Franck−Condon region are mainly governed by the partial planarization of the phen-TPA fragment. In the Franck−Condon region dihedral angles of 55, 57, 55 and 55° are calculated for Re(Br), Re(Cl), Re(dmap) and Re(py); these angles decrease by 8, 10, 7 and 5° in the T1 equilibria, respectively; see Table S20. Transient Absorption and Emission. The nature of the ancillary ligand systematically modulates the ILCT absorption and emission energies of the complexes. However, it also has an unexpected effect on the excited state lifetimes, which range from 6.0 μs to 27 ns at 298 K, for Re(Br) and Re(py), respectively; as shown in Table 1. The emissive lifetimes show sensitivity to the presence of 3O2, indicating that the emitting state is triplet in nature. Unsurprisingly, this effect is more pronounced for the halide complexes due to their longer lifetimes. Similar behavior has been observed previously by Wozna et al. where comparison between [Re(Cl)(CO)3(phen)] and [Re(CH3CN)(CO)3(phen)]+ (alongside other related complexes) showed a change in 3MLCT τem from 0.15 to 3.1 μs, respectively. This was attributed to relative changes in the excited state reorganization energies.37 For the compounds we report here, the inverse trend is observed, with the charged complexes displaying significantly shorter lifetimes than their neutral halide counterpart. This inverse relationship is perhaps not surprising given that for the presented compounds, the emissive state is a 3ILCT rather than a 3MLCT. Transient emission and absorption spectra of the complexes were recorded at a range of wavelengths, to generate transient maps as illustrated in Figure 4. Cuts taken at specific wavelengths, coincident with λem provide kinetic traces, which can be successfully fitted by a monoexponential decay function

Figure 4. Transient emission map of Re(Cl) recorded in CH2Cl2 with 355 nm pulse excitation. Inset shows the spin density of the emissive 3 ILCT state within its equilibrium geometry, where the unpaired electrons are seen to be localized on the phen and TPA moieties.

to obtain τ. The excited state lifetimes obtained from the transient absorption and emission spectra are equal within experimental error range, suggesting that the same electronic state is responsible for both the transient emission and transient absorption signals. The values for the radiative (kr) and nonradiative (knr) decay constants were calculated using the equation below. The complexes possess similar kr values; however, the values for knr are distinctly different between the complexes ranging from 1.62 × 105 s−1 to 3.8 × 107 s−1 for Re(Br) and Re(py), respectively. Akin to many other rhenium systems, knr is greater than kr such that knr ≅ 1/τ. Φ=

kr k r + k nr

τ=

1 k r + k nr

The striking role of the ancillary ligand on the emissive lifetimes, although not unprecedented, is surprising, particularly as the quantum chemical calculations, TR2 and TRIR data indicate that the ancillary ligand is not actively involved in the long-lived excited state (vide inf ra). The effect of temperature on the excited state lifetime was examined in order to explore the possibility of deactivation through close lying excited states (Figure S13, Table S17).38−40 For the Re(Br) and Re(Cl) complexes, this relationship can be fitted with a simple model which did not require inclusion of states other than the emissive triplet and metal centered, dd states. In accordance, the performed quantum chemical calculations reveal no alteration of the electronic character of the low lying 3ILCT states (T1 and T2) within their fully relaxed equilibria for all four complexes, as shown in Figure S15. This suggests that over this temperature range, no other state appears to play a significant role in the deactivation process. Fitting of these data 4537


Article

Journal of the American Chemical Society allowed estimation of the dd state energies above that of the emitting, ILCT state. Interestingly, even though the obtained values are in accordance with other previously reported rhenium complexes,38,41 the calculated energy gap is greater for Re(Br) at 4,300 cm−1 compared to that of Re(Cl) at 3500 cm−1. This is unusual as Br− has a weaker ligand field than Cl− and thus would be expected to destabilize the dd state less. A previous study of [Ru(bpy)2(PPh3)X]n+ complexes effectively outlined how the dd state energy and excited state lifetimes could be affected through modulation of the ligand field strength of X.42 Increasing the ligand field strength on going from X = Br− to Cl− resulted in an increase in emission energy and excited state lifetime. Indeed, in general, the greater the ligand field strength the longer lived the lowest excited state in Ru(II) systems, whether this be due to ancillary ligands43 or through the use of strained ligand architectures.44 However, for our compounds the opposite effect is observed. The increased energy difference between the dd and emissive ILCT state indicates that for Re(Br), the ILCT state exhibits greater stabilization than the dd states, resulting in a larger difference between them compared to Re(Cl). A relationship with temperature was unable to be definitively obtained for the Re(dmap) and Re(py) complexes due to their short-lived nature and the measurement uncertainties (Figure S13). Previous studies of [M(CO)4(N,N)] complexes, where N,N = 1,10-phenanthroline or 3,4,7,8-tetramethyl-1,10-phenanthroline (tmp) have also exhibited similar, significant changes in excited state lifetimes.45,46 For these systems it was conclusively shown that switching of the phen π* LUMO orbital occurs between these complexes, being of b1 and a2 symmetry for the phen and tmp complexes, respectively. The subsequent lifetime differences were, in part, attributed to this change. In the case of the reported compounds, quantum chemical calculations reveal no apparent switching of the frontier orbitals across the complexes studied; see Figure S9.45,46 Another possible explanation for the significant differences in the excited state lifetimes would be if the ancillary ligand altered the type of excited state populated.6,12,47,48 Vlček and coworkers have shown that ancillary ligand variation, where L = Cl−, Br −and I− can alter the nature of the lowest lying emitting state, changing from MLCT to XLCT upon progressing from Cl− to I−.5−7 This increase in halide character was accompanied by an increase of the lifetime. However, in the case of the [Re(L)(CO)3(phen-TPA)]n+ complexes, interpretation of the dynamic data does not require inclusion of an XLCT state specifically, due to the distinct similarity between the vibrational spectra of all four complexes. This is the case for both the ground and excited state RR spectra (vide inf ra). Furthermore, as shown in Figure 5, the linear relation of Eem and ln(knr) appears to follow the energy gap law, whereas the previous work showed a nonlinear relationship due to the nonhomogenous nature of the compounds.6 However, it should be noted that the calculations reveal partial mixing between halogen p and the rhenium dxy orbitals and thus the presence of XLCT contributions cannot be rigorously excluded. The similarity in the lifetime between Re(Br) and Re(Cl) suggests that scalar relativistic effects, i.e. SOC is not a dominating factor. In order to confirm this assumption, SOCs were calculated at the SR-ZORA-TDDFT level of theory for all compounds; the obtained SOCs are summarized in Tables S3 and S4. With the photoexcitation at 355 nm the 1MLCT (S4) is initially

Figure 5. Plot of Eem vs. ln(knr) for the ILCT excited states of the [Re(L)(CO)3(phen-TPA)]n+ complexes, with an indicative experimental slope shown in red, of −21.4 eV−1. The green line indicates the relationship between the calculated emission energies and ln(knr), revealing a similar slope.

populated, which shows a pronounced coupling to several triplet states in the vicinity of its excitation energy. Considerable SOCs of up to 517, 425, 221 and 390 cm−1 were calculated for all complexes: Re(Br), Re(Cl), Re(dmap) and Re(py), i.e., in the spin−orbit picture for Re(Cl), interaction of the bright S4 with the triplet states T5 and T7 gives rise to the spin−orbit (SO) states SO22 and SO25; see Table S1 for more details. For the 1ILCT states (S1 and S2) no pronounced mixing with the energetically close triplet states is predicted by SR-TDDFT for all complexes, which is evident by the contributions of these singlet states, close to unity, in the respective SO states (see Tables S1, S5, S7 and S9). These results are in agreement with previously reported multiconfigurational and TDDFT studies for structurally related Re(I) complexes including SOCs.49−51 Thus, the quantum chemical results reveal no pronounced dependency of SOC with respect to the ligand L. Further evidence for the ILCT nature of the long-lived state across all four complexes is obtained from the transient resonance Raman spectra. With no clear indicators of close lying deactivating states, MO switching, contributions from an XLCT state or dramatic changes in spin−orbit coupling, the influence of the ancillary ligand on the excited state lifetime could be explained in terms of the energy gap law.52 As the energy difference between the emissive state (T1) and the ground state (S0) increases, the rate of knr also decreases as a result of reduced overlap of the vibrational wave functions between the two states and hence a reduction in the Franck−Condon factor. Figure 5 illustrates the approximately linear relationship between ln(knr) and Eem, with a resulting indicative slope of −21.4 eV−1. The presence of only four data points that cluster into two groups, separated in emission energy by approximately 1700 cm−1 means that these data should provide only qualitative insight. However, it should be noted that this grouping of emission energies of the halide and nonhalide species is predicted by TDDFT calculations, which when compared to experimental lifetime data, give a similar slope, as shown in Figure 5. The gradient is greater than that of −11.76 eV−1 reported by Caspar and Meyer, for the MLCT state lifetimes in [Re(L)(CO)3(bpy)]n+ complexes and −7.54 eV−1 for related Os(II)-bpy and Ru(II)-bpy com4538


Article

Journal of the American Chemical Society pounds.1,53,54 The unusually steep nature of the slope may be attributed to the differing nature of the energy accepting vibrational modes involved in the nonradiative decay process. In the case of the Re(CO)3 complexes, the high energy υ(C− O) vibrations act as dominant acceptor vibrations for the deactivation process.1,3,12 As previously discussed, the resonance Raman spectra of the reported complexes show predominant enhancement of the high energy TPA vibrations at 1586 and 1622 cm−1 suggesting they play an important role in the deactivation of the 3ILCT state, along with vibrations of the phen ligand. Transient resonance Raman. The electronic structure of the thermally equilibrated excited (THEXI) state is probed using transient resonance Raman (TR2) spectroscopy. The TR2 spectra are distinctly different to the ground state RR spectra. This implies that the excited state is successfully populated and that the subsequent Raman scattering measured is predominantly from the THEXI state, being the longest lived state. A previous study showed that the MO switching which occurs in [Ru(bpy)2(dppz)]2+ between aqueous and nonaqueous solutions, is manifested through differences in the time-resolved resonance Raman (TR3) spectra.55 However, as shown in Figure 6, the TR2 spectra of the four complexes look almost

four complexes. A TR2 spectrum of the phen-TPA ligand was unable to be obtained due to strong emission. The simulated TR2 spectra of the remaining complexes as well as the normal modes of the most prominent bands contributing to the intensity pattern are collected in the Supporting Information. Time Resolved Infrared Spectroscopy. The TRIR for the complexes show two distinct spectral profiles. In the case of the halide systems, shown representatively for Re(Br) in Figure 7, spectral signatures from 3MLCT and 3ILCT states are clearly

Figure 7. TRIR spectra of the ν(CO) stretching region for Re(Br) in CH2Cl2 (λex = 355 nm). These show the ground state, as measured by FTIR and evolution of the excited state at subsequent time delays.

discernible, most clearly seen by examining the high frequency absorption. For Re(Br) the ground state high wavenumber CO band at 2025 cm−1 is depleted upon excitation, with a MLCT band growing in at 2051 cm−1. Due to the ultrafast nature of the 1MLCT to 3MLCT conversion,59,60 spectral differences between the singlet and triplet MLCT states are not discernible within ps time resolution and are subsequently denoted without a superscript (Figure 8). The MLCT state then decays (τ = 37 ps) to a band at 2015 cm−1, indicative of a 3ILCT state which decays to the ground state on the microsecond time scale. Re(Cl) (see Figure S18) displays distinctly similar spectral features to Re(Br), with comparable dynamics, hence the analogous state assignments are made. In contrast, the Re(py) and Re(dmap) complexes exhibit different behavior to the halide species, as they do not show any spectral features attributable to a MLCT state. This is correlated to the previously discussed increase in the MLCT state energy,1 such that it is not accessible via 355 nm excitation, as depicted in Figure 8. Rather, the Re(py) and Re(dmap) complexes show the establishment of a 3ILCT signature that decays to the ground state with lifetimes of 27 and 44 ns, respectively. The lifetimes and assignments obtained by the ns TRIR data are consistent with the transient

Figure 6. Transient resonance Raman spectra of the Re(L) complexes and the simulated spectrum (red) of Re(Br). Spectra were recorded in CH2Cl2 with 355 nm excitation with approximately 1 mJ per pulse.

identical, providing strong evidence that the same type of longlived state, likely with the same orbital parentage, is populated in each case, consistent with calculations. The excited state bands at 907, 997, 1165, 1574 and 1592 cm−1 are consistent with the formation of a TPA radical cation,56−58 which is further confirmed by means of the simulated TR2 spectra as shown exemplarily for Re(Br) in Figure 6. The quantum chemical calculations performed within the equilibrium structure of the 3ILCT (T1) state are in excellent agreement with the experimental data. Hence, the THEXI state can be characterized as an ILCT state across all 4539


Article


within the lowest energy absorption band. As the electron withdrawing character of the ancillary ligand is increased, going from Br− to py, the MLCT energy blue-shifts and the ILCT energy red-shifts, which is consistent with the calculated data. Likewise, increasing the ligand field strength from Br− to Cl−, results in a decreased energy gap between the dd and emitting state. The similarity of the TR2 spectra across all four complexes reveals that the lowest lying electronic state is consistently of 3ILCT character, with the experimental and simulated TR2 spectra showing excellent correlation. The nature of the ancillary ligand plays a pivotal role in determining the emissive lifetime of the complex without altering the nature of the emitting state. There are no clear indicators for deactivation processes occurring by other means, such as through close lying states. It appears that the influence of the ancillary ligand on the excited state lifetime could be explained in terms of energy-gap law. This study shows that the excitedstate dynamic properties can be altered by orders of magnitude though the use of ancillary ligands that have no obvious direct effect on the molecular orbitals mainly responsible for the lowest energy excited state.

Figure 8. Illustrative representation of the nature, relative energies and dynamics of the excited states present in the halide (left) and nonhalide (right) complexes studied, with the blue arrow indicating λex = 355 nm.

absorption and emission measurements. The somewhat shorter TRIR lifetimes are likely a result of self-quenching effects due to the higher concentrations used.34 The [Re(L)(CO)3(phenTPA)]n+ compounds also exhibit some complicated ultrafast dynamic processes which require further investigation. The quantum chemical simulations reveal for all four complexes very similar TRIR spectra (see Figure S21a−d), whereas the relaxation into the 3ILCT equilibria mimics large delay times associated with the population of the 3ILCT state observed experimentally. Comparison of the calculated totally symmetric in-phase, totally symmetric out-of-phase and asymmetric CO vibrational modes61−63 yields bathochromic shifts of 21−23 cm−1, 28−30 cm−1 and 27−32 cm−1, respectively. The photophysical properties of these systems are unusual as they differ in a number of ways to those of more widely studied materials in inorganic photochemistry. First, the MLCT state is not the lowest excited state and it appears that the metal plays only a minor role in the observed photophysics. The lowest excited state is ILCT in nature, and indeed the MOs involved in those states do not appear to change with ancillary ligand substitution. However, the ancillary ligand has a significant effect on the excited state lifetimes. Interestingly the use of a lower field strength ligand increases the excited state lifetime and, on the basis of variable temperature studies, the gap between the lowest lying 3ILCT state and the dd state increases.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b12868. Synthesis and characterization data, spectroscopic and computational methods, resonance Raman, simulated absorption, molecular orbitals, simulated resonance Raman, vibrational modes, simulated excited state absorption, variable temperature lifetimes, simulated TR2, state energies, simulated TRIR, experimental TRIR (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*[email protected] *[email protected] *[email protected] ORCID

Philipp Traber: 0000-0003-3392-2573 Nigel T. Lucas: 0000-0002-4944-9875 James D. Crowley: 0000-0002-3364-2267 Stephan Kupfer: 0000-0002-6428-7528 Keith C. Gordon: 0000-0003-2833-6166

■

CONCLUSION Alteration of the electron withdrawing ability of the ancillary ligand in a series of fac-[Re(L)(CO)3(α-diimine)]n+ complexes causes dramatic changes in the lowest energy excited state lifetimes, while the character of the emissive excited state (ILCT) remains unaltered. The nature of this lowest excited state is confirmed using transient resonance Raman and timeresolved infrared spectra; in addition to these experimental spectra, the excited state spectra were simulated using DFT and TDDFT methods including relativistic effects. The primary absorbing chromophore is also ILCT in nature, as evidenced by the resonance Raman spectra and TDDFT simulations assessing the RR intensity pattern. The experimental and theoretical analyses point to the presence of two low-lying ILCT states, S1 and S2, with slightly different electronic character resulting in variations of the oscillator strength. The MLCT state S4 remains the highest CT state encompassed

Present Address ⊥

Raphael Horvath: IR Sweep AG, Laubisrütistrasse 44, 8712 Stäfa, Switzerland.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Support from the University of Otago, the Friedrich Schiller University Jena, the MacDiarmid Institute and the University of Nottingham (L.W. and K.E.A.R.) is gratefully acknowledged. S.K. thanks the Abbe Center of Photonics for financial support within the ACP Explore project. All calculations were performed at the Universitätsrechenzentrum of the Friedrich Schiller University of Jena. M.W.G. gratefully acknowledges the award of a Li Dak Sum Char Professorship from the University of Nottingham, Ningbo China. 4540


Article


■

(30) Larsen, C. B.; van der Salm, H.; Clark, C. A.; Elliott, A. B. S.; Fraser, M. G.; Horvath, R.; Lucas, N. T.; Sun, X.-Z.; George, M. W.; Gordon, K. C. Inorg. Chem. 2014, 53, 1339−1354. (31) Shillito, G. E.; Larsen, C. B.; McLay, J. R. W.; Lucas, N. T.; Gordon, K. C. Inorg. Chem. 2016, 55, 11170−11184. (32) Goze, C.; Leiggener, C.; Liu, S. X.; Sanguinet, L.; Levillain, E.; Hauser, A.; Decurtins, S. ChemPhysChem 2007, 8, 1504−12. (33) Jia, C.; Liu, S. X.; Tanner, C.; Leiggener, C.; Neels, A.; Sanguinet, L.; Levillain, E.; Leutwyler, S.; Hauser, A.; Decurtins, S. Chem. - Eur. J. 2007, 13, 3804−12. (34) Adams, B. S.; Shillito, G. E.; van der Salm, H.; Horvath, R.; Larsen, C. B.; Sun, X.-Z.; Lucas, N. T.; George, M. W.; Gordon, K. C. Inorg. Chem. 2017, 56, 12967−12977. (35) Huff, G. S.; Lo, W. K. C.; Horvath, R.; Turner, J. O.; Sun, X.-Z.; Weal, G. R.; Davidson, H. J.; Kennedy, A. D. W.; McAdam, C. J.; Crowley, J. D.; George, M. W.; Gordon, K. C. Inorg. Chem. 2016, 55, 12238−12253. (36) Heller, E. J.; Sundberg, R.; Tannor, D. J. Phys. Chem. 1982, 86, 1822−1833. (37) Wozna, A.; Kapturkiewicz, A. Phys. Chem. Chem. Phys. 2015, 17, 30468−30480. (38) Wallace, L.; Jackman, D. C.; Rillema, D. P.; Merkert, J. W. Inorg. Chem. 1995, 34, 5210−5214. (39) Brennaman, M. K.; Alstrum-Acevedo, J. H.; Fleming, C. N.; Jang, P.; Meyer, T. J.; Papanikolas, J. M. J. Am. Chem. Soc. 2002, 124, 15094−15098. (40) Brennaman, M. K.; Meyer, T. J.; Papanikolas, J. M. J. Phys. Chem. A 2004, 108, 9938−9944. (41) Sacksteder, L.; Lee, M.; Demas, J. N.; DeGraff, B. A. J. Am. Chem. Soc. 1993, 115, 8230−8238. (42) Litke, S. V.; Ershov, A. Y.; Meyer, T. J. J. Phys. Chem. A 2011, 115, 14235−14242. (43) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; von Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85−277. (44) Abrahamsson, M.; Jäger, M.; Kumar, R. J.; Ö sterman, T.; Persson, P.; Becker, H.-C.; Johansson, O.; Hammarström, L. J. Am. Chem. Soc. 2008, 130, 15533−15542. (45) Farrell, I. R.; Hartl, F.; Zalis, S.; Mahabiersing, T.; Vlcek, J. A. J. Chem. Soc., Dalton Trans. 2000, 4323−4331. (46) Farrell, I. R.; van Slageren, J.; Záliš, S.; Vlček, A. n. Inorg. Chim. Acta 2001, 315, 44−52. (47) Stor, G.; Stufkens, D.; Oskam, A. Inorg. Chem. 1992, 31, 1318− 1319. (48) Blanco-Rodríguez, A. M.; Kvapilová, H.; Sýkora, J.; Towrie, M.; Nervi, C.; Volpi, G.; Záliš, S.; Vlček, A. J. Am. Chem. Soc. 2014, 136, 5963−5973. (49) Gourlaouen, C.; Eng, J.; Otsuka, M.; Gindensperger, E.; Daniel, C. J. Chem. Theory Comput. 2015, 11, 99−110. (50) Fumanal, M.; Daniel, C. J. Comput. Chem. 2016, 37, 2454−2466. (51) Harabuchi, Y.; Eng, J.; Gindensperger, E.; Taketsugu, T.; Maeda, S.; Daniel, C. J. Chem. Theory Comput. 2016, 12, 2335−2345. (52) Englman, R.; Jortner, J. Mol. Phys. 1970, 18, 145−164. (53) Caspar, J. V.; Kober, E. M.; Sullivan, B. P.; Meyer, T. J. J. Am. Chem. Soc. 1982, 104, 630−632. (54) Caspar, J. V.; Meyer, T. J. Inorg. Chem. 1983, 22, 2444−2453. (55) Benniston, A. C.; Matousek, P.; Parker, A. W. J. Raman Spectrosc. 2000, 31, 503−507. (56) Oyama, M.; Nozaki, K.; Hatano, H.; Okazaki, S. Bull. Chem. Soc. Jpn. 1988, 61, 4283−4288. (57) Oyama, M.; Okada, M.; Okazaki, S. J. Electroanal. Chem. 1993, 346, 281−290. (58) Kulszewicz-Bajer, I.; Louarn, G.; Djurado, D.; Skorka, L.; Szymanski, M.; Mevellec, J. Y.; Rols, S.; Pron, A. J. Phys. Chem. B 2014, 118, 5278−88. (59) Damrauer, N. H.; Cerullo, G.; Yeh, A.; Boussie, T. R.; Shank, C. V.; McCusker, J. K. Science 1997, 275, 54−57. (60) McCusker, J. K. Acc. Chem. Res. 2003, 36, 876−887. (61) Bredenbeck, J.; Helbing, J.; Hamm, P. J. Am. Chem. Soc. 2004, 126, 990−991.

REFERENCES

(1) Caspar, J. V.; Meyer, T. J. J. Phys. Chem. 1983, 87, 952−957. (2) Wrighton, M.; Morse, D. L. J. Am. Chem. Soc. 1974, 96, 998− 1003. (3) Worl, L. A.; Duesing, R.; Chen, P.; Ciana, L. D.; Meyer, T. J. J. Chem. Soc., Dalton Trans. 1991, 849−858. (4) Gabrielsson, A.; Busby, M.; Matousek, P.; Towrie, M.; Hevia, E.; Cuesta, L.; Perez, J.; Záliš, S.; Vlček, A. Inorg. Chem. 2006, 45, 9789− 9797. (5) Nieuwenhuis, H. A.; Stufkens, D. J.; Vlček, A. Inorg. Chem. 1995, 34, 3879−3886. (6) Rossenaar, B. D.; Stufkens, D. J.; Vlček, A. Inorg. Chem. 1996, 35, 2902−2909. (7) Vlček, A. Ultrafast Excited-State Processes in Re(I) CarbonylDiimine Complexes: From Excitation to Photochemistry. In Photophysics of Organometallics; Lees, A. J., Ed.; Springer: Berlin, Heidelberg, 2010; pp 115−158. (8) El Nahhas, A.; Consani, C.; Blanco-Rodriguez, A. M.; Lancaster, K. M.; Braem, O.; Cannizzo, A.; Towrie, M.; Clark, I. P.; Zalis, S.; Chergui, M.; Vlcek, A. Inorg. Chem. 2011, 50, 2932−2943. (9) Schoonover, J. R.; Strouse, G. F.; Dyer, R. B.; Bates, W. D.; Chen, P.; Meyer, T. J. Inorg. Chem. 1996, 35, 273−274. (10) Sanhueza, L.; Barrera, M.; Crivelli, I. Polyhedron 2013, 57, 94− 104. (11) Striplin, D. R.; Crosby, G. A. Chem. Phys. Lett. 1994, 221, 426− 30. (12) Stufkens, D. J.; Vlček, A. Coord. Chem. Rev. 1998, 177, 127−179. (13) Vaughan, J. G.; Reid, B. L.; Wright, P. J.; Ramchandani, S.; Skelton, B. W.; Raiteri, P.; Muzzioli, S.; Brown, D. H.; Stagni, S.; Massi, M. Inorg. Chem. 2014, 53, 3629−3641. (14) Wallace, L.; Rillema, D. P. Inorg. Chem. 1993, 32, 3836−3843. (15) Walsh, P. J.; Gordon, K. C.; Lundin, N. J.; Blackman, A. G. J. Phys. Chem. A 2005, 109, 5933−5942. (16) Lees, A. J. Chem. Rev. 1987, 87, 711−743. (17) Pomestchenko, I. E.; Polyansky, D. E.; Castellano, F. N. Inorg. Chem. 2005, 44, 3412−3421. (18) Dyer, J.; Blau, W. J.; Coates, C. G.; Creely, C. M.; Gavey, J. D.; George, M. W.; Grills, D. C.; Hudson, S.; Kelly, J. M.; Matousek, P.; McGarvey, J. J.; McMaster, J.; Parker, A. W.; Towrie, M.; Weinstein, J. A. Photochem. Photobiol. Sci. 2003, 2, 542. (19) Kuimova, M. K.; Alsindi, W. Z.; Blake, A. J.; Davies, E. S.; Lampus, D. J.; Matousek, P.; McMaster, J.; Parker, A. W.; Towrie, M.; Sun, X.-Z.; Wilson, C.; George, M. W. Inorg. Chem. 2008, 47, 9857− 9869. (20) Dyer, J.; Creely, C. M.; Penedo, J. C.; Grills, D. C.; Hudson, S.; Matousek, P.; Parker, A. W.; Towrie, M.; Kelly, J. M.; George, M. W. Photochem. Photobiol. Sci. 2007, 6, 741−748. (21) Kuimova, M. K.; Gordon, K. C.; Howell, S. L.; Matousek, P.; Parker, A. W.; Sun, X.-Z.; Towrie, M.; George, M. W. Photochem. Photobiol. Sci. 2006, 5, 82−87. (22) Elliott, A. B.; Horvath, R.; Sun, X.-Z.; Gardiner, M. G.; Müllen, K.; Lucas, N. T.; George, M. W.; Gordon, K. C. Inorg. Chem. 2016, 55, 4710−4719. (23) Horvath, R.; Fraser, M. G.; Clark, C. A.; Sun, X.-Z.; George, M. W.; Gordon, K. C. Inorg. Chem. 2015, 54, 11697−11708. (24) Fraser, M. G.; Clark, C. A.; Horvath, R.; Lind, S. J.; Blackman, A. G.; Sun, X.-Z.; George, M. W.; Gordon, K. C. Inorg. Chem. 2011, 50, 6093−6106. (25) Horvath, R.; Huff, G. S.; Gordon, K. C.; George, M. W. Coord. Chem. Rev. 2016, 325, 41−58. (26) van der Salm, H.; Fraser, M. G.; Horvath, R.; Turner, J. O.; Greetham, G. M.; Clark, I. P.; Towrie, M.; Lucas, N. T.; George, M. W.; Gordon, K. C. Inorg. Chem. 2014, 53, 13049−60. (27) van der Salm, H.; Larsen, C. B.; McLay, J. R.; Fraser, M. G.; Lucas, N. T.; Gordon, K. C. Dalton Trans 2014, 43, 17775−85. (28) van der Salm, H.; Larsen, C. B.; McLay, J. R. W.; Huff, G. S.; Gordon, K. C. Inorg. Chim. Acta 2015, 428, 1−7. (29) Fraser, M. G.; Blackman, A. G.; Irwin, G. I.; Easton, C. P.; Gordon, K. C. Inorg. Chem. 2010, 49, 5180−9. 4541


Article

Journal of the American Chemical Society (62) Dattelbaum, D. M.; Omberg, K. M.; Hay, P. J.; Gebhart, N. L.; Martin, R. L.; Schoonover, J. R.; Meyer, T. J. J. Phys. Chem. A 2004, 108, 3527−3536. (63) Dattelbaum, D. M.; Omberg, K. M.; Schoonover, J. R.; Martin, R. L.; Meyer, T. J. Inorg. Chem. 2002, 41, 6071−6079.

4542


ILCT lifetimes using ancillary ligands in - American Chemical Society

Recommend Documents