Conformational control of electron delocalisation in geometrically-constrained, binuclear ruthenium(ii) bis(2,2′:6′,2″-terpyridine) complexes

On the basis of temperature-dependent emission studies made with the target binuclear compounds, it is concluded that the extent of electron delocalisation at the triplet level depends on the torsion angle imposed by the bridge.

It is well established that the triplet lifetime of ruthenium(ii) bis(2,2′:6′,2″-terpyridine) at ambient temperatures can be prolonged by attaching an alkynylene substituent at the 4′-position.1

This strategy works particularly well for the corresponding binuclear complexes, where the triplet lifetimes at 20 °C exceed that of the parent complex by a factor of at least .10002

Prolongation of the triplet lifetime is due to two complementary effects; namely, decoupling of the lowest-energy triplet state from upper-lying metal-centred states and extended delocalisation of the promoted electron over the substituted ligand.

Electron delocalisation can occur because the lowest-energy triplet is of metal-to-ligand, charge-transfer (MLCT) character and is formed by selective charge injection into the alkyne-substituted ligand.

We now show that the extent of electron delocalisation is subject to conformational control, as imposed by constraining the geometry of the alkynylene-based bridge.

The target molecules comprise the following units: (i) The main chromophore is a ruthenium(ii) bis(2,2′:6′,2″-terpyridine) unit.

(ii) To ensure luminescence at ambient temperature, an ethynylene group is attached at the 4′-position.

(iii) A central biphenylene unit is used as a variable rotor.

(iv) A strap of predetermined length is attached at the 2,2′-positions of the rotor so as to restrict internal twisting around the connecting bond.

The generic molecular structures, and their abbreviations, are shown in Fig. 1.

These compounds were synthesized as outlined previously3 and fully characterised by 1H and 13C NMR, mass spectrometry and elemental analysis.

Each compound shows a prominent MLCT transition centred at 490 nm, with a tail stretching towards 600 nm.

The tail is due to the spin-forbidden MLCT transitions.

At higher energy, the parent terpyridine ligand absorbs around 270 nm whilst the polytopic ligand displays absorption bands between 290 and 400 nm.

Very weak luminescence can be observed in deoxygenated solution at ambient temperature.

This emission is centred around 675 nm and can be assigned to phosphorescence from the lowest-energy MLCT triplet state of the metal complex.4

At room temperature, the emission quantum yields are very low (ΦLUM ≈ 0.0004) and the lifetimes are relatively short (τLUM ≈ 20 ns).

Even so, emission is much more intense than that observed from the parent complex under these conditions.5

Luminescence from the binuclear complexes becomes more pronounced as the temperature is lowered.

In all cases, there is good agreement between the increase in quantum yield and lifetime.

The corrected excitation spectrum matches the absorption spectrum over the entire visible and near-UV regions and all spectroscopic measurements indicate that only a single species emits.

Even in a frozen glass, there is no suggestion of emission from either a ligand-centred triplet or an intramolecular charge-transfer state.6

As such, the temperature dependence is due to perturbation of the photophysical properties of the lowest-energy MLCT triplet state.

The photophysical properties of metal poly(pyridine) complexes are subject to the energy-gap law.7

But, in order to expose this relationship, it is necessary to allow for coupling between the emitting state and higher-lying triplets.

The energy of the lowest triplet excited state (ET) can be derived by detailed analysis of the emission spectrum.

For the target compounds in fluid solution, ET (≈15,800 cm–1) is independent of temperature and insensitive to the length of the strap.

Likewise, the total reorganisation energy (λT ≈ 700 cm–1) accompanying decay of the first triplet excited state remains essentially constant for all the complexes studied here.

The rate constant (kNR) for nonradiative deactivation of the lowest-energy triplet state was found to follow eqn. (1) over a wide temperature range (Fig. 2).

This expression provides for the activationless rate of decay (k0), seen at low temperature, and for coupling to two additional excited triplet states.

Of these two higher-energy triplets, one is accessed by passage over a small barrier (E1).

This upper-lying state is probably a second MLCT triplet but with increased singlet character.

The overall rate constant (k1) associated with reaching this state is rendered complex by the reversible nature of the barrier crossing.

At ambient temperature, a metal-centred (MC) state can be accessed by crossing a more substantial barrier (E2).

In this case, barrier crossing is considered to be irreversible such that the derived rate constant (k2) is set by the rate of decay of the MC state.

This latter process results in direct population of the ground state.

The various values extracted from the kinetic fits are collected in Table 1.

It is seen that only k0 depends on the length of the strap.

It is known from related systems that k0 decreases with increasing extent of electron delocalisation of the promoted electron.8

The fact that k0 depends on the length of the strap is clear indication that the conformation of the bridging unit affects the degree of electronic communication along the molecular axis.

Two possibilities need be considered on the basis of molecular dynamics simulations made for the target compounds in a bath of acetonitrile molecules.

First, the strap controls the torsion angle between the phenylene rings.

As such, the length of the strap should influence the degree of orbital overlap, in the event that the promoted electron extends to the LUMO on the biphenylene group.

Second, the strapped biphenylene unit does not lie coplanar with the covalently-linked terpyridine groups.

The mean torsion angle between nearby phenylene and pyridine rings varies markedly with strap length, especially for optimised structures.

This torsion angle controls the degree of electronic coupling between the terpyridine ligand and the ethynylene group and will have a pronounced effect on the extent of electron delocalisation over the acetylenic unit.

As mentioned above, the variation in k0 should be explainable within the framework of the energy-gap law, eqn. (2).7

The most important variable within this expression is the amount of energy to be dissipated during nonradiative decay.

However, since ET remains invariant throughout this series it is clear that other factors combine to affect the magnitude of k0.

The remaining parameters in eqn. (2) are the Huang–Rhys factor (SM),9 the electron-vibrational coupling matrix element (C) and the averaged medium-frequency vibrational mode (M) coupled to nonradiative decay.

The coefficient γ depends on both the energy gap and the extent of nuclear displacement.

Both S and M can be determined by analysis of the emission spectra recorded in fluid solution.

Usually, the full emission spectrum, at any temperature, can be expressed10 in terms of a weighted average, medium-frequency mode (M) of around 1380 cm–1.

This was not the case for the target compounds studied here since, even in fluid solution, it was necessary to include an additional low frequency mode (L) to properly express the emission spectra.

Thus, the normalised emission spectra could be reconstituted on the basis of eqn. (3)11 and using the data collected in Table 2.

It should be stressed that the inclusion of L in fluid solution is unique to those metal complexes equipped with the constraining strap.

In fitting the experimental data to eqn. (3) (Fig. 3), I(ν) is the normalised emission intensity at wavenumber ν, nM and nL are the vibrational quantum numbers for the averaged medium (M) and low (L) frequency acceptor modes, Δv1/2 is the half-width of individual vibronic bands and E00 is the energy gap for the 0–0 transition.

The quantities SM and SL are the Huang–Rhys factors for medium and low frequency modes, respectively.

Only the half-width depends on temperature and the other parameters were averaged over the entire temperature range (Table 2).

The best fits to the experimental data gave M and L values, respectively, of ca. 1400 and ca. 800 cm–1.

The latter value increases systematically with increasing temperature; e.g. for C4 L varies from ca. 535 cm–1 at 90 K to 1020 cm–1 at 290 K.

The medium-frequency mode corresponds to CC and CN stetching vibrations associated with the terpyridine ligand.12

The low frequency vibration can be assigned to substituted phenylene bending modes.13

It is now clear that motion around the bridging phenylene unit is coupled to nonradiative decay of the lowest-energy triplet state.

Presumably, this effect relates to the system trying to achieve the maximum orbital overlap at the triplet level.

It should be noted that the strap exerts a pronounced effect on k0 since there is a 90-fold increase in rate between C2 and the corresponding complex lacking the strap.

The exact details underlying this behaviour will be described elsewhere.