DFT Investigation of Ligand Photodissociation in [Ru^II(tpy)(bpy)(py)]²⁺ and [Ru^II(tpy)(Me₂bpy)(py)]²⁺ Complexes

Abstract

Photoinduced ligand dissociation of pyridine occurs much more readily in [Ru(tpy)(Me₂bpy)(py)]²⁺ than in [Ru(tpy)(bpy)(py)]²⁺ (tpy = 2,2′:6′,2″-terpyridine; bpy = 2,2′-bipyridine, Me₂bpy = 6,6′-dimethyl-2,2′-bipyridine; py = pyridine). The S₀ ground state and the ³MLCT and ³MC excited states of these complexes have been studied using BP86 density functional theory with the SDD basis set and effective core potential on Ru and the 6–31G(d) basis set for the rest of the atoms. In both complexes, excitation by visible light and intersystem crossing leads to a ³MLCT state in which an electron from a Ru d orbital has been promoted to a π* orbital of terpyridine, followed by pyridine release after internal conversion to a dissociative ³MC state. Interaction between the methyl groups and the other ligands causes significantly more strain in [Ru(tpy)(Me₂bpy)(py)]²⁺ than in [Ru(tpy)(bpy)(py)]²⁺, in both the S₀ and ³MLCT states. Transition to the dissociative ³MC states releases this strain, resulting in lower barriers for ligand dissociation from [Ru(tpy)(Me₂bpy)(py)]²⁺ than from [Ru(tpy)(bpy)(py)]²⁺. Analysis of the molecular orbitals along relaxed scans for stretching the Ru–N bonds reveals that ligand photodissociation is promoted by orbital mixing between the ligand π* orbital of tpy in the ³MLCT state and the dσ* orbitals that characterize the dissociative ³MC states. Good overlap and strong mixing occur when the Ru–N bond of the leaving ligand is perpendicular to the π* orbital of terpyridine, favoring the release of pyridine positioned in a cis fashion to the terpyridine ligand.

Graphical Abstract

INTRODUCTION

There is a highly sustained interest in photoactivatable metal complexes for applications in the broad field of solar energy conversion, including the photocatalytic production of fuels from abundant sources and photovoltaic systems, as well as compounds that have increasing potential as tools in biomedical research.^1–5 Photoactivated compounds that release biologically active species from a nontoxic metal-based chaperone in the presence of light are being developed, so that the release can be accomplished with spatiotemporal control over biological activity. Their potential as selective and specific tools for biological research as well as agents for photoactivated chemotherapy (PACT) has been noted.^6–12

Photoinduced therapies are being developed for the treatments of various disease states, including cancer and microbial infections.^4–14 Active species currently being used include established inhibitors, neurotransmitters, drugs, and their derivatives.^15–17 Therapies that rely on photoactivation overcome the downfalls of those that are currently in use which lack the ability to achieve location-specific inhibition and have low bio-availability, leading to dose escalation, drug resistance, and intensified side effects.¹⁸ A photoreleasable drug or inhibitor has the potential to minimize the risk and side effects by providing noninvasive methods for achieving high levels of control over the effects of drugs in diseased vs normal tissue.

Metal centers of interest include Pt(IV),¹⁹ Re(I),²⁰ and Ir(III)²¹ as well as complexes containing Ru(II), which have all been investigated extensively. Photoactivatable Ru(II)-centered chaperones are typically composed of tridentate or bidentate chelators, such as 2,2′:6′,2″-terpyridine (tpy), 2,2′-bipyridine (bpy), and 1,10-phenanthroline (phen) and their derivatives as ancillary ligands and one or more monodentate ligands as the active species for release. The low ligand exchange in Ru(II) complexes observed in the dark, together with their high photoreactivity, makes these complexes highly attractive as potential PACT agents. In photochemotherapy, absorption of a photon by the complex opens coordination sites on the ruthenium for binding to biomolecules, including DNA and proteins. Alternatively, the photodissociation can be used to release reactive molecules and species while the leftover ruthenium fragment is not toxic, which is a highly desirable property for chemical tools. The efficiency of a PACT agent is typically rated on the basis of its relative efficacy upon irradiation in comparison to dark conditions. Chaperone complexes that can release two or three monodentate active nitrile species have also been previously developed.^4,22,23 In these compounds, only one of the biologically active ligands usually dissociates efficiently upon irradiation.

Whereas much of the initial work in the field focused on the photorelease of nitrile-bound inhibitors and drugs, the Turro group recently designed a photoactive Ru(II) complex able to deliver pyridine and pyridine-bound inhibitors efficiently with low-energy visible light, a requirement for tissue penetration.²⁴ The release of pyridine and other N-heterocycles is important due to the very large number of active agents available that contain these functional groups, opening the field of PACT to include compounds that can achieve a method of cell death independent of oxygen concentration,¹² unlike the case for photodynamic therapy, which requires oxygen. When the octahedral orientation is distorted using 6,6′-dimethyl-2,2′-bipyridine (Me₂bpy), these complexes become more photo-reactive. [Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺ complexes (Scheme 1) are stable in the dark, and the latter releases pyridine efficiently upon irradiation with visible light, whereas pyridine ligand exchange is not observed in the former upon photoexcitation under the same conditions.²⁵

Scheme 1.

[Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺ Complexes with Atomic Numbering

It is generally accepted that photoactivated dissociation and solvolysis occurs because a dissociative triplet metal-centered state is thermally accessible from the observed triplet metal–ligand charge transfer state.^25–37 Through steric crowding about the Ru center, the Ru–N₅ and Ru–N₆ bonds are distorted in [Ru(tpy)(Me₂bpy)(py)]²⁺ relative to [Ru(tpy)-(bpy)(py)]²⁺. As a result, the energy difference between the ³MC and ³MLCT states is smaller in Ru(tpy)(Me₂bpy)(py)]²⁺, allowing for efficient population of the ³MC state and increase Higher resolution plots of the biorthogonal and canonical SOMOs are available in Figure S1 and S2 of the Supporting Information. in the quantum yield for photodissociation. In [Ru(tpy)(bpy)-(py)]²⁺,^29,34 population of the ³MC state and photodissociation are unfavorable because the ³MC state is significantly higher in energy than the ³MLCT state.

For biomedical applications, it is desirable for the complex to absorb red or near-IR wavelengths of light, which penetrate tissue more deeply than shorter wavelengths in the visible and UV regions. It is also necessary that excitation is followed by conversion to an excited state that will promote ligand dissociation. In Ru(II) complexes, excitation to a singlet metal–ligand charge transfer state (¹MLCT) state is followed by ultrafast intersystem crossing to a triplet metal–ligand charge transfer state (³MLCT), and the dissociative triplet metal-centered state (³MC) is known to be thermally accessible from the ³MLCT state.^26–37 In [Ru(tpy)(Me₂bpy)(py)]²⁺, photodissociation consistently leads to the substitution of the pyridine ligand by coordinating solvent molecules. In a previous study, we developed a molecular orbital based explanation for the selectivity of photochemical ligand dissociation in ruthenium nitrile complexes.³⁸ When the orbitals are oriented favorably, mixing between the ligand π* orbital and the Ru dσ* orbital was shown to lead to a low barrier for the conversion of the ³MLCT state to a dissociative ³MC state. The same mechanism explains the selective ligand release in ruthenium tris(2-pyridylmethyl)amine (tpa) complexes. These prior findings motivated us to explore the photodissociation of pyridine in [Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺ complexes, the subject of the present work.

COMPUTATIONAL METHODS

Electronic structure calculations were performed using Gaussian 09³⁹ and the BP86 density functional.^40,41 For a set of Ru(II) polypyridyl complexes, Gonzalez and co-workers⁴² found that BP86 showed the best state ordering and state mixing in comparison to MS-CASPT2 calculations. In preparation for our earlier study,³⁸ we examined a number of different functionals and found that the ³MC state of a RuTQA complex was 3–27 kcal/mol lower in energy than the ³MLCT state. The BP86 functional gave the smallest energy difference between the ³MLCT and ³MC states, whereas the hybrid functionals gave the largest energy differences. The SDD basis set and effective core potential^43–45 were used for the central Ru atom. The 6–31G(d) basis set^46,47 was used for the other atoms. Solvation effects in methanol were incorporated by using the implicit SMD solvation model⁴⁸ and were included during structure optimization. The optimized structures were confirmed to be minima by harmonic vibrational frequency calculations. The ¹S₀, ³MLCT, and ³MC electronic configurations were tested for SCF stability and were characterized by examining the molecular orbital populations and the spin densities. GaussView⁴⁹ was used to generate isodensity plots of the spin densities (isovalue 0.004 au), the canonical orbitals, and biorthogonal/corresponding orbitals⁵⁰ (isovalue 0.04 au). To explore the potential energy surfaces for dissociation, relaxed potential energy surface scans were performed by stretching selected Ru–N bonds while the remaining coordinates were optimized. Transition states were obtained by optimizing the highest energy structures of the relaxed scans and were confirmed to have only one imaginary vibrational frequency.

RESULTS AND DISCUSSION

The photoinduced ligand dissociation in the [Ru(tpy)(bpy)-(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺ complexes (Scheme 1) has been studied experimentally and reported previously.^24,34 Upon irradiation with visible light, the pyridine ligand dissociates from [Ru(tpy)(Me₂bpy)(py)]²⁺ much more readily than from [Ru(tpy)(bpy)(py)]²⁺, and in fact, ligand exchange from the latter is not observed under certain irradiation conditions. Typically, the ¹MLCT excited states of Ru complexes convert rapidly to a lower ³MLCT state by intersystem crossing. It is generally accepted that ligand dissociation occurs via internal conversion of the ³MLCT states to a dissociative ³MC state.^25–37 Therefore, exploration of the triplet potential energy surface is key to the understanding of the photodissociative behavior of these Ru complexes. The molecular orbitals and spin densities for the lowest ³MLCT and the three lowest ³MC states of the [Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺ complex are shown in Figures 1 and 2 and in Figures S1 and S2 of the Supporting Information.

Figure 1.

Biorthogonal orbitals of SOMO₁ and SOMO₂ for the ³MLCT and ³MC states of (a) [Ru(tpy)(bpy)(py)]²⁺ and (b) [Ru(tpy)(Me₂bpy)(py)]²⁺.

Figure 2.

Spin density plots for the ³MLCT and ³MC states of (a) [Ru(tpy)(bpy)(py)]²⁺ and (b) [Ru(tpy)(Me₂bpy)(py)]²⁺.

The difference between the ³MLCT and ³MC states can be discerned from the singly occupied molecular orbitals (SOMOs) and from the different spin densities on Ru. For the ³MLCT state, SOMO₁ is a dπ orbital of Ru and SOMO₂ is a π* orbital of the tpy ligand, resulting in Mulliken spin densities of 0.93 and 0.98 on Ru for [Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺, respectively. The two unpaired electrons in the ³MC states occupy a dπ orbital and a dσ* orbital on Ru, yielding spin densities on Ru ranging from 1.66 to 1.83 for ³MC₁, ³MC₂, and ³MC₃ for the two complexes. Because the SOMOs of the ³MC states have dσ* antibonding character in Ru–N bonds, the various ³MC states can be found by elongating different Ru–N bonds in the ³MLCT excited state structure. The nature of these ³MC states can be understood in terms of the different Ru d orbitals involved in the SOMOs. In the ³MC₁ state, the Ru–N₆ bond dissociates and SOMO₂ is a dσ₁* orbital of Ru which is antibonding with N₅ and N₆, while SOMO₁ is a dπ₁ orbital of Ru (comparable to SOMO₁ of the ³MLCT state). SOMO₁ and SOMO₂ in the ³MC₂ state are similar to those in ³MC₁, since the Ru–N₅ bond is elongated along the same axis as Ru–N₆. In the ³MC₃ state, the Ru–N₄ bond is broken, and SOMO₂ is a dσ₂* orbital of Ru which is antibonding with N₂ and N₄.

The Ru–N bond lengths in the S₀, ³MLCT, and ³MC states of [Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺ are compared in Table 1. Methyl substitution on the bpy ligand results in a more crowded structure and increases the calculated Ru–N₄ and Ru–N₅ bond lengths in the ground state.²⁴ Similar increases in these bond lengths are found in the excited-state structures. The changes in the N–Ru–N angles also reffect this crowding. Only modest changes in the bond lengths are seen on excitation from S₀ to ³MLCT. The S₀ to ³MLCT excitation energies for [Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)-(py)]²⁺ are 41.8 and 40.4 kcal/mol, respectively, indicating that the methyl groups have similar effects on the S₀ and ³MLCT states. Excitation to the ³MLCT state puts an electron in the tpy π* orbital, and the nodal patterns of this orbital explain the bond length changes in the tpy ligand seen on excitation (see Table S1 in the Supporting Information).

Table 1.

Selected Calculated Bond Distances (Å) and Angles (deg) for [Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺ Complexes in the S₀, ³MLCT, ³MC, and ³TS States in Methanol

	Ru–N1	Ru–N2	Ru–N3	Ru–N4	Ru–N5	Ru–N6	N1–Ru–N3	N2–Ru–N4	N5–Ru–N6	ϕc
[Ru(tpy)(bpy)(py)]2+
S0	2.080	1.979	2.088	2.094	2.070	2.130	158.7	175.1	172.6	−2.4
3MLCT	2.086	2.031	2.081	2.103	2.088	2.125	151.8	175.8	175.7	−2.1
3 MC1a	2.096	1.987	2.097	2.095	2.171		157.1	174.0		1.5
3MC2	2.137	2.006	2.090	2.160	4.593	2.215	146.2	175.2	127.5	−141.8
3MC3	2.120	2.203	2.131	4.384	2.142	2.125	147.0	135.9	170.8	137.2
3TS1b	2.097	1.985	2.087	2.115	2.170	3.036	155.0	172.2	163.0	2.7
[Ru(tpy)(Me2bpy)(py)]2+
S0	2.073	1.979	2.106	2.124	2.126	2.132	158.3	179.1	168.3	−1.4
3MLCT	2.084	2.039	2.120	2.131	2.169	2.119	149.6	176.2	169.0	−0.1
3MC1a	2.088	1.982	2.095	2.124	2.225		154.8	166.0		2.5
3MC2	2.151	2.000	2.075	2.197	4.246	2.261	148.2	169.6	132.0	−134.2
3MC3	2.115	2.165	2.118	3.961	2.179	2.152	146.0	152.0	157.3	117.4
3TS1b	2.100	1.993	2.111	2.156	2.182	2.639	154.5	173.8	165.5	−7.6
3TS2 b	2.095	1.990	2.151	2.169	2.439	2.691	149.8	177.7	152.5	−18.8

^aFive-coordinate ³MC₁ structure that has released the pyridine.

^bOptimized transition structures are shown in Figure 5.

^cDihedral angle ϕ is defined in Scheme 1.

Stretching the Ru–N₆ bond in the ³MLCT optimized structure results in conversion to the ³MC₁ state. Geometry optimization of the ³MC₁ structure leads to a weakly bound six-coordinate structure which dissociates to a five-coordinate complex and a free pyridine that is about 2 kcal/mol higher in energy. Because SOMO₂ is a dσ* orbital aligned with both the Ru–N₆ and Ru–N₅ bonds, the latter is also somewhat elongated (Table 1). Figure 3 shows the relative energies of the optimized ³MLCT and ³MC structures. The six-coordinate ³MC₁ structure for [Ru(tpy)(bpy)(py)]²⁺ is 9.8 kcal/mol higher in energy than the ³MLCT state. For [Ru(tpy)-(Me₂bpy)(py)]²⁺; the five-coordinate ³MC₁ structure is only 1.6 kcal/mol higher in energy than the ³MLCT state since elongation of the Ru–N₆ bond releases the strain from interaction between the pyridine and the methyl group of Me₂bpy.

Figure 3.

Relative energies of the ³MLCT and ³MC structures for (a) [Ru(tpy)(bpy)(py)]²⁺ and (b) [Ru(tpy)(Me₂bpy)(py)]²⁺. The arrows indicate the positions of ligand dissociation. The definition of dihedral angle ϕ is shown in Scheme 1.

Stretching of the Ru–bpy bonds leads to the ³MC₂ and ³MC₃ states. The ³MC₂ optimized structures are 15.8 and 8.1 kcal/mol higher in energy than the ³MLCT states for [Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺, respectively. The difference can be attributed to the release of strain from interaction between tpy and the methyl group of Me₂bpy in the latter complex. In the ³MC₂ structure, the Ru–N₅ bond elongates and the N₅ pyridyl group of bpy rotates away from Ru. For the ³MC₃ structures, the N₄ pyridyl group of bpy rotates away from Ru, resulting in structures that are 22.0 and 15.5 kcal/mol higher than ³MLCT for [Ru(tpy)(bpy)(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺, respectively. The fact that ³MC₁ is the lowest-energy ³MC state is consistent with the experimental results that the ligand dissociation occurs at the N₆ position.

Potential Energy Scan on Triplet Surface

The energy barriers for the ligand dissociation on the triplet surface can be estimated by conducting relaxed scans from ³MLCT and stretching various Ru–N bonds. For each scan, one Ru–N bond was chosen, elongated in steps of 0.05 Å, and the energy was minimized with respect to all of the remaining coordinates at each step of the scan. When the Ru–N₆ bond in [Ru(tpy)(bpy)(py)]²⁺ is elongated, there is a smooth transition from the ³MLCT to ³MC₁ state with a barrier of approximately 12 kcal/mol (Figure 4a). After the barrier, there is a weakly bound six-coordinate complex before the pyridine fully dissociates to the five-coordinate ³MC₁ complex. The transition state structure ³TS₁ for [Ru(tpy)(bpy)(py)]²⁺ (Figure 5a) was found by optimizing the highest energy point on the scan, yielding a barrier height of 11.5 kcal/mol.

Figure 4.

Relaxed potential energy scans from the ³MLCT state for stretching the Ru–N₆ bond of (a) [Ru(tpy)(bpy)(py)]²⁺ and (b) [Ru(tpy)(Me₂bpy)(py)]²⁺. The red squares indicate the optimized transition states. The values along the scan show the spin densities on Ru. For (b), the purple line is the IRC in the forward direction from ³TS₁ and the green line is the relaxed scan in the reverse direction from Ru–N6 = 2.67 Å.

Figure 5.

Optimized transition state geometries and spin density plots for the conversion from ³MLCT to ³MC₁: (a) ³TS₁ for [Ru(tpy)(bpy)(py)]²⁺; (b) ³TS₁ for [Ru(tpy)(Me₂bpy)(py)]²⁺; (c) ³TS₂ for [Ru(tpy)(Me₂bpy)(py)]²⁺.

When the Ru–N₆ bond in [Ru(tpy)(Me₂bpy)(py)]²⁺ is elongated, the transition from the ³MLCT to ³MC₁ state occurs at a shorter distance and has a barrier of only approximately 7 kcal/mol (Figure 4b). Transition state ³TS₁ for [Ru(tpy)(Me₂bpy)(py)]²⁺ (Figure 5b) was found by optimizing the highest energy point on the scan. The optimized ³TS₁ is 6.9 kcal/mol higher in energy than the ³MLCT structure and has one imaginary frequency which corresponds to stretching of the Ru–N₆ bond. Following the IRC and the relaxed scan confirms that this transition state connects to the ³MC₁ structure. Structure ³TS₁ has a spin density of 1.38 on Ru (midway between 0.98 in ³MLCT and 1.66 in ³MC₁) and a Ru–N₆ bond length of 2.639 Å. The differences between [Ru(tpy)(bpy)-(py)]²⁺ and [Ru(tpy)(Me₂bpy)(py)]²⁺ in both the barrier heights and the Ru–N₆ bond lengths can be understood in terms of an avoided crossing between the potential energy surfaces of the ³MLCT and ³MC₁ states as the Ru–N₆ bond is stretched (see Figure 3). Because the ³MLCT to ³MC₁ energy difference is smaller for [Ru(tpy)(Me₂bpy)(py)]²⁺ than for [Ru(tpy)(bpy)(py)]²⁺, the avoided crossing between the ³MLCT and ³MC₁ states occurs at a lower energy and shorter bond length for [Ru(tpy)(Me₂bpy)(py)]²⁺. Other coordinates such as the Ru–N₅ bond length and various N–Ru–N angles also indicate that the transition state for [Ru(tpy)(Me₂bpy)(py)]²⁺ occurs earlier along the reaction path with a greater release of strain in comparison to the transition state for [Ru(tpy)(bpy)(py)]²⁺. Because the conversion of ³MLCT to ³MC involves electron transfer from the tpy ligand to Ru, some changes are also observed in the bond length of the tpy ligand (see Table S1 in the Supporting Information).

After the high point on the scan of the Ru–N₆ in [Ru(tpy)(Me₂bpy)(py)]²⁺, there is a 1.4 kcal/mol drop in energy (Figure 4b) and a 0.298 Å lengthening of the Ru–N₅ bond (Figure S3 in the Supporting Information). Continuing the scan in the forward direction leads to the ³MC₁ structure. The energy decreases monotonically while the Ru–N₅ bond shortens and the Me₂bpy ligand twists to a lower energy geometry. Scanning the Ru–N₆ bond in the reverse direction also produces a monotonic decrease in energy (Figure 4b, green line), leading to the ³MLCT structure. Optimizing the highest energy point on this scan results in transition structure ³TS₂ (Figure 5c), which is 5.4 kcal/mol above the ³MLCT state. The Ru–N₆ bond length in ³TS₂ is similar to that in ³TS₁ but the Ru–N₅ bond length is significantly longer and the spin density on Ru is higher. As discussed in the next paragraph, ³TS₂ can also be found by stretching the Ru–N₅ bond in the ³MLCT structure and then stretching the Ru–N₆ bond. Thus, ³TS₂ represents the barrier for a second, lower energy pathway for dissociation of the ³MLCT state to form the ³MC₁ and pyridine.

Elongation of the Ru–N₅ bond perpendicular to the tpy plane results in a smooth increase in the energy from the ³MLCT state to the ³MC₂ state (Figure 6). The increase for is significantly smaller than for [Ru(tpy)(Me₂bpy)(py)]²⁺ [Ru(tpy)(bpy)(py)]²⁺. When the Ru–N₅ bond in [Ru(tpy)-(Me₂bpy)(py)]²⁺ is stretched to 2.62 Å, the potential energy increases by 4.05 kcal/mol, the spin density on Ru increases gradually to 1.62, and the Ru–N₆ bond elongates slightly to 2.26 Å. Because the bpy ligand is tethered to Ru at the N₄ position, the N₅ pyridyl does not dissociate. When the Ru–N₅ bond is stretched beyond 2.62 Å, the potential energy continues to increase because of the twisting of the bpy ligand. If the Ru–N₅ bond is frozen at 2.62 Å and another relaxed scan is conducted by stretching the Ru–N₆ bond, the energy increases monotonically until the pyridine dissociates. Optimization of the highest point on this scan results in transition state ³TS₂ (Figure 5c).

Figure 6.

Relaxed potential energy scans from the ³MLCT state for stretching the Ru–N₅ bond in (a) [Ru(tpy)(bpy)(py)]²⁺ and (b) [Ru(tpy)(Me₂bpy)(py)]²⁺. The red line represents the relaxed scan for stretching the Ru–N₆ bond in [Ru(tpy)(Me₂bpy)(py)]²⁺ when the Ru–N₅ bond is frozen at 2.62 Å. The values along the scan show the spin densities on Ru.

When the Ru–N₄ bond, which is coplanar with the tpy ligand, is stretched, the estimated barrier for the transition to the ³MC₃ state is 23 kcal/mol for [Ru(tpy)(bpy)(py)]²⁺ and 15 kcal/mol for [Ru(tpy)(Me₂bpy)(py)]²⁺, values that are significantly higher than for the conversion to the ³MC₁ and ³MC₂ states (Figure 7). The higher barriers for breaking the Ru–N₄ bond are consistent with experiment, which did not find photodissociation of the bpy ligand.

Figure 7.

Relaxed potential energy scans from the ³MLCT state for stretching the Ru–N₄ bond in (a) [Ru(tpy)(bpy)(py)]²⁺ and (b) [Ru(tpy)(Me₂bpy)(py)]²⁺. The values along the scan show the spin densities on Ru.

MO Analysis along the Relaxed Scans

In our previous study³⁸ we analyzed the SOMOs along the relaxed potential energy scans and found that photodissociation of the nitrile-bound Ru polypyridyl complexes is facilitated by orbital mixing between the ligand π* orbital of the ³MLCT state and the Ru dσ* orbitals of a dissociative ³MC state. Figure 8 shows the corresponding plots for the SOMOs of [Ru(tpy)(Me₂bpy)(py)]²⁺ as the Ru–N₆, Ru–N₅, and Ru–N₄ bonds are stretched. SOMO₂ in the ³MLCT state is a ligand-based π* orbital on tpy. When the Ru–N₆ bond is stretched longer than 2.52 Å, the ligand-based SOMO₂ mixes with the dσ₁* orbital of Ru, which corresponds to SOMO₂ of ³MC₁ (Figure 8a). This orbital mixing promotes dissociation because the dσ₁* orbital involves an antibonding interaction with the pyridine ligand. Stretching the Ru–N₅ bond (trans to Ru–N₆) also leads to similar orbital mixing of the tpy π* orbital and the Ru dσ₁* orbital (Figure 8b). However, the Ru–N₅ bond does not dissociate because the bpy ligand is still tethered by the Ru–N₄ bond. Further elongation of the Ru–N₅ bond results in an increase in energy because of the twisting of the bpy ligand. In contrast to the orbital mixing seen when the Ru–N₅ and Ru–N₆ bonds are stretched, when the Ru–N₄ bond is stretched, the π* orbital of tpy remains orthogonal to the Ru dσ₂* orbital that corresponds to SOMO₂ of ³MC₃, and no mixing occurs (Figure 8c). In addition, the rigidity of the bpy ligand restricts the motion of the Ru–N₄ bond. As a result, the barrier for the transition to the ³MC₃ state is high.

Figure 8.

Isosurface plots of the SOMOs of [Ru(tpy)(Me₂bpy)(py)]²⁺ along the relaxed scan of the ³MLCT state for elongation of (a) Ru–N₆, (b) Ru–N₅, and (c) Ru–N₄.

CONCLUSIONS

Experimental studies of photoinduced ligand dissociation found that the pyridine ligand is released from [Ru(tpy)(Me₂bpy)-(py)]²⁺ significantly more efficiently than from [Ru(tpy)(bpy)-(py)]²⁺. To explore the ligand dissociation reaction on the triplet surface, we have calculated the energies and geometries of the ³MLCT and dissociative ³MC states. In comparison to [Ru(tpy)(bpy)(py)]²⁺, the geometry of [Ru(tpy)(Me₂bpy)-(py)]²⁺ shows significant strain because of interaction of the methyl groups with the other ligands in both the S₀ and ³MLCT states. Transition to the dissociative ³MC states releases this strain, resulting in lower barriers for ligand dissociation for [Ru(tpy)(Me₂bpy)(py)]²⁺ than for [Ru(tpy)-(bpy)(py)]²⁺. By analyzing the molecular orbitals along relaxed scans for stretching the Ru–N bonds, we find that ligand photodissociation is promoted by orbital mixing between the ligand π* orbital of the ³MLCT state and the dσ* orbitals that characterize the dissociative ³MC states. Mixing can occur when the Ru–N₆ bond perpendicular to a π-acceptor ligand is stretched and the π* orbital of tpy and the dσ* orbital of Ru have good overlap. Orbital mixing results in a smooth and continuous transition from ³MLCT to ³MC₁ with a small barrier for photodissociation of the pyridine ligand in [Ru(tpy)(Me₂bpy)(py)]²⁺. In contrast, when the Ru–N₄ bond coplanar with the π-acceptor ligand is stretched, the ligand (tpy) π* and Ru dσ* orbitals remain orthogonal; no mixing occurs, and the barrier for the transition from ³MLCT to ³MC₃ is high. In addition to orbital mixing, ligand dissociation also depends on the rigidity of the ligand. When the Ru–N₅ bond perpendicular to the π acceptor is stretched, orbital mixing occurs but the bpy group does not dissociate from Ru; instead, bpy twists about its central bond in order to break the Ru–N₅ bond in the ³MC₂ state. Nevertheless, stretching of the Ru–N₅ bond followed by elongation of the Ru–N₆ bond can lead to a smaller barrier for transition from ³MLCT to ³MC₁, facilitating dissociation of the pyridine ligand.

This work provides an understanding of the factors that lead to enhancements in photoinduced ligand dissociation and may be used to predict the structures of complexes for drug photorelease with improved properties.