Catalyzing Singlet Fission by Transition Metals: Second versus Third Row Effects

Abstract

The synthesis and characterization of platinum(II) and palladium(II) complexes bearing two (dimers Pt(L_pc)₂Cl₂ and Pd(L_pc)₂Cl₂), one (monomers Pt(L_pc)(L_ref)Cl₂ and Pd(L_pc)(L_ref)Cl₂), or no (reference compounds Pt(L_ref)₂Cl₂ and Pd(L_ref)₂Cl₂) pentacene-based pyridyl ligands are presented. Photophysical properties of the dimers are probed by means of steady-state and time-resolved transient absorption measurements in comparison to the monomer and model compounds. Our results document that despite enhanced spin–orbit coupling from the presence of heavy atoms, intramolecular singlet fission (iSF) is not challenged by intersystem crossing. iSF thus yields correlated triplet pairs and even uncorrelated triplet excited states upon decoherence. Importantly, significant separation of the two pentacenyl groups facilitates decoupling of the two chromophores. Furthermore, the mechanism of iSF is altered depending on the respective metal center, that is, Pt(II) versus Pd(II). The dimer based on Pt(II), Pt(L_pc)₂Cl₂, exhibits a direct pathway for the iSF and forms a correlated triplet pair with singlet–quintet spin-mixing within 10 ns in variable solvents. On the other hand, the dimer based on Pd(II), Pd(L_pc)₂Cl₂, leads to charge transfer mixing during the population of the correlated triplet pair that is dependent on solvent polarity. Moreover, Pd(L_pc)₂Cl₂ gives rise to a stable equilibrium between singlet and quintet correlated triplet pairs with lifetimes of up to 170 ns. Inherent differences in the size and polarizability, when contrasting platinum(II) with palladium(II), are the most likely rationale for the underlying trends.

Introduction

Singlet fission (SF) is a spin-allowed photophysical process, which allows the generation of two pairs of charges from absorption of a single photon, and it provides an avenue to break the detailed balance limit of single junction solar cells from ca. 32% to over 45%.^1,2 When SF occurs, a chromophore in its singlet excited state (S₁) interacts with a nearby chromophore in its ground state (S₀), forming a spin-correlated triplet pair state ¹(T₁T₁).³⁻⁶ Subsequently, the triplet pair can undergo decorrelation, resulting in two noninteracting, free triplet states (T₁ + T₁). These triplets are then available for completing the process of solar energy conversion.^4,7−11

Over the years, many criteria have been identified for chromophores to perform effective SF. The most straightforward and basic criteria are, on one hand, the energetic relationship between the singlet (S₁) and triplet (T₁) excited states. For exothermic SF, the energy level of (S₁) must be close to twice that of (T₁), which also serves as a thermodynamic driving force for SF, namely E(S₁) > 2 × E(T₁). Alternatively, if E(S₁) is equal or slightly below 2 × E(T₁), endothermic SF is possible.^1,12−14 Prominent representatives for exothermic and endothermic SF are pentacene-based¹⁵⁻¹⁹ and anthracene-based²⁰⁻²³ chromophores, respectively. E(S₁) of pentacene is 1.83 eV, which is greater than 2 × E(T₁) with a value of 1.72 eV, thus it is a typical sample capable of SF while the opposite pathway triplet–triplet annihilation up-conversion (TTA-UC) is undetectable. Tetracene, with similar E(S₁) and 2 × E(T₁) of 2.35 eV, exhibits the coexistence of SF and TTA-UC. Conversely, anthracene has E(S₁) and 2 × E(T₁) with values of 3.30 and 3.70 eV, respectively, which is an ideal sample for TTA-UC.^2,24 The energy level of any higher triplet excited states, such as (T₂), should exceed 2 × E(T₁) to avoid its population via triplet–triplet annihilation (TTA) and avoid intersystem crossing (ISC) between (S₁) and (T₂), which is quite common in the case of strongly coupled chromophores.^1,12,25 On the other hand, sufficient electronic interactions between the interacting chromophores need to be assured. Coupling is typically realized by either high concentrations in solution, or through control of the crystal packing in the solid state,^15,26−31 and thus efficient intermolecular SF evolves. An alternative strategy relies on a spacer that provides molecular dimers and oligomers that drive intramolecular interactions for SF (iSF). The merit of oligomers is the control over, for instance, spatial overlap and degree of coupling, which offer the ability to fine-tune and analyze the impact of coupling via the spacer on the intricacies of the mechanism, kinetics, and efficiency of iSF as a whole.^{14,18,32−36} Despite the broad accomplishments toward efficient formation of ¹(T₁T₁), the underlying mechanism of decorrelation toward formation of free triplet states (T₁ + T₁) remains to be tested and optimized.

Conventionally, the design of SF materials excludes heavy atoms, since spin–orbit coupling (SOC) facilitates ISC. As well, the synthesis of heavy-atom chromophores can be more challenging when compared to small organic molecules.^2,37,38 Triplets born from ISC are fundamentally different from triplets of SF and are clearly distinguishable.³⁹⁻⁴¹ Studies have shown that SOC enhancement produced by a heavy atom, the so-called heavy-atom effect (HAE), can have profound effects on the decorrelation of the spin-correlated triplet pair (T₁T₁). Musser et al. have shown that the spin evolution within (T₁T₁) is faster through replacing sulfur in polythienylenevinylene with the heavy atoms selenium and tellurium. They suggest that the presence of heavy atoms may manipulate the dynamics of triplets formed by SF.⁴² Pt-bridged pentacene dimers have been synthesized, and the impact of the HAE has been investigated.^40,43 It is reported that the presence of Pt has no impact on the formation of the singlet correlated triplet pair ¹(T₁T₁) from (S₁S₀), but Pt did influence subsequent transitions (Figure 1). The enhanced SOC on ¹(T₁T₁) caused by Pt leads to a spin-flip of ¹(T₁T₁) into both the triplet correlated triplet pair ³(T₁T₁) and the quintet correlated triplet pair ⁵(T₁T₁). Ultimately, population of the triplet excited states (S₀T₁) and (T₁ + T₁) results from ³(T₁T₁) and ⁵(T₁T₁), respectively. Thus, enhanced SOC produced by a heavy atom may improve the yield of decorrelation. The presence of heavy atoms, however, is not always beneficial to SF. Recently, a tetramer has been synthesized through coordinating pyridyl pentacene-based ligands to Ru(II), and, in this case, the presence of the heavy atom gives rise to ultrafast ISC following photoexcitation rather than SF.⁴⁴ Therefore, the influence of HAE on triplet decorrelation in SF remains unpredictable.

Figure 1.

Comparison of iSF between cis-Pt with SOC and m-2 without SOC.^40,45

Intrigued by these findings, we designed and synthesized two pentacene dimers that use heavy atoms Pt and Pd to link pyridyl pentacene ligands (Scheme 1). The pyridyl group is ubiquitous for bonding with metal ions by coordination and allows the versatile synthesis of pentacene oligomers featuring different metal ions.^44,46 It is important to emphasize that by design, the two dimers, Pd(L_pc)₂Cl₂ and Pt(L_pc)₂Cl₂, are structurally identical except for the metal ion. As second- and third-row metals, Pd and Pt are congeners, and both dimers Pd(L_pc)₂Cl₂ and Pd(L_pc)₂Cl₂ feature metals in the M(II) oxidation state. Thus, while the valence electron structure is the same, the HAE is reduced in Pd compared to Pt as the effective nuclear charge of Pd is lower than Pt.⁴⁷⁻⁴⁹ In comparison to suitable model compounds that cannot undergo iSF, the electronic influence of the heavy atoms and their impact on the mechanism of SF can be carefully studied. In the presence of Pt in dimer Pt(L_pc)₂Cl₂, analyses show that ¹(T₁T₁) forms directly from iSF. Conversely, the incorporation of Pd in dimer Pd(L_pc)₂Cl₂ leads to the formation of ¹(T₁T₁) via a mediating species likely due to the lower spin–orbit coupling.

Scheme 1. Synthesis of Dimers M(L_pc)₂Cl₂, Monomers M(L_pc)(L_ref)Cl₂, and Model Compounds M(L_ref)₂Cl₂.

Reagents and conditions: (a) M = Pt, L_pc, toluene, 80 °C; (b) M = Pd, L_pc, CH₂Cl₂, rt; (c) M = Pt, L_ref, toluene, 110 °C then L_pc, toluene, 110 °C; (d) M = Pd, L_ref, L_pc, CH₂Cl₂, rt; (e) M = Pt, L_ref, toluene, 80 °C; (f) M = Pd, L_ref, CH₂Cl₂, rt. Inset: X-ray crystallographic analysis of Pt(L_pc)(L_ref)Cl₂ (H atoms removed for clarity, ellipsoids at 30% level).

Results and Discussion

Synthesis

The pyridyl-end-capped pentacene ligand (L_pc) and its corresponding reference pyridyl ligand without pentacene (L_ref) were synthesized following a published procedure.⁴⁴ A ligand exchange reaction of L_pc with either Pt(PhCN)₂Cl₂ or Pd(PhCN)₂Cl₂ (1a or 1b) efficiently gives the dimers Pt(L_pc)₂Cl₂ and Pd(L_pc)₂Cl₂ (Scheme 1). Unsymmetrical monomeric complexes Pt(L_pc)(L_ref)Cl₂ and Pd(L_pc)(L_ref)Cl₂ bearing a single pentacenyl unit were prepared by either a stepwise reaction for Pt(L_pc)(L_ref)Cl₂ (via the intermediate Pt(PhCN)(L_ref)Cl₂) or a one-pot reaction in the case of Pd(L_pc)(L_ref)Cl₂. While monomer Pt(L_pc)(L_ref)Cl₂ could be purified through a sequence of column chromatography and recrystallization, Pd(L_pc)(L_ref)Cl₂ was unstable to purification by column chromatography (using either SiO₂ or Al₂O₃). Ultimately, preparative gel permeation chromatography gave pure product Pd(L_pc)(L_ref)Cl₂. Finally, ligand exchange of L_ref with Pt(PhCN)₂Cl₂ or Pd(PhCN)₂Cl₂ gave the reference compounds Pt(L_ref)₂Cl₂ or Pd(L_ref)₂Cl₂. All products are highly soluble in common organic solvents such as CH₂Cl₂, CHCl₃, tetrahydrofuran (THF), and toluene (≥20 mg/mL). The complexes are all stable as solids under normal laboratory conditions. However, all compounds containing a pentacene moiety decompose in solution upon exposure to air (O₂) and light over days.

The successful formation of the desired metal complexes has been confirmed by mass spectrometry combined with ¹H and ¹³C NMR spectroscopic analyses. ¹H NMR spectra of M(L_ref)₂Cl₂ and M(L_pc)₂Cl₂ (M = Pt, Pd) mirror the ¹H NMR spectra of ligands L_ref and L_pc, as the two ligands in the metal complexes are equivalent. It is noted, however, that the signals of pyridyl protons in M(L_ref)₂Cl₂ (M = Pt, δ 8.78; and M = Pd, δ 8.70) and M(L_pc)₂Cl₂ (M = Pt, δ 9.03; and M = Pd, δ 8.95) are shifted downfield by 0.24–0.28 and 0.16–0.20 ppm for the Pt and Pd complexes (Figure S13),⁵⁰ respectively, compared to the analogous protons of the free ligands L_ref and L_pc (δ 8.54 and δ 8.75, respectively).⁴⁴ The observed trends in ¹H shifts for the pyridyl groups are consistent with literature reports, which indicate a decreased electron density upon metal coordination of a pyridyl ligand.⁵¹⁻⁵⁵ The greater deshielding of these protons when bound to Pt(II) in comparison to Pd(II) corresponds to the lower spin–orbit coupling for Pd due to its nuclear charge.^48,49

The molecular structures of monomer Pt(L_pc)(L_ref)Cl₂ and reference compounds Pt(L_ref)₂Cl₂ and Pd(L_ref)₂Cl₂ have been unambiguously established by X-ray crystallography (Scheme 1 and Figure S20). The structures confirm the trans orientation of the ligands as well as the expected square-planar geometry (full details are provided in the Supporting Information). In all cases, the opposing pyridyl ligands are coplanar, while the bulky tert-butyl substituents on the pyridyl ring prevent coplanarity between the pentacenyl and pyridyl moieties of Pt(L_pc)(L_ref)Cl₂ (with a torsion angle of ca. 16°).

Photophysics

Steady-state absorption measurements of compounds without pentacenyl groups (L_ref, Pt(L_ref)₂Cl₂, and Pd(L_ref)₂Cl₂) show absorption bands in the high energy region of the UV spectrum between 300–400 nm (Figure S21). As such, they are transparent to time-resolved transient absorption (TAS) measurements, in which 610 nm serves as the excitation wavelength of the pentacene chromophores. Furthermore, emission is not detected for any of these reference compounds based on steady-state fluorescence measurements.

The steady-state absorption spectra of monomers Pt(L_pc)(L_ref)Cl₂ and Pd(L_pc)(L_ref)Cl₂ show well-resolved vibrational fine structures (Figure 2, Table 1), in contrast to Pt-acetylide complexes bearing acenes that show broadened features.^40,43,56 Compared to the free ligand L_pc, the absorptions for monomers Pt(L_pc)(L_ref)Cl₂ and Pd(L_pc)(L_ref)Cl₂ are subtly red-shifted and slightly broadened. Extinction coefficients of both monomers and L_pc typically range between 30,000 and 40,000 M^–1 cm^–1. Increasing the solvent polarity from toluene to tetrahydrofuran (THF) and benzonitrile (BN) yields only minor differences, with only a slight broadening of the features in BN being the most visible change, especially for Pt(L_pc)(L_ref)Cl₂ (Figures S22–S24). Absorption spectra for pentacene dimers Pt(L_pc)₂Cl₂ and Pd(L_pc)₂Cl₂ show similar features. The shift in the low-energy absorptions is minimal in spite of additional π-conjugated resulting from the second pentacenyl, pointing toward the presence of an effectively decoupled system in those dimers. The extinction coefficients of Pd(L_pc)₂Cl₂ and Pt(L_pc)₂Cl₂ are nearly twice that of the monomers, which further reinforced this fact (Figure 2, Table 1). Interestingly, the change of solvent to BN has a stronger effect on dimers than monomers, with a noticeable drop in extinction coefficient, suggesting possible changes in the electronic ground state of Pt(L_pc)₂Cl₂ and Pd(L_pc)₂Cl₂ (Figures S22–S24).

Figure 2.

Steady-state absorption (toluene) of L_pc (black), Pt(L_pc)₂Cl₂ (blue), Pd(L_pc)₂Cl₂ (green), Pt(L_pc)(L_ref)Cl₂ (red), and Pd(L_pc)(L_ref)Cl₂ (purple).

Table 1. Long-Wavelength Extinction Coefficients (ε) and Fluorescence Quantum Yields (FQYs), Obtained from Steady-State Absorption and Fluorescence Measurements.

Compound	Solvent	λmax/nm	ε/M–1 cm–1	FQY/%
Lpc	toluene	633	40,700	35.3
	THF	660	34,700	23.4
	BN	670	37,000	9.1
Pt(Lpc)(Lref)Cl2	toluene	676	31,800	9.3
	THF	672	31,900	3.0
	BN	685	28,800	0.7
Pd(Lpc)(Lref)Cl2	toluene	675	35,300	12.5
	THF	671	35,800	7.4
	BN	684	27,400	3.3
Pt(Lpc)2Cl2	toluene	678	68,500	6.9
	THF	674	71,500	2.6
	BN	687	51,400	0.6
Pd(Lpc)2Cl2	toluene	677	67,600	10.1
	THF	671	61,700	6.0
	BN	681	46,400	2.5

Steady-state emission spectra provide initial insight into the excited-state deactivation of the metal complexes (Figures S25 and S26, Table 1). The fluorescence quantum yields in toluene drop consistently in the series L_pc (35.3%), Pt(L_pc)(L_ref)Cl₂ (9.3%), and Pt(L_pc)₂Cl₂ (6.9%). The fluorescence quantum yields of Pd(L_pc)(L_ref)Cl₂ (12.5%) and Pd(L_pc)₂Cl₂ (10.1%) show a similar trend. In polar solvents tetrahydrofuran and benzonitrile, fluorescence quantum yields trend to even lower (Table 1), and, for example, the values of dimers Pt(L_pc)₂Cl₂ and Pd(L_pc)₂Cl₂ in benzonitrile drop to 0.6% and 2.5%, respectively.

Following the initial set of steady-state measurements, we turned to analysis by TAS (λ_ex = 610 nm; E = 400 nJ), using fs-TAS to derive information related to the formation of the correlated triplet pair ¹(T₁T₁) on the early time scales of SF, and ns-TAS to probe subsequent decoherence of ¹(T₁T₁) and formation of (T₁ + T₁). The behavior of ligand L_pc was established first (Figures S28–S33). fs- and ns-TAS raw data for L_pc were best fit by global analysis with a kinetic model based on three species. In particular, the first and second species both have singlet excited state character and feature excited state absorptions (ESA) at 450 and 1360 nm, next to ground state bleaching (GSB) at ca. 670 nm. Considering that the initially formed excited state is likely more polar than the ground state, the first and second species are singlet excited states (S₁) prior to and after ((S₁)_rel) vibrational and solvent relaxation. The corresponding lifetimes are 10.7 ps (S₁) and 4.7 ns ((S₁)_rel) in benzonitrile. The third species shows a fundamentally different ESA at around 470 and 510 nm. This species evolves as a product of a slow and spin-forbidden intersystem crossing that is assigned as the triplet excited state (T₁), with a lifetime of 20.2 μs.

Monomer Pt(L_pc)(L_ref)Cl₂ was then probed to establish the effect of metal coordination to L_pc (Figures S34–S39). The kinetic model established for L_pc fits the raw data of Pt(L_pc)(L_ref)Cl₂ well with the same three species, namely, (S₁), (S₁)_rel, and (T₁). In comparison to the L_pc, the lifetimes of (S₁)_rel and (T₁) are, however, influenced by the HAE and reduced by a factor of about three with values of 12.4 and 27.1 μs for L_pc versus 4.9 ns and 7.9 μs for Pt(L_pc)(L_ref)Cl₂, respectively, in toluene. Hand-in-hand with a shorter lifetime for (S₁)_rel is an increased triplet quantum yield for Pt(L_pc)(L_ref)Cl₂ (5.2%), which is about 3.5 times greater than that of L_pc (1.5%) in toluene. The behavior in polar solvents, such as THF and benzonitrile, was similar.⁵⁷

Finally, the photophysics of dimer Pt(L_pc)₂Cl₂ were explored. The development of a kinetic model based on four species was, however, needed to fit the fs- and ns-TAS raw data (Figure 3 and Figures S40–S45). The first and second species show ESA and GSB features consistent with singlet excited state. Conversely, the third and fourth species show ESA and GSB characteristics of triplet excited state. The first (6.3 ps) and second (0.5 ns) species of Pt(L_pc)₂Cl₂ in BN display similar deconvoluted spectra, which are analogous to that of (S₁) and (S₁)_rel of Pt(L_pc)(L_ref)Cl₂. Furthermore, the first species exhibits a short lifetime of 6.3 ps, consistent with the initial formation of (S₁S₀) that transforms into (S₁S₀)_rel. It is noted that the fourth species has a long-lived lifetime (7.0 μs) as well as triplet signatures that are characteristic of (T₁) in Pt(L_pc)(L_ref)Cl₂. Thus, there is no doubt that the fourth species is a free triplet state. This leaves the third species as the only unidentified state. Considering this species is rapidly formed, and combined with triplet signatures in the ESA and GSB, we postulate a fast and spin-allowed triplet formation via iSF that results in a correlated triplet pair ¹(T₁T₁). Subsequently, the final product of uncorrelated triplet excited state (T₁ + T₁), that is fourth species, is formed from decoherence of ¹(T₁T₁) and then deactivates by means of recovery of the ground state. Importantly, this spin-allowed iSF for Pt(L_pc)₂Cl₂ (2.8 ns in toluene) is twice as fast as spin-forbidden ISC (4.9 ns in toluene), which results in an overall ¹(T₁T₁) triplet quantum yield of around 140% in toluene and nearly 200% in benzonitrile.⁵⁸ In Pt(L_pc)₂Cl₂, the final step of iSF is the decoherence of ¹(T₁T₁), by which two separate triplet excited states, namely, (T₁ + T₁), are formed. The electronic interaction between the two triplets at short distances in singlet ¹(T₁T₁) is stronger than that for either the triplet ³(T₁T₁) or quintet ⁵(T₁T₁). Moreover, this additional stabilizing energy has been related to the energy required to dissociate the two triplets.⁵⁹¹(T₁T₁) can be stabilized by configurational interactions with other singlet configurations, while the ⁵(T₁T₁) preserves much larger pure diabatic character. The mixing of ⁵(T₁T₁) with other quintets is limited because these states are generally higher in energy.⁶⁰ Hence, the binding energy of the two triplets can be approximated as E_b = E(⁵(T₁T₁)) – E(¹(T₁T₁)).^59,61 The role of the ⁵(T₁T₁) as an intermediate state in singlet fission has been hypothesized by means of the spin Hamiltonian of interacting triplets^62,63 and has been established by time-resolved electron paramagnetic resonance (TR-EPR) in weakly coupled pentacene dimers.⁷ Along with the dissociation of ¹(T₁T₁), the intermediate state is the quintet form of (T₁T₁), which is denoted as ⁵(T₁T₁). Moreover, the heavy atom effect enhances the spin–orbit coupling leading to the triplet correlated form of (T₁T₁), namely, ³(T₁T₁).⁴⁰ In particular, the mixing of ³(T₁T₁) and ⁵(T₁T₁) determines the fate of SF. The involvement of ³(T₁T₁) strongly depends on the SF chromophore, the coupling between chromophores, as well as spin–orbit coupling.^40,45,64 On one hand, the mixing of ³(T₁T₁) and ⁵(T₁T₁) is not experimentally observed in the case of large J values like those observed in heavy-atom free pentacene dimers.⁴⁵ On the other hand, the small J value in weakly coupled terrylene diimides favors the formation of ³(T₁T₁) that triplet–triplet annihilates to produce a single rather than a pair of uncorrelated triplets.⁶⁴ Here, spin mixing allows for annihilation through the triplet channel to compete with free triplet formation. The formation of ⁵(T₁T₁) is, nevertheless, essential as it shuts down annihilation from ¹(T₁T₁) and ³(T₁T₁) channels.

Figure 3.

(a) fs-TAS (λ_ex = 610 nm, 400 nJ) of Pt(L_pc)₂Cl₂ in BN with time delays between 0–5500 ps, together with the respective (b) deconvoluted spectra representing the singlet (S₁S₀) (black), solvent-relaxed singlet (S₁S₀)_rel (red), and correlated triplet pair (T₁T₁) (blue) excited states as obtained by target analysis. (c) ns-TAS (λ_ex = 610 nm, 400 nJ) of Pt(L_pc)₂Cl₂ in BN with time delays between 0 ns–400 μs, together with the respective (d) deconvoluted spectra representing the solvent-relaxed singlet (S₁S₀)_rel (red), correlated triplet pair (T₁T₁) (blue), and uncorrelated triplet (T₁ + T₁) (green) excited states as obtained by target analysis.

Thus, we refer to the third species as the mixing of a singlet, triplet, and quintet (T₁T₁), namely ^1/3/5(T₁T₁). Subsequently, decoherence dominates the transition to a free triplet pair (T₁ + T₁). Although the decoherence of ^1/3/5(T₁T₁) is rather fast with 15.8, 14.2, and 12.9 ns in toluene, THF, and benzonitrile, respectively, the corresponding (T₁ + T₁) quantum yield is still reasonably high (52% in toluene, 49% in THF, and 46% in benzonitrile).

As a complement to the excited-state behavior of Pt(L_pc)(L_ref)Cl₂ and Pt(L_pc)₂Cl₂, the Pd(II) monomer Pd(L_pc)(L_ref)Cl₂ and the dimer Pd(L_pc)₂Cl₂ were examined. A kinetic model with three species, as used for Pt(L_pc)(L_ref)Cl₂, was also a good fit for the raw data of Pt(L_pc)(L_ref)Cl₂ (Table 2, Figures S46–S51). In particular, the first, second, and third species are assigned to (S₁), (S₁)_rel, and (T₁), respectively. In comparison to Pt(L_pc)(L_ref)Cl₂, the lifetimes of (S₁)_rel and (T₁) are slightly longer, especially in BN with 4.2 ns and 13.6 μs, respectively, as a result of the weaker HAE from Pd(II). The longer (S₁)_rel lifetimes result in triplet quantum yields that are intermediate between those for L_pc and Pt(L_pc)(L_ref)Cl₂, namely, around 4%.

Table 2. Species and Lifetimes of L_pc, Pt(L_pc)(L_ref)Cl₂, Pd(L_pc)(L_ref)Cl₂, Pt(L_pc)₂Cl₂, and Pd(L_pc)₂Cl₂ Were Obtained from fs-TAS and ns-TAS Measurements.

		Species
Compound	Solvent	τ(S1)/ps	τ(S1)rel/ns	τ(T1)/μs
Lpc	toluene	8.0	12.4	27.1
	THF	2.9	9.2	25.6
	BN	10.7	4.7	20.2
Pt(Lpc)(Lref)Cl2	toluene	6.7	4.9	7.9
	THF	3.3	2.2	7.2
	BN	11.3	0.7	7.4
Pd(Lpc)(Lref)Cl2	toluene	6.7	4.2	18.5
	THF	3.7	2.7	8.8
	BN	8.9	4.2	13.6
		τ(S1S0)/ps	τ(S1S0)rel/ns		τ(T1T1)/ns	τ(T1+T1)/μs
Pt(Lpc)2Cl2	toluene	5.3	2.8		15.8	8.0
	THF	2.6	1.3		14.2	8.0
	BN	6.3	0.5		12.9	7.0
		τ(S1S0)/ps	τ(S1S0)rel/ns	τ(M)/ns	τ(T1T1)/ns	τ(T1+T1)/μs
Pd(Lpc)2Cl2	toluene	4.0	3.1	–	138.1	25.7
	THF	2.7	2.0	7.3	169.6	26.4
	BN	10.0	0.9	4.6	122.3	23.8

For dimer Pd(L_pc)₂Cl₂ in toluene, the fs- and ns-TAS required the same kinetic model, which is based on four species as Pt(L_pc)₂Cl₂ to fit the raw data (Table 2, Figures 3 and 4, and Figures S52–S57). The first and second species feature ESA and GSB features of the singlet excited state, as established for Pt(L_pc)₂Cl₂. The lifetime (4.0 ps) of the first species is completely consistent with the unrelaxed state (S₁S₀), while that of the second species (3.1 ns) is assigned to the solvent-relaxed singlet (S₁S₀)_rel. It is noteworthy that the weaker HAE of Pd(II) slightly retards the decay of (S₁S₀)_rel in Pd(L_pc)₂Cl₂ (3.1 ns in toluene) relative to that of Pt(L_pc)₂Cl₂ (2.8 ns in toluene). ESA and GSB spectra confirm the triplet excited state character of the fourth and fifth species, which are thus assigned as ¹(T₁T₁) and (T₁ + T₁). The changes of lifetimes for ¹(T₁T₁) and (T₁ + T₁) are strongly metal dependent. For Pd(L_pc)₂Cl₂, the lifetime of ¹(T₁T₁) is roughly an order of magnitude longer than Pt(L_pc)₂Cl₂, and it leads to decoherence times in the range of hundreds of nanoseconds. Due to the weak HAE of Pd, as well as the weak interpentacene coupling, we posit that the long-lived triplet pair is the result of an unresolvable equilibrium among the approximately isoenergetic states ¹(T₁T₁), ³(T₁T₁), and ⁵(T₁T₁), namely, ¹(T₁T₁) ⇔ ³(T₁T₁) ⇔ ⁵(T₁T₁). Similar relationships have been established for the behavior of other weakly coupled systems.^7,65 The lifetime of (T₁ + T₁) for Pd(L_pc)₂Cl₂ (25.7 μs) is about three times that of Pt(L_pc)₂Cl₂ (8.0 μs) and, as a consequence, comparable to that of (T₁) (Table 2).

Figure 4.

(a) fs-TAS (λ_ex = 610 nm, 400 nJ) of Pd(L_pc)₂Cl₂ in BN with time delays between 0–5500 ps, together with the respective (b) deconvoluted spectra representing the singlet (S₁S₀) (black), solvent-relaxed excited state (S₁S₀)_rel (red) excited states, and mixed singlet/triplet/charge transfer intermediate state (M) (brown) as obtained by target analysis. (c) ns-TAS (λ_ex = 610 nm, 400 nJ) of Pd(L_pc)₂Cl₂ in BN with time delays between 0 ns–400 μs, together with the respective (d) deconvoluted spectra representing the solvent-relaxed singlet (S₁S₀)_Sol (red), mixed intermediate (M) (brown), correlated triplet pair (T₁T₁) (orange), and uncorrelated triplet (T₁ + T₁) (green) excited states as obtained by target analysis.

For Pd(L_pc)₂Cl₂ in the polar solvents THF and benzonitrile, a different picture is observed. The fs- and ns-TAS of Pd(L_pc)₂Cl₂ in polar solvents required a kinetic model that included five species to fit the raw data. It is noted that ¹(T₁T₁) is not formed directly from (S₁S₀)_rel in the case of Pd(L_pc)₂Cl₂ in polar solvents. Rather, it is formed via a third species, which is best described as a spectroscopic superimposition of singlet and triplet excited state features that mix with those of a charge transfer state (M).^17,66 Furthermore, (M) is observed only in the more polar solvents THF and benzonitrile, supporting its charge transfer character. Such an indirect mechanism, that is, the population of an SF intermediate, significantly impacts the quantum yield of ¹(T₁T₁) relative to a direct mechanism. Thus, the quantum yield for Pd(L_pc)₂Cl₂ is around 40%, relative to 140–200% found for Pt(L_pc)₂Cl₂. Turning to the quantum yields for decoherence, comparable values are derived in toluene (58%), THF (54%), and benzonitrile (55%). The kinetic models used for the pentacene dimers Pt(L_pc)₂Cl₂ and Pd(L_pc)₂Cl₂ are shown in Figure 5, which depicts the influence of different metal ions, namely, Pt(II) and Pd(II), on the formation of ¹(T₁T₁).

Figure 5.

Kinetic models used to fit the fs-TAS and ns-TAS data for (a) Pt(L_pc)₂Cl₂ via relaxation, iSF, and decoherence, and (b) Pd(L_pc)₂Cl₂ via relaxation, iSF, and decoherence in toluene, or via relaxation, an intermediate (M), iSF, and decoherence in THF and benzonitrile.

Conclusion

A pyridyl group has been appended to the pentacenyl chromophore, providing a ligand for the formation of complexes with platinum(II) and palladium(II). The third row metal transition metal, Pt(II), facilitates interpentacene electronic communication that leads to iSF in the dimer Pt(L_pc)₂Cl₂ via a direct mechanism. In the polar solvent benzonitrile, iSF is very efficient and gives quantum yields of the correlated triplet (T₁T₁) as high as 200%. For the second row transition metal Pd(II), electronic communication is also present but weaker. Thus, iSF is observed for Pd(L_pc)₂Cl₂, but through a mechanism that requires a mediating species in polar solvents. In view of the fact that the dimers are structurally identical, aside from the metal atom, mechanistic changes are attributed to differences in the size and polarizability of the metal species. Furthermore, it is noteworthy that iSF is operative in both Pt(L_pc)₂Cl₂ and Pd(L_pc)₂Cl₂, despite a rather large interpentacene separation in both dimers as well as enhanced spin–orbit coupling from the presence of heavy atoms.

Methods

Hazards

No unexpected or unusually high safety hazards were encountered during the course of this research.

Synthesis of Pt(L_pc)₂Cl₂

A mixture of L_pc (36 mg, 0.042 mmol) and Pt(PhCN)₂Cl₂ (10 mg, 0.021 mmol) in dry toluene (5 mL) was stirred at 80 °C for 16 h under an atmosphere of argon. The flask was wrapped in aluminum foil during the reaction to limit the light exposure. The reaction mixture was cooled to room temperature, and the solvent was then removed under reduced pressure. The residue was purified by column chromatography (alumina, CH₂Cl₂). The eluent volume was reduced in vacuo to about 1 mL followed by the addition of hexanes (15 mL). The resulting suspension was filtered, and the residue was washed with hexanes (3 × 2 mL), affording Pt(L_pc)₂Cl₂ as a dark blue solid (32 mg, 76%). Mp 322 °C (decomp). R_f = 0.50 (CH₂Cl₂/hexanes 1:2). UV–vis (CH₂Cl₂) λ_max (ε) 271 (128000), 306 (sh, 202000), 315 (544000), 347 (33200), 407 (36700), 441 (15600), 577 (sh, 13200), 626 (34700), 679 nm (63500). IR (CH₂Cl₂, cast) 3086 (w), 3050 (w), 2953 (s), 2927 (m), 2901 (m), 2867 (m), 2168 (m), 2122 (w), 1590 (m), 1461 (m) cm^–1. ¹H NMR (500 MHz, CDCl₃) δ 9.19 (s, 4H), 9.03 (s, 4H), 8.37 (s, 4H), 7.94–7.86 (m, 12H), 7.56–7.50 (m, 12H), 7.41–7.33 (m, 8H), 2.18 (nonet, J = 6.6 Hz, 6H), 1.20 (s, 36H), 1.19 (d, J = 6.6 Hz, 36H), 0.96 (d, J = 7.0 Hz, 12H). ¹³C NMR (125 MHz, CDCl₃) δ 152.5, 151.9, 140.2, 133.0, 132.4, 132.3, 130.5, 130.4, 129.5, 129.2, 128.4, 126.2, 126.1, 125.9, 125.7, 120.3, 116.4, 111.2, 104.6, 102.0, 100.4, 34.8, 31.2, 26.6, 25.5, 25.4 (two signals coincident or not observed). MALDI HRMS (DCTB) calcd for C₁₂₆H₁₃₄³⁵Cl₂N₂¹⁹⁵PtSi₂ (M⁺) 1995.9105, found 1995.9082. DSC: Decomposition, 347 °C (onset), and 353 °C (peak).

Synthesis of Pd(L_pc)₂Cl₂

A mixture of L_pc (45 mg, 0.052 mmol) and Pd(PhCN)₂Cl₂ (10 mg, 0.026 mmol) in dry CH₂Cl₂ (5 mL) was stirred at room temperature for 15 h under an atmosphere of argon. The flask was wrapped in aluminum foil during the reaction to limit light exposure. The reaction mixture was plugged through a pad of alumina with CH₂Cl₂. The eluent volume was reduced in vacuo to ca. 1 mL followed by the addition of hexanes (15 mL). The resulting suspension was filtered, and the residue was washed with hexanes (3 × 2 mL), affording Pd(L_pc)₂Cl₂ as a dark blue-green solid (34 mg, 69%). Mp 309 °C (decomp). R_f = 0.75 (CH₂Cl₂/hexanes 1:1). UV–vis (CH₂Cl₂) λ_max (ε) 271 (118000), 306 (sh, 183000), 315 (507000), 348 (28000), 400 (29600), 441 (12400), 573 (sh, 10700), 625 (31300), 677 nm (57100). IR (CH₂Cl₂, cast) 3080 (w), 3050 (w), 2953 (s), 2927 (m), 2902 (m), 2867 (m), 2166 (m), 2120 (w), 1588 (m), 1461 (m) cm^–1. ¹H NMR (500 MHz, CDCl₃) δ 9.18 (s, 4H), 8.95 (s, 4H), 8.36 (s, 4H), 7.92–7.87 (m, 12H), 7.56–7.50 (m, 12H), 7.41–7.33 (m, 8H), 2.18 (nonet, J = 6.7 Hz, 6H), 1.20 (s, 36H), 1.18 (d, J = 6.7 Hz, 36H), 0.96 (d, J = 7.0 Hz, 12H). ¹³C NMR (125 MHz, CDCl₃) δ 152.5, 151.6, 140.0, 133.1, 132.4, 132.3, 130.8, 130.5, 130.4, 129.5, 129.2, 128.4, 126.2, 126.1, 125.9, 125.7, 120.4, 116.3, 111.2, 104.5, 101.9, 100.3, 34.8, 31.2, 26.6, 25.5, 25.4 (one signal coincident or not observed). MALDI HRMS (DCTB) calcd for C₁₂₆H₁₃₄³⁵Cl₂N₂¹⁰⁶PdSi₂ (M⁺) 1906.8492, found 1906.8510. DSC: Decomposition, 321 °C (onset), and 330 °C (peak).

Spectroscopy

Steady-state absorption measurements were performed on a PerkinElmer Lambda2 UV/vis two-beam spectrophotometer with a slit width of 2 nm and a scan rate of 240 nm/min. Steady-state fluorescence measurements were performed on a Horiba Jobin Yvon FluoroMax-3 spectrometer with a slit width of 3 nm for excitation and emission and an integration time of 0.1 s within a wavelength range between 625–900 nm. Fluorescence quantum yields (FQYs) were determined from corrected emission spectra following the standard methods using zinc 2,9,16,23-tetra-tert-butyl-29H,31H-phthalocyanine (ZnPc – FQY = 0.3 in toluene)⁶⁷ as standard for all the pentacene monomer references and pentacene dimers. Femto- (fs-TAS) and nanosecond (ns-TAS) transient absorption spectroscopy measurements were performed on an amplified CPA-2110 titanium:sapphire laser for excitation (1 kHz; 150 fs pulse width; 400 nJ laser energy) from Clark-MXR, Inc. In the EOS SYSTEM from Ultrafast Systems, a 1 kHz pump laser was used at an excitation wavelength of 610 nm. Probing was executed on a 2 kHz continuous white light fiber laser. Data evaluation of the fs-TAS and ns-TAS raw data has been conducted by means of multiwavelength and global/target analysis using the GloTarAn⁶⁸ package. Target analysis was performed on the TAS raw data sets using the proposed kinetic models. The analytic solution to the coupled differential equations that describe the kinetic model is convoluted with a Gaussian instrument response function. After the least-squares fitting has converged, the raw data matrix is deconvoluted using the specific solution to the kinetic model and parameters from the fit to obtain the species-associated spectra and their populations as a function of time.

Data Availability Statement

Crystallographic data for the structures reported in this article have been deposited at the Cambridge Crystallographic Data Centre, under deposition number CCDC 2288684–2288686. Copies of the data can be obtained free of charge via https://www.ccdc.cam.ac.uk/structures/.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/prechem.3c00082.

• General experimental methods, descriptions of synthetic route, experimental data, and HRMS and ¹H/¹³C NMR spectra; figures and tables illustrating results from ground state and excited spectroscopy, time-resolved transient absorption spectroscopy, and kinetic models (PDF)

Author Present Address

^∥ Department of Applied Chemistry, Graduate School of Engineering, Center for Molecular Systems (CMS), Kyushu University, 744 Moto-oka, Nishi-ku, Fukuoka 819-0395, Japan

Author Contributions

^§ Y.H. and I.P. contributed equally. R.R.T. and D.M.G. designed and oversaw the project. R.R.T and Y.H. designed the molecules. Y.H. synthesized and characterized the molecules. I.P., Y.B., A.-S.W., and L.A.M carried out photophysical characterization. M.J.F. conducted X-ray crystallographic characterization, refinement, and analysis. Y.H, I.P., Y.B., D.M.G., and R.R.T. wrote the paper with contributions from all authors. All authors analyzed the results and commented on the manuscript.

The authors declare no competing financial interest.