Chameleonic Photo- and Mechanoluminescence in Pyrazolate-Bridged NHC Cyclometalated Platinum Complexes.

DFT investigations on the ground (GS) and the first triplet (T1) excited state potential energy surfaces (PES) were performed on a new series of platinum-butterfly complexes, [{Pt(C∧C*)(μ-Rpz)}2] (Rpz: pz, 1; 4-Mepz, 2; 3,5-dmpz, 3; 3,5-dppz, 4), containing a cyclometalated NHC in their wings. The geometries of two close-lying local minima corresponding to butterfly spread conformers, 1s-4s, and butterfly folded ones, 1f-4f, with long and short Pt-Pt separations, respectively, were optimized in the GS and T1 PES. A comparison of the GS and T1 energy profiles revealed that an opposite trend is obtained in the relative stability of folded and spread conformers, the latter being more stabilized in their GS. Small ΔG (s/f) along with small-energy barriers in the GS support the coexistence of both kinds of conformers, which influence the photo- and mechanoluminescence of these complexes. In 5 wt % doped PMMA films in the air, these complexes exhibit intense sky-blue emissions (PLQY: 72.0-85.9%) upon excitation at λ ≤ 380 nm arising from 3IL/MLCT excited states, corresponding to the predominant 1s-4s conformers. Upon excitation at longer wavelengths (up to 450 nm), the minor 1f-4f conformers afford a blue emission as well, with PLQY still significant (40%-60%). In the solid state, the as-prepared powder of 4 exhibits a greenish-blue emission with QY ∼ 29%, mainly due to 3IL/3MLCT excited states of butterfly spread molecules, 4s. Mechanical grinding resulted in an enhanced and yellowish-green emission (QY ∼ 51%) due to the 3MMLCT excited states of butterfly folded molecules, 4f, in such a way that the mechanoluminescence has been associated with an intramolecular structural change induced by mechanical grinding.

Introduction

Cyclometalated complexes of Pt(II) are characterized by outstanding photoluminescent properties,[1] arising from the radiative deactivation of triplet excited states, which are at the origin of many challenging applications such as optical sensors,[2] biological imaging,[3] or light emitting devices.[4] The planar geometry of the mononuclear complexes allows them to assemble through Pt···Pt and π–π interactions. Recent computational investigations demonstrate that mononuclear complexes often possess very different photophysical properties than their aggregates. Thus, while mononuclear complexes are likely to deactivate nonradiatively through triplet metal-centered (3MC) excited states,[5] the formation of the latter states is likely more hindered in condensed phases. Additionally, it was found that the formation of excimers in Pt(II) complexes is more favored than the formation of ground state aggregates.[6] As a result, the nature of the emissive triplet state changes from a monomer-based triplet emission to a triplet metal–metal-to-ligand charge transfer (3MMLCT) like emission in the molecular ensembles, leading to red-shifted emissions.[7,8] This kind of platinum compound very often suffers the so-called aggregation-cause quenching (ACQ)[9] effect, which limits their applications. However, many transition metal complexes exhibit aggregation-induced phosphorescent emission (AIPE),[10−14] and they can successfully be used to achieve white light[15,16] or NIR[17,18] organic-light-emitting diodes by adjusting the doping concentration. The emission color strongly depends on the extent of these intermolecular interactions, and in their turn, on environmental factors able to affect them, such as temperature variations,[8] mechanical force,[12,19] or volatile solvent molecules embedded into the lattice.[2,20] Thus, these compounds become thermo-, mechano-, and/or vapoluminescent complexes, enlarging the technological interest of these smart functional materials.

Compared to mononuclear complexes, binuclear luminescent complexes have been less explored.[21−28] In binuclear Pt(II) cyclometalated compounds, the metallophilic interactions and thereto their luminescent properties can partially be controlled by selecting the bridging ligands, which result in different degrees of rigidity and steric hindrance.[21,23,24] Among them, platinum complexes with bridging pyrazolates have been deeply studied by Castellano,[25] Thompson,[26,27] and Ma[28] and co-workers. It was found that the Pt–Pt distance and the extent of the metallophilic interactions can be tuned by the bulkiness of the pyrazolate unit (butterfly body), in such a way that when the bulkiness increases, the cycloplatinated units (butterfly wings) are pushed closer together. As a result, the emission color of this kind of complexes can be tuned from blue to green or red.[23] In addition, in solution they exhibit sometimes a photoinduced structural change (PSC) on the lowest triplet-state potential energy surface (PES), resulting in a dramatic change of the Pt–Pt bond distance and thereto on the emission.[28] This unique butterfly-like structure allows the contraction of the Pt–Pt distance with temperature, thus leading to solid-state thermochromism and thermoluminescence. This is the case of [{Pt(ppy)(μ-Ph2pz)}2],[29] which at a low temperature exhibits monomer-based 3LC/MLCT emission, and it changes to excimer-like 3MMLCT emission above 160 K.

The cyclometalating groups play also an important role in the stability and the control of the photophysical properties.[30] In this sense, the platinum-butterfly complexes reported by Strassner et al., i.e., [{Pt(C∧C*)(μ-Rpz)}2][31,32] are exemplary ones. They revealed that the cyclometalated N-heterocyclic carbenes (C∧C*), forming two strong metal–carbon bonds, are excellent wingsa for the synthesis of highly efficient blue and orange emitters. In this field, we reported compound [{Pt(C∧C*)(μ-pz)}2] (HC∧C* = 1-(4-(ethoxy-carbonyl)phenyl)-3-methyl-1H-imidazol-2-ylidene; pz: pyrazolate 1) which undergoes two-center, two-electron [2c, 2e] oxidation in the presence of haloforms (CHX3, X = Cl, Br, I).[33] Herein, we report three new complexes, [{Pt(C∧C*)(μ-Rpz)}2] (Rpz: 4-methylpyrazolate (4-Mepz), 2; 3,5-dimethylpyrazolate (3,5-dmpz), 3; and 3,5-diphenylpyrazolate (3,5-dppz), 4) bearing the same wings, C∧C*, but different bodies (Rpz). Besides the experimental synthesis and characterization of compounds 1–4, the intriguing luminescence and mechanoluminescence have been studied and deciphered with density functional theory (DFT) and time-dependent DFT (TD-DFT) investigations. For all of the complexes, two close-lying local minima corresponding to the folded (f) and spread (s) conformers were located on both the ground-state (GS) and the lowest adiabatic triplet excited state (T1) PES. A low energy barrier for the thermal interconversion between both structures in the GS seems to be at the core of the stimuli-responsive luminescence of complex 4 in the solid state.

Experimental Section

Compounds [{Pt(EtO2C–C∧C*)(μ-Cl)}2] (A),[34] [Pt(EtO2C–C∧C*)(4-MepzH)2]ClO4 (B2),[35] and [{Pt(EtO2C–C∧C*)(μ-pz)}2] (1)[36] were prepared as described elsewhere.

Synthesis of syn-/anti-[{Pt(EtO2C–C∧C*)(μ-4-Mepz)}2] (2)

NEt3 (0.5 mL, 3.62 mmol) was added to a solution of B2 (133.5 mg, 0.19 mmol) in acetone (30 mL) at room temperature. After 2 h of reaction, the solvent was removed in vacuo to 2 mL. The solution was treated with H2O (20 mL), filtered, and washed with H2O to give 2-anti (83%)/2-syn(17%) as a yellow solid. Yield: 73.7 mg, 75%. Anal. Calcd for C34H36N8O4Pt2: C, 40.40; H, 3.59; N, 11.08. Found: C, 40.00; H, 3.75; N, 11.06. 1H NMR data for 2-anti (500 MHz, acetone-d6): δ 8.02 (d, 4JH7,H9 = 1.8, 3JH7,Pt = 55.1, 2H, H7), 7.65 (d, 3JH2,H3 = 2.1, 2H, H2), 7.63 (dd, 3JH9,H10 = 8.1, 4JH9,H7 = 1.8, 2H, H9), 7.50 (s, 2H, H3′, 4-Mepz), 7.46 (s, 2H, H5′, 4-Mepz), 7.18 (d, 3JH10,H9 = 8.1, 2H, H10), 7.06 (d, 3JH3,H2 = 2.1, 2H, H3), 4.23 (m, CH2, CO2Et), 3.35 (s, 6H, H4), 2.12 (s, 6H, Me, 4-Mepz), 1.31 (t, 3JH,H = 7.1, 6H, CH3, CO2Et). 1H NMR data for 2-syn: δ 7.97 (d, 4JH7,H9 = 1.8, 2H, H7), 7.68 (d, 3JH2,H3 = 2.1, 2H, H2), 7.58 (dd, 3JH9,H10 = 8.1, 4JH9,H7 = 1.8, 2H, H9), 7.54 (s, 2H, H3′, 4-Mepz), 7.42 (s, 2H, H5′, 4-Mepz), 3.63 (s, 6H, H4), 1.35 (t, 3JH,H = 7.1, 6H, CH3, CO2Et). The rest of the signals appear overlapped with those of the 2-anti isomer. 13C{1H} NMR plus HSQC and HMBC data for 2-anti (125.75 MHz, acetone-d6): δ 161.2 (C1), 152.9 (C5), 139.6 and 138.3 (C3′ and C5′), 136.3 (C7), 133.8 and 126.8 (C6 and C8), 126.3 (C9), 123.3 (C3), 116.2 (C2), 116.1 (C4’), 110.9 (C10), 60.7 (CH2, CO2Et), 36.5 (C4), 14.6 (CH3, CO2Et), 9.5 (Me, 4-Mepz). 13C{1H} NMR plus HSQC and HMBC data for 2-syn (125.75 MHz, acetone-d6): δ = 160.3 (C1), 123.1 (C3), 36.7 (C4). 195Pt{1H} NMR (108 MHz, acetone-d6): δ −3775 (2-anti), −3785 (2-syn) ppm. (MS (MALDI+): m/z 1010.4 [{Pt(C∧C*)(μ-4-Mepz)}2].

Synthesis of syn-/anti-[{Pt(EtO2C–C∧C*)(μ-3,5-dmpz)}2] (3)

Compound A (106.3 mg, 0.12 mmol) was added to a solution containing NaOBu (22.2 mg, 0.23 mmol) and 3,5-dmpzH (22.5 mg, 0.23 mmol) in acetone/EtOH (10 mL/5 mL). After 3 h of reaction at −10 °C, the solvent was removed to 3 mL under reduced pressure, filtered, and washed with 2 × 5 mL of H2O to give 3-anti(80%)/3-syn(20%) as a yellow solid. Yield: 70 mg, 58%. Anal. Calcd for C36H40N8O4Pt2: C, 41.62; H, 3.88; N, 10.79. Found: C, 41.24; H, 4.02; N, 10.76. 1H NMR data for 3-anti (500 MHz, methylene chloride- d2): δ 7.83 (d, 4JH7,H9 = 1.7, 3JH7,Pt = 52.3, 2H, H7), 7.67 (d, 3JH9,H10 = 7.5, 2H, H9), 7.20 (s, br, 2H, H2), 6.94 (d, 3JH10,H9 = 7.5, 2H, H10), 6.67 (s, br, 2H, H3), 6.11 (s, 2H, H4′, dmpz), 4.24 (m, CH2, CO2Et), 3.33 (s, 6H, H4), 2.32 and 2.27 (s, 12H, Me, dmpz), 1.35 (t, 3JH,H = 7.1, 6H, CH3, CO2Et). 1H NMR data for 3-syn: δ 6.13 and 6.04 (s, 2H, H4′, dmpz), 3.58 (s, 6H, H4). The rest of the signals appear overlapped with those of the 3-anti.

13C{1H} NMR plus HSQC and HMBC data for 3-anti (125.75 MHz, methylene chloride-d2): δ 160.7 (C1), 152.2 (C5), 146.7 (C3′ and C5′). 136.9 (C7), 133.4 and 126.7 (C6 and C8), 125.9 (C9), 122.1 (C3), 115.4 (C2), 110.2 (C10), 104.6 (C4’), 60.8 (CH2, CO2Et), 35.7 (C4), 14.7 (CH3, CO2Et), 14.2 (Me, dmpz). 13C{1H} NMR plus HSQC and HMBC data for 3-syn (125.75 MHz, methylene chloride-d2): 160.2 (C1), 35.8 (C4). 195Pt{1H} NMR (108 MHz, methylene chloride- d2): δ = −3771 ppm (3-anti), −3799 (3-syn) ppm. MS (MALDI+): m/z 1038.2 [{Pt(C∧C*)(μ-dmpz)}2].

Synthesis of syn-/anti- [{Pt(EtO2C–C∧C*)(μ-3,5-dppz)}2] (4)

AgClO4 (52.7 mg, 0.25 mmol) was added to a stirred suspension of A (115.8 mg, 0.12 mmol) in acetone (30 mL) in the dark at room temperature. After 2 h of reaction, 3,5-dppzH (110.9 mg, 0.50 mmol) was added to the mixture and allowed to react overnight in the darkness. Then, the resulting suspension was filtered through Celite and concentrated to ca. 20 mL. NEt3 (0.5 mL, 3.62 mmol) was added to the reaction mixture and stirred for 2 h. Then, the solvent was removed in vacuo. The residue was treated with cold MeOH (5 mL) and filtered to give 4-anti (94%)/4-syn (6%) as a yellow solid. Yield: 90.0 mg, 74%. Anal. Calcd for C56H48N8O4Pt2: C, 52.25; H, 3.76; N, 8.71. Found: C, 52.64; H, 3.90; N, 8.82. 1H NMR data for 4-anti (500 MHz, DMSO-d6, 353 K): δ 8.59 (d, 3JH = 7.2, 4H, H), 8.25 (dd, 3JH = 7.2, 4JH = 1.8, 4H, H), 7.84 (s, br, 1H, HC), 7.72–7.53 (m, 4H, HC), 7.50–7.24 (m, 9H, Hpz and HC), 7.20–7.03 (m, 8H, Hpz and HC), 6.99 (s, br, 2H, Hpz), 4.18 (q, 3JH,H = 7.0, 4H, CH2, CO2Et), 3.18 (s, 6H, H4), 1.29 (t, 3JH,H = 7.0, 6H, CH3, CO2Et). 1H NMR data for 4-syn: δ 3.32 (s, 6H, H4). The rest of the signals appear overlapped with those of the 4-anti isomer. 13C{1H} NMR plus HSQC and HMBC data for 4-anti (125.75 MHz, DMSO-d2,353 K): δ 155.5 (C1), 135.1 (CC), 132.7 (Cpz), 132.6 (Cpz), 127.6 (Cpz), 127.0 (Cpz), 126.8 (CC), 126.7 (C), 125.4 (C), 124.7 (CC), 124.5 (CC), 122.1 (CC), 109.7 (CC), 103.2 (CC), 59.3 (CH2, CO2Et), 34.3 (C4), 13.6 (CH3, CO2Et). 195Pt{1H} NMR (108 MHz, DMSO-d6, 353 K): δ −3680 ppm (4-anti). (MS (MALDI+): m/z 1286.5 [{Pt(C∧C*)(μ-3,5-dppz)}2].

Results and Discussion

Compounds [{Pt(C∧C*)(μ-Rpz)}2] (HC∧C* = 1-(4-(ethoxycarbonyl) phenyl)-3-methyl-1H-imidazol-2-ylidene; Rpz: 4-methylpyrazolate (4-Mepz), 2; 3,5-dimethylpyrazolate (3,5-dmpz), 3; 3,5-diphenylpyrazolate (3,5-dppz), 4) were prepared following path a (for 3) or b (for 2 and 4) in Scheme ).

Scheme 1

Synthetic Routes Followed for Compounds 1–4

–2 KCl (NaCl)/–2 H2O (BuOH).

–2 NHEt3Rpz/–2 NHEt3ClO4.

Just the major isomer “anti” appears, represented for clarity along with its numerical scheme for NMR analysis. Compound 1 is included for overview.

The inability to get compound 3 through path b is in agreement with the greater basicity of 3,5-dmpzH[37] with respect to pzH, 4-MepzH, and 3,5-dppzH, which prevents it from being removed from the coordination sphere of the platinum center (experimental details for 2–4 in the SI). All the complexes were obtained as a mixture of syn/anti isomers with respect to the relative orientation of the cyclometalated C∧C* groups, with the anti-isomer being the predominant one, as can be seen in the 1H and the 195Pt{1H} NMR spectra of 2–4 (Figures S1–S4). The single-crystal X-ray diffraction study of 2 and 3 confirmed the expected spread butterfly-like structure (Figure ). Like compound 1,[36] complex 2 showed three different molecules in the asymmetric unit (A, B, C) with intermetallic distances of 3.355(4) Å (2A), 3.224(3) Å (2B), and 3.156(3) Å (2C). However, complex 3 exhibited only one dinuclear molecule with an intermetallic separation of 3.131(17) Å, in the low range of distances observed in other platinum-butterfly complexes with the same body (3,5-dmpz) but bearing different wings (3.128–3.203 Å).[26,38−40] Unfortunately, no good quality crystals were obtained for 4, but we could confirm the atom connectivity. Two different molecules with a Pt–Pt separation of 3.054 and 2.982 Å were found in the asymmetric unit. Therefore, once again it can be established that when the steric demand of the bridging pyrazolate increases the platinum centers are pushed closer together, like in other butterfly-like platinum complexes reported by Thompson et al.,[26] Umakoshi et al.,[40] and Strassner et al.[32] An extended description of these molecules has been included in the SI (see Table S2 and Figures S5–S7).

Figure 1

Molecular structures of 2A (left) and 3 (right). Ellipsoids are drawn at their 50% probability level; solvent molecules and hydrogen atoms have been omitted for clarity.

Theoretical Calculations

DFT calculations on the GS and the lowest adiabatic triplet excited state (T1) PESs for 1–4 were performed, and the geometries of relevant stationary points, such as, e.g., local minima and transition states (TS), were optimized (see SI for computational details) accounting for solvent effects in THF.

For all of the complexes, the geometries of two close-lying local minima were optimized in the GS PES (see Figure S8 in SI) which corresponded to the butterfly spread structures 1s–4s and the butterfly folded ones 1f–4f. Those corresponding to the butterfly spread conformers 1s–4s show long Pt–Pt distances (3.10 Å for 4s, 3.10 Å for 3s, <3.20 Å for 1s, <3.22 Å for 2s) and intramolecular C∧C* separations (>4.5 Å) following the same trend as the one observed in the experimental values. Also, they are characterized by a small Pt–Pt bond order (BO: 0.036 4s, <0.106 3s, <0.110 1s, <0.111 2s). On the other hand, the GS optimized geometries corresponding to the butterfly folded conformers 1f–4f are characterized by shorter Pt–Pt distances (2.96 Å for 4f, <2.97 Å for 3f, 2.97 Å for 1f, <2.98 Å for 2f) and intramolecular C∧C* contacts (<3.8 Å) along with larger Pt–Pt bond orders (BO: 0.170 4f, 0.174 3f, <0.228 1f, <0.233 2f) than those of 1s–4s. The computed energy profiles in the GS PES are shown in Figure . For all compounds, the conformers 1s–4s, featuring longer Pt–Pt distances, are more stable than the conformers 1f–4f (ΔG: 0.076 eV (1.76 kcal/mol) for 4, 0.129 eV (2.97 kcal/mol) for 1, 0.152 eV (3.50 kcal/mol) for 3, and 0.220 eV (5.08 kcal/mol) for 2). Especially, this is remarkable for 2s, which bears the longest Pt–Pt separation and the largest dihedral angle between the two platinum coordination planes. In addition, for complexes bearing bulkier Rpz units (3 and 4), their 3s and 4s minima are stabilized at shorter Pt–Pt distances than 1s and 2s.

Figure 2

Calculated relative energy profile (PCM-M06/6-31G(d) and MWB60(Pt)) in the GS for the interconversion between the 1f–4f and 1s–4s conformers. Values calculated in THF.

Furthermore, for complexes 1 and 2, we have successfully located the transition state (TS) associated with the interconversion between both conformers (see their optimized geometries in Figure S8). These TSs lie exemplarily 0.0077 eV (0.18 kcal/mol) above 1f and 0.037 eV (0.86 kcal/mol) above 2f (see Figure ). Their optimized geometries display Pt–Pt distances which lie in between those found for the butterfly folded and butterfly spread optimized minima. In the case of 1,[33] a small ΔG (1s/1f) value along with a small activation barrier supports, within the experimental error, a fast thermal equilibration in the ground state PES, thus resembling an intramolecular butterfly flapping-like motion.

These results are fully consistent with the presence of both conformers in solution, with the butterfly spread being the predominant one. Attempts to optimize the geometries of the TSs for the interconversion between conformers of complexes 3 and 4 were unsuccessful. In view of this piece of evidence, the flapping process likely occurs in a barrierless manner for the latter complexes.

Let us now discuss the results for the calculations on the lowest adiabatic triplet excited state (T1) PES. The geometries of two local minima, i.e., s/f, were optimized for all of the complexes (Figure and Figure S9 in the SI). The optimized geometries for the butterfly spread conformers, 1s–4s, show Pt–Pt distances (3.02 Å for 4s, <3.09 Å for 3s, <3.20 Å for 2s, <3.21 Å for 1s) and Pt–Pt bond orders (BO: 0.116 3s, 0.110 2s, 0.102 4s, 0.099 1s), similar to those observed for most of them in the GS.

Figure 3

Calculated relative energy profile (PCM-M06/6-31G(d) and MWB60(Pt)) in the lowest adiabatic triplet excited state (T1) for the interconversion between the 1f–4f and 1s–4s conformers. Values calculated in THF. Spin density distribution plots of 3f (left) and 3s (right).

However, the T1 optimized geometries for the butterfly folded conformers, 1f–4f, exhibit intermetallic separations (2.74 Å for 4f, <2.75 Å for 1f–3f), which are shortened by ca. 0.22 Å with respect to those in the GS, and Pt–Pt bond orders (BO: 0.586 4f, 0.591 3f, 0.626 2f, 0.621 1f), which are increased by 0.4 with respect to those in the GS. The calculated spin density distribution for 1s–4s indicates a mixed 3IL/3MLCT [π(C∧C*) → π*(C∧C*)]/[5d(Pt) → π*(C∧C*)] character for their T1 states (see Figures and S10) but a 3MMLCT [dσ*(Pt–Pt) → π*(C∧C*)] character for the T1 states of 1f–4f. Note that the changes in the Pt–Pt distances and the BO values from the GS to T1 states in the butterfly folded conformers 1f–4f almost agree with a one-electron excitation from the dσ*(Pt–Pt) orbital.

Like in the previously reported C,N-cycloplatinated butterfly-like complexes by Ma et al.,[28] as the steric bulk of the Rpz ligand increases, their spread-like minima (3IL/3MLCT) display shorter Pt–Pt bond distances (compare e.g., 1s and 2s vs 3s and 4s in Figure ). Importantly, comparing the 1s–4s and the 1f–4f optimized geometries in their T1 states, there is a considerable shortening of the Pt–Pt distances in the folded-like structures. The change of excited state character when going from the 1s–4s minima (3IL/3MLCT) to the 1f–4f ones (3MMLCT) leads to an extra stabilization of the latter conformers[41] by 0.085 eV (1.95 kcal/mol), 0.015 eV (0.36 kcal/mol), 0.084 eV (1.93 kcal/mol), and 0.166 eV (3.83 kcal/mol) for complexes 1–4, respectively. Note also that a certain amount of Pt–Pt bonding is only possible in the T1 state but not in the GS.

All in all, a comparison of the GS and T1 energy profiles reveals that an opposite trend is obtained in the relative stability of folded and spread conformers, the former being clearly more stabilized in their T1 states, regardless of the steric hindrance of the bridging Rpz, but specially for complex 4. In addition, we located the transition states (TSs) for the interconversion between conformers in the T1 state for 1–3, which are shown in Figure S9. These TSs all bear one imaginary frequency associated with the interconversion between both conformers. These TSs lie 0.115 eV (2.65 kcal/mol), 0.162 eV (3.73 kcal/mol), and 0.132 eV (3.04 kcal/mol) above the local minima 1s–3s, respectively. These energy barriers for PSC are larger than those for the flapping-like intramolecular motion in the GS

The absorption properties of 1–4 were also investigated with PCM-TD-DFT calculations in the presence of THF (see details in the SI). The results are collected in Tables S3 and S4 and Figure S11. The frontier molecular orbitals (Figure S11) for 1s–4s and 1f–4f along with the energies of their lowest singlet excited states were also calculated (see inset of Figure and Table S4 in SI). The lowest singlet excited states have predominant HOMO to LUMO character and can be described mainly as 1MLCT/1IL [5d(Pt) → π*(C∧C*)]/[π(C∧C*) → π*(C∧C*)] for 1s–4s, while some additional 1MMLCT [dσ*(Pt–Pt) →π*(C∧C*)] character is found for those of 1f–4f. The vertical ΔSCF-M06 emission energies from the T1 optimized geometries were calculated as well, rendering values of ca. 510 nm for 1s–4s and of ca. 570 nm for 1f–4f (see Table S4 in SI).

Figure 4

UV–visible spectra (path length: 1 mm) of 1–3 in 2-MeTHF 10–3 M and 4 in 2-MeTHF 10–5 M (path length: 1 cm). Inset: Expanded view of the UV–vis spectra along with the TD-M06/6-31G(d) and MWB60(Pt) S0 → S1 transitions of the butterfly spread (solid bars) and butterfly folded (dashed bars) conformers.

Photophysical Properties

The absorption and emission properties of 1–4 were investigated and explained on the basis of the DFT calculations. The UV–vis spectra of 1–4 (Figure and Table S26 in SI) do not show differences between diluted (10–5 M) and concentrated solutions (10–3 M). They show their lowest-energy absorption bands (ε ∼ 9 × 103 M–1 cm–1) in the range 325–390 nm. These absorptions bands match the S0 → S1 transitions calculated for the butterfly spread molecules 1s–4s (see Figure ), which are the predominant species according to the calculations.

However, in spite of the low contribution of the Rpz to the frontier molecular orbitals (FMOs), this absorption appears clearly red-shifted as the bulkiness of the R groups on the bridging pyrazolate increases. So, for species 3s and 4s, exhibiting shorter intermetallic distances and smaller interplanar angles in the GS, some 1MMLCT [dσ*(Pt–Pt) → π*(C∧C*)] character could be reasonably attributed (see Figures S8 and S11).

Diluted solutions (10–5 M) of 1–4 in 2-MeTHF were fast-cooled to 77 K. Upon excitation at λ ≤ 340 nm, each of their emission spectra were characterized by highly structured emission bands with λmax ∼ 450 nm and vibronic spacings [∼1450 cm–1], likely corresponding to the C=C/C=N bond stretching modes of the cyclometalated NHC ligands (Figure , left). The emission energies are not affected by the nature of the Rpz ligands, and they are very similar to those observed in the mononuclear compounds bearing the same “(C∧C*)Pt” fragment.[34,35,42,43]

Figure 5

Normalized excitation (dotted lines) and emission (solid lines) spectra at 77 K under an Ar atmosphere. Left: 1–4 in 2-MeTHF 10–5 M. Right: 1–3 in 2-MeTHF 10–3 M.

The computed emission energies for the butterfly spread conformers, i.e., 1s–4s, agree better with the experimental findings at 77 K than those calculated from 1f–4f. Thus, these results highlight that the barriers for interconversion between s/f conformers at the T1 state (see Figure ) are large enough to prevent their thermal equilibrium at 77 K.

Complexes 1 and 4 show additional excitation and emission bands at lower energies (λexc ∼ 450 nm, λem > 600 nm), attributable to the butterfly folded molecules (calculated S1 ∼ 426 nm and T1 = 572 nm for 1f; S1 ∼ 429 nm and T1 = 570 nm for 4f), although for 1 they are only perceptible in concentrated solutions (10–3 M; Figure , right). The coexistence of butterfly spread and butterfly folded molecules for 1 and 4 is in accordance with the small ΔG value computed between the two conformers, s/f in the ground state (ΔG: 0.076 eV (1.76 kcal/mol) 4s/4f, 0.129 eV (2.97 kcal/mol) 1s/1f) within the margin of error for the calculation of the energies of similar complexes with the M06 functional (MUE = 2.48 kcal/mol, see SI). For complexes 2 and 3 because of the greater ΔG between them (0.220 eV (5.08 kcal/mol) 2s/2f, 0.152 eV (3.50 kcal/mol) 3s/3f), it seems to be more unlikely and undetectable at 77 K.

In 5 wt % doped PMMA films in the air, excitation of complexes 1–4 at λ ≤ 380 nm affords intense sky-blue emissions with quantum yields of 72.0% 1, 83.4% 2, 79.0% 3, and 85.9% 4 (see Table ). These emissions match with those observed in 2-MeTHF (10–5 M) at 77 K. The slight blue shift in the emission spectra upon cooling is in accordance with an emissive state of 3IL/ 3MLCT character. The excitation of these films at longer wavelengths, up to 450 nm, render less intense but matched emission bands (see Figure ; Table S27 and Figure S12 in the SI). The short radiative decay of these emissions at room temperature should be noted, which are similar to those observed in analogous complexes [{Pt(C∧C*)(μ-Rpz)}2](HC∧C* = 3-dibenzofuran-4-yl-1-methyl-3H-imidazol-2-ylidene, imidazopyridine-2-ylidene; R = H, Me, tBu)[31,32] but clearly shorter than those measured for mononuclear compounds containing the same “(C∧C*)Pt” fragment.[42,43]

Table 1

Photophysical Data for 1–4 in PMMA Films and Solid State in the Air at 298 K

C	media	λexc (nm)	λem (nm)	CIE (x;y)	τ (μs)	φ(%)	krb	knrc
1	PMMAa	390	483max, 517sh, 567sh	0.18; 0.32	3.7	20	5.4 × 104	21.6 × 104
	PMMAa	350	483max,517sh, 567sh	0.18; 0.32		72
	solid	390	469, 527sh, 556max	0.41; 0.52	0.4 (20%)	3	2.5 × 104	79.2 × 104
					1.4 (80%)
2	PMMAa	390	469, 485max, 524sh	0.16; 0.29	3.5	54	15.4 × 104	13.1 × 104
	PMMAa	370	473, 492max, 536sh	0.16; 0.27		83
	solid	390	472, 527sh, 559max	0.41; 0.53	0.3 (22%)	3	3.2 × 104	103.2 × 104
					1.1 (78%)
3	PMMAa	390	464, 484max, 523sh	0.15; 0.25	3.4	53	15.7 × 104	13.8 × 104
	PMMAa	380	464, 484max, 523sh	0.15; 0.25		79
	solid	390	468, 487max	0.19; 0.35	0.3 (32%)	16	30.1 × 104	158.1 × 104
					0.6 (68%)
	ground	390	468, 492max, 519	0.29; 0.47	0.2 (20%)	6	11.2 × 104	175.4 × 104
	solid				0.6 (80%)
4	PMMAa	390	480max	0.14; 0.26	2.2	69	31.7 × 104	14.1 × 104
	PMMAa	380	480max	0.14; 0.26		86
	solid	390	469,482max, 553sh	0.24; 0.37	0.5 (30%)	29	32.9 × 104	80.7 × 104
					1.1 (70%)
	ground	390	553max	0.39; 0.55	1.1 (33%)	51	28.3 × 104	27.2 × 104
	solid				2.2 (67%)

5 wt %.

Radiative decay rate constant given as kr = φ/τexp.

knr = (1 – φ)/τexp.

Figure 6

Normalized emission and excitation spectra of complex 4 in 5 wt % PMMA film in the air, Picture was taken under 365 nm UV light.

All of these pieces of evidence highlight a greater metallic contribution to the excited state and, then, a greater 3MLCT character of the blue emissions of complexes 1–4 in PMMA films compared to those of the mononuclear complexes.[42,43]

Note that the large radiative rate constant values (kr, Table ) support this. Also, compared to 2–4 (kr > 1.0 × 105 s–1), a smaller radiative rate constant was obtained for 1 (kr = 1.0 × 104 s–1), suggesting that the spin–orbit coupling (SOC) efficiency was lower because of a larger energy separation between the manifold of triplet and singlet states.[44] Exemplarily, from the absorption and the PMMA emission data, the energy differences between S1 and T1 (ΔES–T = 0.93 eV 1, 0.81 eV 2; 0.75 eV 3, 0.70 eV 4) were found to follow the same trend observed for the kr.

In solution at room temperature, compounds 1–4 are scarcely luminescent even in an argon atmosphere, a usual feature for blue-emitting Pt(II) complexes since the population of dd* states and formation of exciplexes are very common thermal quenching processes.[43] However, in a fluid solution of 2-MeTHF (10–5 M) at room temperature under an argon atmosphere (Figures S13 in SI), excitation in the low-energy absorption range (λ ≤ 380 nm) renders for complexes 1 and 2 a weak emission from 1s and 2s, for complex 3 a dual emission with maxima at 456 and 552 nm, likely corresponding to 3s and 3f, and for complex 4 an emission with a maximum at 559 nm arising from 4f, according to theoretical calculations.

In summary, photoexcitation of complexes 1–4 at λ < 380 nm allows the major 1s–4s conformers to reach the high-energy 1IL/MLCT excited state; then, by a rapid intersystem crossing (ISC), the 3IL/MLCT (Ts) state will be populated, which is calculated to have a similar Pt–Pt separation than that of its corresponding ground state geometry (see Scheme ). In fluid solution, where the geometries of neither ground states nor those of the excited states are constrained, a photoinduced structural change (PSC) process between Ts and Tf conformers could happen depending on both, ΔG (Tf–Ts) and the PSC energy barrier.[28] In the case of 4, the computed barrierless PSC process along with the large ΔG values (T4f–T4s = −0.166 eV, −3.83 kcal) leads to T4f almost in an exclusive manner. This piece of evidence explains why the emission from T4f is the only one observed experimentally. In the case of complex 3, characterized by a smaller ΔG (T3f–T3s = −0.84 eV, −1.93 kcal) and a non-negligible PSC barrier (0.132 eV, 3.04 kcal), a thermal equilibrium between T3s and T3f is likely at room temperature, thus explaining its dual emission.

Scheme 2

Schematic Diagrams of Photophysical Processes Based on the Steady-State Excitation and Emission Spectra along with the Results of the Theoretical Investigations

In the case of complex 2 with a lower ΔG (T2f–T2s = −0.015 eV, −0.36 kcal) but larger energy barrier (0.162 eV, 3.73 kcal/mol), the PSC seems not to take place and the emission arises only from T2s. For complex 1, with ΔG (T1f/T1s= 0.085 eV, 1.95 kcal) and the PSC energy barrier (0.115 eV, 2.65 kcal) on the same order of magnitude as those calculated for complex 3, the PSC was expected to occur, but, the emission arises only from T1s.

In this case, we recently reported that internal conversion (IC) from 1IL/MLCT to 1MMLCT competes with ISC to 3IL/MLCT.[33] Therefore, a faster quenching of the 1/3MMLCT states of complex 1, as compared to that occurring in complexes 3 and 4, enabled by the lack of steric hindrance of the reactive positions in complex 1, could account for the absence of this low-energy emission.

In rigid media (2-MeTHF 10–5 M at 77 K or PMMA films), photoexcitation of complexes 1–4 at λ < 380 nm leads in an analogous manner to the emission from the higher-lying triplet state 3IL/MLCT, despite the greater stability of the 3MMLCT state. This indicates that the energy barriers to connect the Ts/Tf wells are large enough to prevent the PSC in 2-MeTHF at 77 K, and it is also in agreement with PMMA being a rigid glass at r.t. (Tg = 378 K).[29] Notably, the frozen glass environment leads to a deceleration of the nonradiative pathways, thus leading to large PLQY values in PMMA films (see Table ).

On the other hand, irradiation at λ > 400 nm will populate the low-energy states of the minor 1f–4f conformers, 1IL/MLCT with some 1MMLCT character (see right panel in Scheme ). A fast ISC to the close-lying triplet state 3IL/MLCT[45] would lead to the high-energy emission, which is the only one observed in 5 wt % PMMA films of 1–4. The low PLQYs when compared with those observed by irradiation at λ < 380 nm (see Table S27 and Figure S12 in SI) are in agreement with the low ratio of butterfly folded molecules in the samples but still being significant. Therefore, for complexes 2–4, the existence of close-lying 1IL/MLCT/MMLCT–3IL/MLCT states makes it possible to get intense blue emissions (PLQY: 40%–60%) from doped films by irradiation with wavelengths in the visible region.

Mechanoluminescence in the Solid State

The as-prepared powders of 1 and 2 are scarcely emissive. They exhibit a weak (PLQY < 5%), broad, structureless emission centered at λ ∼ 555 nm (Table and Figure S14) with a shoulder at λ ∼ 450 nm. The shape and energy of the main band could match with the one arising from the low-lying triplet state for the butterfly folded molecules that is the 3MMLCT state. Yet, the role of excimeric 3π–π* states can not be fully disregarded, given the extended and numerous intermolecular π–π interactions observed in the single-crystal X-ray structures of 1 and 2 (see Figure S6) and the lower quantum yield of compounds 1 and 2 in solid state compared to those in 5 wt % PMMA films. Keeping in mind that other mononuclear Pt(II) compounds containing cyclometalated NHCs reported previously[42,46,47] exhibit a similar behavior mainly as a consequence of π–π intermolecular interactions, it seems likely that the emission of 1 and 2 arises from excimeric 3π–π* states with intermolecular π–π interactions affording efficient nonemissive deactivation channels[12] through an aggregation-caused quenching (ACQ) effect.[9,48] Neither the excitation nor the emission spectra of 1 and 2 exhibit changes after grinding the solids with a mortar and pestle (see Figure S15 for 2 as an example). However, complexes 3 and 4 exhibit mechanoluminescence in the solid state. After grinding, the pale-yellow solids do not visually change their colors, but their photoluminescence changes from blue to yellowish-green. Before being ground, a powdered sample of 3 exhibits a sky-blue emission, similar but weaker than that exhibited in PMMA film (5 wt %), which we attribute to 3IL/MLCT. After grinding, the emission becomes green (Figure S16) due to the presence of an intense lower-energy band with λ ∼ 540 nm that could be assigned to the 3MMLCT state of molecules with the butterfly folded configuration in accordance with the theoretical calculations. However, in view of the intermolecular π–π interactions observed in the single-crystal X-ray structure of 3 (see Figure S6) and the decreased PLQY upon grinding, the participation of excimeric 3π–π* states to the low-energy band cannot be ruled out.[9,48]

In case of compound 4, photoexcitation of as-prepared powder leads to a greenish-blue emission with λmax at 480 nm and an incipient shoulder at 553 nm that can be assigned to 3IL/MLCT and 3MMLCT emissions, respectively, in accordance with the theoretical calculations and with the 3MMLCT emission observed for [{Pt(C∧C*)(μ-NPh-CH-NPh)}2] exhibiting a short Pt···Pt distance, of about 2.8 Å.[24] Mechanical grinding resulted in a suppression of the 3IL/MLCT band along with an increase of the 3MMLCT one and, unlike complex 3, an enhancement of the PLQY (see Figure ). As a result, the photoluminescence of powdery samples of 4 is intensified and changed from greenish-blue to yellowish-green upon grinding.

Figure 7

Normalized emission and excitation spectra of complex 4 in the solid state in the air at room temperature. Pictures were taken under 365 nm UV light.

From the excitation and emission spectra, it seems that the as-prepared powder of 4 shows phosphorescence from the two kinds of conformers, 4s and 4f, that from 4s being predominant. This is in agreement with the butterfly spread conformer, 4s being the major one in the GS and the fact that no PSC can take place in rigid media. Mechanical grinding seems to induce changes in the GS that somehow shorten the Pt–Pt distances and enforces the intramolecular Pt–Pt interactions, in such a way that in the ground solid, 4-g, the phosphorescence arises mainly from 4f.

Structural changes involving the intramolecular Pt–Pt separation in the GS as the origin of mechanoluminesce seems plausible on the bases of experimental and theoretical data, and once other causes were dismissed, like desolvation, since there is no solvent embedded in the solid (see CHN elemental analysis and NMR experiments) or intermolecular interactions, we proved that the luminescence spectrum of 4 in 40 wt % doped PMMA films match that at 5 wt % (see Figure S17), and its PLQY drops to 40%.

The ground solid, 4-g, undergoes the reverse change partially upon cooling to 77 K, as deduced from the emission and excitation spectra collected at room temperature and 77 K (Figure S18).

Structural changes involving the intramolecular Pt–Pt separation in the GS were reported for the thermochromic platinum-butterfly compound [{Pt(ppy)(μ-Ph2pz)}2],[29] but those induced by mechanical grinding have never been reported. In this case, like in complex 4, elongation of the Pt–Pt distance occurs when the temperature drops. Also, the transformations resulted to be reversible by the addition of THF, toluene, or diethyl ether to the ground samples of 3 and 4 that led to the recovery of the blue emission (see Figure and S19), and thus presumably restoring the previous structure arrangement.

Figure 8

Photographic images of mechanical grinding samples of 4 in response to solvent treatment taken under 365 nm UV light

Therefore, it could be argued that the bulkiness of the μ-pyrazolates has a strong impact not only on the luminescence, but also on the mechanoluminiscence of these platinum butterflies in the solid state. As the bulkiness increases, the intermolecular π–π interactions become more hindered, affording less efficient nonemissive deactivation channels and consequently a more efficient emission. In addition, as the steric demand of the μ-pyrazolates increases, the Pt–Pt interaction, enhanced by mechanical stimulation, causes a bathochromic shift of the emission (2750 cm–14) along with a remarkable increment of its PLQY.[49−54]

Conclusions

Photoluminescence and DFT calculations on a new series of platinum butterflies, [{Pt(C∧C*)(μ-Rpz)}2] (Rpz: pz 1, 4-Mepz 2, 3,5-dmpz 3, 3,5-dppz 4) containing a cyclometalated NHC in their wings, prove the presence of two conformers in the ground state at room temperature, the butterfly spread, 1s–4s and the butterfly folded, 1f–4f ones, which are characterized by long and short Pt–Pt separations, respectively. DFT calculations revealed that the former are the more stable in the GS but, in most of them, low ΔG (s/f) and low energy barriers in solution of THF support a fast thermal equilibration in the ground state PES, thus resembling an intramolecular butterfly flapping-like motion. By contrast, the butterfly folded, 1f–4f, conformers are the more stable in the T1PES. In 5 wt % doped PMMA films in the air, these complexes show intense sky-blue emissions (PLQY: 72.0–85.9%) upon excitation at λ ≤ 380 nm mainly arising from an 3IL/MLCT excited state, the first triplet state of the major butterfly spread conformer 1s–4s, in accord with no PSC occurring in a rigid matrix. The existence of close-lying 1IL/MLCT/MMLCT-3IL/MLCT states for the 1f–4f species enables the obtaining of intense blue emissions (PLQY: 40%–60%) under excitation with wavelengths in the visible region, up to 450 nm.

In addition, it could be argued that the bulkiness of the μ-pyrazolates has a strong impact on the luminescence and mechanoluminescence of these platinum butterflies in a solid state. In complexes 3 and 4, the intermolecular π–π interactions become more hindered than in complexes 1 and 2, affording more efficient emissions. In addition, in the 3,5-dppz derivative, 4, mechanical grinding causes a bathochromic shift of the emission from greenish-blue to yellowish-green along with a remarkable increment of its PLQY. This mechanoluminescence mechanism has been associated with an intramolecular structural change in the GS that somehow shortens the Pt–Pt distances and enhances the Pt–Pt interactions in such a way that the thermal butterfly flapping can be induced by mechanical grinding.