Ru(II) Oligothienyl Complexes with Fluorinated Ligands: Photophysical, Electrochemical, and Photobiological Properties

Abstract

A series of Ru(II) complexes incorporating two 4,4′-bis(trifluoromethyl)-2,2′-bipyridine (4,4′-btfmb) coligands and thienyl-appended imidazo[4,5-f][1,10]phenanthroline (IP-nT) ligands was characterized and assessed for phototherapy effects toward cancer cells. The [Ru(4,4′-btfmb)₂(IP-nT)]²⁺ scaffold has greater overall redox activity compared to Ru(II) polypyridyl complexes such as [Ru(bpy)₃]²⁺. Ru-1T–Ru-4T have additional oxidations due to the nT group and additional reductions due to the 4,4′-btfmb ligands. Ru-2T–Ru-4T also exhibit nT-based reductions. Ru-4T exhibits two oxidations and eight reductions within the potential window of −3 to +1.5 V. The lowest-lying triplets (T₁) for Ru-0T–2T are metal-to-ligand charge transfer (³MLCT) excited states with lifetimes around 1 μs, whereas T₁ for Ru-3T–4T is longer-lived (~20–24 μs) and of significant intraligand charge-transfer (³ILCT) character. Phototoxicity toward melanoma cells (SK-MEL-28) increases with n, with Ru-4T having a visible EC₅₀ value as low as 9 nM and PI as large as 12,000. Ru-3T and Ru-4T retain some of this activity in hypoxia, where Ru-4T has a visible EC₅₀ as low as 35 nM and PI as high as 2,900. Activity over six biological replicates is consistent and within an order of magnitude. These results demonstrate the importance of lowest-lying ³ILCT states for phototoxicity and maintaining activity in hypoxia.

Graphical Abstract

A series of Ru(II) polypyridyl complexes incorporating two 4,4′-bis(trifluoromethyl)-2,2′-bipyridine (4,4′-btfmb) coligands and a variable thienyl-appended imidazo[4,5-f][1,10]phenanthroline (IP-nT) ligand was characterized and assessed for phototherapy effects toward cancer cells. Ru-4T has a visible EC₅₀ value as low as 9 nM and PI as large as 12,000.

1. INTRODUCTION

Cancer continues to be a leading cause of mortality worldwide.¹ Despite substantial progress in available medical treatments,^2–6 there remains a critical need for innovative therapeutic strategies and adjuvants to complement traditional methods like surgery, radiotherapy, and chemotherapy. In this regard, photodynamic therapy (PDT) is a possible avenue for targeted cancer therapy. It employs a nontoxic photosensitizer (PS), light, and molecular oxygen to produce cytotoxic reactive oxygen species (ROS), importantly singlet oxygen (¹O₂), that destroy tumors and tumor vasculature and can induce antitumor immunity.^7–10 Phototoxic effects are restricted to oxygenated areas where light is delivered and PS has accumulated. The localized action of PDT results in fewer side effects and improved quality of life for patients.^11,12 PDT can be enhanced by tumor-selective PSs and precise delivery of light. In addition, the light protocol can be optimized by altering the drug-to-light interval (DLI) and tuning the wavelength, fluence, and irradiance.

There is also a drive to develop light-responsive compounds that generate ROS at low oxygen tension or function via alternate mechanisms that complement the ¹O₂ pathway. This capacity would allow effective treatment of hypoxic tumors and the use of high light irradiance. Metal complexes, particularly Ru(II) polypyridyl systems, are of significant interest in this regard. The strategic combination of ligands and metals allows access to a range of excited-state configurations with unique photophysical characteristics. Approaches have included the photorelease of bulky ligands to expose phototoxic metals and/or ligands,^13–21 photocaging of chemotherapeutics and enzyme inhibitors,^18,22–43 photoredox reactions,^44,45 and enhancing ROS production even under low oxygen conditions.^17,20,46

Our research group has a strong focus on metal complexes as PSs for their diverse mechanisms of action. Their modular design and relatively straightforward synthesis allow for rapid alteration of their physicochemical, photophysical, and biological characteristics and aligns with our tumor-specific approach to PS development. We advocate that no singular ideal PS exists; rather, PS design should be tailored to the intended application. Our Ru(II) polypyridyl complex, TLD1433, is a prime example, currently in Phase II clinical trials for treating non-muscle invasive bladder cancer (NMIBC) with PDT (Chart 1a).^11,47,48 It exhibits significant phototoxicity toward cancer cells while maintaining low dark toxicity. Clinically, it is activated using green light to prevent harm to underlying muscle tissue.

Chart 1.

(a) Structure of Racemic TLD1433 and (b) Structures of Racemic Complexes in This study.

To expand our understanding of oligothiophene-containing coordination complexes, we are examining variations in their central metals, coordinating ligands, thienyl groups, counterions, and coordination geometries.^{11,14,17,20,46,49–51} We aim to establish structure-activity relationships (SARs) for medicinally active complexes that reconcile their physicochemical, photophysical, electrochemical, and biological properties with the ideal phototherapy profile for a given clinical setting. In this study we introduce a series of Ru(II) PSs (Ru-nT), each featuring two 4,4′-bis(trifluoromethyl)-2,2′-bipyridine (4,4′-btfmb) coligands and a thienyl-containing imidazo[4,5-f][1,10]phenanthroline (IP-nT) ligand with n=1–4 thiophene rings. The members are compared to the parent [Ru(4,4′-btfmb)₂(LL)]²⁺ complexes with LL=IP-0T, 1,10-phenanthroline (phen), or 4,4′-btfmb (Chart 1b). The syntheses and structural characterization of complexes Ru-3T and Ru-4T were previously published.^52,53 In this study the photocytotoxic effects of [Ru(4,4′-btfmb)₂(IP-nT)]²⁺ (n=1–4) and the parent complexes without nT groups are examined on melanoma cells under various light conditions and oxygen levels. We also report on their lipophilicities, steady-state absorption and emission characteristics, excited-state properties, electronic configurations, and electrochemical profiles.

2. MATERIALS AND METHODS

The complexes in this series were characterized by ¹H NMR, high-performance liquid chromatography (HPLC), and ESI⁺ mass spectrometry (MS). They were evaluated for lipophilicity, ground and excited-state characteristics using absorption and emission spectroscopy, electrochemistry, and (photo)cytotoxicity. Further details regarding procedures and characterization data are available in the Supplementary Information (SI).

2.1. Instrumentation

A CEM Discover microwave reactor was used to perform microwave reactions. Flash column chromatography was carried out on the Teledyne ISCO EZ Prep UV model of CombiFlash^® EZ Prep using SILICYCLE SiliaSep 25 g prepacked silica cartridges. Size-exclusion chromatography was performed using a gravity column packed with Sephadex^® LH-20. The NMR spectra were collected on JEOL 500 MHz spectrometers (University of North Carolina at Greensboro, University of Texas at Arlington) operating at 500 MHz for ¹H experiments, and on an Agilent 700 MHz Magnet spectrometer (The Joint School of Nanoscience and Nanoengineering at Greensboro) operating at 700 MHz for ¹H experiments. The chemical shifts are reported in parts per million (ppm) and were referenced to the residual solvent peaks. High-resolution ESI⁺ mass spectra were obtained using a Thermo Fisher Scientific LTQ Orbitrap XL instrument (Triad Mass Spectrometry Laboratory at University of North Carolina at Greensboro) and Shimadzu IT-TOF instrument (Shimadzu Center for Advanced Analytical Chemistry at University of Texas at Arlington). HPLC analyses were carried out on an Agilent/Hewlett Packard 1100 series instrument in 100 μM solutions in methanol using a Hypersil GOLD C18 reversed-phase column with an A→B gradient (98% → 5% A; A=0.1% formic acid in H₂O, B=0.1% formic acid in MeCN). Reported retention times are accurate to within ±0.1 min.

2.2. Synthesis and Characterization

We previously published the synthetic methods and structural characterization of Ru-3T and Ru-4T,^52,53 and Ru(4,4′-btfmb)₃]²⁺ has been reported by others.^54–57 To the best of our knowledge, all other complexes presented in this study have not been reported. All complexes were prepared as racemic mixtures. Solvents and reagents were purchased from commercial sources and used without further purification. Water used for all biological experiments was deionized to a resistivity ≥ 18.2 MΩ using either a Barnstead or Milli-Q^® filtration system. Methanol was purchased from Fisher Scientific (ACS grade for synthesis, HPLC grade for LC eluent, Optima grade for HPLC and MS sample preparation). Deuterated solvents for NMR were purchased from Cambridge Isotope Laboratories. Ruthenium(III) trichloride trihydrate was purchased from Ark Pharm and Acros Organics. Ru(4,4′-btfmb)₂Cl₂•2H₂O⁵⁸ and IP-based ligands⁵⁹ were prepared according to adapted literature procedures. The synthesis of IP-based ligands follows that described below for IP-4T. [2,2′:5′,2″:5″,2‴-Quaterthiophene]-5-carbaldehyde (4T-CHO) was prepared as previously described.^60,61 The final products are synthetically characterized in Figures S1–S21 via ¹H NMR, ¹H–¹H COSY NMR, HPLC, and ESI⁺–MS. The Cl⁻ salts of final complex products were obtained via anion metathesis on HCl-treated Amberlite IRA-410 resin with methanol as eluent and isolated in vacuo. Final complexes are a mixture of Δ/Λ isomers.

rac-[Ru(4,4′-btfmb)₃](Cl)₂ Ru(Cl)₃·xH₂O (58 mg, 0.2 mmol) and 4,4′-btfmb (175 mg, 0.6 mmol) was added to a microwave vessel containing argon-purged ethylene glycol (3 mL), and then the mixture was subject to microwave irradiation at 180 °C for 45 min with stirring. The resulting dark red solution was then transferred to a separatory funnel with deionized water (25 mL) and CH₂Cl₂ (25 mL). After gentle agitation, the CH₂Cl₂ was drained and the remaining aqueous layer was washed with CH₂Cl₂ (25 mL) until the CH₂Cl₂ layer was colorless. Then, CH₂Cl₂ (25 mL) and saturated aqueous KPF₆ (5 mL) was added, and the mixture was shaken gently. The CH₂Cl₂ layer was drained and the product was further extracted from the aqueous layer with CH₂Cl₂ (25 mL) until the aqueous layer was colorless. The CH₂Cl₂ extracts were then combined and concentrated under reduced pressure. The crude product was then purified using silica gel flash column chromatography with a gradient of MeCN to 10% water in MeCN, followed by 7.5% water in MeCN with 0.5% KNO₃. The product-containing fractions were then combined and concentrated under vacuum, then transferred to a separatory funnel with CH₂Cl₂ (25 mL), deionized water (25 mL), and saturated aqueous KPF₆ (1 mL). The resulting mixture was gently agitated and the CH₂Cl₂ layer was drained. Additional CH₂Cl₂ (25 mL) was used to extract the remaining product until the aqueous layer was colorless. The CH₂Cl₂ layers were then combined and dried under vacuum. This was then converted to the corresponding Cl⁻ salt in quantitative yield using Amberlite IRA-410 with MeOH as the eluent, then purifying further using Sephadex LH-20 with MeOH as the eluent, affording a dark red solid (50 mg, 20%). ¹H NMR (400 MHz, MeOD-d₃, ppm): δ 9.41 (d, J=1.9 Hz, 6H), 8.15 (d, J=5.9 Hz, 6H), 7.86 (dd, J=6.0, 1.8 Hz, 6H). HRMS (ESI+) m/z for [M-2Cl⁻]²⁺ calcd: 489.0169; found: 488.9529. HPLC retention time: 22.41 min (99.5% purity by peak area).

rac-[Ru(4,4′-btfmb)₂(phen)](Cl)₂ (Ru-phen) Ru(4,4′-btfmb)₂Cl₂·2H₂O (91 mg, 0.12 mmol) and phen (22 mg, 0.1 mmol) were added to a microwave vessel containing argon-purged ethylene glycol (4 mL) and subjected to microwave irradiation at 180 °C for 15 min. The resulting dark red mixture was then isolated and purified in the same manner as [Ru(4,4′-btfmb)₃](Cl)₂, yielding the desired product as a dark red solid (43 mg, 45%). ¹H NMR (400 MHz, MeOD-d₃, ppm): δ 9.38 (d, J=14.1 Hz, 4H), 8.81 (d, J=8.0 Hz, 2H), 8.36 (s, 2H), 8.28 (d, J=5.9 Hz, 2H), 8.23 (d, J=5.3 Hz, 2H), 7.94–7.84 (m, 6H), 7.67 (dd, J=6.0, 1.8 Hz, 2H). HRMS (ESI+) m/z for [M-2Cl⁻]²⁺ calcd: 433.0295; found: 432.9739. HPLC retention time: 22.41 min (99.5% purity by peak area).

rac-[Ru(4,4′-btfmb)₂(IP)](Cl)₂ (Ru-0T) Ru(4,4′-btfmb)₂Cl₂·2H₂O (91 mg, 0.12 mmol) and IP (22 mg, 0.1 mmol) were added to a microwave vessel containing argon-purged ethylene glycol (4 mL) and subjected to microwave irradiation at 180 °C for 15 min. The resulting dark red mixture was then isolated and purified in the same manner as [Ru(4,4′-btfmb)₃](Cl)₂, yielding the desired product as a dark red solid (81.3 mg, 51%). ¹H NMR (400 MHz, MeOD-d₃, ppm): δ 9.38 (dd, J=16.6, 1.9 Hz, 4H), 9.10 (d, J=8.4 Hz, 2H), 8.68 (s, 1H), 8.29 (d, J=5.9 Hz, 2H), 8.15 (dd, J=5.3, 1.3 Hz, 2H), 7.95 (d, J=6.0 Hz, 2H), 7.92 (dd, J=8.3, 5.3 Hz, 2H), 7.87 (dd, J=6.0, 1.9 Hz, 2H), 7.66 (dd, J=6.0, 1.9 Hz, 2H). HRMS (ESI+) m/z for [M-2Cl⁻]²⁺ calcd: 453.0326; found: 452.9735. HPLC retention time: 20.33 min (>98% purity by peak area).

rac-[Ru(4,4′-btfmb)₂(IP-1T)](Cl)₂ (Ru-1T). Ru(4,4′-btfmb)₂Cl₂·2H₂O (91 mg, 0.12 mmol) and IP-1T (30 mg, 0.1 mmol) were added to a microwave vessel containing ethylene glycol (4 mL) and subjected to microwave irradiation at 180 °C for 15 min. The dark red solution was transferred to a separatory funnel with H₂O (25 mL) and CH₂Cl₂ (25 mL). CH₂Cl₂ layer was used to wash the aqueous layer. CH₂Cl₂ (25 mL) and saturated aqueous KPF₆ (5 mL) were used to extract the product from the aqueous layer. The product was then purified using silica gel flash column chromatography. The product-containing fractions were transferred to a separatory funnel with CH₂Cl₂ (25 mL), H₂O (25 mL), and saturated aqueous KPF₆ (1 mL) and the [Ru(4,4′-btfmb)₂(IP-1T)](PF₆)₂ product was isolated via extraction. The PF₆⁻ salt was then converted to the corresponding Cl⁻ salt in quantitative yield using Amberlite IRA-410 with MeOH as the eluent, followed by further purification using Sephadex LH-20 with MeOH as the eluent, yielding [Ru(4,4′-btfmb)₂(IP-1T)](Cl)₂ as a dark red solid (43 mg, 25%). ¹H NMR (400 MHz, MeOD-d₃, ppm): δ 9.40 (d, J=1.9 Hz, 2H), 9.36 (d, J=1.8 Hz, 2H), 9.19 (s, 2H), 8.29 (d, J=5.8 Hz, 2H), 8.12 (d, J=5.2 Hz, 2H), 8.01 (d, J=3.6 Hz, 1H), 7.97 (d, J=6.1 Hz, 2H), 7.91 (dd, J=8.3, 5.2 Hz, 2H), 7.88 (dd, J=6.0, 1.9 Hz, 2H), 7.74 (d, J=5.0 Hz, 1H), 7.68 (dd, J=6.1, 1.8 Hz, 2H), 7.31 (dd, J=5.0, 3.7 Hz, 1H). HRMS (ESI+) m/z for [M-2Cl]²⁺ calcd: 494.0265; found: 494.0261, m/z for [M-2Cl-H]⁺ calcd: 987.0456; found: 987.0474. HPLC retention time: 21.75 min (99% purity by peak area).

rac-[Ru(4,4′-btfmb)₂(IP-2T)](Cl)₂ (Ru-2T). Ru(4,4′-btfmb)₂Cl₂·2H₂O (91 mg, 0.12 mmol) and IP-2T (38 mg, 0.1 mmol) were added to a microwave vessel containing argon-purged ethylene glycol (4 mL) and subjected to microwave irradiation at 180 °C for 15 min. The resulting dark red mixture was then isolated and purified in the same manner as Ru-1T, yielding the desired product as a dark solid (33 mg, 29%). ¹H NMR (400 MHz, MeOD-d₃, ppm): δ 9.40 (d, J=1.9 Hz, 2H), 9.37 (d, J=1.9 Hz, 2H), 9.17 (s, 2H), 8.29 (d, J=5.8 Hz, 2H), 8.13 (dd, J=5.2, 1.2 Hz, 2H), 7.99 (d, J=6.1 Hz, 2H), 7.94 (d, J=3.8 Hz, 1H), 7.91 (dd, J=8.3, 5.2 Hz, 2H), 7.88 (dd, J=5.9, 1.9 Hz, 2H), 7.71–7.67 (m, 2H), 7.47 (dd, J=5.1, 1.1 Hz, 1H), 7.43–7.38 (m, 2H), 7.13 (dd, J=5.1, 3.6 Hz, 1H). HRMS (ESI+) m/z for [M-2Cl⁻]²⁺ calcd: 535.0203. Found: 535.0201, m/z for [M-2Cl⁻-H]⁺ calcd: 1069.0333; found: 1069.0353. HPLC retention time: 22.86 min (95.2% purity by peak area).

rac-[Ru(4,4′-btfmb)₂(IP-3T)](Cl)₂ (Ru-3T). Ru(4,4′-btfmb)₂Cl₂·2H₂O (151 mg, 0.2 mmol) and IP-3T (76 mg, 0.164 mmol) were added to a microwave vessel containing argon-purged ethylene glycol (4 mL) and subjected to microwave irradiation at 180 °C for 15 min. The resulting dark red mixture was then isolated and purified in the same manner as Ru-1T, yielding the desired product as a dark red solid (55 mg, 27%). ¹H NMR (700 MHz, MeOD-d₃, ppm): δ 9.40 (d, J=2.0 Hz, 2H), 9.37 (d, J=1.9 Hz, 2H), 9.26 (s, 1H), 9.11 (s, 1H), 8.29 (d, J=5.8 Hz, 2H), 8.14 (dd, J=5.2, 1.3 Hz, 2H), 7.99 (s, 2H), 7.95–7.90 (m, 3H), 7.88 (dd, J=6.0, 1.9 Hz, 2H), 7.69 (dd, J=6.1, 1.9 Hz, 2H), 7.40 (d, J=3.8 Hz, 1H), 7.35 (d, J=3.8 Hz, 1H), 7.31 (dd, J=3.5, 1.1 Hz, 1H), 7.23 (d, J=3.8 Hz, 1H), 7.09 (dd, J=5.1, 3.6 Hz, 1H). ¹³C NMR (700 MHz, MeOD-d₃, ppm): δ 163.28, 163.08, 162.89, 159.60, 159.43, 155.07, 154.53, 150.34, 141.88, 140.97, 140.81, 140.77, 140.60, 139.10, 137.67, 135.97, 131.57, 129.82, 129.20, 126.94, 126.39, 125.84, 125.75, 125.40, 125.33, 125.23, 124.45, 124.34, 123.11, 123.02, 122.89, 122.79, 119.88. HRMS (ESI+) m/z for [M-2Cl⁻]²⁺ calcd: 576.0142; found: 576.0141, m/z for [M-2Cl⁻-H]⁺ calcd: 1151.0211; found: 1151.0232. HPLC retention time: 23.80 min (95.4% purity by peak area).

rac-[Ru(4,4′-btfmb)₂(IP-4T)](Cl)₂ (Ru-4T). Ru(4,4′-btfmb)₂Cl₂·2H₂O (114 mg, 0.2 mmol) and IP-4T (90 mg, 0.164 mmol) were added to a microwave vessel containing argon-purged ethylene glycol (4 mL) and subjected to microwave irradiation at 180 °C for 15 min. The reaction mixture was transferred to a 100 mL beaker and diluted with ~30 mL H₂O, then treated with 3 mL of saturated aqueous KPF₆, and stirred for 5 minutes. At this time, a red precipitate formed and was collected using a Büchner filtration apparatus. The product was then purified following the same procedure as described for Ru-1T, yielding [Ru(4,4′-btfmb)₂(IP-4T)](Cl)₂ as a dark red solid (77 mg, 35%). ¹H NMR (700 MHz, MeOD-d₃, ppm): δ 9.40 (d, J=2.0 Hz, 2H), 9.37 (d, J=1.9 Hz, 2H), 9.26 (s, 1H), 9.11 (s, 1H), 8.29 (d, J=5.8 Hz, 2H), 8.14 (dd, J=5.2, 1.3 Hz, 2H), 7.99 (s, 2H), 7.95–7.90 (m, 3H), 7.88 (dd, J=6.0, 1.9 Hz, 2H), 7.69 (dd, J=6.1, 1.9 Hz, 2H), 7.40 (d, J=3.8 Hz, 1H), 7.35 (d, J=3.8 Hz, 1H), 7.31 (dd, J=3.5, 1.1 Hz, 1H), 7.23 (d, J=3.8 Hz, 1H), 7.09 (dd, J=5.1, 3.6 Hz, 1H). ¹³C NMR (175 MHz, MeOD-d₃, ppm): δ 163.28, 163.08, 162.89, 159.60, 159.43, 155.07, 154.53, 150.34, 141.88, 140.97, 140.81, 140.77, 140.60, 139.10, 137.67, 135.97, 131.57, 129.82, 129.20, 126.94, 126.39, 125.84, 125.75, 125.40, 125.33, 125.23, 124.45, 124.34, 123.11, 123.02, 122.89, 122.79, 119.88. HRMS (ESI+) m/z: [M–2Cl]²⁺ calcd for C₄₉H₂₆F₁₂N₈RuS₃: 576.0142; found: 576.0141. [M–2Cl–H]⁺ calcd for C₄₉H₂₅F₁₂N₈RuS₃: 1151.0211; found: 1151.0232. HPLC retention time: 23.80 min (99% purity by peak area).

2.3. Computational Details

The complexes of this series were studied with a combination of DFT and TDDFT methods⁶² as implemented in the Gaussian 16 software⁶³. This computational protocol has been used to describe a variety of metal complex PSs for PDT, including Ru(II) and Os(II) systems.^{17,46,49,51,64–71} The geometries of the lowest energy singlet and triplet excited states were optimized with the PBE0 exchange-correlation functional⁷² combined with the 6–31+G(d,p) basis set for all atoms except Ru, which was described with the Stuttgart-Dresden pseudopotential⁷³. The M06 exchange correlation functional was used to compute the electronic excited states on top of the optimized geometries through the TDDFT formalism making use of the Tamm-Dancoff approximation (TDA).⁷⁴ This level of theory has been employed recently to describe ³MLCT and ³ILCT/³LLCT excited states and to predict emission energies for related oligothiophene-based Ru(II) and Os(II) complexes, providing a more accurate description of the vertical triplet state energies with respect to the conventional TDDFT analysis (which tends to underestimate the energy gaps⁷⁵). The calculations were performed in water within the framework of the integral equation formalism polarizable continuum model (IEFPCM),^76–78 with the dielectric constant of ε = 80 and default Gaussian 16 parameters. The excited state topologies were analyzed by post-processing of the Gaussian 16 output in two ways: (i) determination of the natural transition orbitals (NTOs) with the Chemissian 4.67 software,⁷⁹ and (ii) computation of the charge-transfer descriptors through fragment-based analyses performed with the TheoDORE 3.1.1 toolbox⁸⁰.

2.4. Electrochemistry

Voltammetry was performed in dimethylformamide (DMF, Fisher HPLC grade) that had been dried and deoxygenated with an Inert PureSolv MD7 solvent purification system, with 100 mM tetrabutylammonium hexafluorophosphate (TBAPF₆) (Fisher) as the supporting electrolyte, in a two-compartment low volume cell with the three-electrode configuration under argon. A 3 mm glassy carbon disc was used as the working electrode with a platinum wire counter electrode and a Ag/AgCl/4M KCl reference electrode. Ferrocene (Fc) was used as an internal standard. The complex solutions were approximately 4 mM for oxidation sweeps and 0.25 mM for reduction sweeps.

Measurements were conducted at room temperature using a WaveNow potentiostat (Pine Research Company) with Aftermath software. Cyclic differential-pulse voltammetry (CDPV) measurements used a sweep rate of 2 mV·s⁻¹ with a modulation amplitude varying from 12.5 to 100 mV. For reversible processes, the formal redox potential E°′ was taken as the average of E_pa (anodic peak potential) and E_pc (cathodic peak potential). For quasi-reversible processes, only E_pa or E_pc is reported.

3. RESULTS AND DISCUSSION

3.1. Synthesis and Characterization

[Ru(4,4′-btfmb)₃]⁺² and Ru-nT were synthesized utilizing our established method for related Ru(II) 2,9-dimethyl-1,10-phenanthroline systems.¹⁷ The initially isolated PF6⁻ salts of the complexes were purified using silica gel flash chromatography and then converted to their corresponding Cl⁻ salts on Amberlite IRA-410. Then size-exclusion chromatography was carried out on Sephadex LH-20 to give final yields of approximately 60% for [Ru(4,4′-btfmb)₃]⁺², Ru-0T, Ru-1T, and Ru-3T, around 40% for Ru-2T, and close to 30% for Ru-4T. These complexes underwent thorough characterization by 1D and 2D 1H NMR spectroscopy (Figures 1, and S1–S7), with signal assignments for [Ru(4,4′-btfmb)₃]²⁺ and Ru-0T–Ru-4T conducted using ¹H–¹H COSY NMR. The assignments aligned with those of our previously reported, related compounds.^17,20,46 Additionally, these complexes were characterized by high-resolution ESI⁺ MS (Figures S8–S14). The complexes were ≥95% pure by HPLC (Figures S15–S21).

Figure 1.

¹H NMR spectra showing aromatic region for [Ru(4,4′-btfmb)₃](Cl)₂ and Ru-nT (n=0–4) in MeOD-d₃ (Cl⁻ salts; 298 K). All spectra were collected at 500 MHz, except for Ru-4T, which was collected at 700 MHz.

The lipophilicities of the chloride salts of complexes were determined by measuring their distribution between 1-octanol and 10 mM phosphate buffer (pH 7.4) following the “shake-flask” method as we described previously.⁴⁹ The log D_o/w values for this series span 4 orders of magnitude, with [Ru(4,4′-btfmb)₃]²⁺ being the most hydrophilic at about −2 and Ru-4T most lipophilic near +2 (Figure 2). The three reference compounds lacking thiophene rings ([Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-0T) as well as Ru-1T show a preference for the aqueous buffer and accordingly have negative log D_o/w values. In contrast, complexes with 2–4 thiophenes (Ru-2T, Ru-3T, Ru-4T) have positive log D_o/w values and increase on going from n=2 to 4. The addition of the trifluoromethyl group qualitatively improved the overall aqueous solubility of the complexes with positive log D_o/w values compared to analogous Ru(II) and Os(II) complexes with other coligands where 4T often leads to precipitation at the octanol/buffer interface.^46,49

Figure 2.

Lipophilicity of [Ru(4,4′-btfmb)₃](Cl)₂, Ru-phen, and Ru-nT (n=0–4) in 1-octanol and phosphate buffer using the in-house “shake-flask” method.

3.2. Computation

Singlet states.

The singlet ground-state structures for [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-nT (n=0–4) optimized in water at the DFT/PBEO level of theory are shown in Figure 3. The associated geometric parameters (Table S3) confirm the octahedral arrangement of coordinating nitrogen atoms around the central Ru(II) ion and similar Ru-N bond distances for all of the compounds. The thiophene ring of Ru-1T and the first thiophene ring of Ru-nT (n=2–4) are coplanar with the coordinating IP ligand, but additional thienyl rings are conformationally more flexible. In the case of Ru-4T, the fourth thiophene ring is twisted as much as 18° with respect to the IP plane.

Figure 3.

Optimized geometries of [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-nT (n=0–4) in a water environment at the PBE0/6–31+G(d,p)/SDD level of theory. For the sake of clarity, the 4,4′-btfmb coligands are shown in grey for all the complexes except for [Ru(4,4′-btfmb)₃]²⁺.

The nT chain length impacts the frontier molecular orbitals and the resulting electronic transitions. The systematic reduction of the highest occupied molecular orbital (HOMO)–lowest unoccupied molecular orbital (LUMO) gaps with longer thiophene chains is due to an increase in the HOMO energies as the orbital contribution from nT increases. The LUMOs are primarily bftmb-based (>70%) with similar energies across the series, but the nT contribution to the HOMO reaches 51% for Ru-3T and almost 66% for Ru-4T (Figures 4 and S22, Table S4). The red-shift of the lowest energy spin-allowed singlet–singlet electronic transitions with increasing n is accompanied by a change in the nature of these states toward mixed ¹LLCT/¹ILCT/¹IL character for n>2 as shown by the natural transition orbital (NTO) analysis (Figures 5 and S23), consistent with other oligothiophene-bearing Ru(II) and Os(II) complexes.^{17,20,46,51,67} The NTOs are predominantly ¹MLCT (Ru→4,4′-btfmb) for the complexes lacking thiophene rings and Ru-1T, with electronic transitions computed in the range of 438 to 459 nm (only slightly higher in energy than the experimental bands falling between 456 and 471 nm in Figure 3). The NTOs for Ru-2T are of mixed ¹MLCT (53%)/¹LLCT (26%) character (¹LLCT is nT→4,4′-btfmb), and the lowest-energy transition is computed at 455 nm in close agreement with the experimental band at 462 nm. In contrast, the NTOs for Ru-3T and Ru-4T lack ¹MLCT character and are mixed ¹LLCT/¹ILCT/¹IL. Both complexes have lowest-energy transitions with significant ¹LLCT (nT→IP) character (55% for Ru-3T and 45% for Ru-4T) but differ in the relative contributions of ¹ILCT versus ¹IL. The NTOs for Ru-3T have a larger contribution from ¹IL (IP-based ππ*) compared to ¹ILCT (nT-based CT). The opposite is true for Ru-4T, where the ¹ILCT contribution is slightly more than that of ¹IL. The computed transitions are 469 and 493 nm for Ru-3T and Ru-4T, respectively, which lie within the broadened spectral features in the experimental spectra.

Figure 4.

(a) Calculated HOMO, LUMO, HOMO-1 and LUMO+1 orbital energies. (b) Percent contribution of the nT, bftmb, IP and Ru fragments to the HOMO and LUMO orbitals, for Ru-nT (n=0–4) in the singlet ground state, at the M06/6–31+G(d,p)/SDD level of theory. Additional details can be found in the SI.

Figure 5.

Occupied and virtual NTOs of the computed lowest-energy singlet-singlet transitions in water (λ) with the predominant character indicated. Additional NTOs are reported in Figure S23.

Triplet states.

The lowest energy excited triplet state (T₁) structures for [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-nT are similar to their corresponding ground-state structures except that those involving oligothiophenes adopt a completely planar chain conformation for maximum π-conjugation, with successive nT groups antiplanar with respect to one another. The geometrical parameters of these optimized T₁ states are compiled in Table S3 alongside the data for S₀. All complexes, with the exception of Ru-3T and Ru-4T, have T₁ states that are ³MLCT (Ru→btfmb), as shown by the NTO analyses (Figures 6 and S24, Table 1) and supported by Mulliken spin densities (MSDs) close to 1 for the Ru(II) center (Table S3). In sharp contrast, T₁ is localized primarily to the IP-nT ligand for Ru-3T and Ru-4T and its nature changes to mixed ³ILCT/³LLCT/³IL. The largest contributor to these mixed T₁ states is ³ILCT (CT within the nT chain) and is slightly more for Ru-4T (51% vs 39%), whereas the ³IL contribution (nT-localized) is similar for both at ~20% and ³LLCT (CT from nT to IP) contributes more for Ru-3T (25%) and less for Ru-4T (16%). The ³MC states involving Ru(II) and ³IL states localized to 4,4′-btfmb or IP are higher in energy and do not contribute to the computed NTOs for the lower-energy excited states.

Figure 6.

(a) Computed T₁ adiabatic energies for [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-nT (n=0–4). (b) Molecular fragments (left) defined to quantify the molecular topology of the T₁ excited states and their character (right). The corresponding NTOs are reported in Figure S24.

Table 1.

Adiabatic labels (State) and adiabatic ³MLCT and ³ILCT energies (ΔE_adia) for [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-nT (n=0–4) in a water environment at the TD-M06/6–31+G**/SDD level of theory. Additional details can be found in Table S5.

Compound	3MLCT		3ILCT
	State	ΔEadia(eV)	State	ΔEadia(eV)
[Ru(4,4′-btfmb) 3 ] 2+	T1	2.01	/	/
Ru-phen	T1	1.91	/	/
Ru-0T	T1	1.91	/	/
Ru-1T	T1	1.91	/	/
Ru-2T	T1	1.91	/	/
Ru-3T	T2	1.99	T1	1.54
Ru-4T	T2	1.88	T1	1.42

The computed adiabatic ³MLCT energies lie near 1.9–2.0 eV across the series, regardless of whether this state is T₁ or T₂ (Tables 1 and S5). These ³MLCT energies are lower (by ~0.2 to 0.3 eV) than those for related Ru(II) polypyridyl systems without substituted ligands due to the electron-withdrawing groups of the 4,4′-btfmb coligands in this series. These groups do not impact the ³ILCT/³LLCT/³IL energies, which are very similar to those computed for the related series with 1,10-phenanthroline (phen) coligands.⁵¹ However, the lower ³MLCT energies in the Ru(II) 4,4′-btfmb complexes result in an ³MLCT-based T₁ for Ru-2T (rather than the IP-nT-based triplet that tends to be T₁ in other related families). The mixed ³ILCT/³LLCT/³IL states that are T₁ in the case of Ru-3T and Ru-4T have computed adiabatic energies of 1.54 and 1.42 eV, respectively. The drop in energy of the T₁ state for Ru-3T and Ru-4T is directly related to its organic ³ILCT character and length of the nT chain as we have seen previously.^17,51

3.3. Spectroscopy

3.3.1. Electronic Absorption

The UV-vis absorption spectra of the series are overlaid in Figure 7, and the corresponding molar extinction coefficients for various peak maxima are summarized in Table 2. The spectra display some of the general features of typical Ru(II) polypyridyl complexes⁸¹ but with some distinctions for the complexes bearing IP-nT ligands. The sharp peaks near 295 nm and shorter are due to π‒π* transitions involving the 4,4′-btfmb ligand,⁵⁵ as well as phen and/or IP in the cases of Ru-phen and Ru-nT. The energies of these transitions are not affected by the length of the thiophene chain appended to the IP ligand. The broader and less intense peaks between 400 and 500 nm with a local maximum at 456 nm in [Ru(4,4′-btfmb)₃]⁺² are due to Ru²⁺(dπ)→L(π*) ¹MLCT transitions involving mainly the 4,4′-btfmb ligands as the π* acceptor orbitals. The substitution of one 4,4′-btfmb ligand for phen or IP causes the lowest-energy ¹MLCT transitions to red-shift by 10–20 nm. Assuming that the 4,4′-btfmb π* orbitals remain the acceptor orbitals for the lowest-energy ¹MLCT transitions in all cases, the effect of phen or IP is primarily on the dπ orbital energy. This effect is most evident for Ru-phen, Ru-0T, and Ru-1T. For complexes containing the IP-nT ligand, additional transitions overlap the ¹MLCT transitions. These contributions from the IP-nT ligands can be seen in the absorption spectra of the analogous uncomplexed IP-nT ligands and free oligothiophenes.⁸² The computed NTOs provide information on their contributions and characters in the Ru(II) complexes (Figures 5 and S23). The lowest-energy singlet–singlet transition is mixed ¹MLCT/¹LLCT for Ru-2T wherein ¹LLCT involves nT→4,4′-btfmb CT, while these transitions are mixed ¹LLCT/¹IL/¹ILCT and ¹LLCT/¹ILCT/¹IL for Ru-3T and Ru-4T, respectively. In these cases, ¹LLCT is the major contributor and involves nT→IP CT, ¹ILCT involves nT→nT CT, and ¹IL involves localized ππ* transitions within nT. These mixed transitions shift to longer wavelengths and increase in ¹ILCT character with increasing n as we have observed previously in related compounds.^{11,17,46,47,51}

Figure 7:

UV-vis spectra of [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-nT (n=0–4) normalized to the π‒π* peak near 295 nm

Table 2:

Molar Extinction Coefficients at Selected Wavelengths for the Complexes of This Study.

cmpd	λabs (nm) (log (ε / M−1 cm−1))
[Ru(4,4′-btfmb)3]+2	244 (4.34), 294 (4.84), 427 (4.01), 456 (4.12),
Ru-phen	260 (4.69), 296 (4.80), 353 (4.05), 432 (4.09), 471 (4.12)
Ru-0T	248 (4.68), 296 (4.84), 341 (4.13), 435 (4.12), 470 (4.13)
Ru-1T	245 (4.57), 294 (4.92), 335 (4.51), 462 (4.20)
Ru-2T	250 (4.68), 295 (4.89), 370 (4.69), 462 (4.35)
Ru-3T	251 (4.58), 294 (4.76), 410 (4.57), 426 (4.55)a
Ru-4T	246 (4.57), 297 (4.81), 351 (4.33), 441 (4.70), 468 (4.62)a

^aThe maximum of the longest wavelength singlet–singlet transition is obscured.

3.3.2. Singlet Oxygen Sensitization

The singlet oxygen quantum yields (Φ_Δ) of the complexes were calculated by measuring the intensity of O₂ phosphorescence (¹Δ_g→³Σ_g) centered at 1260 nm against [Ru(bpy)₃]²⁺ as the standard (Φ_Δ=0.56).⁸³ These are reported in Table 3. [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-0T are moderately efficient ¹O₂ generators (Φ_Δ=0.47–0.64) and similar to [Ru(bpy)₃]²⁺. However, the thiophene-containing complexes Ru-1T–Ru-4T exhibit larger differences (Φ_Δ=0.13–0.66) and are less efficient than what has been observed for related compounds.^11,47 The largest ¹O₂ quantum yield was measured for Ru-3T at 66%. While all the complexes show a wavelength-dependence for Φ_Δ, Ru-4T exhibits a notable concentration dependence for the ¹O₂ quantum yield (Figure S25 and Table S6). These differences in Φ_Δ appear to be unrelated to differences in other photophysical parameters such as emission wavelengths and triplet lifetimes (vide infra).

Table 3:

Spectroscopic data for compounds [Ru(4,4′-btfmb)₃](PF₆)₂, Ru-phen, and Ru-0T–Ru-4T as PF₆⁻ salts. Excitation wavelengths are indicated in parenthesis. Emission lifetimes were measured following a <5 ns 355 nm laser pulse.

cmpd	RT emission			77 K emission		ΦΔ (λex / nm)	τTA / μs)
	λem. (λex) / nm	Φem	τem / μs	λem. (λex) / nm	Φem,77K		λex=355 nm	λex=532 nm
[Ru(4,4′-btfmb)3]2+	636 (458)	1.6×10−1	1.5	604,653 (460)	2.5×10−1	0.47 (462)	1.5	1.5
Ru-phen	660 (470)	9.0×10−2	0.89	619, 668 (472)	2.4×10−1	0.64 (470)	0.89–0.94	0.77–0.92
Ru-0T	662 (470)	9.1×10−2	0.83	619, 669 (470)	1.9×10−1	0.50 (469)	0.78–0.86	0.78–0.87
Ru-1T	661 (460)	1.8×10−2	0.79	620, 670 (469)	2.0×10−1	0.13 (462)	0.81–0.93	0.71–0.87
Ru-2T	659 (457)	4.1×10−3	0.62	634, 687 (463)	4.3×10−2	0.28 (466)	0.61–0.72	0.63–0.71
Ru-3T	664 (462)	1.3×10−3	0.79	623, 673 (470)	9.7×10−3	0.66 (470)	20–21	22–24
Ru-4T	650 (435)	v. wk.	0.64	625, 672 (468)	1.4×10−3	0.40 (467)	19–20	20–21

3.3.3. Emission

All of the complexes in the series exhibited a broad, featureless red emission band at room temperature (Figure 8a and Table 3). This emission was centered around 636 nm (τ_em=1.5 μs) for the parent [Ru(4,4′-btfmb)₃]²⁺ complex in MeCN at room temperature (RT) and shifted to shorter wavelengths with vibronic intervals of around 1350 cm⁻¹ at 77 K (Figure 4b).^54–57 Such behavior is consistent with emissive ³MLCT states in Ru(II) complexes with polypyridyl ligands.⁸⁴ The thermally induced Stokes shift (ΔE_S) of around 830 cm⁻¹ is slightly smaller than the related model complex [Ru(bpy)₃]²⁺ (ΔE_S=1127 cm⁻¹)⁸⁵ but in agreement with the assignment.⁸⁶ Similar to what was observed for the ¹MLCT absorption bands, the introduction of a phen or an IP ligand results in bathochromic shifts of up to 30 nm but otherwise similar spectra and lifetimes (τ_em=0.8–0.9 μs). The ³MLCT emission energies and lifetimes (τ_em=0.6–0.8 μs) also do not depend on the number of thiophenes in the IP-nT ligand, suggesting that the π* acceptor orbitals in the Ru(dπ)→L(π*) transitions involve the 4,4′-btfmb coligands. The emission quantum yield (Φ_em) ranges from 16% for [Ru(4,4′-btfmb)₃]²⁺ to 9% for the complexes without thienyl groups. The value of Φ_em drops to around 2% for Ru-1T, and additional thienyl rings further decrease the emission output to <1%. Although Ru-4T produces weak but detectable emission, a value for Φ_em was not calculated due to the extremely low signal-to-noise ratio. The values for Φ_em increase up to 10-fold at 77 K but Ru-4T is still only about 0.14%. While the energy of the emissive ³MLCT state does not appear to change across the series, the emission quantum yields decrease with increasing n and implicate additional excited-state deactivation pathways. The computed NTOs indicate that the nature of T₁ changes substantially with the introduction of three or four thienyl groups. The lowest-energy triplet states are of ³MLCT character for all of the complexes except Ru-3T and Ru-4T, where T₁ becomes mixed ³ILCT/³LLCT/³IL (with ³ILCT involving nT→nT CT, ³LLCT involving nT→IP CT, and ¹IL based on localized ππ* transitions within n). These organic triplets involving the IP-nT ligand are nonemissive, and thus the very weak emission arises from the ³MLCT state that is T₂ for Ru-3T and Ru-4T and of similar energy to the other complexes in the series.

Figure 8:

Normalized emission spectra [Ru(4,4′-btfmb)₃]²⁺ and the Ru-nT series as PF₆⁻ salts at RT and at 77K. The RT emission was measured in MeCN degassed by freeze-pump-thaw (5 cycles). The 77K emission was measured in a 4:1 EtOH:MeOH glass. Excitation wavelengths are noted in parentheses.

3.3.4. Transient Absorption and Excited State Pathways

The triplet excited states were investigated using nanosecond transient absorption (TA) spectroscopy with excitation from a 355 nm (Figure 9) or 532 nm laser (Figure S26) of ≤5 ns pulse width. The responses with the two different excitation wavelengths are similar, suggesting similar excited-state dynamics. The differential excited-state absorption (ESA) spectra were collected using solutions of the PF₆⁻ salts of the compounds in degassed (5x freeze-pump-thaw) MeCN. Early time slices are presented in Figures 9 and S26, and the full set of relaxation spectra over different time points are collected in Figures S27 and S28. Transient lifetimes were measured at the ESA maxima and bleach minima and are listed in Table 3.

Figure 9:

Transient absorption (TA) spectra of [Ru(4,4′-btfmb)₃](PF₆)₂, Ru-phen, and Ru-nT as PF₆⁻ salts in degassed MeCN at RT (λ_ex=355 nm) integrated over the indicated time slice following the excitation pulse. The profiles for λ_ex=532 nm are similar (Figure S26). The horizontal dashed lines indicate ΔO.D.=0 for each compound differentiated by color.

The ESA profiles of [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, Ru-0T, and Ru-1T are similar, consisting primarily of a strong ¹MLCT ground-state bleach near 450 nm superimposed with new ESA characteristic of the ³MLCT state. The stronger ESA near 375 nm involves 4,4′-bftmb⁻ transitions, and the extremely weak and broad absorption past 525 nm is due to 4,4′-btfmb⁻ or LMCT transitions involving Ru(II). The TA lifetimes match the emission lifetimes and support lowest-lying ³MLCT states for these complexes as predicted from the NTO analyses.

The ESA profile of Ru-2T exhibits features consistent with a triplet excited-state localized to the IP-nT ligand, as we have previously reported.^{11,17,50,51,87–89} This broad and rather intense ESA near 450–700 nm is superimposed on the bleach with its minimum near 370 nm due to a strong ground-state absorption contributed by the IP-2T ligand (Figure 9). While Ru-2T has the signature of an nT-based triplet, its TA lifetime matches the emissive ³MLCT lifetime at all wavelengths. This observation and the NTO analyses suggest that the ³MLCT state computed as T₁ and the ligand-localized triplet are close in energy and decay with a common lifetime.⁹⁰ In this regard, Ru-2T behaves differently from some its close analogs containing the IP-2T ligand, where T₁ is of mixed ³LLCT/³ILCT character with a prolonged lifetime,⁵¹ due to the lower ³MLCT energies of Ru(II) polypyridyl complexes incorporating 4,4′-btfmb coligands.

The ESA spectra of Ru-3T and Ru-4T are characteristic of oligothiophene-based ³ILCT states, with longer TA lifetimes. The excited state relaxation pathways are proposed in Scheme 1. Ru-3T shows a bleach corresponding to the ground-state ππ* transition near 410 nm and a broad ESA with its maximum near 610 nm. For Ru-4T these were shifted to around 430 nm and 655 nm, respectively. These ESA features are similar to those of the free IP-3T and IP-4T ligands.^46,49 The decays are monoexponential, with τ=~20 μs for both complexes, indicating that the long-lived ³ILCT is decoupled from the shorter-lived emissive ³MLCT with computed T₂ energies near 1.9–2.0 eV. This is further supported by the very low emission quantum yields for Ru-3T and Ru-4T and computed T₁ energies of 1.54 and 1.42 eV, respectively, with significant ³ILCT character.

Scheme 1.

Jablonski Diagram Illustrating the Excited-State Relaxation Pathways of Complexes Ru-3T and Ru-4T. Energies are not to scale.

3.4. Electrochemistry

The electrochemistry of Ru(II) polypyridyl complexes is typified by single electron processes involving one-electron oxidation of the metal center and three sequential reductions on each of the ligands.⁹¹ Oxidation of the Ru²⁺ center (+0.98 V versus Fc, MeCN) tends to be electrochemically reversible, and the ensuing low-spin 4d⁵ complex is chemically stable. In complexes like [Ru(bpy)₃]²⁺, the first reduction (−1.72 V versus Fc, MeCN) involves the lowest-energy ligand π* orbital. Since the low spin 4d⁶ configuration is thereby unaffected, the complex remains substitutionally inert, and the process is also reversible. The added electron is localized on one ligand, and thus [Ru(bpy)₃]²⁺ exhibits three sequential one-electron reductions under straightforward electrochemical conditions. In a potential window widened by low temperature cyclic voltammetry, [Ru(bpy)₃]²⁺ has been shown to participate in a total of six one-electron reductions.^92,93

There are additional redox processes for some members of the present series due to the presence of the electrochemically active oligothiophene⁹⁴ unit in complexes Ru-2T–Ru-4T. The formal redox potentials of the series were measured by cyclic differential pulse voltammetry (CDPV) to enhance the signal, with ferrocene (Fc) as an internal reference (E_1/2 (Fc/Fc⁺)=0.380 V vs SCE⁹⁵). The potentials are compiled in Table 4 and compared graphically in Figure 10 with tentative assignments. The CDPV plots for oxidation and reduction are shown in Figures S29 and S30, respectively.

Table 4.

Formal redox potentials for the hexafluorophosphate salts of the complexes measured using CDPV at approximately 1.0 mM in MeCN containing TBAPF₆. The potentials are referenced in volts (V) against ferrocene as the internal standard. The working and reference electrodes were glassy carbon and Ag/AgCl/4M KCl, respectively. Overlapping waves were deconvoluted mathematically (error approximately ±0.02 V).

Compound	nT2−→nT3−	nT−→nT2−	nT→nT−	LL2−→LL12−	LL1−→LL12−	LL3→LL3−	LL2→LL2−	LL1→LL1−	Ru2+→Ru3+	nT→nT+
[Ru(4,4′-btfmb)3]2+				−2.75	−2.21	−1.61	−1.39	−1.24	+1.34
Ru-phen				−2.70	−2.30	−2.01	−1.49	−1.29	+1.18
Ru-0T				−2.81	−2.29	−2.14	−1.54	−1.33	+1.18
Ru-1T				−2.75	−2.30	−2.12	−1.51	−1.32	+1.22	+1.02
Ru-2T			−2.81	−2.71	−2.29	−2.12	−1.51	−1.31	+1.26	+0.89
Ru-3T		−2.96	−2.54	−2.86	−2.32	−2.15	−1.54	−1.32	+1.29	+0.69
Ru-4T	−2.93	−2.55	−2.23	−2.83	−2.33	−2.13	−1.52	−1.32	+1.27	+0.58

Figure 10:

Redox potentials (vs. Fc) of [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-nT as PF₆⁻ salts in degassed MeCN at RT and their tentative assignments.

The metal oxidation of [Ru(4,4′-btfmb)₃]²⁺ and the three one-electron reductions agree well with published data⁵⁵ and are around 0.3–0.5 V more positive than the corresponding processes in [Ru(bpy)₃]²⁺, a consequence of the electron-withdrawing nature of the 4,4′-btfmb ligands. The shift of the Ru²⁺→ Ru³⁺ oxidation is slightly attenuated when one 4,4′-btfmb ligand is replaced with phen or IP-nT. The oxidation of nT becomes more favorable with increasing n, consistent with the behavior of free oligothiophenes.⁹⁶ The oxidation potential of the thienyl group ranges from +1.02 V for Ru-1T to +0.58 for Ru-4T and is less positive than the metal center in all cases, indicating that nT is more easily oxidized than Ru(II) regardless of the number of thiophenes. The oxidation of nT does shift the Ru²⁺→ Ru³⁺ oxidation slightly more positive, consistent with a decrease in electron density on the metal. For Ru-3T and Ru-4T there is at least one additional peak between those for nT and Ru²⁺ oxidation that is of lesser intensity and solvent-dependent (Figure S29). This peak, not listed in Table 4, is thought to arise from the oxidation of the σ-dimer radical cation ((Ru-nT)₂^+• → (Ru-nT)₂ + 2H⁺ + e⁻) given the known abilities of certain oligothiophenes to undergo electrochemical σ-dimerization through their α positions. However, this requires further investigation to confirm.

[Ru(4,4′-btfmb)₃]²⁺ and the other complexes without thiophenes (Ru-phen, Ru-0T) as well as Ru-1T exhibit five sequential reductions spanning −1.24 to −2.81 V (Figure 10). For [Ru(4,4′-btfmb)₃]²⁺, this involves sequential one-electron reductions on each of the three 4,4′-btfmb ligands followed by second reductions on two of those ligands (within the experimental potential window). We believe this study is the first to report the second reductions of the two 4,4′-btfmb ligands occurring between −2.2 and −2.9 V for [Ru(4,4′-btfmb)₃]²⁺.

When one of the 4,4′-btfmb ligands is replaced by phen, IP-0T, or IP-1T, the third reduction involves phen or IP and is less favorable by around 0.4–0.5 V (owing to their lack of electron-withdrawing −CF₃ substituents). Ru-2T–Ru-4T exhibit similar reductions involving the 4,4′-btfmb coligands and IP, but they accommodate additional reductions on the oligothienyl groups. For Ru-2T, the 2T group accepts only one electron and this sixth reduction is the least favorable and occurs near −2.81 V. The 3T group of Ru-3T can be doubly reduced (−2.54 and −2.96 V), where the first nT reduction is easier than the last 4,4′-btfmb reduction. Continuing this trend, the 4T group of Ru-4T accommodates three electrons (−2.23, −2.55, and −2.93 V), with the first nT reduction occurring near that of IP at −2.13 V. The effect of the thiophene chain length is dramatic. The nT reduction potential shifts positive by more than 0.5 V on going from two to four thiophenes, and the Ru-4T complex can accommodate at least eight extra electrons in its ground state at room temperature.

3.5. In Vitro Photobiological Activity

The compounds in this series were assessed for their cytotoxicity in the absence of light (dark) as well as their light-triggered cytotoxicity against human skin melanoma cells (SK-MEL-28) and lung carcinoma cells (A549) cultured as 2D monolayers (Figure 11). The experimental details can be found in our previously published work^11,17,20 as well as in the SI. The spectral output of the light sources is shown in Figure S31, their overlap with the absorption profiles of the complexes is shown in Figure S32, and the estimated absorbed photon flux is compiled in Table S7. The results are summarized in Figure 11 and Tables S8, S9.

Figure 11.

In vitro cytotoxicity and photocytotoxicity log (EC₅₀ ± SEM) values (left) and PI values (right) obtained from dose–response curves in the A549 cell line (a) and SK-MEL-28 melanoma cell line (b) with [Ru(4,4′-btfmb)₃](Cl)₂ and Ru-phen–Ru-4T. Treatments include dark (black circles) or light delivered at a fluence of 100 J cm⁻² and irradiance of ~20 mW cm⁻². The light wavelengths were broadband visible (400–700 nm, blue squares), 523 nm (green inverted triangles), or 633 nm (red triangles). Data collected under normoxic (~18.5% O₂) and hypoxic (1% O₂) conditions is represented with closed symbols and open symbols, respectively.

3.5.1. Normoxia.

For each photobiological assay in normoxia, SK-MEL-28 or A549 cells were seeded into two sets of 384-well plates: one for dark cytotoxicity evaluation and the other for photocytotoxicity assessment. After allowing the cells to adhere for 3–5 hours at 37°C, cells were dosed with the compounds (1 nM to 300 μM). The light plates were exposed to light treatments after a 13–22 h DLI, while the dark plates remained in the incubator. The light treatments were delivered at a fluence of 100 J cm⁻² emitted from broadband-visible (400–700 nm, 21 mW cm⁻²) or monochromatic (±2.5 nm) green (523 nm, 18 mW cm⁻²) or red (633 nm, 18 mW cm⁻²) LEDs.

After light treatment, the plates were allowed to incubate in normoxia at 37°C for 24 h. Cell viability was then assessed indirectly using a resazurin-based cell viability assay. Sigmoidal fits of the dose-response curves were used to calculate the effective concentrations required to reduce cell viability by 50% (EC₅₀ values) for both treatment conditions. The amplified cytotoxic effects upon light exposure, known as phototherapeutic indices (PIs), were calculated as ratios of dark EC₅₀ to light EC₅₀ values.

The complexes were mostly nontoxic to both cell lines in the dark. [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-0T exhibited dark EC₅₀ values that exceeded the highest doses used in the assay (>300 μM), indicating a lack of toxicity. They were also inactive against both cell lines under any light condition. Ru-1T–Ru-3T were considered nontoxic toward both cell lines, with dark EC₅₀ values >170 μM. Ru-4T had the lowest dark EC₅₀ values, which were still relatively high at 100 μM and 120 μM in SK-MEL-28 and A549 cells, respectively.

SKMEL28 cells.

Increasing the number of thiophene rings (n=1–4 thienyl groups) increased the potency with visible light, ranging from 22 μM (PI=11) for Ru-1T to as low as 10 nM (PI=10,000) for the most potent compound Ru-4T in SK-MEL-28. Appending two thiophenes (Ru-2T) improved the potency 36-fold, shifting the EC₅₀ values to sub-micromolar values near 0.61 μM (PI=290). About 100-fold enhancement in photocytotoxicity was seen moving to three thiophene rings (Ru-3T; SK-MEL-28: EC₅₀=0.21 μM, PI=920), and over 2000-fold improvement occurred with four thiophenes (Ru-4T; SK-MEL-28: EC₅₀=10 nM, PI=10,000).

The activity of the series was mostly diminished with longer-wavelength green or red light. With green light, Ru-1T lost most of its activity (EC₅₀=103 μM, PI=2). The activity Ru-2T dropped by a factor of 3 but was still single-digit micromolar (EC₅₀=1.9 μM, PI=91), and Ru-4T dropped by 10-fold but remained sub-micromolar (EC₅₀=0.12 μM, PI=815). Ru-3T, on the other hand, maintained its activity (EC₅₀=0.27 μM, PI=731). With red light, Ru-3T (EC₅₀=7.02 μM; PI=28) and Ru-4T (EC₅₀=2.74 μM; PI=37) showed modest activity. This outcome aligns with the anticipated behavior of compounds that exhibit minimal absorption of red light.⁹⁷

A549 cells.

The A549 cell line proved more resistant to the light-triggered compounds than SK-MEL-28 under similar conditions, but the overall trend in activity remained the same. Ru-1T (visible EC₅₀=29 μM, PI=7) was the least active thienyl complex, followed by Ru-2T (visible EC₅₀=1.37 μM, PI=127), Ru-3T (visible EC₅₀=0.70 μM, PI=280), and finally Ru-4T (EC₅₀=0.077 μM, PI=1500). Treatments with red and green light similarly attenuated the overall activity of the series.

3.5.2. Hypoxia.

The assays under hypoxic conditions mirrored the procedure followed for normoxia, with a notable exception: post-cell adhesion, the plates—both for dark and light conditions—were transferred to a hypoxia chamber with 1% O₂ atmosphere for a duration of 2–3 h prior to introducing the compounds. After a DLI of 18 h in the hypoxia chamber, the concentration of dissolved O₂ was verified in the assay wells using an immersive optical probe before sealing the plates to be light treated with transparent qPCR film that has a low gas permeability. Light was delivered outside the hypoxia chamber for approximately 1.5 h, while the dark plates were kept inside an incubator. The films were removed from the light plates at the end of the illumination period. Both dark and light plates were then incubated in normoxia (37°C, 5% CO₂, with relative humidity above 90%) for 20–23 h, after which time the cell viability was assessed.

Hypoxic conditions broadly attenuated the activity of the series in both cell lines, where A549 cells were completely resistant to all treatments and SK-MEL-28 cells were moderately sensitive under certain conditions. Regardless of light-treatment, [Ru(4,4′-btfmb)₃](Cl)₂ and compounds Ru-0T–Ru-2T exhibited no photocytotoxic effect under hypoxic conditions toward SK-MEL-28 cells. On the other hand, Ru-3T showed an EC₅₀ value of ~1 μM with visible and green light treatments. With red light, Ru-4T was less active (EC₅₀=26 μM, PI=4), but otherwise the series was inactive. The dramatic decline in the activity of Ru-3T and Ru-4T, coupled with the complete inactivity of the other compounds in the series under hypoxic conditions, suggests that oxygen plays a pivotal role in their photocytotoxic mechanisms. This observation solidifies the hypothesis that the primary driver behind the observed photocytotoxicity of these compounds in normoxic conditions stems from oxygen-dependent photophysical processes.

3.5.3. Biological replicates.

Figure 11 and Table S9 show representative activity against SK-MEL-28 for one biological replicate based on the mean values from three technical replicates with minimal standard deviation. Since a greater degree of variation is expected over biological replicates, we assessed the activity of the compounds (Ru-3T and Ru-4T) of highest activity over five biological replicates (each carried out in triplicate) with SK-MEL-28 cells, as illustrated in Figure S34 and detailed in Tables S10–S13. Repeat 0 corresponds to the data in Figure 11 and Table S9. The subsequent biological replicates are labeled Repeats 1–5. SK-MEL-28 cells were selected for these studies because this is the cell line we have historically used to rank potency across all compounds made in our laboratory.¹⁷

Ru-3T and Ru-4T were nontoxic over all biological replicates in both normoxia and hypoxia without light activation, with mean EC₅₀ values of 197 μM for Ru-3T and 100 μM for Ru-4T. The visible EC₅₀ values for Ru-3T in normoxia were in the range of 70–700 nM (mean=440 nM), and their visible PIs spanned 305–2700 (PI_avg=1400). The EC₅₀ values for Ru-4T under the same conditions ranged from 9–59 nM with a mean of 17 nM (PI=1900–12000, mean=6500). The results from all 6 replicates varied less an order of magnitude.

Using green light, the EC₅₀ values for Ru-3T fluctuated between 0.22 and 0.66 μM, with PIs between 300 and 1000. The mean values were determined to be 0.44 μM and 548, respectively. For Ru-4T, the results were slightly more variable, with green EC₅₀ values ranging from 54 to 320 nM, and the corresponding PIs falling in the 800 to 2000 range. Even though Ru-4T was generally more active than Ru-3T, the difference in their activities was only around 3-fold.

Under red light illumination in normoxia, the EC₅₀ values for Ru-3T and Ru-4T were reproducibly in the single digit micromolar regime. Ru-3T exhibited red EC₅₀ values between 2 and 7 μM (mean=4.6 μM) and PIs ranging from 23 to 101 (PI_avg=52). The activity of Ru-4T was slightly more consistent than Ru-3T, with red EC₅₀ values falling between 1.3 and 3.3 μM (mean=2.2 μM) with PIs between 21 and 91 (PI_avg=53). The responses with red light for Ru-3T and Ru-4T is notable as both complexes exhibit vanishingly low molar extinction coefficients at 633 nm. This characteristic has been reported previously,⁹⁷ and tends to require lowest-lying ³ππ* triplets with prolonged lifetimes such as those observed for Ru-3T and Ru-4T.

The activity of Ru-3T in hypoxia was mostly consistent but repeats 3 (with red light) and 4 (with vis and green light) were outliers: repeat 3 displayed unusually high activity compared to the other replicates, and no activity at all was observed in repeat 4. In some of the replicates, the visible and green activity of Ru-3T improved when compared to our initial measurements, but all values were still within one order of magnitude. With visible light, the EC₅₀ value was 1.2 μM (PI=170) in repeat 0 but improved to a range of 300–870 nM (PI=230–700, PI_avg=300). Similarly, the initial green EC₅₀ value for Ru-3T was 1.1 μM (PI=180), which lowered to 0.3–1.0 μM (PI=200–700, PI_avg=270) in subsequent assays. The observation of two distinct outliers, and the fact that the activity of Ru-3T was actually greater in subsequent biological replicates, underscores the importance of performing biological replicates when evaluating photobiological efficacy.

The activity of Ru-4T varied slightly more in hypoxia than it did in normoxia, but all EC₅₀ values were still within roughly one order of magnitude. Using visible light treatment, repeats 0 and 2–5 fell between 0.13–0.54 μM (PI=200–400), but repeat 1 presented a slightly lower EC₅₀ of 35 nM (PI=2900). With green and red light treatments, Ru-4T proved to be consistently more active than what was originally observed in repeat 0, with initial EC₅₀ values of 1.0 μM (PI=90) and 26 μM (PI=4), respectively, which improved to a range of 200–900 nM (PI=100–500, PI_avg=210) and 3–9 μM (PI=7–30, PI_avg=21) during replicates 1–5. The average hypoxic red activity (<10 μM) of Ru-4T is particularly notable for this class of PS, and is comparable to the “ubertoxin” ML19C01 (which presents sub-nanomolar normoxic phototoxicity).^17,20

In conclusion, across six biological replicates, Ru-4T has consistently displayed superior activity compared to Ru-3T, particularly when activated with visible light under normoxic conditions. Under these conditions, Ru-4T had an average EC₅₀ of 24 nM (PI_avg=6500). Notably, the values obtained in all biological replicates were within one order of magnitude, making Ru-4T more consistent than ML19C01, which varies by up to 6 orders of magnitude.^17,20 This marked improvement in consistency may be related to the improved aqueous solubility observed in this family of complexes, and the effects of fluorination on biological reproducibility is being investigated further.

4. CONCLUSIONS

Herein we synthesized a family of Ru(II) polypyridyl complexes featuring two 4,4′-btfmb coligands and the IP-nT ligand (n=1–4) and compared these systems to the reference compounds [Ru(4,4′-btfmb)₃]²⁺, Ru-phen, and Ru-0T. This study stands as part of our larger initiative to examine the impact of structural variations on the physicochemical, photophysical, electrochemical, and (photo)biological properties of oligothienyl-containing metal complexes.

The complexes lacking thienyl groups and Ru-1T are relatively hydrophilic, while the lipophilicities of Ru-2T–Ru-4T increase with n. The number of thienyl groups in the Ru-nT complexes also have a striking influence on the ground-state absorption and TA spectra with transitions involving the nT groups (¹LLCT and ¹ILCT) shifting to longer wavelengths and increasing in ¹ILCT character with increasing n. Based on their TA profiles, Ru-3T and Ru-4T have lowest-lying ³ILCT states, while the complexes lacking nT groups and Ru-1T as well as Ru-2T have ³MLCT states as T₁. The TA lifetimes corroborate this assignment with Ru-3T and Ru-4T having triplet lifetimes on the order of 20–24 μs and ESA spectra consistent with ³ILCT states. The other complexes have triplet lifetimes and TA spectra consistent with ³MLCT states near 1 μs. The exception is Ru-2T, which has the signature of an nT-localized triplet but with a lifetime characteristic of the ³MLCT state. All of the complexes have emissive ³MLCT states, but with emission quantum yields that decrease with increasing n. The ¹O₂ quantum yields increase from 13 to 66% from one to three thiophenes but decrease to 40% at four. The parent complexes lacking thienyl groups generate ¹O₂ better than the Ru-nT complexes, with the exception that Ru-3T is slightly more efficient than Ru-phen.

Compared to Ru(II) polypyridyl complexes such as [Ru(bpy)₃]²⁺, the parent [Ru(4,4′-btfmb)₃]²⁺ complex is oxidized less readily but is reduced more easily owing to the electron-withdrawing nature of the −CF₃-substituted bipyridyl coligands. All of the complexes undergo at least five reductions over the potential window investigated, whereas [Ru(bpy)₃]²⁺ undergoes only three. For Ru-1T–Ru-4T, the additional oxidation due to the thienyl group (nT^0/+) occurs before Ru² oxidation. Ru-2T–Ru-4T underwent additional reductions involving their nT groups: one for Ru-2T, two for Ru-3T, and three for Ru-4T. The most redox-active of the series, Ru-4T, undergoes two oxidations and eight reductions over the potential window investigated.

Only complexes with thiophenes are phototoxic toward melanoma cells (SK-MEL-28), with potency increasing with n. Ru-4T demonstrated EC₅₀ values in the low-nanomolar regime and PIs around 10⁴ in normoxia under visible light. Notably, all values fell within one order of magnitude of each other over six biological replicates. Both Ru-3T and Ru-4T retain some of this activity in hypoxia, where Ru-4T has a visible EC₅₀ as low as 35 nM and PI as high as 2,900. These results highlight the significance of lowest-lying ³ILCT states for enhancing phototoxicity toward cancer cells and maintaining activity in hypoxia. This family illustrates that Ru(II) coordinated to IP-3T or IP-4T ligands results in an excellent pharmacophore that can tolerate a variety of changes to the polypyridyl ancillary ligands. In this case, the redox activity of the system is changed substantially by the 4,4′-btfmb coligands (compared to our previously published analogs), and the complexes with three or four thiophene rings were still of high potency. The fluorinated polypyridyl ligands are of ongoing interest given the important role that fluorinated groups play in medicinal chemistry. These findings highlight Ru-4T as an excellent candidate for further investigations.