Investigation of the Effect of the Trifluoropropynyl Ligand on Pt(N^C^N)X (X = Cl, C₂CF₃) Complexes

Abstract

The tuning of the luminescent properties of Pt^II complexes for possible use in organic light-emitting diodes (OLEDs) and sensing applications is commonly achieved by altering the electronic properties of the ligands. Our group recently demonstrated that the trifluoropropynyl ligand is strongly electron-withdrawing and possibly useful for blueshifting emission. Herein, we report the synthesis of two complexes of this trifluoropropynyl ligand, namely PtLC₂CF₃ and PtL^FC₂CF₃ (L = 1,3-di(2-pyridyl)benzene; L^F = 4,6-difluoro-1,3-di(2-pyridyl)benzene). The PtLC₂CF₃ complex crystallized in the monoclinic space group P2₁/n with Z = 4. The PtL^FC₂CF₃ complex crystalized in the triclinic space group P-1 with Z = 2. Changing the tridentate ligand from L to L^F resulted in a change in the packing structure, with the latter showing a metallophilic interaction (Pt-Pt distance = 3.3341(3) Å). The solution photophysics of the trifluoropropynyl complexes is compared with that of the corresponding Cl complexes, PtLCl and PtL^FCl. Replacement of the chloro ligand with the trifluoropropynyl ligand blueshifted the monomer emission by less than 5 nm but blueshifted the excimer emission peaks by 15–20 nm. The complexes of the trifluoropropynyl ligand also favor the excimer emission more than the complexes of the chloro ligand. The excimer emission is quenched by dissolved oxygen significantly more than the corresponding monomer emission. The excimer emission and monomer emission are well separated, and the ratio of monomer to excimer emission is strongly dependent on oxygen concentration.

1. Introduction

Recently, there has been substantial research into improving the performance of organic light-emitting diodes (OLEDs) for use in digital displays and lighting applications. A variety of organometallic Ir^III- and Pt^II-luminophores have been studied for use in OLEDs because the high spin–orbit coupling (SOC) constants associated with these metals [1] enable the harvesting of both singlet and triplet excitons [2]. In particular, these organometallic luminophores are actively researched for use in the manufacture of efficient deep-blue OLEDs, which has proven to be challenging [3,4]. Additionally, deep-blue emitters are desirable for white OLED (WOLED) applications, as deep-blue light can be mixed with orange light and appear white to a human observer [5,6].

The ground electronic state of Pt^II complexes is a singlet ground state (¹GS), i.e., all electrons are paired. Spin-allowed excitation results in a singlet excited state (¹ES) which can undergo intersystem crossing (ISC) to a triplet excited state (³ES) (Figure 1). To produce deep-blue-emitting Pt^II luminophores, it is necessary to increase the energy gap between the ¹GS and the emissive ³ES, which often has some triplet metal-to-ligand charge-transfer (³MLCT) character [7,8,9,10,11,12]. Thus, it is helpful to utilize electron-withdrawing ligands to stabilize the HOMO. However, it is sometimes the case that increasing the energy of the emissive ³ES causes the metal-centered ³d-d state to become thermally accessible through internal conversion (IC), which diminishes the efficiency of emission because the ³d-d state undergoes rapid nonradiative (nr) decay (Figure 1) [4,13,14,15]. Therefore, it is beneficial to utilize ligands that are both high-field and electron-withdrawing to maintain the energy gap between the emissive ³ES and the ³d-d state, while increasing the energy of emission.

Our group reported that the trifluoropropynyl ligand (:C≡C–CF₃) is more electron-withdrawing and higher field than the pentafluorophenylethynyl ligand (:C≡C–C₆F₅) [16], a ligand commonly used for its electron-withdrawing behavior. Recently, we showed that for Pt^II complexes with bidentate bipyridine ligands, utilizing the trifluoropropynyl ligand instead of the pentafluorophenylethynyl ligand blueshifted the emission and raised the ³MLCT energy beyond that of the ³π-π*, leading to a change in emissive ³ES character [7]. Therefore, it may be advantageous to synthesize Pt^II luminophores with the trifluoropropynyl ligand to achieve efficient deep-blue emission.

The cyano-ligand has also been used to blueshift emission and maintain emission efficiency. The Shinozaki group showed that replacement of the Cl ligand with the CN ligand contributed to a blueshift in emission for Pt(Fmdpb)CN (FmdpbH=4-fluoro-1,3-di(4-methyl-2-pyridyl)benzene; Figure 2) and reported that additionally Pt(Fmdpb)CN emits more efficiently when compared to Pt(Fmdpb)Cl [17]. Because the :C≡C–CF₃ ligand has been shown to have similar electronic properties to the CN ligand [18], replacing the chloro ligand of N^C^N-chelated Pt^II complexes with the trifluoropropynyl ligand may indeed lead to more efficient and bluer emission. Thus, PtLC₂CF₃ and PtL^FC₂CF₃ (Figure 2; L = 1,3-di(2-pyridyl)benzene; L^F = 4,6-difluoro-1,3-di(2-pyridyl)benzene) were synthesized. Pt^II complexes with L and L^F ligands were chosen for this study as PtLCl and PtL^FCl have been investigated extensively by the Shinozaki group [17,19] and the Williams group [11,13,20,21] and emit blue-green (491 nm) and turquoise (472 nm) light, respectively.

Herein, we present the synthesis and characterization of PtLC₂CF₃ and PtL^FC₂CF₃. Solution-phase absorption and emission spectra as well as the related photophysical constants are reported. Excited-state aggregation into excimers (a common phenomenon among square planar Pt^II complexes) [7,20,22,23] is observed, and the excimer formation is explored. Additionally, the photophysical effects of rigidification in poly (methyl methacrylate) (PMMA) films are discussed, and the practicality of utilizing PtLC₂CF₃ and PtL^FC₂CF₃ in OLED applications is addressed.

In addition, in the course of these investigations, we found that for PtLCl, PtLC₂CF₃, PtL^FCl, and PtL^FC₂CF₃ the well-separated excimer emission and monomer emission each exhibit different sensitivities to oxygen concentration in the form of oxygen quenching. Thus, it is possible to establish a ratiometric relationship between the oxygen concentration and the ratio of excimer/monomer emission intensity. Differential sensitivities of monomer and excimer to oxygen for Pt(dtbpy)(CN)₂ (dtbpy = 4,4′-di-tert-butyl-2,2′-bipyridine) and Pt(bathophen)(CN)₂ (bathophen = 4,7-diphenyl-l,10-phenanthroline) have previously been noted and these complexes suggested as oxygen-sensing materials [24]. Thus, the ratiometric response of the excimer/monomer emission ratio of PtLCl, PtLC₂CF₃, PtL^FCl, and PtL^FC₂CF₃ was probed and is discussed herein.

2. Materials and Methods

2.1. General Methods

UV–Vis: Absorption spectra were collected using a Cary-50 UV–vis spectrophotometer (Agilent Technologies, Santa Clara, CA, USA).

Emission Spectra: Emission spectra were collected using a Horiba Scientific (Piscataway, NJ, USA) Fluorolog-3 spectrofluorometer equipped with a FL-1013 liquid nitrogen dewar assembly for 77 K measurements and a J-1933 solid-sample holder for film measurements. All emission spectra were corrected for the response factor of the R928 photomultiplier tube using the factory correction files. Emission spectra involving excimer emission were additionally corrected with blank subtraction.

Excitation spectra: Excitation spectra were collected using a Horiba Scientific Fluorolog-3 sprectrofluorometer.

Emission Lifetime: Emission lifetimes were measured using a Photon Technology International (PTI, Lawrenceville, NJ, USA) GL-3300 pulsed nitrogen laser fed into a PTI GL-302 dye laser as the excitation source. The resulting data set was collected on an OLIS (Athens, GA, USA) SM-45 EM fluorescence lifetime measurement system using a Hamamatsu (Bridgewater, NJ, USA) R928 photomultiplier tube fed through a variable feed-through terminator into a LeCroy (Chestnut Ridge, NY, USA) WaveJet 352A oscilloscope and analyzed using OLIS Spectral Works.

NMR spectra: ¹H NMR spectra were obtained using a JEOL (Peabody, MA, USA) JNM-ECZR (500 MHz) spectrometer.

Elemental analysis: Elemental analyses were performed by Midwest Microlabs (Indianapolis, IN, USA).

Quantum Yields of Photoluminescence (Φ_PL): Relative solution-state photoluminescence quantum yields (Φ_PL) in CH₂Cl₂ were determined using solutions that were absorbance-matched with solutions of quinine sulfate (Φ_PL = 0.546) in 0.1 M H₂SO₄.

Thin Films: PMMA films for spectroscopy were prepared by dissolving PMMA (123 mg, M_w ~ 120,000) and the appropriate complex in CH₂Cl₂ (1 mL) solution. The mass of the complex was chosen to achieve the desired complex/PMMA mass percent. The solution was pipetted onto a quartz slide that was set at a 45° angle in a vial. The vials were covered with a Kimwipe and the film allowed to cure overnight. These film-coated quartz slides were mounted in the solid-sample holder for emission measurements with the film facing the excitation source at a 112.5° angle.

Monomer/Excimer Emission Ratio Under Different [O₂] Concentrations: A solution of the desired platinum complex (≈1 × 10⁻⁴ M) was prepared in CH₂Cl₂ in a cuvette with a septum top without exclusion of air, and the emission spectrum was recorded. The sample was then sparged with O₂ for 10 min and the emission spectrum was recorded. The sample was then sparged with Ar for 10 min and the emission spectrum was recorded. The sparge gases were presaturated by bubbling through CH₂Cl₂ in order to minimize solvent evaporation.

2.2. Synthesis

PtLCl [25] and PtL^FCl [26] were prepared according to the literature procedures, with the exception that PtL^FCl was sufficiently pure, without the need for recrystallization. 1,3-di(2-pyridyl)benzene for the synthesis of PtLCl was purchased from AmBeed (Arlington Heights, IL, USA). 4,6-difluoro-1,3-di(2-pyridyl)benzene for the synthesis of PtL^FCl was prepared by a modification of the literature procedure [25]. In particular, the purification following the synthesis of the L^F ligand using a Stille coupling did not require column chromatography; rather, the initial solid was simply washed with a 3:1 hexanes/ether mixture, resulting in product that was sufficiently pure for coordination of the Pt. 3,3,3-trifluoro-1-propyne was obtained from Synquest Laboratories (Alachua, FL, USA). Sodium methoxide (25 wt % in methanol) was obtained from Sigma Aldrich (St. Louis, MO, USA). Poly(methyl methacrylate) was obtained from Sigma-Aldrich. All reactions were performed in oven-dried glassware under an argon atmosphere using an argon manifold.

General Synthetic Method: An oven-dried two-necked round-bottom flask under a positive pressure of Ar was charged with sodium methoxide (25% in CH₃OH) and degassed CH₃OH/CH₂Cl₂ (3:1 v:v). The valve to the Ar manifold was then closed and a positive pressure of Ar was maintained using an Ar-filled balloon attached to the flask through a short cannula. While the solution was stirring, 3,3,3-trifluoropropyne gas was slowly injected (2 min) into the solution through a septum and allowed to react for 2 h. A separate single-necked round-bottom flask under Ar was charged with the appropriate Pt precursor in degassed CH₃OH/CH₂Cl₂ (3:1 v:v). This Pt precursor solution/suspension was injected into the trifluoropropyne solution, and the resulting mixture was stirred for 24 h. The solvent was removed under reduced pressure and the remaining solid was purified using column chromatography (silica gel, 2.5 cm × 15 cm), eluting with 4:1 dichloromethane/hexanes. Due to the emissive nature of the product, the column was monitored with long-wavelength UV irradiation. The first band was collected, and the solvent was removed under reduced pressure. The resulting solid was suspended in diethylether and the product was collected using vacuum filtration.

PtLC₂CF₃: The general procedure was followed using PtLCl (0.1027 g, 0.222 mmol, 1 eq) in 5 mL of solvent, and 3,3,3-trifluoropropyne(g) (15 mL, 0.61 mmol, 3.5 eq) and NaOMe (0.10 mL, 0.46 mmol, 2.7 eq) in 12 mL of solvent, yielding 0.0898 g (77%). Anal. Calc (found) for PtC₁₉H₁₁N₂F₃: H, 2.13 (2.09); C, 43.94 (43.64); N, 5.39 (5.39)%. ¹H NMR (500 MHz, DMSO-d₆). δ 8.96 (m, 2H), 8.22 (m, 2H), 8.11 (m, 2H), 7.79 (d 2H), 7.52 (m, 2H), 7.27 (t, 1H). UV–Vis (CH₂Cl₂): λ_max/nm (ε/Lmol⁻¹cm⁻¹) 337 (7930), 367 (7890), 381 (9820). Rf 4:1 dichloromethane/hexanes = 0.5.

PtL^FC₂CF₃: The general procedure was followed using PtL^FCl (0.1996 g, 0.40 mmol, 1 eq) in 20 mL of solvent, and 3,3,3-trifluoropropyne(g) (40 mL, 1.63 mmol, 4 eq) and NaOMe (0.40 mL, 1.85 mmol, 4.6 eq) in 24 mL of solvent, yielding 0.0824 g (37%). Anal. Calc (found) for PtC₁₉H₉N₂F₅: H, 1.63 (1.84); C, 41.09 (41.55); N, 5.04 (5.17)%. ¹H NMR (500 MHz, CDCl₃) δ 9.19 (m, 2H), 7.97 (m, 2H), 7.91 (d, 2H), 7.26 (m, 2H) (obscured), 6.64 (t, 1H). UV–vis (CH₂Cl₂) λmax/nm (ε/L mol⁻¹ cm⁻¹) 323 (9540), 337 (13,500), 359 (8230). Rf 4:1 dichloromethane/hexanes = 0.6.

2.3. Single-Crystal X-ray Diffraction

Single-crystal X-ray diffraction data were collected at 100 K using a Bruker (Madison, WI, USA) D8 Venture diffractometer. The data were collected using phi and omega scans (0.5° oscillations) with a Cu Kα (λ = 1.54178 Å) microfocus source and a Photon 2 detector. Data were processed (SAINT) and corrected for absorption (multi-scan, SADABS), using the Apex3 suite [27]. The structures were solved by intrinsic phasing (SHELXT) and subsequently refined by full-matrix least squares on F² (SHELXL) [28,29]. All non-hydrogen atoms were refined anisotropically. Hydrogen atoms attached to carbon atoms were refined in calculated positions using riding models where U_eq(H) = 1.2U_eq(C). Crystallographic data are summarized in

Table 1. Crystallographic data and refinement details for PtLC₂CF₃ and PtL^FC₂CF₃.

	PtLC2CF3	PtLFC2CF3
Empirical Formula	C19H11F3N2Pt	C19H9F5N2Pt
F. W. (g/mol)	519.39	555.37
Temperature (K)	100(2)	100(2)
Crystal System	Monoclinic	Triclinic
Space group	P21/n	P-1
a (Å)	11.7293(5)	8.0971(7)
b (Å)	7.8617(4)	8.9919(7)
c (Å)	18.1664(8)	11.5368(9)
α (°)	90	98.714(2)
β (°)	107.5583(14)	103.405(2)
γ (°)	90	100.939(2)
Volume (Å3)	1597.12(13)	785.26(11)
Z	4	2
D(calcd) (g/cm3)	2.160	2.349
Wavelength (Å)	1.54178	1.54178
μ, mm−1	16.769	17.306
F(000)	976	520
Crystal Size (mm)	0.05 × 0.23 × 0.26	0.04 × 0.11 × 0.12
θ range (°)	5.316 to 72.243	5.777 to 70.259
Reflections collected	32,402	26,668
Independent reflections	3158	2969
R(int)	0.0526	0.0370
No. of parameters	226	244
No. of restraints	0	0
R indices (I > 2σ(I))	R1 = 0.0283 a wR2 = 0.0855 b	R1 = 0.0168 a wR2 = 0.0425 b
R indices (all data)	R1 = 0.0288 a wR2 = 0.0863 b	R1 = 0.0168 a wR2 = 0.0425 b
Goodness-of-fit on F2	1.118	1.157
Largest diff. peak/hole (e/Å3)	2.864, −1.175	1.549, −0.685

^a R₁ = Σ||F_o| − [F_c||/Σ|F_o|. ^b wR₂ = {Σ[w(F_o² − F_c²)²]/Σ[wF_o²]²}^1/2.

. CCDC 2364563-2364564 contain the complete supplementary crystallographic data for this paper and can be obtained from the Cambridge Crystallographic Data Centre.

2.4. Computational Methods

Gaussian 16 [30] was used for all DFT and TDDFT calculations. Computational models involved the functional B3LYP [31] and the basis sets 6-31G(d) [32] and LANL2DZ [33]. GaussView version 6 [34] was used for all orbital imaging. GaussSum3 [35] was used for Mulliken population analysis.

3. Results and Discussion

3.1. Syntheses and Structural Characterization

The precursor complexes, PtLCl [25] and PtL^FCl [26], were prepared using slight modifications of the literature procedures. Ligand substitution of an alkynyl ligand for chloride on Pt^II complexes typically follows a CuI-catalyzed procedure derived from Sonogashira et al. [8,36,37]. For these complexes, however, we found this method to result in a product mixture that was not easily purified. Several reports of substituting alkynyl ligands for Cl on Pt(N^C^N) complexes have used an alternate procedure that did not use a catalyst, but instead used a strong base such as NaOMe or NaOH (Figure 3) [11,38,39]. This procedure gave the desired products that were easily purified by column chromatography. Only one equivalent of the alkynyl anion is necessary for the reaction but an excess (2–5 eq) was used in all attempted syntheses. The equivalents of the alkynyl anion that can form can be controlled by both the amount of NaOMe and trifluoropropyne added, and it is not necessary that all of the alkyne be deprotonated. The reaction did not appear to be significantly impacted by whether NaOMe or trifluoropropyne was in excess. Though the syntheses are performed under air-free conditions, the final products are air- and moisture-stable and can be handled as solids and in solution under ambient atmosphere.

Crystals of PtLC₂CF₃ and PtL^FC₂CF₃ were grown from the diffusion of diethyl ether into a CH₂Cl₂ solution of the complex, and slow evaporation of THF, respectively. For both complexes, the Pt is in a distorted square planar coordination environment (Figure 4, Table 1 and

Table 2. Selected interatomic distances (Å) and plane-to-plane distances (Å) for Pt complexes.

	Pt-CNCN	Pt-N1	Pt-N2	Pt-X	Pt⋯Pt	Pt⋯Plane
PtLCl a	1.907(8)	2.033(6)	2.041(6)	2.417(2)	4.85	3.40
PtLC2CF3	1.942(4)	2.040(3)	2.034(3)	2.034(4)	5.18	3.27
PtLFCl b	1.910(6)	2.028(5)	2.043(5)	2.4124(15)	3.42	3.38
PtLFC2CF3	1.939(3)	2.036(2)	2.037(2)	2.042(3)	3.34	3.33

^a From [25], X-ray crystallography performed at 296 K. ^b From [26], X-ray crystallography performed at 298 K.

), with similar angles enforced by the L and L^F ligands. The C–Pt–C angles show some deviation between the complexes (179.52(17)° for PtLC₂CF₃ versus 174.86(11)° for PtL^FC₂CF₃). This deviation from linearity for PtL^FC₂CF₃ continues along the alkynyl bond angles, with Pt–C≡C and C≡C-CF₃ bond angles of 172.3(3)° and 170.6(3)°, respectively (compared to 177.1(4)° and 177.0(5)° in PtLC₂CF₃). In this way, the :C≡C–CF₃ ligand is noticeably bent in PtL^FC₂CF₃. The C≡C (C1-C2) bond lengths for both PtLC₂CF₃ (1.221 Å) and PtL^FC₂CF₃ (1.208 Å) are in the normal range for alkynes, giving no indication that the distortion from linearity for PtL^FC₂CF₃ is due to a reduction in the C≡C bond order. Thus, weak packing forces are likely the cause for these deviations from linearity. Indeed, the CF₃ groups of PtL^FC₂CF₃ participate in three C–H···F short contacts with favorable distances and geometries enabled by the bent nature of the :C≡C–CF₃ ligand (Supplementary Materials, Figure S4). In PtLC₂CF₃, only one C–H···F contact is made (Supplementary Materials, Figure S3), as the packing of neighboring molecules restricts favorable C–H···F angles so that a bent :C≡C–CF₃ ligand is not imposed. Deviations from nonlinearity for RHgC₂CF₃ complexes have also been observed and attributed to intermolecular interactions within the crystal [40].

In PtLC₂CF₃, the complexes have a classic herringbone packing arrangement. Stacks of offset complexes have plane–plane separations of 3.27 Å, suggesting moderate pi⋯pi interactions (Figure 5). Neighboring stacks interact through C–H⋯pi interactions (Supplementary Materials, Figure S3). Because of the offset within a given stack, the Pt atoms do not line up right over one another, so there is no short Pt⋯Pt contact. For PtL^FC₂CF₃, on the other hand, neighboring complexes stack along the a-axis (Figure 5). This occurs through alternating Pt⋯Pt and offset pi stacking interactions with plane-to-plane separations of 3.33 Å, with neighboring stacks connected through C–H···F interactions (Supplementary Materials, Figure S4). The resulting Pt⋯Pt distance (3.3441(3) Å) is less than the sum of van der Waals radii (3.50 Å) [41], indicating a metallophilic interaction between Pt centers (Figure 6) [42,43]. This presence of solid-state Pt⋯Pt interactions for PtL^FC₂CF₃ likely contributes to the color of the crystals being red/orange, whereas crystals of the corresponding PtLC₂CF₃ complex are yellow. This is due to lower-energy absorptions to the singlet metal–metal-to-ligand charge transfer (¹MMLCT) state that are not present in the absence of such Pt⋯Pt interactions [44].

A structural comparison with the corresponding chloro analogues (Table 2) demonstrates that replacement of Cl with C₂CF₃ increases the Pt–C_NCN bond distance (Pt-C10 for PtLC₂CF₃ and Pt-C14 for PtL^FC₂CF₃) by approximately 0.03 Å for both analogues, likely an indication of the stronger trans influence of the :C≡C–CF₃ ligand. Furthermore, the presence of Pt⋯Pt contacts less than the sum of van der Waals radii appears to be more dependent on the identity of the N^C^N ligand than the X ligand (X = Cl, C₂CF₃), as such contacts are only present for the L^F complexes.

3.2. Photophysical Characterization

3.2.1. Absorption Spectra

The absorption spectra of PtLC₂CF₃ and PtL^FC₂CF₃ are similar in shape to their chloro analogues, with a slight blueshift of the onset of the lowest-energy absorptions (Figure 7, Supplementary Materials, Table S3). These lowest-energy absorptions have been ascribed to excitation to an excited state of mixed orbital parentage, namely one with ¹MLCT/¹IL(π-π*) character for many Pt^II complexes [8,11,22,37,45] including the chloro analogues [9,21]. This latter component can involve both localized ¹π-π* and ¹ILCT transfer from the phenyl to the pyridine. By analogy, the peaks in this range for PtLC₂CF₃ and PtL^FC₂CF₃ have been assigned as representing excitation into a mixed ¹MLCT/¹ILCT (π-π*) excited state. In addition, both complexes show sharp weak absorptions on the red tail of these CT absorptions (Supplementary Materials, Figure S7). These have been previously observed for the chloro analogues and have been assigned to spin-forbidden excitation into the ³ES [20,21]. Lastly, plots of absorbance vs. concentration (measured up to ~1 × 10⁻⁴ M) are linear (Supplementary Materials, Figure S8), indicating a lack of ground-state aggregation.

3.2.2. Monomer Emission Spectra

The emission spectra of PtLC₂CF₃ and PtL^FC₂CF₃ in a room-temperature (RT) CH₂Cl₂ solution (≈2.0 × 10⁻⁶ M) have well-defined peaks between 450 and 600 nm, with a clear vibronic progression (Figure 8, Supplementary Materials, Table S4). For each complex, the excitation spectrum agrees well with the absorbance spectrum (Supplementary Materials, Figure S9) indicating that the recorded emission is not due to an impurity. The vibronic progressions in the emission spectra are between 1230 and 1300 cm⁻¹. Such progressions have been attributed to emission from a ³ILCT state in related complexes [37,45,46,47]. The replacement of the chloro ligand with the trifluoropropynyl ligand resulted in less than a 5 nm blueshift for both complexes. Thus, for this series of complexes the trifluoropropynyl ligand is not a particularly impactful candidate for significantly blueshifting emission.

For both complexes, the lowest-energy, spin-forbidden absorption peak (Supplementary Materials, Figure S7) is <10 nm lower in wavelength than the corresponding emission peak. This very small Stokes shift is indicative of an excited state that is relatively undistorted from the ground-state geometry [14]. In addition, there is less than a 5 nm blueshift of the emission in a rigid medium (77 K glass, 1:4 CH₂Cl₂:2-MeTHF) compared to the spectra in solution (Supplementary Materials, Figure S10). This very small rigidochromic shift is also consistent with a relatively undistorted excited state [7].

3.2.3. Computational Investigation

To further investigate the nature of the emissive ESs, time-dependent density functional theory (TDDFT) experiments were performed. Vertical S₀ → T₁ absorption energies are considered good models for 0-0 emission energies, especially in a rigid medium where structural relaxation has little effect on the emission energy [48,49]. A recent review suggested that density functional theory (DFT) geometry optimization using B3LYP/LANL2DZ followed by TDDFT calculations of the transition from the ¹GS to the lowest energy ³ES gave the best predictions for emission energies [50]. Yersin’s group also demonstrated that the B3LYP function handled calculations of MLCT and LC excited states well [48]. Finally, our recent benchmarking on a similar set of trifluoropropynyl complexes indicated that for this class of Pt^II complexes, TDDFT calculations using B3LYP/6-31G*/LANL2DZ gave accurate predictions of emission energies and orbital character of the ES [7]. Thus, PtLC₂CF₃, PtLCl, PtL^FC₂CF₃, and PtL^FCl were modeled using B3LYP/6-31G*/LANL2DZ//B3LYP/LANL2DZ (TDDFT//Geometry-Optimization). All calculations employed a Tomasi polarizable continuum model assigned the dielectric constant for CH₂Cl₂ [51]. There is excellent agreement between the computationally predicted lowest-energy triplet energy and the experimentally determined 0-0 band (77 K glass), suggesting that the computational model utilized is a good fit for simulating these Pt^II systems (

Table 3. Emission wavelength, TDDFT-calculated S₀ → T₁ wavelength, and population analysis ^a.

Complex	Experimental λPL b (77K)	Predicted λPL c	Pt	Phenyl	Pyridines	X d	Orbitals e
PtLCl	486 nm	484 nm	32 → 6 (−26)	38 → 26 (−12)	21 → 68 (47)	9 → 0 (−9)	78 → 79 (49%) 77 → 80 (33%)
PtLC2CF3	483 nm	483 nm	25 → 6 (−19)	43 → 25 (−18)	23 → 67 (44)	9 → 2 (−7)	92 → 93 (47%) 91 → 94 (44%)
PtLFCl	467 nm	467 nm	30 → 6 (−24)	36 → 27 (−9)	23 → 67 (44)	11 → 0 (−11)	86 → 87 (44%) 85 → 87 (38%)
PtLFC2CF3	466 nm	465 nm	23 → 6 (−17)	41 → 27 (−14)	24 → 65 (41)	11 → 2 (−9)	99 → 102 (43%) 100 → 101 (41%)

^a TDDFT and Mulliken population analysis used the B3LYP/6-31G*/LANL2DZ computational model. All structures were optimized as singlets with the B3LYP/LANL2DZ computational model. Columns 4–7 indicate Mulliken population changes for each molecular region. ^b Wavelength of 0-0 emission band. ^c Computationally predicted energy of the lowest-energy triplet state. ^d X is the non-chelating ligand. ^e Frontier molecular orbitals involved in the transition as depicted in Supplementary Materials, Figure S11.

).

The population analysis (Table 3) and MO images (Supplementary Materials, Figure S11) are consistent with the character of the emissive excited states assigned based on the spectroscopy, namely that they are of mixed ³ILCT(π-π*)/³MLCT character. For example, for PtLC₂CF₃, the calculations show that for the lowest-energy triplet there is a decrease in electron density at both the platinum and the phenyl rings (19% and 18%, respectively), along with an increase in electron density at the pyridine rings (44%). The same general trend is observed for PtLCl, but importantly, the contribution from the platinum to the pyridine rings (26%) played a more dominant role than the transition from the phenyl to the pyridine rings (12%). This is consistent with the :C≡C–CF₃ ligand acting as more electron-withdrawing than the chloro ligand. The same general observations are true of the corresponding L^F complexes. Note that these CT components account for only half of the electronic transition, indicating a significant amount of localized π-π* character.

3.2.4. Excimer Formation and Excited-State Kinetics

As the [Pt] is increased to ≥1 × 10⁻⁵ M in the RT CH₂Cl₂ solution, broad, unstructured emission peaks appear for PtLC₂CF₃ (668 nm) and PtL^FC₂CF₃ (659 nm) that are not significant in more dilute solutions (Figure 9). This same behavior has been observed for the corresponding chloro analogues, PtLCl and PtL^FCl, and is attributed to excimer emission from a ³MMLCT excited state [17,21]. Recall that ground-state aggregation was ruled out by the Beer–Lambert Law plot. Thus, this emission must be the result of the formation of an excimer (excited-state dimer) from a ground-state monomer and an excited-state monomer. In addition, there is a more pronounced blueshift (15–20 nm) of the excimer emission of the :C≡C–CF₃ complexes relative to their corresponding Cl complexes (Figure 9) than for monomer emission. Similar behavior was observed for PtLCl and PtLI, where the more electronegative Cl resulted in a 15 nm blueshift of the excimer emission relative to the iodo complex, whereas there was only a 2 nm blueshift for the monomer emission [19]. This supports the hypothesis that :C≡C–CF₃ is more electron-withdrawing than Cl. This is confirmed by the Mulliken charge distribution obtained from DFT calculations. Chiefly, for PtLX the charge on the Pt increases from +0.457 to +0.890 upon replacement of Cl with :C≡C–CF₃. Likewise, for PtL^FCl the charge on the Pt increases from +0.505 to +0.945. Furthermore, the greater blueshift of the excimer relative to the monomer emission indicates that the electron density on the X ligand has a greater impact on the HOMO of the excimer (σ*d_z2) than on the HOMO of the monomer (d_z2).

As the [Pt] increases, the intensity of monomer emission decreases while the intensity of excimer emission increases (Supplementary Materials, Figure S12). This is accompanied by a decrease in the quantum yield (Φ_PL) and lifetime of monomer emission with increasing platinum concentration. A kinetic scheme proposed by Shinozaki and coworkers for PtLCl and analogous complexes [17,22] is adopted herein (Scheme 1). For some complexes, at even higher concentrations, trimer formation complicates the overall kinetic scheme [17]. Here, k_sq represents the rate constant for self-quenching (i.e., the excimer formation rate constant), k_−sq is the rate constant for the reverse process, k_M is the intrinsic rate constant of the deactivation of the excited-state monomer in the absence of self-quenching, and k_E is the intrinsic rate constant for the deactivation of the excimer in the absence of contributions from k_−sq and trimer formation.

Given that both excimer and monomer emission can be observed and are well separated, it is possible to measure the total deactivation rate constant, k_d, at both the monomer emission wavelength (k_dM) and the excimer emission wavelength (k_dE). At very low concentrations, the equilibrium lies toward M*, and a Stern–Volmer plot based on k_dM yields a slope of k_sq and an intercept of k_M (Equation (1) and Supplementary Materials, Figure S13). The values for the monomer lifetimes (τ_M = 1/k_M) are determined from the intercept of these plots (

Table 4. Photophysical constants of Pt(N^C^N) complexes in CH₂Cl₂.

Compound	Emission RT		QYlum		τM
	Monomer λmax	Excimer λmax	Ar	Air	Ar	Air
PtLCl a	491	691	0.60	0.04	7.2	0.5
PtLC2CF3	487	668	0.53 b	0.061 b	7.1 c	0.8 c
PtLFCl a	472	677	0.85	0.09	7.9	0.7
PtLFC2CF3	471	659	0.22 b	0.062 b	4.2 c	0.9 c

^a Reference 20. ^b Calculated relative quantum yield value using quinine hemisulfate monohydrate in 0.1 M H₂SO₄ as a standard. Absorbances were matched at 360 nm for [Pt] ≈ 2.5 × 10⁻⁶ M. ^c The k_dM of Pt solns was plotted against [Pt] over a range of conc. from ≈5.0 × 10⁻⁶ M to ≈2.5 × 10⁻⁵ M. τ_M is the inverse of k_M, which is the y-intercept.

). In addition, the emission quantum yields for the monomer

$$k_{\mathrm{dM}}=k_{\mathrm{sq}}\left[\mathrm{Pt}\right]+k_{\mathrm{M}}$$

(1)

were measured at concentrations where self-quenching is negligible. Lastly, the emission is quenched in the presence of oxygen. The relevant lifetimes and quantum yields for the C₂CF₃ complexes in deaerated and air-saturated solutions reported herein are compared to those of the analogous Cl complexes reported in the literature. Perhaps the most unexpected result is that the photoluminescence quantum yield for PtL^FC₂CF₃ in deaerated solution is around a quarter of that for the corresponding Cl complex, whereas the corresponding lifetimes differ by only a factor of two. Measurements on different samples and repeated purifications of the complex did not result in a change in Φ_PL; thus, we believe this lower quantum yield is not due to a quenching impurity, but rather an increase in the non-radiative rate constant along with a decrease in the radiative rate constant for this complex. We do not currently have an explanation for this difference.

WOLED applications depend on controlling the excimer-to-monomer emission intensity ratio to achieve white emission. Thus, we investigated whether the trifluoropropynyl ligand impacts this ratio relative to the chloro ligand. A strict comparison requires that corresponding emission spectra must be collected under conditions where both the initial ground-state and excited-state concentrations of the Pt monomer ([M] and [M*], respectively; Scheme 1) are identical. This requires absorbance-matching solutions of the complexes at wavelengths where the two complexes have identical molar absorptivities, followed by excitation at that wavelength. The spectra in Figure 9 are indeed performed under these conditions, and then normalized at the maximum emission intensity of the monomer in order to show the relative monomer/excimer emission. Clearly, the trifluoropropynyl ligand increases the relative intensity of the excimer emission and may indicate that the trifluoropropynyl ligand increases the equilibrium constant for excimer formation.

3.3. Emission in PMMA Film

Details relevant to excited-state reorganization and possible use in OLED devices can be gained from the investigation of emission in films. Thus, the emission in PMMA films of PtLC₂CF₃ and PtL^FC₂CF₃ were investigated and compared to emission in solution (Supplementary Materials, Figure S14). For PtLC₂CF₃, the monomer emission in the RT PMMA film is nearly identical to that in the RT CH₂Cl₂ solution (λ_max = 487 and 488 nm, respectively). Likewise, for PtL^FC₂CF₃, λ_max = 470 nm in both media. The fine structure is also unaffected by this change in medium. This lack of rigidochromic behavior is consistent with the conclusion that the excited-state geometry of the monomer is close to the ground-state geometry [7,52,53,54,55,56], and has been reported for PtL^FCl [57,58] and other similar Pt^II complexes [8,36,37].

The excimer emission peaks, on the other hand, show significant rigidochromic behavior. For example, PtLC₂CF₃ (650 nm) and PtL^FC₂CF₃ (641 nm) in PMMA film are both blueshifted by 18 nm relative to those in CH₂Cl₂ solution (668 nm and 659 nm, respectively; Supplementary Materials, Figure S14). This rigidochromic shift of the excimer emission suggests that the geometry of the excimer may be significantly distorted from the geometry of the corresponding ground-state dimer. A similar rigidochromic shift in excimer emission has been reported for PtL^FCl [57,58] and a spiro-flourine-chelated Pt^II complex [59], which both emit from ³MMLCT excited states. Additionally, the Castellano group reported solution-phase calculations of a 0.19 Å shortening of the Pt---Pt bond of the excimer when compared to the corresponding ground-state dimer [44]. Such distortion is likely inhibited in film. Therefore, the observed rigidochromic blueshift is consistent with the classification of the emissive ³ES of the excimer as a ³MMLCT state and this blueshift may be explained by the inhibition of the Pt---Pt bond contraction in film.

A single-component WOLED device using a film of 8% (w/w) PtL^FCl with notably white CIE chromaticity coordinates (0.33, 0.35) has been reported [58]. Thus, attempts were made to prepare films of PtL^FC₂CF₃ at 8% Pt w/w. However, the C₂CF₃ complexes proved to be insoluble at the [Pt] concentrations required for solution processing of 8% Pt w/w film. Therefore, films were prepared at lower [Pt] concentrations. A PMMA film of PtL^FC₂CF₃ (0.7% Pt w/w) showing a good balance of both monomer and excimer emission resulting in a warm orange-white emission with CIE chromaticity coordinates of (0.39, 0.38) was prepared and characterized (Figure 10). At that same concentration of [Pt] (0.7% Pt w/w), PtL^FCl exhibits very little excimer emission. Further, in films at greater concentrations that typify OLED devices, the perceived color of emission from PtLC₂CF₃ and PtL^FC₂CF₃ grows redder at such a magnitude as to render the C₂CF₃ complexes ineffective for WOLED devices.

3.4. Oxygen Quenching

As [O₂] increases in the RT CH₂Cl₂ solution, the intensity of the emission spectra of PtLCl, PtLC₂CF₃, PtL^FCl, and PtL^FC₂CF₃ decreases due to quenching by energy transfer. This has previously been reported for monomer emission for PtLCl and PtL^FCl as well as other similar Pt^II complexes [20,21,60,61]. It is noteworthy that the excimer emission is quenched by oxygen to a larger extent than the monomer in the RT CH₂Cl₂ solution (Figure 11 and Figure 12). This is particularly evident in spectra where the monomer emission has been normalized (Figure 12).

In order to probe the suitability of this class of Pt^II complexes as ratiometric sensors of [O₂], the ratio of the peak intensity of the excimer to that of the monomer, I_E/I_M, was plotted against the concentration of oxygen in CH₂Cl₂ solutions (Figure 13). Three [O₂] values were used: 0 M (Ar-purged), 2.2 × 10⁻³ M (air-saturated), and 1.07 × 10⁻² M (O₂-purged) [62]. Similar behavior was observed for all complexes investigated herein. A greater change in I_E/I_M occurs between Ar-purged and air-saturated solutions than between air-saturated and oxygen-saturated solutions. This suggests that the emission ratios of these complexes are more sensitive to changes in [O₂] under hypoxic conditions.

The conditions under which such complexes demonstrate the greatest change in excimer/monomer emission ratio were further evaluated using PtL^FCl and PtL^FC₂CF₃ test cases. To achieve this, the factor of increase in the excimer/monomer ratio (Equation (2)) between Ar- and air-saturated solutions was used as a gauge of sensitivity and was plotted against either excitation wavelength or [Pt] concentration. Whereas there was no observable impact of excitation wavelength across the range from 300 to 375 nm for either PtL^FCl or PtL^FC₂CF₃ (Supplementary Materials, Figure S15), there was a significant dependence on concentration, with the sensitivity going through a maximum at ≈30 µM for PtL^FC₂CF₃ (Figure 14) with a similar trend observed for PtL^FCl (Supplementary Materials, Figure S16). Such concentrations provide a good balance of excimer to monomer emission in the absence of oxygen. Significantly higher [Pt] concentrations are dominated by the excimer and significantly lower concentrations are dominated by the monomer, lowering the sensitivity.

$${\displaystyle \raisebox{1ex}{$\mathrm{E}$}\!\left/ \!\raisebox{-1ex}{$\mathrm{M}$}\right.}\mathrm{Factor}\mathrm{of}\mathrm{Increase}={\displaystyle \frac{{\left(\raisebox{1ex}{${\mathrm{I}}_{\mathrm{E}}$}\!\left/ \!\raisebox{-1ex}{${\mathrm{I}}_{\mathrm{M}}$}\right.\right)}_{\mathrm{Ar}}}{{\left(\raisebox{1ex}{${\mathrm{I}}_{\mathrm{E}}$}\!\left/ \!\raisebox{-1ex}{${\mathrm{I}}_{\mathrm{M}}$}\right.\right)}_{\mathrm{air}}}}.$$

(2)

It is noteworthy that a differential dependence of monomer vs. excimer emission intensity on oxygen concentration requires that equilibrium between the excited monomer and excimer not be rapidly maintained under these conditions. This is indeed the case, as demonstrated by the following analysis. If the excimer/monomer equilibrium were rapid under all conditions, then the apparent deactivation rate constants for monomer and excimer k_dM and k_dE would be equal under all conditions. Using PtL^FC₂CF₃ as an example, we have observed that deaerated CH₂Cl₂ solutions at ~5 × 10⁻⁵ M have a k_dM (3.95 × 10⁵ s⁻¹) that is 19% greater than the k_dE (3.31 × 10⁵ s⁻¹). However, whereas both rate constants increase with an increase in concentration (Table S3), they do so with different slopes such that at [Pt] ≈ 1 × 10⁻³ M, k_dE (2.54 × 10⁶ s⁻¹) is only half of the value for k_dM (4.95 × 10⁶ s⁻¹). This same trend is observed for PtLC₂CF₃ (Supplementary Materials, Table S5). This suggests that equilibrium is not rapid relative to excited-state deactivation, in agreement with reports that k_–sq < k_E for a range of similar Pt^II complexes [17,19,22,23], and given that the monomer/excimer equilibrium is not rapid in the absence of oxygen, the presence of oxygen would only exacerbate this. Chiefly, the rate constants for the establishment of equilibrium (k_sq, and k_–sq) are independent of [O₂], but the deactivation rate constants of both the monomer and excimer increase in the presence of oxygen.

4. Conclusions

The trifluoropropynyl ligand was investigated for its impact on the structure and emission characteristics of Pt complexes of two N^C^N ligands. Namely, the trifluoropropynyl complexes were compared with their chloro analogues. There was a slight increase in the Pt-C_NCN bond length for the trifluoropropynyl complex vs. the chloro complex; otherwise, the structures were largely similar. The chief reason for replacing the Cl ligand with the trifluoropropynyl ligand was to attempt to blueshift the monomer emission for potential OLED use as a blue emitter, or for WOLED use. The <4 nm blueshift of the monomer, and the propensity for PtL^FC₂CF₃ to form excimers at a lower concentration, suggests that these complexes are not useful in either of these applications. However, it was observed during these investigations that for both the Cl and :C≡C–CF₃ complexes, there was a differential sensitivity of the monomer and excimer emission to oxygen, raising the possibility that these and other related Pt complexes that show excimer emission might be useful in ratiometric oxygen sensors. Clearly, to be used in a device, these complexes would need to be suspended in an oxygen-permeable rigid substrate as has been carried out in sol–gel for Pt-porphyrin complexes [63,64,65,66], but these gels rely on an additional doped reference dye, and thus become less accurate over time as the Pt-porphyrin degrades. Avoiding this issue, ratiometric oxygen sensors that rely solely on singlet and triplet emission from [(dppe)Pt{S₂C₂(CH₂CH₂–N–2-pyridinium)}](BPh₄)] have been manufactured in cellulose acetate film [67,68]. Thus, further research into the oxygen sensitivity of excimer and monomer emission from N^C^N-chelated Pt^II complexes in cellulose–acetate films will be helpful in determining the suitability of these complexes to ratiometric oxygen-sensing applications.