Structure Based Modulation of Electron Dynamics in meso-(4-Pyridyl)-BODIPYs: A Computational and Synthetic Approach

Abstract

The effects of structural modification on the electronic structure and electron dynamics of cationic meso-(4-pyridyl)-BODIPYs were investigated. A library of 2,6-difunctionalized meso-(4-pyridyl)-BODIPYs bearing various electron-withdrawing substituents was designed, and DFT calculations were used to model the redox properties, while TDDFT was used to determine the effects of functionalization on the excited states. Structural modification was able to restructure the low-lying molecular orbitals to effectively inhibit d-PeT. A new meso-(4-pyridyl)-BODIPY bearing 2,6-dichloro groups was synthesized and shown to exhibit enhanced charge recombination fluorescence. The fluorescence enhancement was determined to be the result of functionalization modulating the kinetics of the excited state dynamics.

Graphical Abstract

INTRODUCTION

Over the past 30 years, there has been a considerable amount of research focused on exploring the chemistry, properties, versatility, and applications of BODIPY dyes that range from tunable laser dyes to probes for biological fluorescence imaging.^1–9 One aspect that has been underexplored is the development of low molecular weight water-soluble BODIPY derivatives. The first report of a small water-soluble BODIPY was by Wories et al. in 1985, which used sulfonate groups to solubilize the fluorophore,¹⁰ but by 2007 only a handful of water-soluble BODIPYs had been reported.⁶ Since then, the main strategy used for water solubilization of BODIPY derivatives consists of the introduction of water-solubilizing groups, particularly at the boron center, including polyethylene glycols,^9,11–18 hydroxyls and ethers,^19–21 amines,¹⁹ sulfona-mides,²¹ carboxylates,^{9,11,17,21–23} sulfonates,^9,23–30 phosphonates,^9–11 quaternary ammonium salts,^{8,9,24,28,31,32} carbohydrates,^33–35 and peptides.^36,37 Most of these groups increase the size of the dye, decrease its stability,^38,39 or utilize negative charges which tend to decrease the cell membrane permeability. However, cationic dyes such as rhodamine⁷ and (poly)cationic porphyrins⁴⁰ are able to electrostatically interact with the negative charges present on cell membranes, increasing their permeability. Therefore, the investigation of small, cationic, and electron-deficient BODIPYs such as meso pyridylBODIPYs is of interest, particularly the 1,3,5,7-tetramethyl-8-(4-pyridyl)-BODIPY which has C₂ symmetry and is readily available from the condensation of 2,4-dimethylpyrrole with 4-pyridylcarboxyaldehyde, followed by oxidation and boron complexation. Previous studies on meso- (4-pyridyl)-BODIPY using transient absorption spectroscopy found that following BODIPY excitation, the dye undergoes donor-photoinduced electron transfer (d-PeT) to a charge-transfer state, which is quenched by subsequent intersystem crossing (ISC),⁴¹ and the fluorescence quantum yield drops from 0.30 to <0.001.⁴² Hence, these pyridinium BODIPYs have found applications as nonfluorescent heavy-atom free singlet oxygen generators, and several groups have reported on their ISC enhancement, via bromination and iodination.^42–47

However, for bioimaging applications, the aforementioned electronic properties of meso-pyridinium BODIPYs are undesirable, and no studies have been reported so far on the manipulation of the electronic structure of BODIPYs to restore the fluorescent properties of this type of dye. Herein, we describe the synthetic and computational studies on restoring the fluorescence of meso-pyridinium BODIPYs by inhibiting either the d-PeT process or the ISC. In other BODIPY platforms, there are reports of d-PeT inhibition^48–50 and reduced photobleaching⁵¹ (i.e., increased photostability) achieved by installation of electron-withdrawing groups at the 2,6-positions. Using this strategy, a small library of BODIPYs bearing various electron-withdrawing groups at the 2,6-positions was designed (shown in Figure 1) and a combination of density functional theory (DFT) and time-dependent DFT (TDDFT) calculations were used to determine whether this reported methodology applies to the meso-pyridinium BODIPY dyes.

Figure 1.

Structures of the BODIPYs series A used in the calculations. The corresponding N-methylated BODIPY cations are denoted “cat”.

COMPUTATIONAL METHODS

All calculations were performed using the NWChem 6.5 software package.⁵² Molecular properties were calculated for gas phase structures at 0 K. The ground state structures were optimized using density functional theory with the hybrid B3LYP^53,54 functional, the 6−31+G** basis set, and confirmed by subsequent frequency calculations. The energies of the HOMO were then used to calculate the vertical ionization potentials (VIPs) as a means of evaluating the electron deficiency of the BODIPY, which influences the d-PeT to the pyridinium ring. Furthermore, these VIPs can be used to then calculate the experimental oxidation potentials as shown by Zhan and co-workers.⁵⁵ Electronic transitions were calculated for each optimized structure using TDDFT with the B3LYP and the range-separated CAM-B3LYP⁵⁶ functionals. Here, the range-separated functional is used to describe the long-range interactions necessary to capture transitions between the BODIPY and the pyridinium ring. To evaluate the effect of functionalization on the quenching of the charge-transfer state, structures were optimized on their first singlet excited state (S₁). This yields the singlet—triplet energy gap involved in the intersystem crossing.

EXPERIMENTAL METHODS

Materials.

All solvents and reagents were obtained from Sigma-Aldrich, Tokyo Chemical Industry, and Honeywell International Incorporated. The solvents were dried over 4 Å molecular sieves as needed. The reagents were used as received without purification.

Synthesis and Characterization.

Reactions were monitored using 0.2 mm silica gel plates (with indicator, polyester backed, 60 Å, precoated) and UV lamp. Liquid chromatography was performed on preparative TLC plates or via silica gel column chromatography (60 , 230–400 mesh) and all solvent systems were buffered with 0.1% triethylamine. NM R spectra were obtained on 400 (¹H) and 500 MHz (¹³C) spectrometers at room temperature. Chemical shifts (δ) are given in parts per million (ppm) in CDCl₃ (7.27 ppm for ¹H NMR, 77.0 ppm for ¹³C NMR) or C₂D₆SO (2.50 ppm for ¹H NMR, 39.5 ppm for ¹³CNM R); coupling constants (J) are given in hertz. High-resolution mass spectra were measured on an ESI-TOF mass spectrometer in positive mode. All absorption spectra were recorded on a Varian Cary 50 Bio and emission spectra were recorded on a PerkinElmer LS 55 Luminescence Spectrophotometer, at room temperature. Spectrophotometric grade solvents and quartz cuvettes (10 mm path length) were used. For the determination of the optical density (ε), solutions with absorbance at λ_max between 0.5 and 1 were used. For the determination of quantum yields, dilute solutions with absorbance between 0.03 and 0.05 at the particular excitation wavelength (462, 473, 488, 496 nm for 1A, IcatA, 3A, and 3catA, respectively) were used, and all measurements were taken within 8 h after solution preparation.^57,58 The external standards employed were rhodamine 6G in methanol and water and Ru(bpy)₃Cl₂ in water. BODIPY 1A was prepared following a reported procedure⁵⁹ with a modified purification using 1:4 EtOAc/dichloromethane and 4% EtOAc/dichloromethane instead of 1:1 EtOAc/hexanes due to the increased solubility. BODIPY 1catA was prepared in the same way as 3catA and its physical data matched literature reports.⁴¹

2,6-Dichloro-8-(4-pyridyl)-l,3,5,7-tetramethyl-BODIPY(3A).

To an oven-dried flask charged with BODIPY 1A (0.0118 g, 0.04 mmol) in dichloromethane (10.0 mL) was dropwise added trichloroisocyanuric acid (0.0076 g, 0.03 mmol, 2.62 equiv) in dichloromethane (4.0 mL), and the mixture was stirred at room temperature under a nitrogen atmosphere. After TLC showed consumption of the starting material (about 30 min), the solution was purified over silica by column chromatography eluting with 2% methanol/dichloromethane, yielding 10 mg, 69% yield (red solid). ¹H NMR (400 MHz, CDCl₃): δ = 8.85–8.84 (dd, J = 5.9 and 1.5 Hz, 2H), 7.31–7.30 (dd, J = 5.9 and 1.6 Hz, 2H), 2.61 (6H, s), 1.42 (6H, s) ppm. ¹³C NMR (500 MHz,CDCl₃) δ = 153.7, 150.9, 142.8, 138.2, 137.5, 128.6, 123.3, 123.1, 12.5, 12.2 ppm. HRMS (ESI): m/z calcd (%) for C₁₈H₁₉BF₂N₃Cl₂, 393.0891 [M + H]⁺; found, 393.0902. UV/vis:(CH₃CN) λ_max = 518, λ_em = 546 nm; (CH₂Cl₂): λ_max = 534, λ_em = 563 nm.

2,6-Dichloro-8-(N-methyl-4-pyridyl)-1,3,5,7-tetramethyl-BODIPY (3catA).

In a 10 mL round-bottom flask wrapped in foil, BODIPY 3A (0.0111 g, 0.028 mmol) was dissolved in anhydrous M eCN (2.0 mL) and methyl iodide (2.0 mL). After refluxing for 1 h, TLC showed consumption of starting material and the solvent was removed under reduced pressure. Yield: 15 mg, 100% (greenish-gray solid). ¹H NMR (400 MHz, DMSO-d₆): δ = 9.24–9.23 (d, J = 6.6 Hz, 2H), 8.44–8.42 (d, J = 6.7 Hz, 2H), 4.47 (3H, s), 2.55 (6H, s), 1.30 (6H, s) ppm. ¹³C NMR (500 MHz, DMSO-d₆): δ = 154.0, 149.6, 147.5, 138.0, 136.7, 128.2, 128.1, 123.0, 49.0, 13.2, 12.9 ppm. HRMS (ESI): m/z calcd (%) for C₁₉H₁₉BF₂N₃Cl₂, 407.1048 [M*]⁺; found, 407.1048. UV/vis: (CH₃CN) λ_max = 530, λ_em = 596 nm; (CH₂Cl₂) λ_max = 548, λ_em= 570, 645 nm.

RESULTS AND DISCUSSION

Quantum Chemical Calculations.

When orbital energy levels are evaluated, molecular orbitals are calculated for the whole system to illustrate which parts of the molecule are involved in electronic transitions (e.g., BODIPY core or pyridinium ring). This shows how functionalization affects transitions without the need to partition the molecule into fragments, which can introduce artifacts into the calculations.^49,60–62 It is worth noting that these calculations do not take into account spin-orbit coupling. Derivatives 4A and 5A both include large halogens known to induce the heavy-atom effect of quenching their fluorescence.⁴⁵ They are included in this study for the difference in the halogens’ electronegativities, which will alter the vertical ionization potentials and add additional data points for potential results. In discussing the LUMOs of the cationic derivatives, the LUMOs localized on the pyridinium and BODIPY cores are denoted as LUMO_Py and LUMO_BDP, respectively.

Calculated Oxidation Potentials.

The vertical ionization potentials used to calculate the oxidation potentials for the neutral library are shown in Table 1. Here, the VIPs increase as the BODIPYs become more electron deficient. In addition, the ΔVIPs are also shown, given by the difference between the VIP of a given molecule and that of 1A. The ΔVIPs show how the VIP changes with the functionalization. For the halogenated systems 3A–5A, the VIPs decreased for lower halogen electronegativity, as expected. The difluoro derivative (2A), however, gave the smallest VIP increase, which can be explained by looking at the molecular orbital compositions. Fluorine is quite small, having high electronegativity and relatively stable valence electrons (i.e., low energy). In 2A, the highest energy orbital involving the fluorine atoms is the HOMO–7, which is 3.281 eV lower in energy than the HOMO. The chlorines in 3A contribute to the HOMO–1 while the halogens in 4A and 5A contribute to the HOMO. Since 3A–5A halogens contribute to orbitals near the HOMO, they have a larger influence on respective VIPs.

Table 1.

Calculated Orbital Energies and Vertical Ionization Potentials (eV) for Neutral BODIPYs

BODIPY	εHOMO	VIP	ΔVIP(XA−1A)
1A	−5.814	8.144
2A	−6.034	8.433	0.289
3A	−6.072	8.483	0.339
4A	−6.051	8.456	0.312
5A	−6.043	8.444	0.300
6A	−6.170	8.611	0.467
7A	−6.487	9.027	0.883
8A	−6.276	8.751	0.607
9A	−6.319	8.807	0.663
10A	−6.240	8.704	0.560
11A	−6.566	9.131	0.987
12A	−6.742	9.363	1.218

The carbonyl derivatives are also included in Table 1. Unlike the halogenated systems, this group of carbonyl derivatives (6A–10A) does not show a direct correlation to the strength of the withdrawing group. Instead, the VIP depends more on how electrons contribute to higher energy orbitals. Since several electrons are involved in these functional groups, VIP values can be close in energy. As such, ΔVIPs are also calculated to better capture the energy differences (Table 1). For the carbonyl systems, the chloroformyl (7A) had the largest increase, followed by diamide (9A), dialdehyde (8A), diester (10A), and monoaldehyde (6A). If these had been ranked by decreasing electron-withdrawing strength, the expected order would be 10A, 9A, 8A, 7A, and 6A. The large increase in the chloroformyl (7A) is best explained by the contributions of the chlorine to the HOMO, which shifts the VIP by +0.416 eV relative to 6A monoaldehyde. The other outlier 10A (diester) has no contributions to the HOMO, HOMO –1, or HOMO –2 orbitals, each of which had significant contributions from other withdrawing groups. Since the ester groups do not participate in the high-energy orbitals of the valence, they have a small impact on the VIP. Continuing with the remaining derivatives (11A and 12A), the values increased consistent with electron-withdrawing strength. This trend is expected with the dicyano derivative (12A) giving the largest overall increase of 1.218 eV.

Time-Dependent DFT Calculations.

The first two excitations from the TDDFT results for the neutral library are summarized in the Supporting Information Table S1. These results show the HOMO → LUMO transition to be the first excitation with a strong oscillator strength (f) for all 12 dyes. The second excitation was HOMO–1 to LUMO, with a few exceptions.

The TDDFT analysis of the first cationic derivatives (Supporting Information Table S2) revealed that the B3LYP hybrid functional is insufficient to accurately describe their excited states. DFT is known to fail for charge transfer in molecules and range-separated functionals have been developed to address this.⁵⁶ The typical BODIPY is known to have a strong S₀ → S₁ transition and adding the pyridinium with a low-lying LUMO should introduce a single electronic transition lower in energy than the HOMO → LUMO transition of the BODIPY.^63–65 The failure can be seen in the degeneracy of both the LUMO_BDP and LUMO_Py+1 as well as the extra transitions from sub-HOMO orbitals that appeared between the HOMO → LUMO and HOMO → LUMO+1 transitions. In these cases, the first (lowest energy) transition was HOMO → LUMO where the LUMO is localized on the pyridinium ring. Although the second transition should have corresponded to the BODIPY core excitation (HOMO → LUMO+1), it was calculated to be much higher in energy as the fifth transition. Collectively, this indicated the B3LYP functional was not sufficient to describe the long-range interactions, prompting the use of a range-separated functional (CAM-B3LYP).⁶⁵ Using the TDDFT and the CAM-B3LYP functional, the vertical excitation energy (VEE), oscillator strength (f), and predominant transition character (e.g., mostly H → L) were determined for the first two transitions. These results are summarized in Table 2 and are consistent with experimental results for these fluorophores.^42,65,66

Table 2.

Calculated CAM-B3LYP S₁ and S₂ Vertical Excitation Energy (VEE, eV), Oscillator Strengths (f, arbitrary units), and Predominant Transition Character for Cationic BODIPYs

	S1			S2
BODIPY	VEE	f	trans	VEE	f	trans
1catA	1.9148	0.00024	H → L	2.8746	0.55350	H → L+1
2catA	2.0246	0.00037	H → L	2.7642	0.47242	H → L+1
3catA	2.0216	0.00045	H → L	2.7234	0.57025	H → L+1
4catA	2.0089	0.00034	H → L	2.7113	0.60764	H → L+1
5catA	1.9881	0.00016	H → L	2.6823	0.67528	H → L+1
6catA	2.0170	0.00032	H → L	2.8712	0.66292	H → L+1
7catA	2.0949	0.00043	H → L	2.8465	0.78719	H → L+1
8catA	2.0628	0.00045	H → L	2.7969	0.67844	H → L+1
9catA	2.2614	0.00068	H → L	2.8505	0.68369	H → L+1
10catA	2.1953	0.00065	H → L	2.8657	0.80877	H → L+1
11catA	2.2759	0.00015	H → L	2.9184	0.68027	H → L+1
12catA	2.3255	0.00008	H → L	2.8084	0.72367	H → L+1

Figure S1 of the Supporting Information shows the HOMO, LUMO, and LUMO+1 orbitals, localized on the BODIPY, pyridinium, and BODIPY, respectively, for all cationic 12 dyes. The first transition was found to be a strictly HOMO → LUMO_Py dark state, while the second transition was HOMO → LUMO_BDP with a strong oscillator strength. These results indicate that the BODIPYs’ strong absorption will decay to the charge-transfer state. As such, it can be concluded that the BODIPY core oxidation potential has a negligible impact on the d-PeT charge-transfer state for this BODIPY platform.

To interpret these results, consider both the molecular orbitals of the BODIPY core and the excited state potential energy surface. For the BODIPY systems, the HOMO is partially localized on the 2,6-positions; thus, functionalizing these positions with electron-withdrawing groups decreases the energy of the HOMO. With the BODIPY and pyridinium orbitals orthogonal to each other, changes to the BODIPY core orbitals due to functionalization do not significantly affect the orbitals on the pyridinium, which explains the minimal impact on the LUMO_Py. The excited state PES can be used to depict the BODIPY as a donor—acceptor system and interpret how the states should be ordered to prevent formation of a long-lived charge-transfer state. After excitation to the S₂ (LUMO_BDP), the system will nonradiatively decay to the LUMO (i.e., LUMO_Py) forming the charge-transfer state that quenches the fluorescence. From this, the excitation itself can be viewed as the donation, so the only way to inhibit the d-PeT is to either raise the energy of the undesired acceptor (LUMO_Py) or, alternatively, lower the energy of the desired acceptor (LUMO_BDP).

Since it is of interest to raise the energy of the LUMO_Py, methoxy groups were used to donate electron density into the ring. Although dimethylamino groups are better electron donors, their basic nature and large size would have sterically hindered the pyridine methylation during synthesis. To evaluate the best positions for the methoxy groups, the orbitals of the N-methylpyridinium cation (without the BODIPY) were calculated to determine where the LUMO is localized using DFT and the CAM-B3LYP functional. From this, the LUMOPy is located primarily on the 3,5-and 2,6-positions shown in Figure S2 in the Supporting Information, and the latter was chosen for synthetic practicality. Proof of concept was obtained from a small test case by modeling the N-methylpyridinium (N-MePy⁺, A), 2-methoxy-N-methylpyridinium (2-OMe-N-MePy⁺, B), and 2,6-dimethoxy-N-methylpyridinium (2,6-DiOMe-N-MePy⁺, C) cations (Supporting Information, Table S3). It was found that one methoxy group could increase the LUMO energy by 0.509 eV and a second could raise it further by an additional 0.488 eV with an overall increase of 1.003 eV.

In addition to raising the energy of the LUMO_Py to reorder the states, we can also consider lowering the energy of LUMO_BDP. The LUMO_BDP is largely localized on the meso-position and to a smaller extent on the 1,7-positions. Since the2,6-dicyano-BODIPY (12) had the largest oxidation potential increase, it has the greatest shift in the orbital energy. Thus, the1.7-dicyanoBODIPY (13A) was chosen to model the effects of lowering the LUMO_BDP. The new additions to the original library and their calculated VIPs are shown in Figure 2 and Table 3, respectively. The orbital energies of BODIPYs 1catA—C, 12catA—C, and 13catA–C (collectively referenced as Py′-BODIPYs) are shown in Table 4 while the TDDFT results and MO plots are summarized in the Supporting Information Figure S3 and Table S4, respectively. Moving the cyano groups from HOMO (12catA) to LUMO (13catA) structural positions decreased the VIP by 0.227 eV but was able to effectively reorder the LUMO_Py and LUMO_BDP, providing a 0.273 eV energy gap.

Figure 2.

Structures of the Py′-BODIPY series A–C used in quantum calculations.

Table 3.

Calculated Neutral Py′-BODIPY Orbital Energies and Vertical Ionization Potentials (eV)

BODIPY	εHOMO	VIP	ΔVIP(X−1A)
1A	−5.814	8.144
1B	−5.795	8.119	−0.025
1C	−5.775	8.094	−0.050
12A	−6.742	9.363	1.218
12B	−6.719	9.333	1.188
12C	−6.693	9.298	1.154
13A	−6.569	9.135	0.991
13B	−6.525	9.078	0.934
13C	−6.380	8.888	0.743

Table 4.

Calculated Cationic Py′-BODIPY Orbital Energies (eV)

	orbital energies (εX)			ΔE(Xcat–XcatA)
BODIPY	HOMO	LUMOPy	LUMOBDP	HOMO	LUMOPy	LUMOBDP
1catA	−9.472	−5.420	−4.545
1catB	−9.399	−4.983	−4.464	+0.072	+0.437	+0.081
1catC	−9.332	−4.561	−4.383	+0.140	+0.858	+0.162
12catA	−10.298	−5.889	−5.449
12catB	−10.139	−5.609	−5.149	+0.159	+0.280	+0.300
12catC	−10.227	−5.359	−5.439	+0.071	+0.530	+0.010
13catA	−10.162	−5.001	−5.274
13catB	−10.073	−4.737	−5.530	+0.089	+0.264	−0.256
13catC	−10.010	−4.340	−5.455	+0.152	+0.661	−0.181

The addition of methoxy groups to the pyridine unit slightly decreased the overall oxidation potential of the dyes by 0.025 eV per OMe. It should be noted that for series 12A–C, each OMe group decreased the VIP by ~0.03 eV. However, for series 13A–C, there was an additive effect where the first OMe caused a 0.057 eV decrease while the second caused an additional 0.191 eV decrease.

Alteration of the LUMO_Py alone was shown to be an ineffective strategy by the lack of change in the orbitals of 1catA—C. The combination of the most effective LUMO_Py increase and BODIPY VIP increase functionalizations (12catC) was able to raise the LUMO_Py energy above that of the LUMO_BDP, though the energy difference is rather small at 0.080 eV. As such, d-PeT is still possible due to vibrational excitation. Adjusting the LUMO_BDP (l3catA —C), however, effectively reordered the orbitals and separated them by energy differences of 0.273, 0.793, and 1.115 eV, respectively, which prevents d-PeT from occurring. The TDDFT calculations also indicate this with the first transition being HOMO → LUMO for 12catC and 13catA–C and corresponding to the typical BODIPY excitation.

Intersystem Crossing (ISC).

Since most of the BODIPYs studied will still undergo d-PeT, the effects of 2,6-difunctionalization on ISC are of interest. To this end, 1catA and 12catA were studied further. Both structures were optimized along their S₁ excited states using CAM-B3LYP after which the energies of the S₁ and the triplet excitation with the same transition character (T₂ for both dyes) were compared. The results, summarized in Table 5, show that the nature of the transitions, the orbital populations, the states involved, and the singlet—triplet energy gaps are identical for both dyes. These results indicate that the electronic possibility of ISC is unchanged in the 2,6-difunctionalized derivatives.

Table 5.

Calculated S₁ and Analogous Triplet State Excitations for S₁ Relaxed Structures of BODIPYs 1catA and 12catA

BODIPY	excitation	VEE (eV)	trans	population	Δ(S1−T2) (eV)
1catA	S1	1.4688	H → L	0.984261	0.0319
	T2	1.4369	H → L	0.980000
12catA	S1	1.8740	H → L	0.974631	0.0319
	T2	1.8421	H → L	0.973013

Synthesis and Spectroscopic Properties.

BODIPYs 1A and 1catA were synthesized using the previously reported methodology.⁵⁹ In summary, 2,4-dimethylpyrrole reacted with 4-pyridylcarboxaldehyde in the presence of TFA to afford the dipyrromethane, which after DDQ oxidation and BF₃ complexation produced BODIPY 1A in 46% yield. Chlorination of 1A using trichloroisocyanuric acid in dichloro-methane at room temperature, gave BODIPY 3A in 68% yield. The cationic BODIPYs 1catA and 3catA were obtained by methylation using methyl iodide in acetonitrile, as previously reported.⁴⁴

The absorption and emission spectra of BODIPYs 1A, 1Acat, 3A, and 3Acat were obtained at room temperature in acetonitrile while the spectra for 1catA and 3catA were also measured in water (Table 6 and Figure 3). The absorption spectra displayed the characteristic BODIPY profile with a sharp peak due to the S₀ → S₁ (π → π*) transition with a higher energy shoulder corresponding to the first vibrational mode. The absorption and emission spectra of 1catA and 3catA did not change upon switching the solvent from acetonitrile to water. In both cases, methylation caused about 10 nm bathochromic shift in the absorption while the emission shifted substantially to about 600 nm in both acetonitrile and water, and exhibited a broad, poorly resolved profile characteristic of luminescent charge recombination. These results are in agreement with previous observations of the spectroscopic behavior of 1A and 1catA in dichloromethane.^41,44 With the chlorines on the 2,6-positions being mildly electron-withdrawing groups, they had a small to moderate impact on the fluorescence quantum yield of both the neutral and ionic derivatives in acetonitrile, resulting in a 2-fold increase for 3A and about a 2.5-fold increase for 3catA. The quantum yield has been previously observed to increase upon 2,6-dichlorination of meso-arylBODIPYs.,^2,67,69 After accounting for the 2-fold increase in Φ_f from 1A to 3A, 3catA had an additional 26% increase from 1catA. The fluorescence quantum yields decreased upon switching from acetonitrile to water due to the increased solvemt polarity stabilizing the charge-transfer state which facilitated the subsequent deactivation by ISC; however, this resulted in an order of magnitude increse from 1catA to 3catA.

Table 6.

Photophysical Properties of Synthesized BODIPYs in CH₃CN and H₂O

		λmax(nm)
solvent	BODIPY	abs	em	Stokes Shift (nm)	Φf	ε (M−1 cm−1)
CH3CN	1A	501	515	14	0.31a	72100
	1catA	509	596	87	0.019b	26000
	3A	528	546	18	0.58a	50800
	3catA	538	600	62	0.048b	23400
H2O	1catA	509	600	91	0.004b	31600
	3catA	540	605	65	0.038b	4100

^aRelative quantum yields determined using rhodamine-6G (Φ_f = 0.86) in methanol as the standard, λ_ex = 473 nm.⁵⁸

^bRelative quantum yields determined using rhodamine-6G (Φ_f = 0.86) in Ru (bpy)₃Cl₂ in water as the standard (Φ_f = 0.028), λ_ex = 436 nm.⁶⁸

Figure 3.

Normalized (a) absorption and (b) emission spectra of 1A (green), 1catA (blue), 3A (red), and 3catA (yellow) in acetonitrile along with the normalized (c) absorption and (d) emission spectra of 1catA (blue) and 3catA (yellow) in acetonitrile (dashed) and water (solid) at room temperature.

With the d-PeT acceptor orbitals being unaffected by the functionalization, the fluorescence enhancement observed for 3catA in acetonitrile is attributed to an alteration of the excited state lifetime. This is the result of the increased oxidation potential increasing the rate of charge recombination by enhancing the electron—hole electrostatic interaction. Recalling the calculated VIPs, the dichlorination only resulted in a 0.339 eV increase while installation of the 2,6-dicyano groups resulted in a 1.218 eV increase. This suggests that 12A should display photophysical properties with a dramatic improvement over 3A. The synthesis of such a derivative is currently underway in our laboratories.

CONCLUSIONS

Several electronic structure-based molecular properties for a library of meso-(4-pyridyl) BODIPYs have been calculated and shown that the combination of increasing the BODIPY core oxidation potential as well as increasing the pyridinium reduction potential, can reorder the low-lying excited states which inhibits d-PeT; however, decreasing the BODIPY core reduction potential has a greater capacity for both reordering those states and sufficiently separating them. The states involved in the ISC were found to be unaffected by the 2,6-functionalization with various electron-withdrawing groups. Additionally, a new fluorescent meso-(4-pyridium)-BODIPY was synthesized, and a novel mechanism of fluorescence enhancement was demonstrated. In our work, structure based modulation of excited state electron dynamics was demonstrated and manipulation of fluorescence behavior was observed.