syn-Diarylphthalimidoporphyrins: Effects of Symmetry Breaking on Two-Photon Absorption and Linear Photophysical Properties

Abstract

Aromatically π-extended porphyrins possess exceptionally intense one-photon (1P) and sometimes two-photon (2P) absorption bands, presenting interest for construction of optical imaging probes and photodynamic agents. Here we investigated how breaking the molecular symmetry affects linear and 2PA properties of π-extended porphyrins. First, we developed the synthesis of porphyrins fused with two phthalimide fragments, termed syn-diarylphthalimidoporphyrins (DAPIP). Secondly, the photophysical properties of H₂, Zn, Pd and Pt DAPIP were measured and compared to those of fully-symmetric tetraarylphthalimidoporphyrins (TAPIP). The data were interpreted using DFT/TDDFT calculations and Sum-Over-States (SOS) formalism. Overall, the picture of 2PA in DAPIP was found to resemble that in centrosymmetric porphyrins, indicating that symmetry breaking, even as significant as by syn-phthalimido-fusion, induces a relatively small perturbation to the porphyrin electronic structure. Collectively, the compact size, versatile synthesis, high 1PA and 2PA cross-sections and bright luminescence makes DAPIP valuable chromophores for construction of imaging probes and other bio-applications.

Graphical Abstract

Introduction

Tetrapyrroles capable of efficient two-photon absorption (2PA) present interest for several areas of technology and medicine, including optical power limiting,^1,2 up-conversion lasing,³ three-dimensional optical data storage,⁴ biological imaging^5–10 and two-photon photodynamic therapy (PDT).^11–15 Over the past years porphyrin-based systems have been developed with 2PA cross-sections reaching into thousands of Göppert-Mayer (GM) units, including covalently linked porphyrin arrays, porphyrins with expanded electronic system as well as self-assembled non-covalent porphyrin oligomers.^16–23 However, large size, inherently poor solubility, lack of synthetic methods for targeted functionalization and, in many cases, the presence of intramolecular quenching pathways render many of these otherwise spectacular systems of low practical utility, especially in view of those applications where precise control over excited state dynamics is required. In contrast, compact monomeric porphyrins can be constructed with tailored functional groups and high solubility, while rational synthetic strategies exist for manipulation of their excited states in order to reach optimal performance in applications. In view of these advantages, compact porphyrinoids capable of efficient 2PA present an attractive synthetic target, and their development remains an important area of porphyrin chemistry.

Porphyrins in which the pyrrolic fragments are fused via β,β’-positions with external aromatic rings (e.g. benzene, naphthalene, anthracene) are known as aromatically π-extended porphyrins.^24–27 π-Extended porphyrins possess intense one-photon (1P) and in some cases strong two-photon (2P) bands in the near infrared spectral region and exhibit high photo- and chemical stability. In recent years synthetic chemistry of π-extended porphyrins has been significantly advanced,^28–38 making many of these once exotic systems available for practical applications. However, most attention has been devoted to the synthesis of porphyrins with symmetric π-extension, while asymmetric molecules, such as syn-dibenzo-, syn-dinaphtho- etc porphyrins^39–48 have been reported and studied on much fewer occasions. Asymmetric π-extension induces different perturbation to the macrocycle electronic structure compared to the symmetric fusion,^47,48 resulting in different photophysical properties, especially in relation to 2PA.

Pt and Pd complexes porphyrins are used for the construction of phosphorescent sensors for oxygen,^49,50 including probes with enhanced 2PA for two-photon phosphorescence lifetime microscopy (2PLM)^6,10,51,52 - an imaging technique that has gained attention in neuroscience,^53–56 stem cell biology^57,58 and some other areas.^59,60 Enhancement of 2PA in phosphorescent porphyrins while preserving their high emissivity presents an interesting problem. Metalloporphyrin molecule belongs to the D_4h symmetry class and thus possesses center of inversion symmetry. In closed-shell systems with centrosymmetric (gerade) ground state wavefunctions the parity selection rules governing electric dipole transitions are mutually exclusive, rendering 1P-allowed ungerade (u) B (Soret) and Q states inaccessible by 2PA. At the same time, 2P-active gerade (g) states in regular (non-extended) porphyrins lie at such high energies that their 2PA excitation is overshadowed by linear (1P) absorption to low-lying vibronic sublevels of the Q-state or, in some cases, directly to the triplet state.⁶¹ While conjugation of porphyrins into arrays and other assemblies has been shown to greatly increase 2PA,^{16,21,62–67} the triplet emissivity of the resulting constructs is usually diminished. Recently, we found that g-states in π-extended porphyrins can be stabilized upon addition of electron-withdrawing groups to the peripheral extension rings.^68,69 For example, in symmetric tetraarylphthalimidoporphyrins (TAPIP)^69,70 the g-states may occur even below the B state level, while the rigidity of the molecular skeleton causes meso-unsubstituted TAPIP to be highly emissive.⁶⁹ In view of these findings it was of interest to explore asymmetric counterparts of TAPIP, i.e. porphyrins in which phthalimido-fragments are annealed only on one side of the macrocycle, decreasing its symmetry to C_2v. Our previous studies^48,68 have shown that syn-π-extension results in porphyrins with absorption bands that are sharp, strong and only minimally overlapping with those of the symmetrically extended counterparts. From the practical viewpoint, these features are advantageous for the construction of multichromophoric sensor systems.

Here we present the synthesis, structural characterization and photophysical properties of asymmetric syn-diarylphthalimidoporphyrins (DAPIPs), including their 2PA spectra. Our results delineate the effects of asymmetric π-extension on the tetrapyrrolic electronic system, showing that spectroscopic properties of porphyrins are highly resistant to symmetry breaking. From the practical perspective DAPIP were found to be valuable chromophores for the construction of imaging probes and other biological applications.

Results and Discussion

Synthesis.

The asymmetric syn-diarylphthalimidioporphyrins (DAPIP) synthesized in this study are shown in Chart 1.

Chart 1.

syn-Diarylphthalimidoporphyrins (DAPIP).

In the past in collaboration with the group of A. Cheprakov^25,34 we developed a synthetic approach to π-extended porphyrins, in which the key step is oxidative aromatization of precursor porphyrins annealed with nonaromatic rings.^32,33 This method allowed us to generate a variety of syn-dibenzo- and syn-dinaphthoporphyrins with solubilizing substituents.⁴⁸ In such asymmetric systems the lowest energy absorption bands (Q bands) are positioned between those of non-extended (regular) porphyrins and porphyrins with the corresponding symmetric π-extension. Here we set out to explore the synthesis of meso-unsubstituted and meso-substituted DAPIP employing the same general strategy.

The developed synthesis (Schemes 1 and 2) consists of the assembly of ditetrahydrophthalimido-dipyrromethanes 6, 7 (Scheme 1), [2+2] condensation of 6 and 7 with unsubstituted dipyrromethanes to form porphyrininc precursors with fused non-aromatic rings (Scheme 2) and subsequent oxidative aromatization of the latter (Scheme 2). The synthesis of tetrahydrophthalimido-fused dipyrromethanes (6, 7) begins with the preparation of pyrrole-ester 3 via Diels-Alder cycloaddition of sulfoleno-pyrrole benzyl ester 1 to maleimide 2.²⁸ 1 was synthesized using an earlier published method,⁷¹ but with a slight modification. Slow addition of reagents (during 4-5 h) and keeping the temperature low (−70°C) allowed us to improve the yield by minimizing competing degradation of intermediates. The target pyrrole-ester 3 was isolated in 78% yield.

Scheme 1.

Synthesis of Dipyrromethanes.

Reagents and conditions: (a) 1,2,4-trichlorobenzene, reflux, 4h (78%); (b) CH₂(OMe)₂ or PhCHO, p-TsOH, AcOH, rt, 96h, R=H (74%), R=Ph (79%); (c) H₂, Pd/C, THF, rt, 24h; (d) ethylene glycol, 150°C.

Scheme 2.

Synthesis of Diarylphthalimidoporphyrins.^‡

Reagents and conditions: (a) (i) Zn(OAc)₂ 2H₂O, benzene, Ar, reflux, 2-3h; (ii) air, reflux, 12-14h; (b) DDQ, ethyl benzoate, 140°C, 6h: Zn-12 (65%); (c) 50% HCl: H₂-12 (86%), H₂-10 (17%)^†, H₂-11 (14%)^† (d) M=Pt: Pt(acac)₂, PhCN, reflux, 8h: Pt-12 (65%)^‡, Pt-13 (63%)^‡; M=Pd: Pd(OAc)₂, PhCN, reflux, 1h: Pd-12 (63%)^‡, Pd-13 (72%).^{‡ †} over steps a and c. ^‡ over steps d and b. ^‡The identity of compounds M-10 and M-11 (M=Zn, Pd, Pt) was confirmed by MALDI-TOF spectroscopy, after which they were introduced into the subsequent transformations without further characterization.

Condensation of 3 with benzaldehyde or dimethoxymethane was carried out in acetic acid, following the literature procedures,^34,48,72 affording tetrahydrophthalimido-fused dipyrromethanes-esters 4 and 5 in 74% and 79 % yields, respectively. The condensation required rather long time reaction times (72-96 h) to come to completion, likely as a consequence of the large steric bulk of pyrrole 3. The resulting dipyrromethanes-esters 4 and 5 were stable for months when we stored at ~0°C.

Hydrogenolysis of pyrrole-esters 4 or 5 (H₂, Pd/C) gave the corresponding carboxylic acids. However, subsequent decarboxylation upon treatment with TFA led to significant scrambling, and the target α,α’-unsubstituted dipyrromethanes 6 and 7 could be isolated only in low yields (~20%). A more successful procedure entails heating of the esters to 150°C in ethylene glycol,³² giving the target pyrroles in 40-50% yields. Note that 6 and 7 are unstable at ambient temperatures. Therefore, the decarboxylation should be carried out immediately prior to the porphyrin synthesis.

The assembly of the non-aromatized porphyrin-precursors via [2+2] condensation followed the method developed by Lindsey et al.⁷³ Propylimine substituents in α,α’-positions of dipyrromethanes 8, 9 are synthetically equivalent to formyl groups and have been shown to minimize scrambling in the porphyrin condensation.^73,74 The α,α’-propyliminodipyrromethanes were obtained by reacting pyrrole with suitable aldehydes (or their derivatives) in the presence of InCl₃, followed by formylation of the α,α’-positions and subsequent condensation with Propylamine ^73,74

The [2+2] condensation was carried out under the conditions developed previously for similar syntheses,⁴⁸ i.e. as a two-step process, which helped to minimize scrambling. First, 6 or 7 were condensed with 8 or 9 in the presence of Zn(OAc)₂ in a refluxing non-polar solvent (benzene or toluene) under inert (argon) atmosphere. The formation of the corresponding dihydroporphyrinogens was monitored by MALDI-TOF spectroscopy, and after their concentrations peaked (usually after ~3 h), the air was let in, and the mixtures were refluxed for 12-15 h. Our previous results indicate that the above conditions are optimal for efficient oxidation of dihydroporphyrinogens into porphyrins, while avoiding oxidation of dipyrromethanes to dipyrrins and other side reactions.⁴⁸ In the present case the method was successfully used to synthesize both meso-unsubstituted and 5,15-diphenyldicylcohexenoporphyrins, which were obtained as Zn complexes in 17-20% yields.

Aromatization of dicylohexenoporphyrins into the target DAPIP was performed using DDQ as an oxidant. meso-Unsusbtituted Zn-12 was obtained in 82% yield upon refluxing of Zn dicyclohexenoporphyrin (Zn-10) with DDQ in ethyl benzoate for 6 h. Use of ethyl benzoate was found to greatly facilitate aromatization of tetrahydrophlalimido-fused porphyrins with pendant ester groups, as both starting materials and products are high soluble in this solvent unlike in more commonly used toluene.⁴⁸ The product formation was observed by change in the color from red to deep green. The free-base DAPIP (H₂-12) was obtained upon facile demetalation of Zn-12 with HCl.

The synthesis of Pt and Pd complexes of DAPIP followed the same order: first the corresponding metal complexes of the precursor dicylohexenoporphyrins were obtained and then oxidized using DDQ. Zn-10 and Zn-11 were demetalated into free-bases H₂-10 and H₂-11 upon treatment with HCl. Insertion of Pt occurred upon refluxing of H₂-10 and H₂-11 with 3-4-fold molar excess of Pt(acac)₂ in benzonitrile for 6-8 h.⁷⁵ In addition, two other methods of Pt insertion were tested, using benzoic acid melt as a solvent⁷⁶ and microwave-assisted synthesis,⁷⁷ however, the traditional benzonitrile method in the present case gave superior results. Similarly, Pd complexes were prepared by treating the free-bases with Pd(OAc)₂ (3-4 eq) in benzonitrile for 1 h. It is noteworthy that insertion of both Pt and Pd led to partial oxidation of the peripheral cyclohexeno rings.⁶⁹ Consequently, the metal complexes M-10 and M-11 (M=Pd, Pt) were isolated as crude mixtures and subjected to the final aromatization without purification.

Aromatization of M-10 and M-11 was carried out in ethyl benzoate, since more traditional solvents, e.g. toluene or dioxane,³² gave poor results due to the solubility issues: aromatization could not be completed even after 18-20 h of reflux. In contrast, oxidation in ethyl benzoate upon heating to 140°C for 6-7 h afforded Pt and Pd complexes of DAPIP, M-12 and M-13 (M=Pd, Pt) in 63-65% yields.

Structural features.

All porphyrins were characterized by the standard analytical methods. In addition the molecular structure of the Pt complex of 5,15-diphenyl-DAPIP (Pt-13) was determined by X-ray analysis (Fig. 1; SI 3).

Figure 1.

X-ray crystallographic structure of Pt-13. The structure was analyzed by the Normal Mode Structural Decomposition Analysis (NSD)⁷⁸ revealing rather small out-of-plane distortion (D_oop=0.83Å) dominated by saddling (B_2u).

The meso-aryl rings in the 5,15-positions of Pt-13 are tilted by ~70° relative to the mean-square plane of the macrocycle, which is standard for regular nominally planar porphyrins. The aryl groups attached to the phthalimide nitrogens are forced by the 2,6-ester substituents into almost orthogonal positions relative to the porphyrin plane. Such orthogonal orientation is critical for maintaining solubility of DAPIP and similar porphyrins.⁶⁹ In the absence of orthogonal aryl groups, DAPIP, especially those without meso-aryls (such as M-12), would aggregate and become difficult to characterize and/or use in subsequent reactions.

To relieve the steric strain between the meso-aryl ring and the proximal hydrogen atoms in the benzo groups, the macrocycle in Pt-13 undergoes slight in-plane distortion, leaving more space between the two phthalimide substituents. Consequently, the distance between the two pyrrolic rings fused with phthalimides (a=5.07Å) is slightly larger than the analogous distances on the adjacent sides of the molecule (b=4.97Å). The in-plane distortion is a result of a small inward rotation of the phthalimide groups, which affects only minimally the rest of the macrocycle. A similar although larger effect was observed in 5,15-diaryltetrabenzoporphyrins.⁷⁹

In order to quantify the distortion of the macrocycle in Pt-13, its structure was subjected to the Normal-mode Structural Decomposition (NSD) analysis, as described by Shelnutt and co-workers.⁷⁸ In this method the macrocycle’s deviation from the pure D_4h symmetric configuration is expanded in the basis of in-plane and out-of plane distortion modes, corresponding to the vibrational modes of distinct symmetry types, such as saddling (B_2u), ruffling (B_1u), doming (A_2u) etc. The square roots of the mean squared deviations of all in-plane and all out-of-plane distortions are designated as D_ip (in-plane) and D_oop (out-of-plane), respectively. The macrocycle in Pt-13 exhibits moderate out-of-plane (D_oop=0.83Å) and a small in-plane (D_ip=0.15Å) distortion. The out-of-plane distortion is dominated by the saddling mode with small contribution of waving (see SI 4 for NSD details). Out-of-plane distortions in porphyrins are known to be associated with enhancement of non-radiative relaxation pathways,^80–82 however in π-extended porphyrins this effect is less pronounced.⁷⁹ The fact that Pt-13 is only moderately saddled (for comparison, in tetraaryltetrabenzoporphyrins D_oop values can be as high as 3.0-3.2Å) suggests that its emission quantum yield should decrease only slightly compared to that of the presumably completely planar counterpart Pt-12.

The structures of Pt-12 and other DAPIP complexes were not determined experimentally, but computed using DFT (see Experimental for details). The calculations, which were performed initially with B3LYP functional and LanL2DZ basis set, as implemented in Gaussian 16,⁸³ revealed completely planar geometries for all DAPIP complexes (Pd-12,13 and Pt-12,13). Nearly identical planar structures were obtained with two other functionals, CAM-B3LYP and wB97XD, showing that DFT fails to reproduce the non-planarity of the macrocycle seen in the experimental structure. On the other hand, experimental X-ray structures of many other non-planar porphyrins, including π-extended porphyrins,^79,84 have been well-reproduced by DFT. Our present result suggests that the non-planarity in the crystal structure of Pt-13 may be in fact a result of crystal packing forces, while in solution the average structure might be almost planar, as predicted by DFT.

Photophysical properties.

In this study we were particularly interested in the photophysics of the phosphorescent Pt and Pd complexes of DAPIP, for which we measured their linear (1P) absorption spectra, phosphorescence spectra along with phosphorescence decay times and quantum yields as well as the 2PA spectra (see below). The photophysical properties of meso-unsubstituted free-base DAPIP (H₂-12) and its Zn complex (Zn-12) were also examined. All of the synthesized porphyrins are well-soluble in organic solvents (e.g. CH₂Cl₂, THF, DMF), showing no sign of aggregation at the concentrations required for optical measurements. The photophysical data for the studied compounds are summarized in Table 1, and the absorption and phosphorescence spectra of Pt and Pd DAPIP are shown in Fig. 2.

Table 1.

Selected photophysical data for the studied porphyrins.^a.

Compd.	λmax B(nm)	λmax Q(nm)	B ε (M−1cm−1) σ(2) (GM)	Q ε (M−1cm−1)	σ(2)max (GM)b	λmax(em)c (nm)	ɸd	τe(s)
Pd-12	430	578	166540 33.4	83520	118.4	696 (p) 583 (df)	0.14 0.007	3.8×10−4
Pt-12	415	566	127950 74.9	110490	234.7	675 (p)	0.46	7×10−5
Zn-12	457	611	178280	47270	-	616 (f)	0.11	2.3×10−9
H2-12	439	647	-	-	-	647 (f)	0.16	1.1×10−8
Pd-13	441	586	232350 24.9	63850	138.3	703 (p) 592 (df)	0.12 0.008	3.6×10−4
Pt-13	429	573	164940 16.8	84600	94.8	683 (p)	0.43	6.1×10−5

^aAll measurements were performed in DMF at 22° C. Phosphorescence quantum yields and decay times were measured in deoxygenated solutions.

^bMaximal 2PA cross-section measured at the blue edge of the spectrum (see Fig. 4 for details).

^cem - emission type: f - prompt fluorescence, p - phosphorescence, df- thermally activated (E-type) delayed fluorescence.

^dɸ - emission quantum yield

^eτ - emission decay time.

Figure 2.

Absorption spectra of Pd (a) and Pt (c) complexes of porphyrins 12 and 13 in DMF. Emission spectra of Pd (b) and Pt (d) complexes of porphyrins 12 and 13 in deoxygenated DMF: p - phosphorescence; df - delayed fluorescence. The emission spectra are scaled by the respective phosphorescence quantum yields (Table 1).

Generally, the spectroscopic features of DAPIP resemble those of syn-dibenzo and syn-dinaphthoporphyrins^48,68 as well as of fully symmetric tetrabenzo- (TBP), tetranaphtho- (TNP) and tetraarylphthalimidoporphyrins (TAPIP).^69,70,79,84 The absorption spectra of DAPIP in the UV-visible range consist of two major bands, Q and B (Soret) bands, which are formed by the configuration interaction of single-electron excitations between two HOMO’S and two nearly degenerate LUMO’s.⁸⁵ The Q and B bands comprise pairs of orthogonally polarized transitions, Q_x, Q_y and B_x, B_y, which in fully symmetric D_4h porphyrins are pair-wise degenerate. In DAPIP, which have lower symmetry, the degeneracy is expected to be lifted. Indeed, TDDFT calculations predict existence of non-degenerate transitions, however the energies of the individual lines are too close to be resolvable in solution spectra at room temperature.

The TDDFT spectra and the frontier orbitals of Pd-12 are shown in Fig. 3 as an example (Fig. 3b). The summary of the TDDFT results is presented in Table 2. The data for Pd-13, Pt-12 and Pt-13 can be found in the SI (SI. 5). Here and in all calculations the pendant O(CH₂)₃CO₂Et residues were substituted by OMe groups.

Figure 3.

(a) TDDFT (B3LYP/6-31+G(d,p) 1PA spectra of porphyrins Pd-12 and Pd-13. Transitions (vertical bars) are broadened by Lorentzian shapes (Γ=0.07 eV) to facilitate visual comparison with the experimental spectra (Fig. 2a). (b) The shapes and energies (eV) of the frontier Kohn-Sham orbitals in Pd-12. The symbols in the parentheses (a_1u, a_2u, e_g) correspond to the symmetry types of the analogous MO’s in D_4h-symmetric porphyrins. The blue arrows depict the single-electron excitations, whose linear combinations dominate the lowest energy Q and B transitions (see Table 2). The polarization axes for these transitions are shown on the image of LUMO1.

Table 2.

TDDFT energies, oscillator strengths and contributions of single-electron excitations to the lowest excited states in porphyrin Pd-12.^a

Stateb	Energy (eV)	fc	Excitationd	Coefficientd
Qx	2.3197	0.1261	220→223	0.3945
			221→222	0.5809
Qy	2.3309	0.0842	220→222	−0.4069
			221→223	0.5707
Bx	2.9381	0.6906	215→222	0.1551
			220→223	0.4676
			221→222	−0.3395
			221→224	−0.3302
By	3.0038	0.6407	220→222	0.53321
			220→224	0.19836
			221→223	0.37291
			221→225	0.14758

^aCalculations were performed using TDDFT at the B3LYP//6-31+G(d,p) level.

^bTransitions are designated by symbols x and y according to the polarizations axes shown in Fig. 3b.

^cOscillator strength.

^dSingle-electron excitations and the corresponding coefficients that give rise to the Q and B transitions via configuration interaction. Note that in TDDFT the normalization condition for the case of restricted reference reads: ‹A+B|A-B›=1/2, where A and B are the vectors of the coefficients for the excitation and de-excitation determinants, respectively.

The main effect of the addition of the phthalimide residues to the two adjacent pyrrolic units is the destabilization of the HOMO whose density is localized on the β-pyrrolic carbons (MO 221), relative to HOMO-1 (MO 220), which resides primarily on the meso-carbons and nitrogen atoms. In regular non-extended porphyrins these two orbitals belong to the a_1u and a_2u symmetry types and are quasi-degenerate, while the e_g LUMOs are degenerate.^85–87 As in other π-extended porphyrins,^84,88 π-extension in DAPIP leads to a substantial red shift of the absorption bands and an increase in the oscillator strength of the Q band vs B band via intensity borrowing.⁸⁵ The molar extinction coefficients at the peaks of the Q-bands of Pt and Pd DAPIP are nearly as high as those in the analogous TBPs,⁶⁸ while the transition energies fall between those of the TBP and regular non-extended porphyrins. From the applications perspective, having sharp and strong transitions non-overlapping with transitions of TBPs may be advantageous for construction of sensor systems with orthogonal excitation of a pair of chromophores.

While the Q_x and Q_y transitions (designated according to their polarization along the axes shown in Fig. 3b) are dominated by the excitations between the frontier orbitals, B_x and B_y bands have non-negligible contributions from both lower and higher MOs. Similar situation was seen in symmetric TAPIP⁶⁹ as well as in other large π-extended porphyrins (e.g. TNP⁸⁴), revealing that the four-orbital approximation breaks down as the π-extended system increases in size.

Addition of two aryl groups to the 5,15-meso-positions in DAPIP destabilizes HOMO-1 (Table S2), which has its density on the meso-carbons (Fig. 3b), leading to further red shifts in the absorption of 13 vs 12 (Figs. 2,3a). This effect of meso-substitution is known.⁸⁹ The same MO is engaged in the interaction with the central ion, which makes the spectra sensitive to the nature of the metal.⁸⁵ Thus, the spectra of Pd complexes are red-shifted relative to those of their Pt counterparts, and the energy separation between the B_x and B_y transitions in PtDAPIP is larger, making the apparent B bands look broader.

Due to the presence of the heavy metal ions, the S₁→T₁ intersystem crossing in both Pt and Pd DAPIP is extremely efficient. None of these complexes exhibit prompt fluorescence. In contrast, free-base and ZnDAPIP (H₂-12 and Z_n-12) are fluorescent (SI.6), although their emission quantum yields are only moderate. Moderate fluorescence suggests that the intersystem crossing in both H₂-12 and Z_n-12 still efficiently competes with the radiative decay of the singlet state, but not nearly as successfully as in Pt and Pd porphyrins.

Pd and Pt complexes of both 12 and 13 emit remarkably bright phosphorescence at ambient temperatures (Fig. 2b,d). The phosphorescence quantum yields are record high (e.g. ɸ_p=0.46 for Pt-12), and the emissivity is practically unaffected by presence of the meso-aryl group (ɸ_p=0.43 for Pt-13), as expected from the very small degree of distortion seen in Pt-13 (see above). The ability to include a meso-aryl substituent without jeopardizing emissivity is very useful from the practical viewpoint, as it makes DAPIP much more amendable to synthetic modifications, e.g. attachment of linkers, inclusion into multichromophoric assemblies etc. Similar to other Pd and Pt porphyrins,⁷⁵ the phosphorescence quantum yields and emission rates are considerably higher for Pt complexes of DAPIP. Pt(II), being a heavier ion, induces stronger spin-orbit coupling and promotes faster T₁→S₀ decay with favorable for phosphorescence distribution between the radiative and non-radiative decay channels.

An interesting feature of the Pd complexes (Pd-12, 13) is the presence of a relatively strong thermally activated delayed fluorescence, which is emitted in parallel with phosphorescence (Fig. 2b) and has the same decay time. Such E-type delayed fluorescence has been seen previously in PdTAPIP as well as in some Pd tetrabenzoporphyrins.^69,90 Thermal exchange between the triplet and singlet manifolds is consistent with the narrow singlet-triplet energy gap (2J) due to the increase of the π-system. Thermal sensitivity of delayed fluorescence may be useful in temperature sensing applications.⁹⁰

Two-photon absorption (2PA).

syn-Fusion of the macrocycle with phthalimide groups leads not only to the destabilization of one of the HOMOs (MO 221 in Fig. 3b) relative to the other (MO 220), but also decreases its symmetry. In symmetric D_4h porphyrins, such as regular porphyrins, TBP, TAPIP etc, the corresponding HOMO is of a1u symmetry type, while in DAPIP HOMOs and LUMOs are neither symmetric (g) nor antisymmetric (u) with respect to the center of inversion of symmetry. Since the Q and B states comprise linear combinations of excitations between the frontier MOs, they cannot be classified as u or g states either. Consequently, Q and B transitions, which are 2PA-forbidden in symmetric porphyrins, should become allowed in DAPIP.

The experimental 2PA spectra of the studied Pd and Pt DAPIP are shown in Fig. 4. The spectra were measured by the method of 2P-excited time-resolved phosphorescence in time domain,⁶⁸ which allows complete suppression of the scattered excitation and other sources of noise. At every wavelength the power dependence, expressed as a log/log plot of the phosphorescence intensity vs excitation flux, was confirmed to be quadratic (slope 2.0±0.05). The blue edge (750 nm for all porphyrins) was reached when the slope became less than 1.95. Below that limit (λ_ex<750 nm) 2PA was contaminated by 1PA into the lower vibronic components of the Q state and/or by direct spin-forbidden S₀→T₁ excitation, which becomes partially allowed in porphyrins with strong spin-orbit coupling.⁶¹ The red edge of the 2PA spectrum (1080 nm) was set by the tunability range of the excitation laser.

Figure 4.

2PA spectra (blue lines) of Pd and Pt DAPIP in DMF at 22°C measured by the method of 2P phosphorescence excitation. The corresponding 1PA spectra (black lines) are arbitrarily scaled, but the relationship between the spectral intensities (molar extinction coefficients) between the compounds is kept unchanged to facilitate visual comparison. The 2PA spectra are truncated at the blue edge (750 nm; shown on the spectrum of Pd-12) where the slope of the log/log plot, representing the dependence on the signal on the excitation flux, becomes <1.95. The red edge (1080 nm) was determined by the tunability range of the laser.

To interpret the experimental findings we carried out calculations of 2PA using the TDDFT/Sum-Over-States (SOS) methodology, as described previously^68,69,91 (see Experimental and SI. 7 for details). The calculated 2PA spectra of porphyrins Pd-12 and Pd-13 are shown in Fig. 5. The spectra of the Pt complexes can be found in SI. 8.

Figure 5.

(a) 2PA spectra (blue bars) of porphyrins Pd-12 and Pd-13 computed using TDDFT/Sum-Over-States (SOS) methodology. The spectra are broadened by Lorentzian shapes (blue line, Γ=0.1 eV). Proportionally scaled 1PA spectra are shown for comparison (black line), (b) Leading single-electron excitations comprising 72% and 85% of the transition amplitudes for the lowest 2P-active states in PdDAPIP (marked with asterisk in the spectra) and PdTAPIP, respectively, along with the respective orbitals.

First we note that the calculations (Fig. 5) reproduce the experimental results (Fig. 4) quite well. The computed values of the 2PA cross-sections are 2-3 times higher than those determined experimentally. However, this disagreement can be viewed as minor, considering that experimental cross-sections routinely vary by orders of magnitude depending on the measurement method. More importantly, qualitatively all the experimental features are reproduced by the calculations correctly.

It can be seen (Fig. 4) that in all cases 2PA for the B state is indeed non-negligible, as expected from the symmetry breaking by the syn-diphthalimido-extension. The 2PA cross-sections for the B states in the studied DAPIP are higher than in the similarly syn-π-extended porphyrins with ester groups.⁶⁸ However, it is also clear that much stronger 2P-active states in DAPIP are positioned above the B state level, as evidenced by the steep rise of 2PA at the blue edge.

2PA for a transition from the ground state (0) to a state designated as final (f) comprises coherent action of multiple excitation channels, whereby each channel contributes to the total value of the 2P transition strength δ⁽²⁾ as specified by the SOS expression:

$${\text{δ}}^{\left(2\right)}~{\left|\sum _i\frac{{\text{μ}}_{0i}{\text{μ}}_{if}}{{\text{E}}_i-\hslash \text{ω}}\right|}^2,{\text{σ}}^{\left(2\right)}~{\left(\hslash \text{ω}\right)}^2{\text{δ}}^{\left(2\right)}$$ (1)

where ħω is the photon energy, i runs over all of the energy eigenstates of the system, μ_nm, are transition dipole moments connecting states n and m, and E_i, are the states’ energies. 2PA cross-section σ⁽²⁾ is proportional to the rotationally averaged δ⁽²⁾.^92–95 Expression (1) is written for degenerate 2PA, where the two photons have the same energy and the same linear polarization, and a simplified case when the molecule is fixed and all transitions are polarized in the same direction. Under these conditions µ_ij in (1) are simply the projections of the transition dipole moments on the photon polarization axis. For an ensemble of randomly oriented molecules and photons with different energies and polarizations the expression for δ⁽²⁾ becomes more bulky⁹³ (see SI.7), however the qualitative relationship between the cross-section value and the electronic structure of the molecule remains the same.

Each term in (1) represents an excitation channel, 0→i→f where i denotes an intermediate state on the pathway between the ground and the final state. The contribution of each channel is weighted by the energy difference, E_i-ħω, known as the detuning factor. There are two unique channels, 0→0→f and 0→f→f for which the corresponding terms in (1) contain products μ₀₀μ_0f and μ_0fμ_ff. The values μ₀₀ and μ_ff, or simply μ₀ and μ_f, are the diagonal elements of the electric dipole operator, i.e. the static (or permanent) dipole moments in the ground and final states. The combined contribution of these two channels, Σ₀+Σ_f, is expressed in (1) as ${\text{μ}}_{0f}\left({\text{μ}}_f-{\text{μ}}_0\right)/\hslash \omega$, i.e. it is proportional to the difference between the static dipole moments μ_f and μ₀. In non-centrosymmetric molecules the change in the static polarization upon excitation is generally non-zero, and it is common to approximate 2PA to the lowest excited state (f) in such dyes using only the states 0 and f - an approximation known as the two-state model.⁹⁵ However, as we show below, for DAPIP this approach is not valid.

Using (1) it is straightforward to define the contribution of all channels encompassing a certain intermediate estate i, or simply the contribution Σ_i, of state i, to the total transition strength δ⁽²⁾ (see SI.7).⁹¹ The contributions of states Q and B and of the lowest 2P-active state Δ (marked with asterisk in Fig. 5) for PdDAPIP are presented in Table 3 as an example. The contributions of the analogous states in D_4h-symmetric PdTAPIP are shown for comparison. In the latter case, the only state with non-zero 2PA strength is the g-state Δ, which in PdTAPIP lies below the B state level.⁶⁹ Excitation of the Δ state is distributed between four channels involving u-states Q_x, Q_y and B_x, B_y, while these states themselves are parity-forbidden for 2PA.

Table 3.

Calculated 2PA cross-sections (σ⁽²⁾) and fractional contributions Σ_i, of the excitation channels 0→i→f encompassing states 0 (ground state), Q, B and Δ (the lowest 2P-active state) as intermediate states (i) to the total 2P transition strength δ⁽²⁾ in PdDAPIP.^a Contributions of the analogous states in a D_4h-symmetric PdTAPIP are shown for comparison.

Final state (f)	σ(2) (GM)	Σf+Σ0b	ΣQx	ΣQy	ΣBx	ΣBy	ΣΔ
Qx	11.5	0.22	-	0.13	0.36	0.46	−0.26
Qy	41.1	0.06	0.15	-	0.88	−0.04	−0.19
Bx	144.0	0.39	0.15	0.32	-	0.00	−0.07
By	112.5	0.65	0.21	−0.08	0.00	-	−0.14
Δc	578.7	−0.04	0.30	0.20	0.17	0.33	-
ΔpdTAPIPd	1944.0	0.00	0.33	0.33	0.17	0.17	0.00

^aThe electronic structure calculations were performed using TDDFT (B3LYP//6-31G(d,p)) as implemented in Firefly 8.2.0⁹⁶ on the DFT-optimized structures.

^bThe sums of the contributions of channels 0→0→f and 0→f→f representing the term ${\text{μ}}_{0f}\left({\text{μ}}_f-{\text{μ}}_0\right)/\hslash \omega$ in the SOS. μ₀ and μ_f are the static dipole moments in states 0 and f respectively.

^cΔ designates the lowest 2P-active state (marked with asterisk in Fig. 5a).

^dLowest 2P-active gerade state Δ in D_4h-symmetric PdTAPIP. The Q and B states in PdTAPIP (ungerade states) possess zero 2PA activity.

Deviation from D_4h in DAPIP enables all of its states to gain some 2PA strength. In each case, the distribution of the transition amplitude over the channels is governed by the orientation of the moments μ_0i, and μ_if and the respective detuning factors. It can be seen (Table 3) that for all states an appreciable fraction of 2PA is due to the static dipoles terms (Σ_f+Σ₀), reflecting the fact that DAPIP molecule is polar, having its static moments oriented along the y-axis. Nevertheless, in spite of the absence of the center of inversion symmetry, the overall picture of 2PA in PdDAPIP still resembles that in TAPIP and other centrosymmetric porphyrins, indicating that the symmetry breaking, even as significant as by syn-phthalimido-fusion, induces only a relatively minor perturbation of the overall highly symmetric electronic structure of the porphyrin. For example, the maximal 2PA in PdDAPIP is predicted for state Δ, whose orbital origin is essentially the same as that in fully-symmetric PdTAPIP (Fig. 5b). The excitation of this state is also distributed between the Q and B channels, although not as evenly as in PdTAPIP. In both DAPIP and TAPIP the 2P-active states Δ are mixtures of multiple orbital configurations, but the leading terms are due to the respective HOMO’S and LUMO+2 (Fig. 5b), where the electron density shifts from the center of the macrocycle out to the peripheral phthalimide groups. It appears that such a delocalization more evenly over a larger molecular framework leads to narrowing of the HOMO-LUMO+2 gap, stabilizing the Δ-state and increasing 2PA.

The B and Q states in DAPIP are much less 2P-active than Δ, however the static dipole terms in each case are supplemented by large contributions of the complementary Q and B channels. The computed transition strengths are different for B_x and B_y, which is consistent with the experimentally measured spectral shapes (Fig. 4): the 2PA peaks in the B state region are slightly red-shifted relative to the respective 1PA maxima. Also noteworthy is the fact that for all states the channels involving state Δ interfere destructively with other excitation pathways, which is reflected by the negative signs of the respective contributions. Negative interference shows that state Δ effectively acts to diminish the 2PA strength for the Q and B states, while gaining its own intensity through the constructive interference of the channels enabled by these states. This phenomenon resembles the B-to-Q intensity borrowing effect, described by Gouterman,⁸⁵ where lifting of the degeneracy between the two porphyrin HOMOs leads to redistribution of intensity between the Q and B bands. A similar effect of Δ state was observed recently in the case of three-photon absorption in π-extended porphyrins.⁹¹

Conclusions

In this contribution we developed an efficient synthesis of syn-diphthalimidoporphyrins (DAPIP) and performed their structural and photophysical characterization, including measurements of their 2PA spectra. Using quantum-chemical analysis based on the TDDFT/SOS approach we were able to quantify the contributions of individual excitation channels to 2PA in DAPIP and trace the effects of symmetry breaking to the orbital origin of these channels. From the practical viewpoint, the combination of symmetry breaking with the stabilization of the lowest 2P-active state (Δ) significantly enhances 2PA of DAPIP in the B-state region. The exceptionally bright phosphorescence of the Pd and Pt complexes of DAPIP makes them valuable as luminescent markers for biological applications. The ability to concurrently excite DAPIP and their fully-symmetric analogs TAPIP, while differentially sampling their emission may be utilized in the construction of orthogonal optical sensing systems for advanced imaging applications.