Six-Coordinate Zinc Porphyrins for Template-Directed Synthesis of Spiro-Fused Nanorings

Five-coordinate geometry is the standard binding mode of zinc porphyrins with pyridine ligands. Here we show that pseudo-octahedral six-coordinate zinc porphyrin complexes can also be formed in solution, by taking advantage of the chelate effect. UV–vis–NIR titrations indicate that the strength of this second coordination is ca. 6–8 kJ mol–1. We have used the formation of six-coordinate zinc porphyrin complexes to achieve the template-directed synthesis of a 3D π-conjugated spiro-fused array of 11 porphyrin units, covalently connected in a nontrivial topology. Time-resolved fluorescence anisotropy experiments show that electronic excitation delocalizes between the two perpendicular nanorings of this spiro-system within the experimental time-resolution of 270 fs.


A. General Methods
All reagents were purchased from commercial sources and solvents were used as supplied unless otherwise noted. Dry solvents (CHCl 3 , CH 2 Cl 2 and toluene) were obtained by passing through alumina under N 2 . Diisopropylamine (DIPA) was dried over calcium hydride, distilled and stored under N 2 over molecular sieves. NMR data were recorded at 500 MHz using a Bruker AVII500 (with cryoprobe) or DRX500, or at 400 MHz using a Bruker AVII400 or AVIII400, or at 700 MHz using a Bruker AVIII700 (with cryoprobe) at 298 K. Chemical shifts are quoted as parts per million (ppm) relative to residual CHCl 3 (δ H 7.27 ppm for 1 H NMR and at δ C 77.2 ppm for 13 C NMR) and coupling constants (J) are reported in Hertz. MALDI-ToF spectra were measured at the EPSRC National Mass Spectrometry service (Swansea) using the Applied Biosystems Voyager DE-STR or at the University of Oxford using Waters MALDI Micro MX spectrometer (with trans-2-[3-(4-tert-butylphenyl)-2-methyl-2propenylidene]malononitrile (DCTB) as matrix). UV/vis/NIR absorbance measurements were recorded at 25 °C with a Perkin-Elmer Lambda 20 photospectrometer using quartz 1 cm cuvettes. UV/vis/NIR titrations were analyzed by calculating the difference in absorptions and plotted using Origin™ software or fitted to models using SPECFIT TM software. S1 Size exclusion chromatography (SEC) was carried out using Bio-Beads S-X1, 200-400 mesh (BioRad). Analytical and semi-preparative GPC were carried out on Shimadzu Recycling GPC system equipped with LC-20 AD pump, SPD-M20A UV detector and a set of JAIGEL 3H (20 × 600 mm) and JAIGEL 4H (20 × 600 mm) columns in toluene/1% pyridine as eluent with a flow rate of 3.5 mL/min.

Cross pentamer synthesis: Due to solubility and stability issues, fully deprotected compounds xEt-P1a and x-P1a from (TIPS) 4 xEt-P1 and (TMS) 4 x-P1
were not isolated but directly engaged in reaction after a short silica plug (see text below).

Preparation of P1 2H
P1 (74 mg, 0.04 mmol) was dissolved in CHCl 3 (10 mL). Trifluoroacetic acid (0.1 mL) was added dropwise to the porphyrin solution and the reaction mixture was stirred at room temperature for 15 minutes. After completion of the reaction followed by TLC, pyridine was added (0.5 mL). The reaction mixture was immediately passed through a short plug of silica gel (CHCl 3 ) and concentrated to dryness to give the title compound as a dark solid (69 mg, 95%).

Preparation of P1a 2H
P1 2H (69 mg, 0.03 mmol) was dissolved in CH 2 Cl 2 (5 mL) under nitrogen. TBAF solution (1.0 M in THF) (0.1 mL, 0.1 mmol) was added dropwise to the reaction mixture which was stirred at room temperature for 15 minutes. Then acetic acid (50 µL) was added and the crude reaction mixture was immediately passed though a short plug of silica gel (CHCl 3 ). Fraction contained the desired product was concentrated to dryness and used in the next step without further purification (61 mg, 99%).

C1. Characterization of s-P11·(T6) 2 : NMR analysis and Mass Spectra
Based on the D 2d symmetry of s-P11·(T6) 2 complex, only one quarter of the molecule needs to be considered in the interpretation of the 1 H NMR spectrum ( Figure S2). The chemical structure in Figure  S2 represents the four porphyrin units (P S where s is for spiro, P A , P B and P C ) contained in this quarter and the portion of the template that coordinates the porphyrin units. The full assignment of the 1 H NMR spectrum for the s-P11·(T6) 2 complex is outlined in the following section ( Figure S2). This was completed by comparison with similar template-ring systems S5 and with the assistance of COSY and NOESY NMR experiments. Figure S2. Representative quarter of the spiro-fused ring structure for the 1 H interpretation (top, protons 3 A-C are pointing to the inside of the ring (towards the template) and protons 3′ A-C to the outside of the ring, arrows for keys NOEs) and 1 H NMR spectrum of s-P11·•(T6) 2 (bottom, 500 MHz, CDCl 3 , 298 K,* for dichloromethane).
The proton assignments for s-P11·(T6) 2 is based on the initial assumption that the porphyrin β-protons between 9.20 and 9.70 ppm, correspond to protons 1 A-C , 1′ A-C and 1 S which are adjacent to the butadiyne bridge in s-P11·(T6) 2 . This assumption was later confirmed by COSY and NOESY experiments.
Due to ring current effect, the α-protons a, a′, a″ and a‴ on the T6 template pyridine moieties are strongly shielded and appear at very low chemical shifts (around 2.40 ppm for a, a′ and a″ and 2.70 ppm for a‴).
In the NOESY spectrum ( Figure S4), the porphyrin β-protons 1,1′ A and 1,1′ B-C show NOE with the template α-protons a, a′ and a″ and the porphyrin β-protons 1 S show NOE with the template α-protons a‴. Additionally, the porphyrin β-protons 1,1′ A show NOE with protons 1 S of the spiro porphyrin P S which confirm that β-protons 1,1′ A belong to the four porphyrins P A directly bound to the spiro porphyrin. S11 The template β-protons b, b′, b″ and b‴ are less affected by the ring current and are therefore less shielded than α-protons a, a′, a″ and a‴. Based on their coupling with protons a, a′, a″ and a‴ in the COSY spectrum, their assignment can be confirmed ( Figure S5). Based on their integration and the symmetry of the molecule, template β-protons b, b′, b″ can be readily assigned.
In the NOESY spectrum, template protons b, b′, b″ and b‴ show NOEs with protons c, c′, c″ and c‴ respectively ( Figure S6, left. The NOE between b‴ and c‴ was difficult to analyze because of the overlap of protons c″ and c‴). This helps clarify the assignment for protons d, d′, d″ and d‴ that couple to protons c, c′, c″ and c‴, respectively, in the COSY spectrum ( Figure S6, right).

S12
The porphyrin β-protons between 8.65 and 8.85 ppm correspond to protons 2,2′ A-C based on their correlations with protons 1,1′ A-C in the COSY spectrum ( Figure S7, left). The protons of the aryl solubilizing side groups can also be assigned based on their correlations with protons 2,2′ A-C in the NOESY spectrum ( Figure S7, right).

D. UV-Vis-NIR titrations
All titrations were performed in chloroform (containing ca. 0.5% ethanol as stabilizer) at 298 K and at constant porphyrin concentration (by adding porphyrin to the ligand stock solution before titrations were started). A key assumption underlying the analysis presented here is that each step in the titrations can be analyzed as an all-or-nothing two-state equilibrium. This assumption is supported by the isosbestic nature of the UV/vis titration for each step ( Figure S14-25) and by the good fits to the calculated binding curves. Moreover, the measured binding constants are in good agreement with similar systems, previously reported. S4

D1. Determination of K 1 values by titration of xEt-P5·(T3), x-P5·(T3) and l-P3 Zn3 with pyridine.
Determination of K 1 was studied by break-up titration for complexes xEt-P5·T3 and x-P5·T3 with pyridine, because of the high binding constant for the formation of these complexes makes it difficult to determine their stabilities by direct titration. As a reference, complex l-P3 Zn3 ·T3 was also studied under identical conditions. The complexes xEt-P5·T3, x-P5·T3 and l-P3 Zn3 ·T3 were firstly prepared by adding one equivalent of T3 to a solution of xEt-P5, x-P5 or l-P3 Zn3 . Then, a large excess of pyridine was titrated into solutions of the 1:1 complex (ca. 10 -6 M in CHCl 3 at 298 K) to displace the multidentate ligand. The denaturation constant K d was determined from this titration and used to calculate the formation constants K 1 via the corresponding thermodynamic cycle shown below (the same thermodynamic cycle can be drawn for xEt-P5 and l-P3 Zn3 ). S5 Figure S15. Thermodynamic cycle relating the formation constant of the template complex x-P5·(T3) (K 1 ) to the denaturation constant (K d ) and binding constant of each porphyrin unit for pyridine to form x-P5·(py) 3 (K py 3 ). Note that pyridine will also bind to the other two sites on x-P5, but this process occurs in a different range of pyridine concentrations, so it can be ignored when analyzing the denaturation.

Where:
K 1 is the formation constant for the complex x-P5·T3 (also xEt-P5·T3 and l-P3 Zn3 ·T3) from the corresponding cross-pentamer (or trimer). K py 3 is the cube of the binding constant for coordination of pyridine to zinc porphyrin monomer (K py = 3280 M -1 under our conditions), measured as previously described. S5,S6 K d is the denaturation constant determined by fitting the titration data using the Origin™ software; S4 see K d values, curves and spectra in Figure S16-S21.
For linear trimer l-P3 ZnH2Zn , UV/vis/NIR titrations were analyzed by calculating the difference in absorptions and plotted using Origin™ software ( Figure S26-27). Due to the relatively weak binding constant between l-P3 ZnH2Zn and T3, direct titration experiment could be done to measure the corresponding association constant, K 1 , displayed in Table S2.  Solutions of s-P11·(T6) 2 and sEt-P11·(T6) 2 in chloroform (containing ca. 0.5% ethanol as stabilizer) were titrated with quinuclidine to displace the T6 template (at 298 K at constant porphyrin complex concentration by adding porphyrin complex to the ligand stock solution before titrations were started). These UV-vis-NIR titrations show two distinct processes, quantified by the equilibrium constants K d1 and K d2 , which can be interpreted in terms of the thermodynamic cyclic in Figure S28. Figure S28. Thermodynamic cycle relating the formation constants of the template complexes s-P11·(T6) (K f1 ) and s-P11·(T6) 2 (K f2 ) to the denaturation constants K d1 and K d2 and the binding constant of each porphyrin unit for quinuclidine. K Q is the binding constant of one porphyrin unit with quinuclidine, approximated from the binding constant of quinuclidine to cyclic hexamer c-P6 (K Q = 3.6 × 10 5 M -1 ). K f1 = K Q 6 /K d2 and K f2 = K Q 5 /K d1 .

F. Photoluminescence Upconversion Spectroscopy
The photoluminescence (PL) upconversion technique was used to investigate fluorescence dynamics of sample solutions held in quartz cuvettes as described in detail elsewhere. S10,S11 An excitation pulse was generated by a mode-locked Ti:Sapphire laser with pulse duration of 100 fs and a repetition rate of 80 MHz. Fluorescence is collected and optically gated in a beta-bariumborate (BBO) crystal by a vertically polarized time-delayed gate beam. The upconverted signal, which consists of sum-frequency photons from the gate pulse and the vertical component of the fluorescence, was collected, dispersed in a monochromator and detected using a nitrogen-cooled CCD. Using a combination of a half-wave plate and a Glan-Thompson polarizer, the polarization of the excitation pulse was varied and fluorescence intensity dynamics were recorded separately for components polarized parallel ! ∥ and perpendicular ! ! to the excitation pulse polarization. The fluorescence anisotropy is defined using ! = (! ∥ − ! ! )/(! ∥ + 2! ! ) and calculated from the measured components. The full-width-half-maximum of the instrumental response function was measured to be 270 fs. No anisotropy decay was observed, indicating that any polarization decay must occur within 270 fs for this level of signal-to-noise.
The concentration of the sample solution was on the order of 10 -4 M. The excitation fluence was kept low (0.32 µJ/cm 2 ) to avoid fluorescence quenching via exciton-exciton annihilation. The detection wavelength was 915 nm for s-P11·(T6) 2 and 987 nm for sEt-P11·(T6) 2 . Figure S40. Time-resolved fluorescence decay (components polarized parallel I∥ and perpendicular I⊥ to the excitation pulse polarization, black and red circles, respectively, for experimental data and straight lines for corresponding fitted curves) for s-P11·(T6) 2 and sEt-P11·(T6) 2 upon excitation at 825 and 835 nm and detection at 915 and 987 nm, respectively, together with the corresponding fluorescence anisotropy dynamics (γ , blue circles) in toluene with 1% pyridine. (The pyridine has no role in this experiment, but it was added because we use 1% pyridine in toluene as a standard solvent, for comparison with other porphyrin oligomers, some of which require pyridine to avoid aggregation.)