Porphyrin–Polyyne [3]- and [5]Rotaxanes

Porphyrin–polyyne [3]- and [5]rotaxanes have been synthesized by condensing aldehyde–rotaxanes with pyrrole or dipyrromethane. The crystal structure of a [3]rotaxane shows that the macrocycles adopt compact conformations, holding the hexaynes near the porphyrin core, and that the phenanthroline units form intermolecular π-stacked dimers in the solid. Fluorescence spectra reveal singlet excited-state energy transfer from the threaded hexayne to the porphyrin, from the phenanthroline to the porphyrin, and from the phenanthroline to the hexayne.


S1.1 -General Synthetic Details
Solvents were purchased from commercial suppliers (generally Aldrich or Fisher). THF, CH 2 Cl 2 , CHCl 3 , Et 2 O and toluene were dried by passing through activated alumina columns before use. Reagents were used as supplied with the exception of pyrrole, which was freshly distilled. CDCl 3 was stored over K 2 CO 3 and CaCl 2 to remove traces of acid and moisture.
Air-and moisture-sensitive reagents and reactions were handled and performed under an atmosphere of nitrogen or argon. Air-sensitive reactions were degassed by the freeze-pumpthaw method a pump-purge method or by bubbling with nitrogen gas.
Flash column chromatography was performed using VMR silica gel 60 and a positive pressure of nitrogen. Analytical thin-layer chromatography was performed using Merck silica gel 60 F 254 Al-backed plates. Plates were visualized using UV light (254 nm and 365 nm) and occasionally KMnO 4 .
ESI mass spectra were collected using a Waters LCT Premier (LRMS) operating in positive or negative mode. MALDI-TOF mass spectrometry was conducted using a Micromass MALDI micro MX spectrometer in positive reflectron mode. Dithranol was used as the matrix unless stated otherwise.
UV-vis spectra were recorded on a Perkin Elmer Lambda 20 or a Perkin Elmer Lambda 1050. Fluorescence Spectra were recorded on a Fluoromax LS55 at 298K in air-saturated CH 2 Cl 2 .
Microwave reactions were performed in a Biotage Initiator in 2.5 mL vials for 15 min on high power.

S1.2 -Synthetic Schemes
The polyyne component was synthesized as previously. S1 The polyethylene glycol macrocycles used in this work are larger than those previously synthesized in the Sauvage group but based on the same skeleton as Sauvage and Saito's macrocycle. S2 In our hands, the yield was also better than in the case of the smaller previously reported macrocycle.

S3.1 -Absorption and Emission Spectra
A Perkin-Elmer Lambda 25 UV-vis spectrometer was used to record absorption spectra in the range 250-600 nm in CH 2 Cl 2 . Molar absorption coefficients were measured by taking three samples of different concentrations; each sample was made by weighing out approximately 1 mg of the compound with an accuracy of ±0.01 mg.
A Fluoromax LS55 was used to record emission and excitation spectra. The model compounds, Zn-TPP and Zn-TBP were used to check that the correction file was consistent across the spectral range. These reference compounds were also used to estimate the absorption contribution of the porphyrin component of the more complex diaryl and tetraaryl porphyrins. First, all compounds were excited at the Soret band to record emission spectra. The emission maxima was then noted and used as a point to measure excitation spectra.
The concentrations of the samples used to measure emission and excitation spectra were adjusted to give a peak optical density of 0.1 to avoid intermolecular fluorescence quenching, reabsorption effects and saturation of the detector.    Zn-PM2b 0.017 Table S1 -Fluorescence quantum yields measured in air-saturated CH 2 Cl 2 for excitation at the peak of the Soret band (λ max for each compound, about 415 nm) using Zn-TPP and Zn-TBP as standards. S7,S8

S3.2 -Analysis of Absorption Spectra
To facilitate the comparison of the absorption spectra of Zn-PM4a, Zn-PM2a, Zn-P3Ra, Zn-PM4b, Zn-PM2b, Zn-P3Rb and Zn-P5Rb, all these spectra, and those of Zn-TPP and Zn-TBP, were normalized to a value of 1.0 at the Q-band maximum (at around 550 nm).
Comparison of the absorption spectra of this family of compounds shows that there is only weak ground-state electronic coupling between the three types of connected chromophore units: the polyyne, the phenanthroline and the porphyrin; i.e. the absorption bands due to the polyyne at around 300 nm are essentially identical in Zn-P3Ra, Zn-P3Rb These three difference-spectra are compared with the absorption spectrum of the isolated hexayne dumbbell Tr*-C 12 -Tr* in Figure S42. This subtraction demonstrates that the absorption spectra of the macrocycle, porphyrin and polyyne components are essentially identical in these compounds. A red-shift of 2 nm is observed in all of the polyyne vibronic bands in the rotaxanes, in comparison to free Tr*C 12 Tr*, as observed in other polyyne rotaxanes. S6 Figure S42. Calculation of the polyyne contribution to absorbance of different porphyrin rotaxanes and free Tr*C 12 Tr* dumbbell. Normalized at maximum intensity (solvent: CH 2 Cl 2 ).

S49
Similarly the absorption spectrum of the phenanthroline component can be calculated by: (a) Subtraction of the absorption spectrum of Zn-TPP from that of Zn-PM4a. (b) Subtraction of the absorption spectrum of Zn-TBP from that of Zn-PM2a. (c) Subtraction of the absorption spectrum of Zn-TPP from that of Zn-PM4b. (d) Subtraction of the absorption spectrum of Zn-TBP from that of Zn-PM2b.
These four difference-spectra are compared with the absorption spectra of Ma and Mb in Figure S43, showing that the absorption spectra of the macrocycle are essentially identical in all these complexes. These absorption spectra demonstrate that it is valid to split the absorption and excitation into components originating individually from each type of chromophore. Figure S44 shows the different components of the absorption band that are used to determine EET efficiency in Zn-P5Rb. The same type of analysis was performed for Zn-PM4a, Zn-PM2a, Zn-P3Ra, Zn-PM4b, Zn-PM2b and Zn-P3Rb.

S3.2 -Analysis of Excitation Spectra
To facilitate the analysis of the absorption and excitation spectra of Zn-PM4a, Zn-PM2a, Zn-P3Ra, Zn-PM4b, Zn-PM2b, Zn-P3Rb and Zn-P5Rb, all these spectra, and those of Zn-TPP and Zn-TBP, were normalized to a value of 1.0 at the Q-band maximum (at around 550 nm).
We assume that direct excitation of the Q-band of each porphyrin derivative generates the S 1 excited state of the porphyrin unit with 100% efficiency. Thus if the excitation spectrum of a given compound is divided by its absorption spectrum, then the value of the normalized excitation/absorption spectrum at any wavelength provides an estimate of the efficiency with which excitation at that wavelength generates the S 1 excited state of the porphyrin unit.
The interaction between the chromophore components is weak in the ground state (as demonstrated above), but the interaction between the components is stronger in the excited states, which makes it more complicated to calculate of the EET efficiency from a specific component of the molecule, because one component can affect the EET efficiency of another component. This means that we need to split the contributions of the components to the excitation spectra, as discussed above for absorption spectra.
Comparison of the excitation spectra of Zn-P5Rb and Zn-PM4b ( Figure S45) illustrates the fact that the effect of one component on another cannot be ignored. These excitation spectra would be identical at 350 nm if the components acted independently, as the polyyne component has negligible absorption at this wavelength. However the excitation spectra (normalized at the Q-band) have very different intensities at 350 nm. The excitation spectrum of the rotaxane (Zn-P5Rb) is less intense than that of the polyyne-free compound (Zn-PM4b) because of the quenching effect of the polyyne. Figure S45. Comparison of excitation spectra of Zn-P5Rb and Zn-PM4b (normalized at the Q band; solvent: CH 2 Cl 2 ; same data as Fig. S31).

S3.3 -Calculation of EET Efficiency
The absorption and excitation spectra of Zn-P5Rb, Zn-PM4b and Zn-TPP (all normalized at the Q band) are compared in Figure S46a. The absorption coefficient of Zn-TPP is negligible compared to those of Zn-P5Rb and Zn-PM4b in the wavelength range 250-375 nm ( Figures S41, 44 and 46b), which leads to the following two conclusions: (a) The intensity of the excitation spectrum of Zn-PM4b in this region (250-375 nm) can be attributed entirely to fluorescence originating from absorption by the phenanthroline macrocycle followed by energy transfer to the porphyrin.
(b) The intensity of the excitation spectrum of Zn-P5Rb in this region can be attributed to two components: (i) fluorescence originating from absorption by the phenanthroline macrocycle followed by energy transfer to the porphyrin, and (ii) fluorescence originating from absorption by the polyyne followed by energy transfer to the porphyrin.
The absorption coefficient of the polyyne (and isolated Tr*C 12 Tr* hexayne) is negligible compared to those of Zn-P5Rb and Zn-PM4b in the wavelength range 350-400 nm ( Figures  S42 and S44). Thus the excitation spectrum of Zn-P5Rb in this region (350-400 nm) in entirely assigned to fluorescence originating from absorption by the phenanthroline macrocycle followed by energy transfer to the porphyrin, in competition with energy transfer to the polyyne. We assume that the ratio of energy transfer macrocycle → porphyrin and energy transfer macrocycle → polyyne is constant across the wavelength region 250-375 nm. A scaling factor was calculated by dividing the value of excitation spectrum intensity for Zn-P5Rb at 350 nm by that of Zn-PM4b. The excitation spectrum of Zn-PM4b was multiplied by this scaling factor to give a new spectrum for the macrocycle contribution to the excitation S52 spectrum of Zn-P5Rb ( Figure S46b). The area under this curve is marked "macrocycle" in Figure S46b; this area represents the contribution of the macrocycle to the excitation spectrum of Zn-PM4b. Subtraction of this area "macrocycle" from the excitation spectrum of Zn-P5Rb yields the area marked "polyyne", which is attributed to polyyne → porphyrin energy transfer.
Quantum yields for excited-state energy transfer (EET) were calculated from the ratios of excitation spectra divided by absorption spectra, for macrocycle-specific or polyyne-specific spectral components, according to equations S1 and S2, after dissecting absorption and excitation spectra as described in Sections 3.1 and 3.2 respectively. Excitation/absorption rations were integrated across the absorption bands: 275-375 nm in Zn-PM2a, Zn-PM2b and Zn-PM4b; 350-370 nm for φ EET (macrocycle → porph) in Zn-P3Ra, Zn-P3Rb and Zn-P5Rb and 280-330 nm for φ EET (polyyne → porph) in Zn-P3Ra, Zn-P3Rb and Zn-P5Rb Zn-PM2a 0.67 ± 0.08 -Zn-P3Ra 0.32 ± 0.03 0.16 ± 0.02 Zn-PM2b 0.88 ± 0.11 -Zn-P3Rb 0.22 ± 0.03 0.10 ± 0.03 Zn-PM4b 0.68 ± 0.06 -Zn-P5Rb 0.25 ± 0.02 0.09 ± 0.02 Table S2 -EET efficiency of each component There are several key conclusions from this data: 1) The polyyne quenches the macrocycle's S 1 state, thus diminishing the efficiency of the EET 2) The polyyne can participate in an EET to the porphyrin, this is not very efficient, probably because of the competing formation of a triplet state 3) The distance between the components affects the EET (Zn-P3Ra has a more efficient EET from the polyyne due to the shorter distance between these components). Chemistry Central Journal 2015, 9:30) enabled the location of many of the missing atoms, however, the poor resolution of the data and the fact that atoms in these regions were poorly resolved (due to disorder) meant that the location of some of these atoms had to be inferred. Same distance, angle and thermal restraints were then applied to ensure the geometry and displacement ellipsoids remained sensible on refinement. Efforts were made to model the disorder in the oligoethyleneglycol chains, however, the lack of resolution meant that, in general, minor components were not distinguishable in the maps. For this reason, disorder was not fully modeled, but this led to some of the hydrogen atoms of the oligoethyleneglycol chains adopting positions unfeasibly close to other (often hydrogen) atoms. Anti-bumping restraints were used to minimize this, but it is clear that if it had been possible to use a multi-S54 component model this would have been less of an issue. Nevertheless, the phenanthroline groups and the central porphyrins are clearly visible in the initial solution and their close proximity enforce an often gauche conformation on the linking chains.
The distances between the polyyne and porphyrin components are relatively well defined because these components exhibit less disorder than other parts of the structure, as demonstrated from the displacement ellipsoids, and by the fact that these components were immediately visible in the preliminary solution, Figure S47.