Self-Assembly of Russian Doll Concentric Porphyrin Nanorings

Electronic communication between concentric macrocycles with wave functions that extend around their circumferences can lead to remarkable behavior, as illustrated by multiwalled carbon nanotubes and photosynthetic chlorophyll arrays. However, it is difficult to hold one π-conjugated molecular ring inside another. Here, we show that ring-in-ring complexes, consisting of a 6-porphyrin ring locked inside a 12-porphyrin ring, can be assembled by placing different metals in the two rings (zinc and aluminum). A bridging ligand with carboxylate and imidazole binding sites forms spokes between the two rings, resulting in a highly cooperative supramolecular self-assembly process. Excitation is transferred from the inner 6-ring to the outer 12-ring of this Russian doll complex within 40 ps. These complexes lead to a form of template-directed synthesis in which one nanoring promotes formation of a larger concentric homologous ring; here, the effective template is an eight-component noncovalent assembly. Russian doll templating provides a new approach to amplifying the size of a covalent nanostructure.


Benzyl 3,5-bis(2-((tert-butyldiphenylsilyl)oxy)ethyl)benzoate:
The preparation of this compound was adapted from literature procedures. S6,S7 To a solution of 9- BBN (0.5 M in THF,29 mL,14.7 mmol,8.0 equiv) was added benzyl 3,5-divinylbenzoate (372 mg, 1.84 mmol, 1.00 equiv) as a solution in THF (9.0 mL, 0.20 M) and the resulting mixture was stirred at room temperature overnight (16 h). After cooling the reaction mixture to -40 °C, H 2 O 2 (30% in H 2 O, 1.6 mL, 14.4 mmol, 8.0 equiv) was added slowly followed by 3.0 M NaOH (2.6 mL, 8.1 mmol, 4.4 equiv). The reaction mixture was then brought to 5 °C and left to stir for 4 h at this temperature. At this point, the mixture was warmed to room temperature, neutralized by addition of 1 M HCl and finally concentrated under reduced pressure. Purification by column chromatography on silica gel (gradient of 80% EtOAc in PE 40-60 to 100% EtOAc) afforded 180 mg of a mixture of products containing the corresponding diol. This mixture was dissolved in dry DMF (3.0 mL, 0.2 M) and used without further purification.
Aluminum cyclic porphyrin hexamer c-P6•(Ar'CO 2 ) 6 : To a solution of free-base c-P6(H2) (12.4 mg, 2.82 µmol, 1.00 equiv) in dry toluene (0.56 mL, 5.0 mM) and under an atmosphere of argon was added AlMe 3 (2.0 M in hexanes, 9.3 µL, 19 µmol, 6.6 equiv) dropwise. The reaction mixture was left to stir at room temperature for 30 min, after which point the UV-vis-NIR spectrum indicated full conversion of the starting material. CHCl 3 was then added (approximately 3 mL). The mixture was stirred for 20-30 minutes until a precipitate formed, then filtered over celite into a flask containing 3,5-dimethylbenzoic acid (2.54 mg, 16.9 µmol, 6.00 equiv). After stirring at room temperature for 30 min, the 1 H NMR spectrum of the material indicated clean conversion to the desired product. This material was used without further purification. T6•c-P6•(Ar'CO 2 ) 6 : To the sample of c-P6•(Ar'CO 2 ) 6 in dry CHCl 3 (2.0 mL) was added hexapyridyl template T6 (2.53 mg, 2.54 µmol). The solution was stirred at room temperature for 90 minutes, at which point the reaction appeared to be complete (monitored by UV-vis-NIR spectroscopy). The reaction mixture was purified by size exclusion chromatography on Biobeads SX-1 using CHCl 3 to yield 10.3 mg (57% yield from c-P6(H2)) T6•c-P6•(Ar'CO 2 ) 6 as a brown solid.

C1) Russian Doll Structure
The Russian doll complex was prepared by ligand exchange of T6•c-P6•(Ar'CO 2 ) 6 with L1 in the presence of c-P12. A solution of c-P12 in CDCl 3 ([c-P12] = 4.2 × 10 -4 M) was gradually added to an NMR tube containing a solution of T6•c-P6•(Ar'CO 2 ) 6 in CDCl 3 ([T6•c-P6•(Ar'CO 2 ) 6 ] = 1.9 × 10 -4 M). Once a 1:1 ratio of the nanorings was reached, ligand L1 was gradually added as a solution in CD 3 OD ([L1] = 0.16 M) until the signals corresponding to the individual components disappeared and a new product appeared to have been formed ( Figure  S1). The crude mixture was purified by size exclusion chromatography on Biobeads SX-1 using CHCl 3 , concentrated under reduced pressure and dissolved in CD 2 Cl 2 for further characterization ( Figure S2 and Section D3). The Russian Doll complex is soluble in toluene-d 8 , CDCl 3 and CD 2 Cl 2 however the sharpest NMR spectrum was obtained in CD 2 Cl 2 .   6 •c-P12 after purification by size exclusion chromatography on Biobeads SX-1 using CHCl 3 (400 MHz, CD 2 Cl 2 , 298 K). See Section D3 for full characterization and peak assignments.

C2) Control Mix
The control mix was prepared in a similar fashion to the Russian doll structure. A solution of benzyl-protected ligand Bn-L1 in CD 2 Cl 2 ([Bn-L1] = 4.8 × 10 -2 M) was gradually added to an NMR tube containing a solution of c-P12 in CD 2 Cl 2 ([c-P12] = 1.7 × 10 -4 M). Once a 6:1 ratio of the protected ligand/nanoring was reached, a solution of T6•c-P6•(Ar'CO 2 ) 6 in CD 2 Cl 2 ([T6•c-P6•(Ar'CO 2 ) 6 ] = 5.7 × 10 -4 M) was gradually added until a 1:1 ratio of the nanorings was reached ( Figure S3). S9 Figure S3. 1 H NMR titration of a 6:1 mixture of Bn-L1 and c-P12 with T6•c-P6•(Ar'CO 2 ) 6 (400 MHz, CD 2 Cl 2 , 298 K). Figure S4. 1 H NMR spectra of the control mix (top spectrum) and the Russian doll complex (bottom spectrum) (400 MHz, CD 2 Cl 2 , 298 K). The signals corresponding to the larger 12-porphyrin nanoring are highlighted in red while the signals corresponding to the aluminum 6-porphyrin nanoring are highlighted in green (see Section D for the full assignment of the 1 H NMR spectra of the various components). The protons from the inner Al-nanoring in the Russian doll are shielded compared to the free T6•c-P6•(Ar'CO 2 ) 6 in the control mixture, indicating that the smaller ring is nested within the larger ring.

C3) Comparison of the 1 H NMR Spectra for the Russian Doll and the Control Mix
The control mix (1.7 × 10 -4 M in CD 2 Cl 2 ) can be titrated with ligand L1 ([L1] = 0.16 M in CD 3 OD) to yield the Russian Doll complex ( Figure S5). The signals are slightly broad at the titration endpoint but become sharper after removal of excess ligands by size exclusion chromatography in CHCl 3 .

D1) Characterization of T6•c-P6•(Ar'CO 2 ) 6
Due to the high degree of symmetry in the aluminum 6-ring, only a "slice" corresponding to one sixth of the structure needs to be considered for the interpretation of the 1 H NMR spectrum ( Figure S6). The portion of the aryl side-group drawn in bold points towards template T6. The assignment of the 1 H NMR spectrum is presented in the following section and was carried out using COSY, NOESY and HSQC NMR experiments. The similar zinc porphyrin system T6•c-P6(Zn) was also used as a reference. S3 Figure S6. 1 H NMR spectrum of T6•c-P6•(Ar'CO 2 ) 6 (400 MHz, CD 2 Cl 2 , 298 K).
Our assignment of the 1 H NMR spectrum begins with the β-protons a and b ( Figure S7). Proton a, which is adjacent to the butadiyne bridge, is at the highest chemical shift (9.62 ppm). Proton a shows a NOE to proton b. Proton b also shows NOEs to protons c and c' (broad signals at 8.33 and 7.34 ppm). S12 Figure S7. Region of the NOESY spectrum of T6•c-P6•(Ar'CO 2 ) 6 corresponding to protons a, b, c and c' (400 MHz, CD 2 Cl 2 , 298 K, 600 ms mixing time).
NOEs are also observed between the tert-butyl group protons and protons b, c and c' (Figures S8). Strong NOEs are also observed between the tert-butyl group protons and the para proton d on the aryl side-group. This suggests that the aryl side-groups are rotating quickly on the T 1 timescale. In the COSY spectrum, correlations are observed between the CH 3 (protons l) group of the carboxylic acid ligand and protons j and k ( Figure S9). Protons j and k are shielded due to ring current effects. Figure S9. Region of the COSY spectrum of T6•c-P6•(Ar'CO 2 ) 6 corresponding to the protons on the carboxylic acid ligand (400 MHz, CD 2 Cl 2 , 298 K).
The HSQC spectrum also confirms our assignments for the CH 3 and t-Bu protons, since these protons correlate with signals at 20.3 and 31.4 ppm, respectively ( Figure S10). Figure S10. Region of the HSQC spectrum of T6•c-P6•(Ar'CO 2 ) 6 corresponding to the CH 3 and t-Bu protons (400 MHz, CD 2 Cl 2 , 298 K).

S14
The HSQC spectrum also supports our assignments for the aromatic protons in T6•c-P6•(Ar'CO 2 ) 6 ( Figure S11). Protons a and b correlate with signals at 130.3 and 133.4 ppm, respectively. Protons c and c' correlate with signals at 128.7 and 129.6 ppm, respectively, while proton d correlates with a signal at 121.7 ppm. Protons j and k from the carboxylic acid ligand correlate with signals at 124.9 and 130.8 ppm, respectively. Due to ring current effects, the α-pyridyl proton h on template T6 is shielded and appears at low chemical shift. The HSQC spectrum helped to assign proton h at 1.46 ppm due to its correlation with the signal at 141.5 ppm, which is characteristic of the α-carbon in a pyridine ring ( Figure S12). Figure S12. Region of the HSQC spectrum of T6•c-P6•(Ar'CO 2 ) 6 corresponding to template proton h (400 MHz, CD 2 Cl 2 , 298 K).

S15
Template protons e and f were assigned based on their correlation in the COSY spectrum ( Figure S13) and the relative strength of their NOEs with proton g in the NOESY spectrum ( Figure S14).  Finally, the assignments for protons e, f and g were confirmed based on their correlations with signals at 130.6, 123.7 and 118.2 ppm, respectively, in the HSQC spectrum ( Figure S11). S16 D2) Comparison of Δδ for T6 protons in T6•c-P6•(Ar'CO 2 ) 6 and T6•c-P6(Zn) Figure S15. Complexation-induced changes in the chemical shift (Δδ) in the 1 H NMR of free template T6 and bound template T6 (CDCl 3 , 298K) in T6•c-P6•(Ar'CO 2 ) 6 (green) and the reference T6•c-P6(Zn) complex (black). The Δδ was calculated from δ freeδ bound .
The change in chemical shift (Δδ) for template T6 protons upon binding the 6-porphyrin nanoring can be calculated by subtracting the chemical shift of the proton in the complex (δ bound ) from the chemical shift of the proton in free T6 (δ free ). These have been previously calculated for T6•c-P6(Zn) and are indicated by the black numbers in parentheses in Figure S15. S3 The change in chemical shift upon complexation of T6 in c-P6•(Ar'CO 2 ) 6 (green numbers in Figure S15) are very similar to those observed for T6•c-P6(Zn).
In the case of the aluminum 6-ring, the Δδs are slightly larger, suggesting that the hexapyridyl template is closer to the plane of the porphyrin. The Al-N and Zn-N distances in related metalloporphyrin crystal structures (where N corresponds to the nitrogen in a pyridyl group) are very similar (2.215 and 2.15 Å, respectively). S4,S9 In hexacoordinate Al-porphyrins, the metal is in the plane of the porphyrin. S9,S10 However, in T6•c-P6(Zn), the mean distance between the porphyrin plane and the Zn atom is 0.24 ± 0.06 Å, thus pushing the template further from the plane of the porphyrin and decreasing the observed Δδ. S4

D3) Characterization of T6•c-P6•(L1) 6 •c-P12
Due to the high degree of symmetry in the Russian doll complex, only a "slice" corresponding to one sixth of the structure needs to be considered for the interpretation of the 1 H NMR spectrum ( Figure S16). The chemical structure of this "slice" contains one aluminum porphyrin and two zinc porphyrins as well as one equivalent of the bridging ligand L1 and one "leg" of the hexapyridyl template T6. For the aluminum 6-ring, the portion of the aryl side-group drawn in bold points towards template T6 while the rest of the aryl side-group points toward the bridging ligand and the larger 12-porphyrin ring. For the zinc 12-ring, the part of the aryl side-group drawn in bold points towards the inside of the Russian doll (i.e. towards the aluminum 6-ring). The nearly complete assignment of the 1 H NMR spectrum for the Russian doll complex is detailed in the following section. These assignments are based on COSY, ROESY, NOESY and HSQC NMR experiments, as well as by comparison with similar porphyrin systems and the assignments obtained for T6•c-P6•(Ar'CO 2 ) 6 in the previous section. Protons l and q could not be unambiguously identified. Our assignment of the 1 H NMR spectrum of the Russian doll complex begins with the βprotons from c-P12. Protons a and d, which are adjacent to the butadiyne, are at a higher chemical shift than protons b and c. Proton a was distinguished from d based on its NOE with ligand proton p (see Figure S29). Proton a shows a NOE with proton b while proton d shows a NOE with proton c ( Figure S17). Coupling between a/b and c/d was also observed in the COSY spectrum. In the HSQC spectrum, protons a and d correlate with signals at 130.1 and 129.9 ppm, respectively ( Figure S18). The third proton that correlates with a signal at 129.9 ppm must be proton g from c-P6. Similarly, protons b, c and h all correlate with signals of similar chemical shift (132.9, 132.9 and 132.8 ppm, respectively). The assignment of β-protons g and h from c-P6 is further confirmed by the presence of a NOE between these two signals ( Figure S19). Coupling between g and h was also observed in the COSY spectrum. Protons b and c show NOEs with aryl side-group protons e and e' ( Figure S20). The cross peak between protons e and e' probably arises from chemical exchange (rotation of the aryl groups, which is slow on the chemical shift timescale) . Protons a, b, c, d, e and e', which are all on the 12-porphyrin nanoring, show NOEs with a signal at 1.57 ppm. This must correspond to the tert-butyl groups on c-P12. This t-Bu 12 group also has a NOE with aryl side-group proton f ( Figure S21).  In the HSQC spectrum, protons e and e' correlate with signals at 129.7 and 130.1 ppm ( Figure S22). Proton f correlates with a signal at 121.0 ppm and the other proton that correlates with a signal at 121.4 ppm must be para-proton j from c-P6.   From the ROESY spectrum, the assignment of ortho aryl protons e and e' in c-P12 is confirmed. They are in slow exchange on the NMR time scale. A second set of signals in slow exchange can be seen in the same chemical shift range. These must correspond to ortho aryl protons i and i' in c-P6. Proton i' also shows a NOE to β-proton h ( Figure S25).

S22
The assignment of aryl protons i and i' was confirmed in the NOESY spectrum. NOEs are observed between protons g and i/i' as well as protons h and i/i' (Figure S26). A weak NOE is also observed between protons h and j. Having assigned all of the porphyrin protons in the Russian doll, we next turned our attention to the bridging ligand protons in this system. Protons m and n were the clearest to identify, based on their strong coupling in the COSY spectrum. These protons also correlated with signals in the HSQC spectrum corresponding to CH 2 groups. Based on the chemical shifts in the HSQC, proton n corresponds to the signal at 1.95 ppm (correlates to CH 2 group at 46.8 ppm in the HSQC) and proton m corresponds to the signal at 1.06 ppm (correlates to CH 2 group at 35.1 ppm in the HSQC) ( Figure S27). In the NOESY spectrum, protons m and n show strong NOEs to proton k or l ( Figure  S28). There is also a weak NOE between proton k (or l) and t-Bu 6 .

S24
Due to ring-current effects, the imidazole protons p and q in the bridging ligand are shielded and appear at low chemical shift. Proton p was assigned based on its NOE with the c-P12 β-protons a and b ( Figure S29). Proton p correlates with proton o (at 4.88 ppm) in the COSY spectrum. Proton o shows strong NOEs to protons m, n and p. Hexapyridyl template T6 protons t and u were assigned to the multiplet at 5.42 ppm based on the assignments made for T6•c-P6•(Ar'CO 2 ) 6 in the previous section. This was supported by the HSQC spectrum, which shows a correlation between t and u and signals at 130.5 and 123.7 ppm, respectively ( Figure S30). These chemical shifts are in agreement with previously reported nanoring-T6 systems. Similar to observations made for the imidazole protons p and q, the α-pyridyl proton r in T6 is shielded and appears at very low chemical shift due to ring-current effects. Proton r was assigned based on the presence of weak NOEs between c-P6 protons g and h and a proton at 1.55 ppm ( Figure S31). Proton r also has a NOE with proton s at 4.88 ppm ( Figure S32). These assignments correlate nicely with the assignments made for T6•c-P6•(Ar'CO 2 ) 6 in the previous section and with previously reported nanoring-T6 system.     6 with quinuclidine Q was analyzed assuming that each Bn-L1 unit binds independently, so that the equilibrium can be treated as the displacement of a 2-site ligand L from a 2-site receptor P, with the initial concentration of PL being, [P] 0 , six times the concentration of c-P12•(Bn-L1) 6 .

The titration of c-P12•(Bn-L1)
The denaturation equilibrium constant, K dn , defined by equation (S1) was determined by fitting the binding isotherm to equation (S2), from ref. S11, where A is the absorption at any point in the titration, A 0 is the initial absorption,  (S3), which gives logK f = 6.6 ± 0.1.

E3) Calculation of Effective Molarity
The effective molarity can be calculated using equation (S6), where K f is the equilibrium constant for binding of c-P12 with T6•c-P6•(L1) 6 , ! ! is a statistical factor, K 1 is the microscopic binding constant of the corresponding reference ligand Bn-L1 in c-P12•(Bn-L1) 6 and N is the number of binding sites (N = 6).
The statistical factor was calculated as shown in Figure S37.
σ ext = 12 σ ext = 12 σ = 12 σ = 12 σ = 12 With the values of logK 1 = 6.6 ± 0.1, logK f = 40 ± 1, ! ! = 12, and N = 6, the average effective molarity for Russian doll formation is given by log EM = -0.13 ± 0.23, which means that the effective molarity is in the range 1.3 to 0.4 M. Absorption spectra were recorded at room temperature for the Russian doll, the control mix, c-P12•(Bn-L1) 6 and T6•c-P6•(Ar'CO 2 ) 6 in CHCl 3 ( Figure S40). All of the samples show a strong Soret band (400-550 nm) and a split Q band (700-950 nm). The Q band for the c-P12•(Bn-L1) 6 complex is red-shifted compared to T6•c-P6•(Ar'CO 2 ) 6 due to an increased conjugation length (Figure S40c,d). The absorption spectrum of the control mix consists of the sum of the contributions from its components, c-P12•(Bn-L1) 6 and T6•c-P6•(Ar'CO 2 ) 6 , whereas the Russian doll shows a further red-shift in the Q band, due to the rigidity of the c-P12 component in the assembled complex. Fluorescence spectra were measured at room temperature using a Horiba FluoroLog fluorometer for excitation in the Soret band at 500 nm. The fluorescence maxima of T6•c-P6•(Ar'CO 2 ) 6 and c-P12•(Bn-L1) 6 are observed at 897 nm and 917 nm, respectively. Both components contribute to the fluorescence in the control mix. As a result, the peak intensity of the control mix lies at 910 nm. The fluorescence spectrum of the Russian doll peaks at 928 nm, further red-shifted from c-P12•(Bn-L1) 6 because the complex is more rigid than its unbound constituent components.

G3) Fluorescence Quantum Yields.
The fluorescence quantum yields (Φ F , i.e. the photoluminescence quantum efficiency) is calculated from the fluorescence intensity integrated over the entire spectrum and the absorbance at the excitation wavelength, with l-P6 used as reference sample (Φ F = 28 %). S12 The quantum yield of T6•c-P6•(Ar'CO 2 ) 6 (Φ F = 2.1 %) is lower than that of c-P12•(Bn-L1) 6 (Φ F = 8.4 %), which is consistent with previous studies that have shown more strongly suppressed emission for smaller nanorings for which the excitonic wavefunction was delocalized over most of the ring. S12 Both components contribute to the fluorescence efficiency of the control mix, which lies in between, at Φ F = 5.4 % (with excitation at 500 nm). The Russian doll has Φ F = 2.4 %, which is higher than that of T6•c-P6•(Ar'CO 2 ) 6 , which could be attributed to energy transfer from the inner c-P6 ring to the outer c-P12 ring. The c-P12 emission component in the Russian doll is more rigid than c-P12•(Bn-L1) 6 resulting in the observed lowered Φ F compared to c-P12•(Bn-L1) 6 . The control mix comprises two emitting species, and for each emitter the fluorescence excitation is proportional to its absorption spectrum. While the absorption spectrum of the control mix is a sum of the contributions from the two components according to their amount, the fluorescence intensity also depends on the quantum yield of each component. As the Φ F of c-P12•(Bn-L1) 6 is much higher than that of T6•c-P6•(Ar'CO 2 ) 6 , the excitation spectrum of the control mix is not expected to be proportional to its absorption spectrum. For the control mix, the calculated excitation spectrum based on the measured absorption spectra and the Φ F of T6•c-P6•(Ar'CO 2 ) 6 and c-P12•(Bn-L1) 6 agrees well with the experimentally determined excitation spectrum, as shown in Figure S42a. The absorption spectrum deviates from the excitation spectrum, especially at the red-edge of the Q band, as a result of the different fluorescence quantum yields of the components that make up the mix. For the Russian doll, the excitation spectrum matches well with the absorption spectrum ( Figure S42b), indicating that the whole complex acts as a single emitter, even though there are distinct features in the absorption spectrum arising from the different ring components. Evidently, there is energy transfer between the two porphyrin rings in the Russian doll complex. The fluorescence decay dynamics were investigated using the time-correlated single-photon counting (TCSPC) technique, and the fluorescence lifetime was extracted by fitting the experimental data to a single exponential decay model. The dependence of the fluorescence lifetime on the excitation wavelength was examined at two different detection wavelengths for the Russian doll and the control mix as shown in Figure S43. The fluorescence lifetimes of T6•c-P6•(Ar'CO 2 ) 6 and c-P12•(Bn-L1) 6 are 455 ps and 376 ps respectively, measured at a detection wavelength of 760 nm (not shown). The measured fluorescence lifetime for the control mix varies with increasing excitation wavelength from around 450 ps to 420 ps as the contribution from T6•c-P6•(Ar'CO 2 ) 6 and c-P12•(Bn-L1) 6 to the fluorescence changes with excitation wavelength. The fluorescence lifetime of the Russian doll, on the other hand, shows no dependence on excitation wavelength, indicating that the emission originates from a single emitter. The mean fluorescence lifetime of the Russian doll is 364 ps when detected at 760 nm and 375 ps when detected at 860 nm. A comparison with the individual component lifetimes suggests that the emitting species in the Russian doll is most likely the outer c-P12, suggesting fast energy transfer from c-P6 within the time-resolution of the TCSPC system (40 ps).
Note: The aryl solubilizing side-group on the linear tetramer l-P4 used in these reactions was 3,5bis(octyloxy)phenyl.

H3) UV-Vis-NIR Spectra
A strong change in UV-vis-NIR absorption spectrum was observed for the Russian doll templated synthesis of c-P12 upon reaction completion. The appearance of the distinct band at 869 nm is indicative of porphyrin oligomer formation (including c-P12) and marked the completion of the reaction. After the alumina plug -during which the longer insoluble polymeric materials are removed -the absorption pattern from T6•c-P6 as well as from the porphyrin oligomers are clearly visible. In contrast, at the end of the control reaction, the band originating from porphyrin oligomers is less sharp in the UV-vis-NIR spectrum. The control reaction contained a substantial amount of insoluble material which is reflected in the UV-vis-NIR spectrum after the alumina plug, which shows that there is hardly any porphyrin oligomer material left but mainly T6•c-P6•(Ar'CO 2 ) 6 which is seen by the characteristic three peaks at 822, 785 and 752 nm. Figure S44. UV-vis-NIR spectra of the Russian doll templating reaction (left) and the control reaction (right). Absorption spectra of: starting materials after stirring for 1 h (black traces), after reaction completion (red traces) and after the alumina plug (blue traces). All spectra were recorded in CHCl 3 and normalized (at 831 nm).

H4) Analytical GPC Traces
The analysis of the product mixtures of the Russian doll templating reaction and the control reaction was performed by gel permeation chromatography (GPC). It has been previously shown that GPC is a powerful tool for separating and at the same time characterizing large porphyrin nanostructures. S14 To be able to compare relative yields, for both reactions all obtained Zn-porphyrin material was injected. In Figure S45 the GPC traces of both reactions are depicted at their absolute scale (i.e. both have the same range on the y-axis). From this representation it is evident how little material was obtained from the control reaction since most material was lost as insoluble polymer. Figure S45. GPC traces (toluene/1% pyridine) of the Russian doll templated synthesis of c-P12 (top trace) and the corresponding control reaction (red trace). The identity of the products was determined by calibrated retention times, MALDI-TOF analysis and 1 H-NMR spectroscopy for c-P12 and c-P24.