Retaining individualities: the photodynamics of self-ordering porphyrin assemblies† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c5cc09095d Click here for additional data file.

Retained photochemical properties – a simple porphyrin–polyDMA conjugate with the ability to self assemble into large (∼1 μm) vesicles in water. The photodynamics are remarkably preserved despite the extensive aggregation.


Materials
Spectroscopic grade 1,4-dioxane was purchased from VWR. HPLC grade solvents were purchased from Fisher. Water for synthesis, spectroscopy and self assembly was purified to a resistivity of 18.2 MΩ·cm using a Millipore Simplicity Ultrapure water system. 4-(3-hydroxy-3-methylbut-1-ynyl) benzaldehyde (1) was synthesised according to procedures by Stulz  Addition-Fragmentation chain Transfer (RAFT) polymerisation of dimethylacrylamide (5, DP ≈ 60) with CTA 4 was performed as previously reported. 4 Pyrrole was purchased from Aldrich and distilled over CaH 2 under vacuum (0.5 mBar) and stored under N 2 protected from light prior to use. All other chemicals and solvents were purchased from Sigma, Aldrich, Fluka or Acros and used as received unless stated otherwise.
Dialysis was performed using Spectra/Por of appropriate molecular weight cut off (MWCO), purchased from VWR or Fisher. Preparatory size exclusion chromatography (prep-SEC) was performed with Bio-Beads TM S-X Resin in appropriate solvents, purchased from Bio-Rad. All dry-state transmission electron microscope (TEM) samples were prepared on graphene oxide (GO)-coated carbon grids (Quantifoil R2/2). 5 Generally, a drop of sample (20 μL) was pipetted on a grid, blotted immediately and left to air dry. For cryogenic electron microscopy (cryo-TEM), the samples were prepared at ambient temperature by placing a droplet on a TEM grid. The extra liquid was then blotted with a filter paper and the grid was inserted into liquid ethane at its freezing point. The frozen samples were subsequently kept under liquid nitrogen.

Time Resolved Transient Electronic Absorption Spectroscopy (TEAS)
The detailed experimental procedures for TEAS can be found in previous reports. [6][7][8] Briefly, a commercially available Ti:sapphire oscillator and amplifier system (Spectra-Physics) produces 3 mJ laser pulses of ≈ 40 fs duration centered around 800 nm with a repetition rate of 1 kHz. For TEAS, a 1 mJ/pulse 800 nm laser beam is split into two beams of (i) 0.95 and (ii) 0.05 mJ/pulse. Beam (i) is used to generate the pump pulse centered around 400 nm (2-5 mJ·cm -2 ) through second harmonic generation using a β-barium borate crystal. Beam (ii) is used to generate the probe pulse, a white light continuum (330-725 nm). Pump-probe polarizations are held at the magic angle (54.7 • ) relative to one another. Changes in optical density (∆OD) of the sample are calculated from probe intensities, collected using a spectrometer (Avantes, AvaSpec-ULS1650F). The delivery system for the samples is a flow-through cell (Demountable Liquid Cell by Harrick Scientific Products, Inc.).
The sample is circulated using a PTFE tubing peristaltic pump (Masterflex), recirculating sample from a 50 mL reservoir, in order to provide each pump-probe pulse pair with fresh sample.
The solution was protected from light. After stirring the solution for 45 minutes at RT, 2,3-dichloro-5,6dicyano-1,4-benzoquinone (DDQ, 1.532 g, 6.75 mmol, 0.9 eq.) was added to the reaction. The reaction was stirred for one hour and the crude mixture was filtered through a neutral aluminium oxide patch (7.5 cm) and washed with 5% methanol in CH 2 Cl 2 until the eluent was colourless. The crude product was dried in vacuo and re-dissolved in 150 mL of toluene. A fresh batch of DDQ (1.702 g, 7.5 mmol, 1 eq.) was added and the mixture was heated to reflux for 3 hrs.  Figure   Although two metallated porphyrins were synthesised, the Zn(II)-porphyrin (2bB) was much more stable in storage. Thus, it was chosen to undergo step iii from Figure 2 the deprotection of the alkyne function group (2cA). NaOMe (24 mg, 432 mmol, 30 eq.) and 2bB (10.4 mg, 14.4 mmol, 1 eq.) was disolved in 15 mL of toluene, degassed with N 2 for 20 minutes before heating to reflux (125 C). The reaction was left over night. Solvent was removed in vacuo, crude product was extracted with 20 mL DCM, washed with water (3 ⇥ 100 mL) and brine (2 ⇥ 100 mL), dried with MgSO 4 . Basic alumina was added to the filtered solution, dried in vacuo before loading onto column (neutralised silica, eluent : DCM ! 0.5% Methanol in DCM). 10 3a Figure S1 1 H NMR spectra of 3a in CDCl 3

Infra-Red (IR) Spectra
The IR spectra of the starting porphyrins (3, 3a and Zn-dPP), polymers (5 and 6) and the final product Zn-dPP-pDMA are shown in Figure S3. In particular, the appearance of the terminal-alkyne stretch in Zn-dPP (blue dot) and azide stretch in 6 (orange dot) indicated the successful de-protection and azide functionalisation, respectively, of the starting compounds. Their subsequent disappearance demonstrated the successful coupling reaction.  Zn-dPP-pDMA Figure S3 IR spectra of the starting and final products. Regions of interest are highlighted by dots of corresponding colours. 8

Size Exclusion Chromatography
The SEC spectra of the starting pDMA (RI) and Zn-dPP-pDMA (UV absorption at 414 nm) are shown in Figure S4. The anomalous shoulder at approximately double the molecular weight (Mw ≈ 28 kDa), which is attributed to the dimer of Zn-dPP-pDMA, is also observed, in accord with previously reported porphyrin-polymer conjugate systems. 9

Light Scattering
To determine the morphologies of the assembled system, light scattering measurements were performed.
Initially, the results were inconsistent between each measurement and we were unsure whether the strong absorption of the Zn-dPP cores affected the readings. However, as the instrument uses 632 nm laser, it is outside the absorption range of Zn-dPP. We therefore examined our procedures more carefully and did measurements for different filtered as well as unfiltered systems ( Figure S5). All samples were assembled as described in Section 1.5, followed by dilution from 3 mg/mL to 0.5 mg/mL prior to filtration and analysis.
The radius of gyration (R g ) and hydrodynamic radius (R h ) as well as the R g /R h ratio of each measurements are summarised in Table S1.   As shown in Table S1, all radii of gyration (R g ) were approximately a quarter of the filter pore-sizes.
This strongly suggested that the assembled systems underwent rearrangement upon filtration. The universal R g /R h of 1.6 suggested that the system under study were either elongated or of irregular formation undergoing fusion/fission rearrangement. 10

UV-Visible Absorption of Filtrated Samples
Since the unfiltered sample seemed to be multi-modal, we measured the UV-Visible spectra of the filtrated and unfiltered samples to investigate whether Zn-dPP were present in the large particles, shown in Figure   S6. To our surprise, the filtration process seemed to have removed a rather large amount of Zn-dPP from the assembled systems. The Soret peak was reduced by ca. 40% and 50% post filtration, through 0.45 and 0.22 μm pores respectively. Although it is not unusual for samples to 'stick' to the filter membranes, removal of up to 50% of samples was 20 fold higher than the 2.5% observed decrease at the Q-band region in the test measurement with unimeric Zn-dPP-pDMA in dioxane at 6 mg/mL through the 0.22 μm filters (data not shown). This unusual loss of material indicated that Zn-dPP were indeed part of the large aggregates and were removed in the filtration process. However, the same slight red-shifts were present in all the spectra, suggesting that the filtration did not affect the micro-environment of the Zn-dPP core.
0.45 0.22 Figure S6 UV-Visible absorption spectra of filtrated samples. Samples were assembled in conditions detailed in Section 1.5 and diluted from 3 mg/mL to 0.5 mg/mL prior to filtration and measurements.

Transmission Electron Microscropy Studies
In addition to observing the unfiltered samples under cryo-TEM as shown in the main text, we also observed the filtered samples under dry state TEM. In agreement to our filtrated SLS/DLS studies, the size of the observed particles were roughly half the diameter of the filter pore-sizes. These are shown in Figures S7 and   S8. Figure S7 TEM images of Zn-dPP-pDMA assemblies filtered through 0.22 μm pores on GO coated girds. Figure S8 TEM images of Zn-dPP-pDMA assemblies filtered through 0.45 μm pores on GO coated grids.
As observed in these filtrated samples, even though some large spherical structures remained, the overall assemblies were damaged during the filtration processes, evidenced by the small circular fragments (0.22 μm filtered, Figure S7) and needle like structures (0.45 μm filtered, Figure S8). This, together with the SLS/DLS data (Section 2.1), strongly suggests that the assembled structures were extremely sensitive to shear forces applied during the filtration processes. Hence, in an effort to avoid both the loss of material and disturbance to their native assembled structures, all photochemical studies were performed with unfiltered samples. 12 3 Photochemical Information

Static Fluorescence
Heat maps for the measured fluorescence spectra of all systems are presented in Figure S9. All spectra were recorded with identical settings (excitation slit-width = 2.5 nm; emission slit-width = 5 nm; and photomultiplier tube voltage = 800 V). The spectra were uncorrected in energy, as the emission wavelength extends beyond 600 nm, which corresponds to the upper limit to which our instrument can correct for.
Signals greater than the maximum fluorescence intensity (arising from instrument scattering) are all set to 0 for clarity. The emissions demonstrate similar features to those previously reported for Zn-meso-tetraphenylporphyrins. [11][12][13][14] In particular, modest emission following S 2 →S 0 can be observed in both Zn-dPP and Zn-dPP-pDMA in dioxane; in the assembled system however, this feature appear to be relatively weakened, fully solvated systems demonstrated almost identical ratio between the Q(0,0) and Q(0,1) peaks, similar to measurements in previous studies. [11][12][13] However, the ratio between these peaks were altered in the assembled system, the Q(0,0) showing stronger intensity relative to the Q(0,1) peak. This observation may indicate subtle differences in the geometry of the porphyrin in the S 1 and S 0 states, which lead to differences in the Franck-Condon factors. A more refined explanation of this difference would require vibrational frequency calculations that are likely to be prohibitively expensive in computational time and also beyond the scope of the present work.

Global Fitting and Error Analysis
A global fitting procedure is used to determine the set of lifetimes which characterise a function of four exponential decays convoluted with a Gaussian instrument response. 15 The full width at half maximum of the instrument response is measured to be ∼ 150 fs (determined through solvent only transients, data not shown). Representative fitted traces at 416 nm of the experimental data are shown in Figure S10. We use support plane analysis to determine a 95% confidence interval on the lifetimes determined from global fitting. 16 One of these lifetimes is attributed to small spectral shifts in the TAS which is required for global fitting convergence. It is determined to be < 100 fs and therefore not considered as a resolvable dynamical process. Another lifetime behaves as a long-lived baseline offset returning a lifetime of 2 ns which we attribute to intersystem crossing (τ 3/ISC ). 11 As a result, support plane analysis is only used for τ 1/IC and τ 2/IET . Briefly, the goodness of fit, χ 2 , for the globally fitted lifetimes is χ 2 min . The values of τ 1/IC and τ 2/IET are systematically varied, and for each pair of values, the fitting procedure reoptimises and returns a goodness of fit χ 2 (τ 1/IC , τ 2/IET ). The ratio χ 2 (τ 1/IC , τ 2/IET ) χ 2 min is calculated, and a 95% confidence 15 interval for the lifetimes is defined as: 17 where p is the number of parameters used in the global fitting procedure, ν is the number of degrees of freedom and F -1 is the inverse-F cumulative distribution function . An upper bound on the uncertainty for the lifetimes is taken to be the value which satisfies the following two conditions; (i) is the largest deviation from the global lifetimes and (ii), satisfies equation 1.

Transient Absorption Spectra (TAS)
The TAS of the Zn-dPP-pDMA , both fully solvated and assembled are presented below. Almost identical features to the TAS of Zn-dPP are observed and are explained in detail in the main text.  Figure S11 TAS of (a), Zn-dPP-pDMA (3 mg/mL, 250 µM) solvated in dioxane; and (b), Zn-dPP (3 mg/mL, 250 µM) assemebled in water. All TAS are recorded following excitation to S 2 state with 400 nm pump pulse.

Lifetime Uncertainties
Due to the dominant long time delay dynamics, the confidence level of τ 2/IET extending to longer time delays is over estimated by our algorithm. We therefore quoted the error of τ 2/IET as the distance from the origin to the furthest point towards τ 1/IC in the main text (main text, Table 1). The 95% confidence level for each of the systems is highlighted by the bold black lines shown in Figure S12.   Figure S12 Chi ratio (χ 2 (τ 1/IC ,τ 2/IET )/χ 2 min ) for τ 1/IC , τ 2/IET of (a), Zn-dPP solvated in dioxane; (b), Zn-dPP-pDMA solvated in dioxane; and (c), Zn-dPP-pDMA assembled in 18.2 MΩ·cm.