How molecular architecture defines quantum yields

Understanding the intricate relationship between molecular architecture and function underpins most challenges at the forefront of chemical innovation. Bond-forming reactions are particularly influenced by the topology of a chemical structure, both on small molecule scale and in larger macromolecular frameworks. Herein, we elucidate the impact that molecular architecture has on the photo-induced cyclisations of a series of monodisperse macromolecules with defined spacers between photodimerisable moieties, and examine the relationship between propensity for intramolecular cyclisation and intermolecular network formation. We demonstrate a goldilocks zone of maximum reactivity between the sterically hindered and entropically limited regimes with a quantum yield of intramolecular cyclisation that is nearly an order of magnitude higher than the lowest value. As a result of the molecular design of trifunctional macromolecules, their quantum yields can be deconvoluted into the formation of two different cyclic isomers, as rationalised with molecular dynamics simulations. Critically, we visualise our solution-based studies with light-based additive manufacturing. We formulate four photoresists for microprinting, revealing that the precise positioning of functional groups is critical for resist performance, with lower intramolecular quantum yields leading to higher-quality printing in most cases.


Bruker 600 MHz nuclear magnetic resonance (NMR)
1 H and 13 C were recorded on a Bruker System 600 Ascend LH, equipped with a BBO-Probe (5 mm) with z-gradient ( 1 H: 600.13 MHz, 13 C 150.90 MHz).Resonances are reported in parts per million (ppm) relative to tetramethylsilane (TMS).The δ-scale was calibrated to the respective solvent signal of CHCl3 or DMSO for 1 H spectra and for 13C spectra on the middle signal of the CDCl3 triplet or the DMSO quintet.The annotation of the signals is based on HSQC-, COSYand DEPT-experiments.

UV/Vis spectroscopy
UV-Vis spectra were recorded on a Shimadzu UV-2700 spectrophotometer equipped with a CPS-100 electronic temperature control cell positioner.Samples were prepared in acetonitrile with a concentration of 2.18 µM and measured in Hellma Analytics quartz high precision cell cuvettes at room temperature.

LED characterisation
LED emission spectra were recorded using an Ocean Insight Flame-T-UV-Vis spectrometer, with an active range of 200-850 nm and an integration time of 10 ms.LED output energies were recorded using a Thorlabs S401C thermopile sensor, with an active area of 100 mm2 and a wavelength range of 190 nm -20 µm, connected to a Thorlabs PM400 energy meter console.The emitted power from each LED was measured for 60 seconds at a fixed distance from the sensor, after which the mean and standard deviation of the emission could be determined.LEDs were cooled during measurement to minimise any thermal effects on the emission power or sensor performance.

Size-exclusion chromatography coupled with high resolution mass spectrometry (SEC-MS) 1
Spectra were recorded on a Q Exactive Plus (Orbitrap) mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) equipped with an HESI II probe.The instrument was calibrated in the m/z range 74-1822 using premixed calibration solutions (Thermo Scientific) and for the high mass mode in the m/z range of 600-8000 using ammonium hexafluorophosphate solution.A constant spray voltage of 3.5 kV, a dimensionless sheath gas and a dimensionless auxiliary gas flow rate of 10 and 0 were applied, respectively.The capillary temperature was set to 320 °C, the S-lens RF level was set to 150, and the aux gas heater temperature was set to 125 °C.The Q Exactive was coupled to an UltiMate 3000 UHPLC System (Dionex, Sunnyvale, CA, USA) consisting of a pump (LPG 3400SD), autosampler (WPS 3000TSL), and a temperature controlled column department (TCC 3000).Separation was performed on three mixed bead size exclusion chromatography columns (PSS, SDV micro columns 3μm 1000Å 4.6 x 250mm) with a precolumn (SDV micro precolumn 3μm 4.6x30mm) operating at 30 °C.THF at a flow rate of 0.30 mL•min -1 was used as eluent.
The mass spectrometer was coupled to the column in parallel to an UV detector (VWD 3400, Dionex), and a RI-detector (RefractoMax520, ERC, Japan) in a setup described earlier.
[1] 0.27 mL•min -1 of the eluent were directed through the UV and RI-detector and 30 µL•min -1 were infused into the electrospray source after post-column addition of a 50 µM solution of sodium iodide in methanol at 20 µL•min -1 by a micro-flow HPLC syringe pump (Teledyne ISCO, Model 100DM).A 100 µL aliquot of a polymer solution with a concentration of 2 mg•mL -1 was injected into the SEC system.
For experiments using tandem MS, normalised collision energy of 40 was used to fragment the selected mass of the target macromolecule's sodium adduct

Direct laser writing (DLW)
DLW was performed with a NanoScribe Photonic Professional GT with a Ti-Sapphire light source (760 nm) producing 100 fs, 80 MHz pulses.All structures were printed in galvo scan mode using a 63x oil immersion objective (NA 1.4)

Scanning electron microscopy
SEM images were captured using a Tescan MIRA3 scanning electron microscope operating at 3 kV with a beam intensity of 8. Samples were coated with 4 nm platinum prior to imaging.

Nanoindentation
Nanoindentation measurements were performed using a Hysitron TI950 Nanoindentor.A tip with the Berkovich geometry was used to perform all measurements.All measurements were carried out under depth control.

10 W, 445 nm LED Emission Spectrum
Supplementary Fig. 1.Emission profile for the 445 nm 10 W LED

Synthesis of 6-hydroxyhexanoic acid (1)
6-Hydroxyhexanoic acid was synthesised as described in the literature. 2,3Sodium hydroxide (4.33 g, 108 mmol, 1.2 eq.) was dissolved in 200 mL Milli-Q water.ε-caprolactone (10 mL, 90 mmol, 1.0 eq.) was added and stirred overnight.Subsequently, concentrated hydrochloric acid (32% in water, 13 mL) was added to the solution to acidify to pH = 1.The aqueous solution was extracted with ethyl acetate (3 x 150 mL)the combined organic phases were dried over anhydrous sodium sulphate and the solvent was removed under reduced pressure to give 8.6 g (72% yield) of a clear colourless oil that crystallised into white crystals upon refrigeration.The crude product was used without further purification.Supplementary Fig. 5. 1 H NMR spectrum of (3) in d6-DMSO

Synthesis of (3-hydroxypropoxy)pyrene-chalcone (4)
(3-Hydroxypropoxy)pyrene-chalcone was synthesised as described in the literature. 43 (5.5 g, 24 mmol, 1.0 eq.) and 1pyrenecarboxaldehyde (6.2 g, 27 mmol, 1.1 eq.) were suspended in 65 mL ethanol.Aqueous sodium hydroxide solution (3M, 20 mL, 2.5 eq.) was added, and the solution stirred rapidly in the dark overnight.The slurry was poured into 1 L water, then 20 mL of concentrated hydrochloric acid was added and stirred for 5 minutes.The slurry was filtered and the crude product recovered as a yellow solid.The product was purified by silica gel chromatography eluting 5 → 20% acetone in toluene 7.5g of yellow solid were recovered (70% yield).

Synthesis of acetonide-2,2-bis(methoxy) propionic acid (5)
Acetonide-2,2-bis(methoxy) propionic acid was synthesised as described in the literature. 52,2bis(hydroxymethyl)propionic acid (10 g, 75 mmol, 1 eq.), acetone dimethyl acetal (13.8 mL, 112 mmol, 1.5 eq.), and para-toluenesulphonic acid (0.7 g, 4 mmol, 0.05 eq.) were dissolved in acetone (50 mL) and stirred for 2 h under ambient conditions.Then, triethylamine (0.6 mL, 4.4 mmol, 0.06 eq.) was added.The volatile organics were removed under reduced pressure, leaving a white solid as the crude product.The crude product was subsequently dissolved in dichloromethane (100 mL) and washed with water (2 x 20 mL), the organic fraction was dried over magnesium sulphate, and after solvent evaporation, 9.73 g (75% yield) of a white crystalline solid were recovered.extracted with 1 M hydrochloric acid (2 x 20 mL) and brine (1 x 20 mL).The organic phase was dried over sodium sulphate and the dichloromethane removed to give a yellow solid that was subsequently suspended in a slurry of DOWEX resin (500 mg) in methanol (50 mL) and was stirred for 4 h at 50 °C.The slurry was filtered and washed with dichloromethane until all yellow solid had passed through the filter.The crude solution was extracted with NaHCO3 (2 x 20 mL), and brine (1 x 20 mL).The organic phase was dried over sodium sulphate and the solvent removed under reduced pressure.Finally, (6) was isolated by column chromatography eluting 1 → 5 % methanol in dichloromethane to yield 512 mg (81% yield) as a yellow solid.

Synthesis of (butyric acid)pyrene chalcone (8)
(butyric acid)pyrene chalcone was synthesised by an adaptation of literature procedures. 4Ethyl 4-(4-acetyl-2methoxyphenoxy)butanoate (4.47 g, 15.95 mmol, 1.0 eq.) and 1-pyrenecarboxaldehyde (4.04 g, 17.54 mmol, 1.1 eq.) were suspended by sonication in a mixture of 3 M sodium hydroxide (13.29 mL, 39.87 mmol, 2.5 eq.) and ethanol (45 mL).The suspension was stirred rapidly in the dark, under argon for 48 h at room temperature.The slurry was subsequently poured into milli-Q water (900 mL) to which concentrated hydrochloric acid (7.5 mL) was added and stirred for 30 minutes.The solid was isolated by filtration, and the cake was dried under reduced pressure at 40°C for a week.The crude product was purified by recrystallization from toluene, using drops of methanol as an antisolvent.5.26 g of a yellow powder was obtained (71 % yield).Supplementary Fig. 19. 1 H NMR spectrum of (8) in d6-DMSO

Iterative sequential growth procedures 2.4.1 General procedure for addition steps
The diol (either (6) for the first step, or the deprotected oligomer for subsequent steps, 1 eq.), (2) (4 eq.), 1-ethyl-3-(3dimethylaminopropyl)carbodiimide (8 eq.) and 4-dimethylaminopyridine (1 eq.) are dissolved in the minimum amount of dichloromethane under argon and stirred for 4 h, or until TLC indicates complete consumption of the starting material.The dichloromethane solution is subsequently extracted with equal volumes of 1 M hydrochloric acid twice and then brine.The organic layer is dried over sodium sulphate and the solvent removed under reduced pressure.The crude mixture is used in the next deprotection step without further purification.

General procedure for deprotection steps
The chain-end protected oligomer (1 eq.) is dissolved in a 1 M tetrabutylammonium fluoride solution in THF (10 eq.) under argon.The solution is stirred for 2 h, or until TLC shows complete consumption of the starting material.Then, it is dissolved in a 3:1 volume:volume mixture of diethyl ether:dichloromethane and extracted with an equal volume of saturated ammonium chloride solution twice, followed by saturated sodium bicarbonate solution twice, and finally a single wash with brine.The organic phase is dried over sodium sulphate and the solvent removed under reduced pressure.The crude product is used in the next addition step without further purification or purified by silica gel chromatography eluting 5 → 20 % acetone in toluene before chain-end functionalisation with (8).

Chemical structures of T0-5
Supplementary Fig. 49.conversion versus number of photons plots for each of the three molecules as monitored Supplementary Table 3. Tabulated quantum yield data for each macromolecule.

Molecular dynamics simulations
Simulation parameters for the central pyrene-chalcone (PyChal, ATB molid: 1423694), terminal PyChal (ATB molid: 1423693) and methyl 6-hydroxyhexanoate linker (ATB molid: 1423695) subunits of T0, T1, T3 and T5 were developed individually using the Automated topology Builder 7 and compatible with the GROMOS 54A7 molecular dynamics force field 8 for each subunit.Each subunit was capped at its functionalization points to incorporate an overlap region of neighboring units, at least 3 bonds deep.Atomic coordinates and molecular topologies for the complete macromolecule were developed in Chimera 1.17.1 9 by computationally ligating the appropriate number of monomer units at their overlapping functionalization points to reproduce each experimental trifunctional molecule.In T1, T3 and T5, the two terminal PyChal are connected to the central PyChal through linker fragments.In contrast, the central PyChal of T0 is directly connected to the two terminal PyChal units.
All simulations were prepared and performed using GROMACS 2023 10 in conjunction with the GROMOS 54A7 forcefield. 8Four different simulations systems were prepared, corresponding to the macromolecules T0, T1, T3 and T5.
In each simulation, a single macromolecule was placed in a cubic box and solvated in acetonitrile (ATB molid: 913062).The box was of sufficient size such that the polymer was at least 1.0 nm from any box edge at the start of the simulation.Each simulation system was energy minimized using a steepest descent algorithm 11 and equilibrated under NPT condition using 1 fs and 2 fs timestep for 1 ns in each case.During the equilibration process, the pressure was maintained at 1 bar using a Berendsen barostat 12 with an isothermal compressibility of 4.5x10 -5 bar; and the temperature was maintained at 300 K using the Bussi-Donadio-Parrinello velocity-rescaling thermostat 13 with a coupling constant of 0.1 ps.Non-covalent interactions were calculated with a 1.0 nm cutoff in all simulations.Following equilibration, 500 ns production simulations were carried out in triplicate using a 2 fs timestep.
Analysis was performed on trajectory frames spaced at 200 ps intervals.The pairwise distance between photoactive groups in each arm were calculated using the first carbon atom of the double bond in each PyChal arm (i.e., the double bond carbon atom closest to the carbonyl moiety, shown in SI Figure 8.2) and the corresponding carbon atom in the opposing PyChal arms.
To determine the relative populations of specific conformations, the replicate trajectories for the T0, T1, T3 and T5 simulations, respectively, were clustered using the Gromos clustering algorithm.The Gromos algorithm uses an RMSD cut-off value to count the number of neighbors for the conformation sampled in each frame of the trajectory.The structure with the largest number of neighbors is then chosen as the central conformation of the first cluster, and all of its neighbors are included in this cluster.This first cluster is then removed from the pool of conformations and clustering continues until all conformations have been assigned to a conformational cluster.The most populated cluster contains the conformation with the largest number of neighbors, and is the predominant cluster sampled, representing at least 86 % of the trajectory.Due to the size differences between T0, T1, T3 and T5, the RMSD cut-off values were adjusted to maintain the cluster population size of the predominant conformation for each macromolecule.This corresponded to RMSD cutoff values of 0.70, 0.85, 0.98 and 1.15 nm for the T0, T1, T3 and T5 systems, respectively.
Plots of the simulation data were generated using Matplotlib 3.5.2. 14Trajectory visualization, analysis and image rendering were done using VMD version 1.9.4. 15