Light-fueled dynamic covalent crosslinking of single polymer chains in non-equilibrium states

While polymer synthesis proceeds predominantly towards the thermodynamic minimum, living systems operate on the reverse principle – consuming fuel to maintain a non-equilibrium state. Herein, we report the controlled formation of 3D macromolecular architectures based on light-fueled covalent non-equilibrium chemistry. In the presence of green light (525 nm) and a bivalent triazolinedione (TAD) crosslinker, naphthalene-containing polymers can be folded into single chain nanoparticles (SCNPs). At ambient temperature, the cycloaddition product of TAD with naphthalene reverts and the SCNP unfolds into its linear parent polymer. The reported SCNP is the first example of a reversible light triggered folding of single polymer chains and can readily be repeated for several cycles. The folded state of the SCNP can either be preserved through a constant supply of light fuel, kinetic trapping or through a chemical modification that makes the folded state thermodynamically favored. Whereas small molecule bivalent TAD/naphthalene cycloaddition products largely degraded after 3 days in solution, even in the presence of fuel, the SCNP entities were found to remain intact, thereby indicating the light-fueled stabilization of the SCNP to be an inherent feature of the confined macromolecular environment.

Wavelength / nm P2 + BisTAD 1 min 3 min 5 min 10 min 20 min Figure S2: UV/Vis spectrum of P2 (0.05 mg mL -1 , blue), BisTAD (0.02 mg mL -1 , red) and the reaction mixture of P2 with 3 eq. BisTAD (black) in acetonitrile. The inset between 400-500 nm provides an indication of the successful removal of the RAFTend-group. Figure S3: UV/Vis spectra recorded during the unfolding process of SCNP2 in acetonitrile (c = 0.05 mg mL -1 ). Note that a quantification of the naphthalene consumption and recovery is not possible through UV/Vis spectroscopy, since TAD absorbs significantly in the same region (compare Figure S2) and is prone to a variety of degradation reactions (compare chapter 1.4). While the naphthalene recovery increases absorption, the solvent impurity dependent TAD degradation of both released and excess TAD decreases the absorption in this region. Wavelength / nm SCNP2 2h 4h 6h 8h 10h 12h 14h 16h 18h 20h 22h 24h Figure S4: UV/Vis spectra recorded during the irradiation process of pure BisTAD with a green LED (10 W, "#$ = 525 nm) in acetonitrile.

Reversible cycloaddition at low concentration
The influence of concentration on the reversibility of the TAD/naphthalene cycloaddition/cycloreversion is not negligible. Whereas higher concentrations of the low molecular weight derivatives N1 and BisTAD lead to a good reversibility and a consistent hysteresis, lowering the concentration (to 0.2 mg mL -1 ) resulted in significant lower amounts of cycloadduct being formed with more TAD consumed in side reaction, thus resulting in a lower extent of reversibility. This is indeed observed from the UV/vis spectra depicted in Figures S5 and S6, which indicate only a minor fraction of TAD being regenerated at the end of each cycle.  0s  10s  20s  40s  60s  120s  180s  300s  600s  1200s  1800s  6000s  6300s  7200s  9900s 11700s 14100s Figure S5: UV/Vis spectra recorded during the cycloreversion process of naphthalene N1 and BisTAD with concentrations four times higher than on the SCNP scale (c = 0.2 mg mL -1 ). Compared to the same experiments with an 18-fold higher concentration (cf. Figure 2, main paper, c = 3.6 mg mL -1 ), the reversibility of the TAD/naphthalene system is considerably lower. Figure S6: Plot of UV/Vis absorbance over three consecutive irradiation and reversion cycles, derived from Figure S5. Left absorbance intensity at λ = 526. 5    Whereas with increasing amount of crosslinker (i.e. from 1 eq. to 3 eq.), an increase in elution time is evident, further increasing the amount of crosslinker (i.e. to 6 and 9 eq. BisTAD) apparently resulted in more inter-chain crosslinking, as evidenced from a broader dispersity and a more pronounced shoulder at shorter elution times. Normalized Intensity Volume / mL P2 1 eq. TAD 2 eq. TAD 3 eq. TAD 6 eq. TAD 9 eq. TAD Figure S11: Normalized UV-THF-SEC trace at 250 nm of P2 (back) and folded SCNP2 with dichloromethane (DCM) and hexane as solvent. While DCM leads to a small increase in molecular weight, hexane and ethyl acetate as a solvent appear not to effect the folding process.

Reversible folding with addition of new BisTAD
As shown in Figure 4 in the main paper, the SCNP can be reversibly folded and unfolded with a slight hysteresis throughout the process. To control whether the hysteresis occurs because of the formation of side products on the polymer or as a result of TAD degradation, additional experiments were carried out. Specifically, the reaction mixture of P1' (1 mL, prepared as described in section 4.2.5) was treated with 1 eq. of new BisTAD (0.025 mg). After purging with argon, the solution was irradiated for 20 min, the solvent was removed and subjected to SEC measurement. The same procedure was applied to the reaction mixture of P1''. Whereas for SCNP1' with new BisTAD no increase in elution time is observed, the resulting SCNP1'' with new BisTAD has a significantly higher elution time, indicating a higher contraction and thus evidencing that the hysteresis most likely can be attributed to TAD-degradation occuring over time.  Figure S13: Chemical structures of plausible TAD-based side products generated through TAD homo-polymerization, hydrolysis (reduction to urazole and amine) or dimerization detected by HR-ESI-MS analysis, which might have formed during consecutive cycloaddition/cycloreversion cycles. Table S1: Masses of plausible TAD-based side products (displayed in Figure S13), identified from the HR-ESI-MS spectrum.  (

ESI-HR-MS analysis of plausible TAD-derived side products
For P1 this leads to: (

Calculation of M1 per polymer chain of P2
Using the percentage of M1 in P2 obtained from the 1 H-NMR spectra and the & ' obtained from the SEC, the average number of M1 units and MMA monomers per polymer chain were calculated as follows: ( For P2 this leads to: ( 1 x 100

1D NMR Measurements
1 H-and 13 C-spectra 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,). All measurements were carried out in deuterated solvents. Resonances are reported in parts per million (ppm) relative to tetramethylsilane (TMS). The δ-scale was calibrated to the respective residual solvent signal. 1 The measured coupling constants were calculated in Hertz (Hz). The spectra were analysed using the MESTRENOVA 11.0 software. The signals were abbreviated as follows: s = singlet, bs = broad singlet, d = doublet, t = triplet, dd = doublet of doublets and m = multiplet.

Diffusion Ordered Spectroscopy (DOSY) NMR
All samples were measured in amber NMR-tubes to prevent exposure to ambient light. DOSY experiments based on 1 H-NMR were performed in THF-d8 at 279.15 K on a Bruker 400 UltraShield spectrometer equipped with a Quattro Nucleus Probe (QNP) with an operating frequency of 400 MHz ( 1 H). A sequence with longitudinal eddy current delay (LED) using bipolar gradients was employed in order to compensate eddy currents. Bipolar gradient δ and a diffusion delay Δ were determined separately for each sample. Gradient strength was linearly incremented from 2% at 0.96 G to 95% at 45.7 G in 32 steps. The obtained data was processed with TopSpin 4.0.6 and Dynamics Center 2.5.3. After Fourier transformation of the 1D spectra, the signal decay along the gradients G was fitted to: Gcd e f e g e hiG g j k • 10 l With the gyromagnetic ratio γ and the full signal intensity I0.
Hydrodynamic diameters DH were calculated from the Stokes-Einstein equation: Where kB is the Boltzmann constant, T the temperature and η the solvent viscosity (THF at 6 °C; extrapolated from literature: 0.572 mPa s). 2

UV-VIS Spectroscopy
UV/vis spectra were recorded at ambient temperature on a Shimadzu UV-2700 spectrophotometer equipped with a CPS-100 electronic temperature control cell positioner. Samples were prepared in THF and measured in Hellma Analytics quartz high precision cells with a path length of 10 mm.

LC-MS Analysis
LC-MS measurements were performed on an UltiMate 3000 UHPLC system (Dionex, Sunnyvale, CA, USA) consisting of a pump (LPG 3400SZ, autosampler WPS 3000TSL) and a temperature-controlled column department (TCC 3000). Separation was performed on a C18 HPLC-column (Phenomenex Luna 5 μm, 100 Å, 250 × 2.0 mm) operating at 40 °C. A gradient of MeCN:H2O 10:90 -> 80:20 v/v (additive 10 mmol·L -1 NH4CH3CO2) at a flow rate of 0.20 mL·min -1 during 15 min was used as the eluting solvent. The flow was split in a 9:1 ratio, where 90 % (0.18 mL·min -1 ) of the eluent were directed through the UV-detector (VWD 3400, Dionex, detector wavelengths 215, 254, 280, 360 nm) and 10 % (0.02 mL·min -1 ) were infused into the electrospray source. Spectra were recorded on a LTQ Orbitrap Elite 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). A constant spray voltage of 3.5 kV, a dimensionless sheath gas and a dimensionless auxiliary gas flow rate of 5 and 2 were applied, respectively. The capillary temperature and was set to 300 °C, the S-lens RF level was set to 68, and the aux gas heater temperature was set to 125 °C.

Irradiation experiments with green light (l max = 525 nm):
P1 (0.05 mg mL -1 ) and BisTAD (0.025 mg mL -1 ) were dissolved in dry acetonitrile (3 mL). The reaction mixture was deoxygenized by purging with argon for 10 min. The vial was placed into a photoreactor ( Figure S19) and irradiated with a green LED provided by Future Eden Ltd.

Naphthalene Monomer (M1)
2-Naphtoyl chloride (5.00 g, 26.2 mmol, 1.0 eq.) was dissolved in dry DCM (50 mL) and cooled to 0 °C. Triethylamine (4.02 mL, 28.9 mmol, 1.1 eq.) was added to the cooled solution, followed by the dropwise addition of 2-hydroxyethyl methacrylate (3.50 mL, 28.9 mmol, 1.1 eq.). The reaction mixture was stirred overnight and was allowed to slowly warm to ambient temperature. Subsequently, 20 mL of an aqueous saturated ammonium chloride solution were added to quench the reaction and the organic phases were separated. The organic phase was dried over MgSO4 and the solvent removed in vacuum. The crude product was purified with flash chromatography using a gradient (silica, ethyl acetate:cyclohexane 0:100 à 15:85). 2-(Methacryloyloxy)ethyl 2-naphthanoate was obtained as a white powder (6.81 g -91%).

Synthesis of methyl 2-naphthoate (N1)
2-Naphtoyl chloride (0.50 g, 2.62 mmol, 1.0 eq.) was dissolved in dry DCM (5 mL) and cooled to 0 °C. Triethylamine (0.69 mL, 5.25 mmol, 2.0 eq.) was added to the cool solution, followed by the dropwise addition of methanol (1 mL). The reaction mixture was stirred overnight and was allowed to slowly warm to ambient temperature. Afterwards, the solution was filtered through silica to remove the base and DCM was removed in vacuum. To the dry residue, methanol (5 ml) and sulphuric acid was added, and the reaction solution was again stirred overnight, followed by the addition of DCM (20 mL) and extraction with water (2x10 mL). The organic phase was dried over MgSO4 and the solvent removed in vacuum. The crude product was purified with flash chromatography using a gradient (silica, ethyl acetate:cyclohexane 0:100 à 15:85). Methyl 2-naphthoate was obtained as a white powder (0.298 g -75%).

Synthesis of P1
2-Cyanopropan-2-yl benzodithioate (15.00 mg, 0.067 mmol, 1.0 eq.), methyl methacrylate (542.80 g, 5.42 mmol, 80.0 eq.), M1 (385.63 mg, 1.36 mmol, 20.0 eq.) and AIBN (2.23 mg, 0.013 mmol, 0.2 eq.) were dissolved in 1.2 mL toluene. The reaction mixture was purged with argon for 10 min and subsequently stirred at 70 °C overnight. The reaction mixture was precipitated from cold methanol, followed by dissolution in THF and precipitation in cold nhexane. Afterwards, the precipitation was repeated in diethyl ether and one more time in nhexane. The polymer was dissolved in a large excess of THF, the solution purged with air and stirred overnight in an open flask to remove the RAFT end groups. 4 The residue was dissolved in THF and precipitated out of cold n-hexane.

Synthesis of P2
2-Cyanopropan-2-yl benzodithioate (30.00 mg, 0.135 mmol, 1.0 eq.), methyl methacrylate (1.22 g, 12.20 mmol, 90.0 eq.), M1 (385.63 mg, 1.36 mmol, 10.0 eq.) and AIBN (4.45 mg, 0.0271 mmol, 0.2 eq.) were dissolved in 1.2 mL toluene. The reaction mixture was purged with argon for 10 min and afterwards stirred at 70 °C overnight. The reaction mixture was precipitated from cold methanol, followed by resolving in THF and again precipitation in cold n-hexane. This was repeated in diethyl ether and again in n-hexane. The polymer was dissolved in a high excess of THF, the solution purged with air and stirred overnight in an open flask to remove the RAFT end groups. 4 The residue was dissolved in THF and precipitated out of cold n-hexane.

General procedure for the folding of P1 with BisTAD
To a solution of P1 (0.3 mg, 0.036 µmol) in 6 mL (c = 0.05 mg mL -1 ) dry acetonitrile BisTAD (0.150 mg, 0.54 µmol, 3 eq.) synthesized according to a literature procedure) 3 was added. The solution was purged with argon for 10 min, followed by irradiation with a green LED for 20 min. After completion, the solution was cooled to 0 °C and the solvent removed under reduced pressure maintaining the cold temperature. The residual solid was dissolved in THF (with 0.05% Toluene as SEC standard) and injected into the SEC without further purification and delay to measure the SCNP at the maximum possible folded state.

General procedure for the folding of P2 with BisTAD
To a solution of P2 (0.3 mg, 0.0268 µmol) in 6 mL (c = 0.05 mg mL -1 ) dry acetonitrile BisTAD (0.125 mg, 0.44 µmol, 3 eq.) synthesized according to a literature procedure) 3 was added. The solution was purged with argon for 10 min, followed by irradiation with a green LED for 20 min. After completion, the solution was cooled to 0 °C and the solvent removed under reduced pressure maintaining the cold temperature. The residual solid was dissolved in THF (with 0.05% Toluene as SEC-standard) and directly injected into the SEC without further purification or queue time to measure the SCNP at the maximum possible folded state. For 1 H-NMR measurements a sephadex LH20 column using acetonitrile as eluent was performed after irradiation and cycloreversion to remove the BisTAD excess. To increase the quality of NMR measurements, Shigemi NMR-tubes for CDCl3 were used. Figure S23: 1 H NMR spectra (in CDCl3) of P2 (bottom), SCNP2 after irradiation with green light for 20 min (middle) and unfolding of SCNP2 after 24h in the dark (top). The dashed red marked area displays the resolved naphthalene signal of the starting material, while the bridge protons of the cycloadducts are highlighted within the green dashed lines. Whereas the starting polymer P2 displays the specific naphthalene pattern and no signals are visible between 6 and 7 ppm, the cycloproduct resonances become notable upon formation of the SCNP, with 45% of naphthalene units consumed (cf. red marked proton). Note that the obtained NMR of the SCNP does not represent the maximum folded state, since the sample preparation after irradiation required more than one hour, hence delaying the NMR measurement, which moreover ran for one more hour. After a total of 24 h in the dark, 96% of naphthalene side chains are present, whereas 3% of cycloproduct remain intact.

Procedure for the light stabilization of naphthalene N1
Naphthalene N1 (2.5 mg, 0.13 mmol, 1.0 eq) and BisTAD (1.5, 0.0054 mmol, 0.4 eq.) were dissolved in deuterated acetonitrile (0.7 mL, dried one week over molecular sieves) and the solution transferred into an NMR-tube. Argon was purged through the reaction solution for deoxygenation and the NMR tube was closed. Under exclusion from light, 1 H-NMR was measured followed by irradiation for 40 min (2 cm distance, 10 W, "#$ = 525 nm). The NMRtube was removed from the light source and 1 H-NMR was measured immediately. After completed measurement, the samples were put back into the setup and irradiation was proceeded. Using the same procedure, 1 H-NMR measurements were undertaken at 16 h, 24 h respectively. The percentage of dimer was calculated from the integration of proton 1 of the staring material N1 relative to the bridge protons 2a and 2b of the Diels-Alder cycloadducts. The overall product mixture can contain three different regio-isomers, as displayed in the reaction scheme. An exemplary 1 H-NMR spectrum, obtained after the first irradiation, is shown here below.

Cycloaddition/cycloreversion of naphthalene N1 with BisTAD
Naphthalene N1 (2.5 mg, 0.013 mmol, 1.0 eq) and BisTAD (1.5, 5.35 mmol, 0.4 eq.) were dissolved in deuterated acetonitrile (0.7 mL, dried one week over molecular sieves) and the solution transferred to an NMR-tube. Argon was purged through the reaction solution for deoxygenation and the NMR tube was sealed. Under exclusion from light, 1 H-NMR was measured followed by irradiation for 40 min (2 cm distance, 10 W, "#$ = 525 nm), leading to a loss of color ( Figure S25). The NMR-tube was removed, and 1 H-NMR was measured immediately. After complete measurement, the samples were covered with aluminium foil and stored at ambient temperature for 24 h, during which the color was regained ( Figure S25). 1 H-NMR was measured afterwards, and the samples were irradiated again for 40 min, followed by the 24 h of darkness. After three consecutive cycles, the sample was stored for 4 additional days in the dark to investigate the final reaction outcome (8% Dimer, see figure  S26). The percentage of dimer to naphthalene N1 was calculated by comparing proton 1 of staring material N1 to the bridge protons 2a and 2b of the Diels-Alder adducts.
A first NMR tube was subjected to green LED irradiation (λ = 515 -525 nm, 3 x 3 W) for 45 min, at which complete photobleaching was observed. Immediately after irradiation, the resulting clear and colorless solution was submitted for 1 H NMR analysis (see Figures S31a for the corresponding spectrum). Two regioisomers CA and CB (CA:CB = 42:58) were identified, i.e. the cycloadduct formed upon addition of TAD onto the substituted and non-substituted ring of N1, respectively. The recovered samples was then kept in the dark (wrapped in aluminium foil) and stored at 25 °C for 60 h, after which a 1 H NMR spectrum was recorded. Integration of the well-resolved signals in the 1 H NMR spectra allowed to determine the concentration of remaining TAD/naphthalene cycloadduct (see Figure 31b).
The second NMR tube was kept in the dark at 25 °C (wrapped in aluminium foil) throughout the cycloreversion study. It hence served as a non-irradiated reference to assess the cleanliness of the cycloreversion process (see Figure S31c). ). Both regioisomers CA and CB, i.e. the cycloadduct formed onto the substituted and non-substituted ring, respectively as well as the complete consumption of TAD can be detected. b) 1 H NMR spectrum (in Me2CO-d6) of the irradiated TAD/naphthalene mixture upon standing in the dark for 60 h, indicating a clean cycloreversion process with less than 7 % cycloadduct remaining. c) 1 H NMR spectrum (in Me2CO-d6) of a N1/4-n-butyl-TAD reference sample, kept in the dark for 60 h without being subjected to green light.
The resulting mixture was submitted for 1 H NMR analysis (see Figures S32a for the corresponding spectrum) and divided over two equal samples, whereby a first was immediately re-subjected to green LED light for an additional 5 hours, while a second one was placed in the dark (wrapped in aluminium foil) hence serving as a reference sample. Whereas the former sample remained colorless throughout the irradiation process, the latter slowly regained a pink color over time.
Both samples were submitted for 1 H NMR analysis (see Figures S31b-c) in order to identify whether an exchange of cycloadducts occurred under continuous irradiation, and thus whether the initial TAD/naphthalene cycloadduct remained their dynamic behaviour. Figure S32: 1 H NMR analysis of the exchange experiment of the methyl 2-naphthoate (N1)/4-n-butyl-TAD system with plain naphthalene under continuous irradiation. a) 1 H NMR spectrum (recorded in Me2CO-d6) of the initial cycloadduct mixture, formed upon 45-min green light irradiation (λ = 515 -525 nm, 3 x 3 W LEDs) of 4-n-butyl-TAD (15 mM, Me2CO-d6) in the presence of methyl 2-naphthoate (1.2 eq.), following the addition of 1.2 eq. plain naphthalene. b) Whereas upon standing in the dark for 5 h, only the cycloreversion of the N1/4-n-butyl-TAD system is observed, c) continued irradiation with green light results in the formation of a new TAD-cycloadduct with plain naphthalene. d) Comparison with a reference spectrum of the latter cycloadduct indeed evidenced the initially formed cycloadduct to retain their dynamic behavior, allowing for an exchange reaction to take place.

General procedure for light-fueled stabilization of P2.
After SCNP2 was obtained (see general procedure for folding 4.2.6) one sample was taken out under argon gas flow for SEC measurement and the residual solution continuously irradiated. This was repeated at 8 h, 48 h and 72 h.

Procedure for UV/Vis measurements of reversible cycloaddition/cycloreversion of naphthalene N1
4.2.11.1 Relatively high concentrations (NMR-concentrations) Naphthalene N1 (5.5 mg, 0.0030 mmol, 1.0 eq.) and BisTAD (3.5, 0.013 mmol, 0.4 eq.) were added to a 10 mL Schlenk flask and dissolved in dry acetonitrile (1.5 mL). Argon was bubbled through the reaction solution for deoxygenation and the sample was irradiated for 40 min (2 cm distance, 10 W LED, "#$ = 525 nm). Under argon flow, 0.1 mL of the now colorless reaction solution was taken out using an Eppendorf pipette and added to a UV/Vis cuvette. The flask was then covered in aluminum foil and stored for 24 h at ambient temperature. 2 mL acetonitrile was added to the cuvette and UV/Vis absorbance was measured over a period of 24 h. After 24 the retrieved pink reaction solution in the Schlenk flask was irradiated again for 40 min and submitted for UV/Vis measurement, following the described procedure. The procedure was repeated for three consecutive cycles.

Diluted concentrations
To investigate the influence of diluted concentration on the reversibility, which is critical for the typically used highly dilute concentrations of SCNP folding, the previously described UV/Vis experiments were repeated at lower concentrations. Therefore, naphthalene N1 (0.40 mg, 0.0022 mmol, 1.0 eq) and BisTAD (0.24, 0.0008 mmol, 0.4 eq.) were dissolved in dry acetonitrile (2 mL) and transferred into a cuvette. The solution was deoxygenated, and UV/Vis was measured, followed by irradiation for 40 min. After irradiation, UV/Vis was measured for 24 h.

General procedure for chemical SCNP stabilization upon reduction
The procedure was carried out following a literature procedure. 5 After the folding process (see general procedure for folding 4.2.5), the dry polymer was dissolved in a small amount of ethyl acetate and cooled to 0 °C. To this, acetic acid and potassium azodicarboxylate (PAD, synthesized according to literature) 6 were added to the solution. The reaction mixture was stirred for 2 h and then allowed to warm up to room temperature for SEC analysis.

General procedure for light-fueled stabilization of P1.
The obtained SCNP (see general procedure for folding 4.2.5) was continuously irradiated and samples were taken under inert gas flow.

General procedure for the kinetic trapping of SCNPs below ambient temperature
The SCNP obtained after irradiation (see general procedure for folding 4.2.5) was immediately covered in aluminium foil to prevent it from being exposed to light and cooled in an ice bath. Subsequently, the sample was transferred into the fridge or freezer, respectively. For sample analysis, the solvent was removed under reduced pressure, maintaining temperatures below ambient temperature, and submitted to SEC measurement.