Metal-free atom transfer radical polymerization with ppm catalyst loading under sunlight

Organocatalytic atom transfer radical polymerization (O-ATRP) is recently emerging as an appealing method for the synthesis of metal-free polymer materials with well-defined microstructures and architectures. However, the development of highly effective catalysts that can be employed at a practical low loading are still a challenging task. Herein, we introduce a catalyst design logic based on heteroatom-doping of polycyclic arenes, which leads to the discovery of oxygen-doped anthanthrene (ODA) as highly effective organic photoredox catalysts for O-ATRP. In comparison with known organocatalysts, ODAs feature strong visible-light absorption together with high molar extinction coefficient (ε455nm up to 23,950 M–1 cm–1), which allow for the establishment of a controlled polymerization under sunlight at low ppm levels of catalyst loading.

The ultraviolet−visible (UV−vis) spectra were obtained using a Perkins Elmer Lambda 900 spectrometer equipped with a PTP-1 Peltier temperature controller and the photoluminescence (PL) spectra were recorded at room temperature on an Edinburgh Instruments, FLS980 spectrometer equipped with a 450 W Xe lamp for excitation and detected by a photomultiplier (PMT R928P). UV−vis measurements were carried out using anhydrous DCM solution at sample concentration of 0.04, 0.08, 0.12 mM respectively. (Transparent cuvette on four sides: 1×1×5 cm 3 ); PL measurements were carried out using anhydrous DCM solution at sample concentration of 0.08 mM. Fluorescence decay measurements were carried out by the time-correlated single photon counting (TCSPC) technique. TCSPC event timer with 1 ns time resolution was used to measure the PL decay. The excitation source was a 340nm pulsed light emitting diode (EP-LED Edinburgh Instruments) of pule width (FWHM) 835.5 ps. The decay time fitting procedure was carried out by using the F980 software (Edinburgh Instruments). Smallest residual values were obtained in the fitting procedure.

Electrochemical measurements
Cyclic voltammetry experiments were carried out with a CHI660 D electrochemical workstation (Shanghai Chenhua Instrument Plant, China) using a one compartment electrolysis cell consisting of a typical glassy carbon working electrode (3 mm diameter), a platinum wire counter electrode, and a Ag/AgCl reference electrode. Before performing electrochemical cleaning, the electrode should be sonicated in ethanol and deionized water for 1~3mins respectively to obtain a clean electrode. There is no graininess on the electrode surface when polishing on a Microcloth polishing fleece coated with 1μM and 0.05μM alumina powder (both purchased from Shanghai Chenhua) and the polishing can be stopped. For Step1, scan CV until no obvious oxidation peak existence. Subsequently, CV was swept in a 1 mM K3[Fe(CN)6]/0.1M KCl solution (sweep rate: 50 mV/s, potential window: -0.1-0.5 V, number of turns: 4). Calculate the potential difference peak potential difference of the last circle of CV, the potential difference is required to be close to 59 mV (optimum: 64 mV, acceptable: 65~72 mV). Specifically, the electrode is a silver wire that is coated with a thin layer of silver chloride and an insulated lead wire connects the silver wire with measuring instrument. The electrode also consists of a porous plug on the one end which will allow contact between the field environment with the silver chloride electrolyte. Saturated potassium chloride is added inside the body of the electrode to stabilize the silver chloride concentration and in this condition the electrode's reference potential is known to be +0.197 V at 25 o C. The measurements were done in 1.0 mM DCM solution with 0.1 M tetrabutylammonium hexafluorophosphate (n-Bu4NPF6, TCI chemicals) as supporting electrolyte at a scan rate of 50 mV/s. The redox potential was calibrated after each experiment against the ferrocenium/ferrocenecouple (Fc + /Fc), which allowed conversion of all potentials to the aqueous saturated calomel electrode (SCE) scale by using E 0 (Fc + /Fc) = 0.42 V vs. SCE in CH3CN.

DFT calculation
We carried out density functional theory, DFT, calculations using the Gaussian09 program package. Geometries optimization calculations were carried out by a meta-GGA hybrid functional PBE0 with 6-31G* basis set for all atoms. Vibrational frequencies were calculated analytically at the same level to obtain the thermodynamic corrections. For details, see: Computational Details.

The setup of photocatalytic ATRP polymerization
Supplementary Figure 1 | Reactors with 6 W Purple LEDs (λmax = 400 nm), 6 W blue LEDs (λmax = 460 nm) and under sunlight irradiation 6 W purple LEDs and 6 W blue LEDs reactors were purchased from GeAo Chemical (see: www.geaochem.com/) and were used as shown above ( Figure S1). All reactions were conducted in a 6 W purple LEDs reactor placed 1 cm from light. At this distance, we estimate the light intensity of 6 W purple LEDs and 6 W blue LEDs to be ~25 mW/cm 2 and ~30 mW/cm 2 respectively.

Experimental determination of excited state reduction potentials
Using photoluminescence maximum and E ox , the excited state reduction potential was estimated for OPCs (Ered (OPC •+ / OPC*)) according to the following equations 3  were turned off and the reaction vial was wrapped entirely in aluminum foil. After same time of dark period, another aliquot was taken for NMR. The lights were turned back on and the sample was irradiated for two hours, at which point an aliquot was taken for characterization. The lights were turned off and the samples were subjected to another two hour dark period. This light on-off cycle was repeated several times until over 80% conversion of the monomer was achieved.

General methods for analysis of kinetics and molecular weight growth
A typical procedure of kinetics experiments were performed in glovebox using a predetermined times after the start of the polymerization as indicated (when the reaction mixture was exposed to light). Specifically, the operation that each aliquot is taken in this manner to ensure no further introduction of air throughout the polymerization. The aliquot was analysed by 1 H NMR spectroscopy to determine the conversion at that time.
After NMR analysis, the sample was dried under air, re-dissolved in DCM and dropt into CH3OH for precipitation, and the Mn and Mw/Mn were analysed by GPC. Analysis of kinetics and molecular weight growth of other catalysts loading can be found in the supplementary details below.

Synthesis of PMMA-b-PBA
A Schlenk tube with a PTFE stirring bar was charged with 0.25 mg of 5d (0.427 ×10 -6 mol, 0.05 eq.) and 28 mg of the PMMA macroinitiator described above (Mn = 4.1 kDa, 1.0 eq.) which were dissolved in 1.50 mL of DCM. Then 1.05 mL of BA were added (7.29 mmol, 1070 eq.), reacted according to the above general polymerization procedure for 9 hours. The resulting polymer was isolated according to the above general polymerization procedure and analyzed. After 9 hours, the reaction mixture as loaded into a syringe and slowly dripped into room temperature methanol to precipitate the polymer. After stirring for 2 h, a yellow oil crashed out, and the solution was placed into a freezer (ca. -20 °C) for 1 h. The methanol was then decanted off and the residual solvent was removed under reduced pressure. This process was repeated once to yield 0.770 g of a yellow oil (80% yield). The resulting PMMA-b-BA) copolymer was found to have Mn = 228 kDa, Ɖ = 1.63 (GPC trace in Figure 4., blue line).

MALDI-TOF analysis of polymer
MMA (1.00 mL, 9.35 mmol, 1000 eq.), DBMM (18.0 μL, 93.5 μmol, 10 eq.), and 5d (2.5 mg, 9.35 μmol, 0.5 eq.) were dissolved in 1.20 mL DCM and reacted according to the above general polymerization procedure for 10 hours. At this time, the reaction was removed, dripped into 150 mL methanol and stirred for 2 h. The resulting precipitate was then isolated by vacuum filtration and washed with excess methanol. The polymer was then re-dissolved in a minimal amount of DCM again and dripped into 100 mL of methanol and stirred for 1 h to fully remove unreacted monomer, initiator or catalyst.
The product was again collected by vacuum filtration and dried under reduced pressure to reveal a slight yellow powder (Mn = 5.20 kDa, Đ = 1.14).

Synthesis of PMMA-b-PBnMA
Inside a glovebox, a Schlenk tube with a PTFE stirring bar was charged with 0.055 mg of 5d (1.0 ×10 -7 mol, 0.005 eq.) (stock solution of 5d: 0.188 μmol/mL in anhydrous DCM was used for accurate addition) and 126 mg of the PMMA macroinitiator described above (Mn= 6.30 kDa, 1.0 eq.) which were dissolved in 1.30 mL of DCM.
Then 0.65 mL of BnMA were added (3.8×10 -3 mol, 191 eq.), reacted according to the above general polymerization procedure for 10 hours. The resulting polymer was isolated according to the above general polymerization procedure and analyzed. After 10 hours, the reaction mixture as loaded into a syringe and slowly dripped into room temperature methanol to precipitate the polymer. After stirring for 4 h, the polymer was collected via vacuum filtration, washed multiple times with excess methanol and dried in vacuum oven until a constant weight at 30 °C to yield 0.477 g of polymer (52% yield).

Synthesis of PMMA-b-PBA
Inside a glovebox, a Schlenk tube with a PTFE stirring bar was charged with 0.020 mg of 5d (0.37 ×10 -7 mol, 0.005 eq.) (stock solution of 5d: 0.188 μmol/mL in anhydrous DCM was used for accurate addition) and 47 mg of the PMMA macroinitiator described above (Mn = 6.30 kDa, 1.0 eq.) which were dissolved in 1.

Computational detail
All of the theoretical calculations were performed in Gaussian09 package. Geometries optimization calculations were carried out by a meta-GGA hybrid functional PBE0 with 6-31G* basis set for all atoms. Vibrational frequencies were calculated analytically at the same level to obtain the thermodynamic corrections. No imaginary frequency was obtained at optimized geometries for all species. The CPCM solvation model using the self-consistent reaction field (SCRF) method with the solvents of acetonitrile was employed to account the solvent effect. The changes in Gibbs free energy are reported in the content. The redox potentials of triplet state were calculated by the energy differences of triplet states and cation radical,