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

Organocatalytic atom transfer radical polymerization (O-ATRP) is recently emerging as an

1 Key Laboratory of Molecule Synthesis and Function Discovery (Fujian Province University), State Key Laboratory of Photocatalysis on Energy and Environment, College of Chemistry, Fuzhou University, Fuzhou 350108, China. 2 College of Chemistry and Molecular Engineering, Zhengzhou University, Zhengzhou 450001, China. 3Grubbs Institute, Southern University of Science and Technology, Shenzhen 518055, China. 4Beijing National Laboratory of Molecular Science (BNLMS), Beijing 100190, China.* E-mail: shliao@fzu.edu.cn 1

Metal-free atom transfer radical polymerization: reaching ppm catalyst loading under sunlight
Qiang Ma, 1 Jinshuai Song, 2 Xun Zhang, 1 Yu Jiang, 1 Li Ji, 3  Since the discovery in the 1990s, atom transfer radical polymerization (ATRP) has evolved into one of the most versatile and utilized polymerization methods for the synthesis of polymer materials with well-defined structures and architectures, and widely employed in a variety of industrial applications including coatings, adhesives, cosmetics, inkjet printings, etc. [1][2][3] However, conventional ATRPs rely on transition metal catalysts [i.e., Cu(I), Ru(II)], 3 which will result in transition metal contaminations in the final products, and thus raise concerns when applied to fields sensitive to metal contaminants.9][20][21] Until now, several frameworks/core structures 21 including phenothiazine (1), 15,22,23 dihydrophenazine (2), [24][25][26] phenoxazine (3), 27,28 etc. [29][30][31][32][33][34][35] have been successfully identified as efficient photocatalysts for O-ATRP (Figure 1a).However, a 1000ppm level of catalyst loading was typically required to reach a satisfactory control over the polymerization.Controlled polymerization with narrow dispersity at a practical low catalyst loading (<10 ppm) could not only eliminate the need of residual catalyst removal or further product purification, but could also decrease the cost of commercial production. 13,2623,26 According to the oxidative quenching mechanism (Figure 1c), organic photoredox catalysts possessing a highly reducing excited state are required to reduce the alkyl bromides via a single electron transfer (SET) to initiate the polymerization. 367][38][39] Whereas, many mechanistic aspects in O-ATRP remain unclear so far, and there is still a lack of general guidelines for the catalyst design. 21,39othiazine, 15 dihydrophenazine, 24 and phenoxazine 27 are among the most efficient and widely used catalyst frameworks that could achieve a well-controlled O-ATRP polymerization with low dispersity. 21To further increase the catalyst efficiency, much effort has been dedicated to the modifications of the catalyst structures such as introducing aryl groups, which could enhance the visible light absorption, but the red shift of absorption maximum (λmax) was often quite limited. 20,21.27Therefore, we questioned whether there is a possibility to find a suitable and tuneable O-ATRP catalyst framework with their absorption maximum locating in the visible light region, and a stronger light absorption could probably lower the catalyst loading.The three photocatalysts (1-3) can be recognized as derivatives obtained by modifying the corresponding core chromophoric structures: phenothiazine, dihydrophenazine, and phenoxazine, respectively (Figure 1a).Recently, we were thinking heteroatom doping 40,41 of small polycyclic arenes might be a feasible catalyst design logic for the O-ATRP photocatalyst development (Figure 1b).
The integration of heteroatoms such as N, O, S etc. into aromatic hydrocarbons could regulate the photophysical and photoredox properties of these polycyclic arene catalysts 36,40 and thus probably lead to an improvement in their catalytic performance on polymerization and suppress underside catalyst decomposition or catalyst-initiation. 16,29,30 With this idea in mind, we thus decided to practice this heteroatom-doping strategy on graphene, and narrowed down the parental pi system to antratherene (Figure 1b, left), which possesses the desired strong absorption in visible light region. 425][26] Herein, we wish to report our efforts toward this goal, and the discovery of oxygen-doped antratherene (ODA) as a new organocatalyst framework for O-ATRP, which exhibits strong absorption at the visible light region (ε450 nm > 20000 M -1 cm -1 ), and allows for a ppm/sub-ppm level of catalyst loading only to deliver a metal-free, controlled polymerization under visible light or even sunlight (Figure 1d).
To improve the catalyst solubility, we synthesized the n-butyl substituted catalyst 5b, while 5c and 5d were prepared to examine the influence of aryl substituents (for synthetic procedures and spectra, please see the Supplementary Information).As anticipated, 5a did show a strong absorption in the visible light region (Figure 2b), with a red shift of the absorption maximum from 437 nm to 443 nm (comparing with the parental antratherene 42 ).This absorption profile is in sharp contrast to that of the known photocatalysts such as 1-3, which possess a smaller conjugation in the core structure and absorption maximums appearing below 400 nm.Both the absorption maximums (λmax = 454 and 455 nm) of 5c and 5d shown a further red shift (ca.12 nm) as well as enhanced light absorption (εmax = 22580 and 23950 vs 17450 M -1 cm -1 ).In addition, a notable feature of the ODA catalysts substantially different from the known high performance O-ATRP photocatalysts 23,26 is the lack of a charge-separation in their excited states (Figure 2c).Based on the fluorescence emission and the cyclic voltammetry (CV) data, we could assess the reducing capability of the excited singlet states (E 0 *(PC • + / 1 PC*) vs SCE) of these catalysts, ranging from -1.76 V to -1.84 V vs SCE (Figure 2d).Density functional theory (DFT) was used to estimate the reduction potentials of the triplet excited-states (E 0 (PC • + / 3 PC*), which are in the range of -1.58 to Although the excited states of ODA 5a-5d are less reductive than that of N-phenyl phenothiazine 15 but all are more negative than -0.7 V that is required to reduce the initiator EBP (ethyl α-bromophenylacetate, E 0 (EBP/EBP •-) = -0.74V SCE). 24bly, the radical cations (PC •+ ) of ODA catalysts are much more oxidizing (up to +0.82 V vs. SCE), in comparison with the most efficient dihydrophenazine-based photocatalyst (+0.38 V vs.We conducted the initial evaluation of these oxygen-doped catalysts in the polymerization of methyl methacrylate (MMA) by using EBP as initiator, dimethylacetamide (DMA) as solvent under the irradiation of purple light-emitting diodes (LEDs, λmax 400 nm).To our delight, 5a could afford a controlled polymerization with a moderate dispersity (entry 1).The polymerization can also be conducted with control in other solvents, such as dimethylformamide (DMF, entry 2), dichloromethane (DCM), toluene, tetrahydrofuran, etc. (Supplementary Table 2).DCM gave the lowest dispersity (Đ = 1.25) together with a good agreement between the experimental and theoretical Mn (entry 3).Remarkably, ODA 5a is also effective with other initiators such as ethylbromopropanoate (EBrP), ethyl-2-bromoisobutanoate (EBiB), diethyl 2-bromomalonate (DBM) and diethyl 2-bromo-2-methylmalonate (DBMM) (Đ 1.19-1.27,entries 4-7).5b-5d with different substituents were then compared under the standard conditions by using EBP as the initiator.

SCE). 26
Pleasingly, all the three new catalysts are effective for the polymerization (entries 8-10), and ODA 5b could achieve a dispersity (Đ) lower than 1.20 (entry 8).Catalysts 5b and 5d were further examined with DBMM (entry 12 & 13), and a remarkable narrow dispersity (Đ = 1.12, entry 12) was obtained with 5d.Irradiation with lower energy blue LEDs was also effective (entry 13 & 14).A prominent feature of the O-ATRP is that the polymerization can be regulated by light. 46To examine the temporal control ability of the current system, a light on/off experiment was performed with light on-off cycle repeating for several times until over 90% conversion was achieved.As depicted in Figure 3a, the polymerization only preceded in the presence of light irradiation, while no polymerization was observed in dark.This strict control by light over the whole process indicates an efficient activation and deactivation mechanism of the polymerization.
The polymerization was then followed by 1 H NMR to gain some insights about the polymerization kinetics, which unveiled a first-order kinetics through the course of the reaction (Figure 3b).The Mn,GPC values were plotted against the monomer conversion, as shown in Figure 3c, and a linear increase of molecular mass and narrow dispersity throughout the polymerization process were observed.Notably, the y-intercept of the number average molecular weight (Mn) versus conversion plot was 870 Da, suggesting the control of the polymerization was achieved after the initial addition of ~6 MMA (Figure 3c). 24This strict control may benefit from a more efficient deactivation process 25,35,43 due to the radical cations of ODA photocatalysts possessing a better oxidizing ability.
Regarding the chain-end fidelity, a PMMA sample obtained via 5d-mediated metal-free ATRP was subjected to MALDI-TOF analysis (Supplementary Figure 10), which shows a consistence of observed molecular weight with the expected values for the individual PMMA polymers, based on the presence of the initiating unit at one chain end and a bromine atom at the propagating chain end with each peak separated by the mass of one monomer unit (100 Da).An advantage of high chain-end fidelity is it enables the synthesis of block polymers.As shown in Figure 4, chain extension (PMMA-b-PMMA) and block copolymerization products (PMMA-b-PBnMA and PMMA-b-PBA) can all be prepared with this catalytic system.Notably, the GPC traces, all clearly show an obvious shift to higher molecular weight species with little tailing in the homo-polymer regime, giving further support to the high alkyl bromide chain-end fidelity in the PMMA macroinitiators (purified via re-precipitation from methanol) and also a high re-initiation efficiency.
Of note, this result is also in consistence with the high initiator efficiency observed in the polymerizations with freshly distilled DBMM (Supplementary Table 8).Moreover, triblock synthesis is also viable with this system as demonstrated with the preparation of PMMA-b-PBnMA-b-PBA (see Supplementary Figure 13).
The high molar extinction coefficiency (εmax upto 23950 M -1 cm -1 ) of ODA catalysts at the visible light region encouraged us to examine their performance at extremely low catalyst loadings.
As shown in Table 2, decreasing the catalyst loading from 500 ppm to 10 ppm, both photocatalyst 5a and 5d could maintain their catalytic efficiency, giving a controlled polymerization with low dispersity (entries 1-3).To our delight, in the present system, very narrow dispersity can be achieved with 10 ppm 5d only (entry 3).Remarkably, the catalyst loading of 5d can be further decreased from 10 ppm to 0.1 ppm (entry 3-6), and even to 50 ppb by more than two orders of magnitude (entry 7), which is among the lowest catalyst loadings so far in O-ATRP. 26It is worth mentioning that a catalyst loading less than 10 ppm could be very meaningful, which may eliminate the need of the catalyst removal process for industrial and biomedical applications.Importantly, white polymer products can be obtained at the catalyst loading less than 50 ppm, even though the photocatalysts are normally colored compounds (Figure 1d).The high catalytic performance may benefit from their strong visible-light absorption (5d, ε455nm = 23950 M -1 cm -1 ).In fact, the visible light absorption is approximately one order of magnitude stronger than that of fac-Ir(ppy)3 (ε458nm ), which is a commonly used organometallic photoredox catalyst. 47Remarkably, sunlight is also suitable to drive the polymerization with 10 ppm catalyst only (entry 8 & 9).To the best of our known, 10 ppm represents the lowest catalyst loading achieved so far in controlled O-ATRP under sunlight.
Furthermore, other methacrylate monomers such as 2,2,2-trifluoroethyl methacrylate (TFEMA) and benzyl methacrylate (BnMA) can also be polymerized with low dispersity at 10 or 50 ppm The polymerizations were performed under standard conditions.Conversions were measured by 1 H NMR. Mn and Ɖ were determined using GPC with PMMA standards.Ɖ in parentheses were determined using GPC coupled with MALS.Purple LEDs (λmax 400 nm); Blue LEDs (λmax 460 nm).Sunlight experiments were performed outside in a sunny winter day at Fuzhou University (26 °N, 119 °E, 25 ℃).
catalyst loading (entries 10-12).Surprisingly, ODA catalysts could also deliver a controlled polymerization with low dispersity in the polymerization of n-butyl acrylate (entry 13 & 14, also see Supplementary Table 6).It is worth mentioning that acrylate monomers proved to be very challenging to achieve a controlled polymerization with low dispersity by O-ATRP methods due to their high polymerization rate. 24,26,43To our delight, this control on low dispersity is even better than that with dihydroacridine photocatalysts which were developed especially for this type of monomers very recently. 43In addition, ODA photocatalysts could also mediate the reversible addition-fragmentation chain transfer (RAFT) polymerization [48][49][50] of MMA at low ppm catalyst loading (1-10 ppm) under sunlight, with excellent control and narrow dispersity (Đ < 1.10, entries 15-17), which are comparable to the precious iridium photocatalyst 51 (Supplementary Table 7).
Notably, ODA 5d also represent a rare example of organic photocatalysts that could promote the polymerization of both MMA and BA, as well as RAFT polymerization of MMA with good controls. 43,49ummary, a catalyst design logic based on heteroatom-doping of polycyclic arenes has been successfully introduced into the development of new photocatalysts for organocatalyzed atom transfer radical polymerization (O-ATRP).Oxygen-doped antratherene (ODA) has been identified as a novel catalyst framework with unprecedented high efficiency, featuring its strong visible light absorption and excellent performance at extremely low catalyst loadings in both organocatalyzed ATRP and RAFT polymerization.This framework also represents the first example of non-Nheterocyclic organocatalysts lacking a CT character that could mediate the atom transfer radical polymerization with good control and low dispersity (Ɖ < 1.20) at a ppm level of catalyst loading.
We anticipate that this new class of photoredox catalysts will find further applications in polymer synthesis and other photocatalysis-related fields.
The polymerization was conducted with MMA (1.0 mL, 9.35 mmol, 100 eq.) as the model monomer, DBMM (18 μL, 93.5 μmol, 1.0 eq.) as the ATRP initiator, organic photocatalyst (4.70 μmol, 0.5 eq.) and DCM (1.0 mL) as the solvent in a Schlenk tube with a PTFE stirring bar.The mixture was deoxygenized by freeze-pump-thaw cycle three times, backfilled argon and sealed up subsequently.And then the polymerization was occurred under purple LED or blue LED or sunlight irradiation at room temperature.
After the desired time, the tube was opened under argon and 20.0 μL of mixture were syringed out and quenched into CDCl3 containing 250 ppm BHT to determine the monomer conversion by 1 H NMR. The reaction mixture was then diluted with 0.5 mL dichloromethane and dissolved completely, then dripped into 75 mL methanol and stirred for 2 hours.The precipitates was then collected by suction filtration with a Buchner funnel and dried in vacuum oven until a constant weight at 30 °C to give the purified polymers.

Figures
Figure 1 The

Supplementary Files
This is a list of supplementary les associated with this preprint.Click to download. OATRPSuppInfoMQ.pdf Saihu Liao 1,4,* 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 led to the discovery of oxygendoped antratherene (ODA) as a new class of 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 23950 M -1 cm -1 ), which allow for the establishment of a controlled polymerization under sunlight at low ppm levels of catalyst loading.

Figure 1 |
Figure 1 | The development of catalysts for O-ATRP.a, Catalyst development via core structure modification.b, O-ATRP photocatalyst design based on a heteroatom-doping logic (this work).c, Catalytic cycle for a light-mediated O-ATRP.d, Polymerization under sunlight and the product color (PMMA).

Figure 3 |
Figure 3 | Temporal control and kinetic study on ODA 5d-catalyzed ATRP of MMA.a, Light on/off experiments and the plot of monomer conversion versus time.b, Kinetic plot for the metal-free ATRP.c, Plot of Mn and Đ versus monomer conversion for the polymerization of MMA under continuous irradiation.d, GPC traces of each polymer depicted in c (color coded).All polymerization were performed at a [MMA]0/[DBMM]0/[5d]0 ratio of 200/1/0.05under purple light irradiation.

Figure 4 |
Figure 4 | Block polymer preparation.Left: chain-extension from a PMMA macro-initiator (black) to produce block copolymers after a further polymerization with MMA (green), BnMA (red), and BA (blue).Right: GPC traces of the corresponding polymers depicted in left by using the same color coded.
development of catalysts for O-ATRP.a, Catalyst development via core structure modi cation.b, O-ATRP photocatalyst design based on a heteroatom-doping logic (this work).c, Catalytic cycle for a lightmediated O-ATRP.d, Polymerization under sunlight and the product color (PMMA).

Figure 4 Block
Figure 4