Supported Cu0 nanoparticles catalyst for controlled radical polymerization reaction and block-copolymer synthesis

The synthesis of Cu0 nanoparticles on different supports and their activity in controlled living radical polymerization processes is presented. The type of support influences the final size of the copper nanoparticles as well as their adhesion to the support. These aspects have a direct influence on the characteristics of the polymers obtained. The best results were obtained for SiO2 particles, which afforded a good molecular weight distribution (Mw/Mn = 1.25). The activity, recovery and recycling of the catalyst was explored for ultrafast polymerization reaction of butyl acrylate. Further, the terminal bromine reactivity was used for the synthesis of a block poly(n butyl acrylate-block-styrene). The influence of ligand type on the control of the reaction was studied. Also, a straightforward polymerization procedure without any ligand afforded a polydispersity value of 1.38.

This study presents the synthesis of catalysts based on copper nanoparticles immobilized on different supports such as silica, titania and alumina employed in the initiation of controlled polymerization reactions. The novelty of our approach consists in the use of ex situ synthesized supported copper nanoparticles with high surface area and demonstrated reusability. The size of copper nanoparticles and support type controls the efficiency of the immobilization on the support. These parameters affect also the polymerization reaction by modifying both the molecular weight attained and polydispersity index. The activity, recovery and reuse of the catalyst was explored for butyl acrylate ultrafast polymerization reaction. Standard polymerization reaction conditions involving Me 6 TREN as ligand agent (without Cu II salts), were used to study the catalysts. The reaction in the absence of any ligand was also studied. The catalyst, and synthesized polymer were employed the synthesis of a block-copolymer. The kinetics of the ligand free polymerization reaction was studied in order to calculate the k p app and to establish whether the process has a "living" characteristic.

Methods
Nanoparticles synthesis. The SiO 2 particles were synthesized according to a modified Ströber method 36 .
Briefly, 0.22 g CTAB were dissolved in 90 mL H 2 O, followed by the addition of 0.7 mL NaOH 2 M. The solution was stirred at 80 °C for 30 minutes prior to the addition of 0.9 mL of TEOS, and then stirring at 80 °C was continued for 2 h.
The CuNPs/SiO 2 catalyst was prepared by the dissolution of (0.15 g or 0.3 g) CuSO 4 •5H 2 O in the SiO 2 dispersion (90 mL) which was further subjected to sonication (using an ultrasonic processor Heischler UP50H) prior (for 20 minutes) and during the reduction process (for 30 minutes) which was performed by using an equimolecular amount of ascorbic acid. The mixture was then stirred at 80 °C for 2 h. The catalyst was separated through centrifugation, washed with water several times and dried under vacuum. The amount of CuSO 4 •5H 2 O was varied in order to obtain two CuNPs/SiO 2 catalysts containing 18%, respectively 36% (weight %) Cu 0 .
The CuNPs/Al 2 O 3 and the CuNPs/TiO 2 were prepared by dispersing 0.25 g of support (aluminum oxide 90 active neutral ; or titanium dioxide P25) in 90 mL H 2 O with 0.1 g CTAB and 0.7 mL NaOH 2 mol/L. To the support dispersion were then added 0.15 g CuSO 4 •5H 2 O. This solution was sonicated for 20 minutes prior to the addition of ascorbic acid (stoichiometric vs copper salt) and continued for another 30 minutes. The mixture was stirred for 2 h at 80 °C. Then, the catalyst was separated through centrifugation, washed with water several times and dried under vacuum.
The polymerization procedure was based on literature reports for ultrafast SET-LRP reaction 25,30 . Briefly, in a reaction vial were introduced 4 mL of BA and 1. For the recovery of the catalysts, the polymerization reaction mixture was centrifuged, the catalyst being then dispersed in DMSO and centrifuged several times, and dried under vacuum.
The block copolymer synthesis was realized by the addition of 1 mL styrene and 3 mL DMSO to the BA polymerization reaction mixture. Nitrogen was introduced in this new mixture and the reaction was activated by addition of 3 mg of NaBH 4 . The reaction mixture was stirred for 2 h, after which it was precipitated in methanol, filtrated and dried until a constant mass was attained.
Characterization. The molecular weights of the resulted polymers and oligomers were analyzed using PL-GPC 50 Integrated GPC/SEC System (Agilent Technologies) using a 1 mL/min THF flow rate and a column oven temperature of 30 °C using polystyrene as standard. The GPC traces were treated according to Gavrilov et al. 37 .
The high-resolution transmission electron microscopy (HRTEM) studies were performed on an atomic resolution analytical JEOL JEM-ARM 200 F electron microscope, operating at 200 kV. Specimens for HRTEM were prepared in the following way: a drop of the solution was put on holey carbon TEM grid and dried at 100 °C for 5 min.

Results and Discussion
Currently there is still a real controversy regarding the atom transfer radical polymerization (ATRP) between the single electron transfer living radical polymerization SET-LRP 25,38 and supplemental activator and reducing agent SARA-ATRP [39][40][41][42] .
This study presents an ATRP polymerization process starting from CuNPs deposited on different supports with or without the presence of complexing amine species. The aim is to present a catalyst that can be recycled and to bring more information related to the mechanism of the reaction.
Considering that the SET-LRP 26 mechanism implies Cu 0 species as the activating agent, the first approach of this study consisted in the synthesis and characterization of a catalyst comprised of copper nanoparticles generated on the surface of SiO 2 particles. The aim of the synthesis is to obtain a catalyst with high surface area able to initiate a SET-LRP process and also allows its recovery, re-activation and re-utilization.
In order to characterize the morphology of the CuNPs/SiO 2 catalyst, transmission electron microscopy (TEM) analyses were employed. In Fig. 1, one can observe that on the surface of the SiO 2 particles with ellipsoidal shapes, and sizes around 100 nm, the Cu nanoparticles with dimensions smaller than 10 nm are uniformly spread. Some of the silica particles present the characteristic mesoporous structure.
In order to confirm the oxidation state of the CuNPs, Fig. 2 shows a region of the specimen where larger Cu nanoparticles where also present. Here it was possible to obtain an electron diffraction pattern (insert of Fig. 2) which was indexed with the metallic copper (Cu 0 ) structure. Therefore, the nanoparticles are in a reduced state. This aspect suggests that we can use the catalyst for the initiation of the polymerization reaction in a SET-LRP mechanism.
As presented in the literature 25,30,38,43 , the polymerization initiation requires activation with NaBH 4 for the generation of the SET-LRP active species. The high specific surface area of our catalyst compared to literature examples such as Cu wire, allows a fast reaction. Therefore, a conversion of 90% is reached in 30 minutes. In the SET-LRP mechanism, the dominant reactions are the activation of alkyl halides by Cu 0 , the deactivation of radicals by Cu II and the disproportionation of Cu I species to regenerate Cu 0 and Cu II . In the SET-LRP reaction,  there is minimal activation of alkyl halides by Cu I , due to an instantaneous disproportionation, and negligible deactivation of radicals by Cu I or comproportionation.
The use of a catalyst with high surface area, easily separable and reusable can facilitate the scale-up of the process. To this end, we have recovered the catalyst after the first reaction and recycled it in another polymerization step under the same conditions. The GPC analyses of the first and second BA polymerization experiment are presented in Table 1 and Fig. 3.
The comparison of the molecular weights obtained reveals only a slight change of the values for the second experiment which afforded a lower molecular weight and a larger polydispersity (PD). This change can be explained by the increase of the surface area due to the consumption of Cu 0 which leads to higher number of active sites. Thus, the lower molecular weight and the increased polydispersity of the second experiment, can both be explained by an increased number of active sites in the case of the recovered catalyst.
The support can play an active role in heterogeneous catalysis; therefore, we have synthesized copper nanoparticles on different supports and at different level of concentration. We have opted for Al 2 O 3 and TiO 2 as supports in addition to the SiO 2 particles. All the supports have in common the presence of hydroxyl functional groups at the particle surface, which can interact with the copper ions prior to the reduction process and facilitate the nucleation and adhesion of the Cu nanoparticles. Also, we have prepared a catalyst on SiO 2 particles with a double the amount of CuNPs (weight %) in order to study the influence copper concentration. The GPC analyses of the polymers obtained using the different catalysts are presented in Table 2 and Fig. 4.
The analysis of Table 2 and Fig. 4 reveals that the molecular weight is only slightly influenced by the support. In order to investigate these characteristics TEM analysis was used to characterize each catalyst (see Supplementary  Information Figs S1-S3). The TEM analysis of CuNPs/Al 2 O 3 catalyst (See Supplementary Information Fig. S1) revealed a high dimensional polydispersity of the CuNPs with a considerable amount of CuNPs (larger than 30 nm), which do not adhere to the support. These aspects, can justify the PD value of the GPC analysis, if we consider the different reactivity of the active sites on small and larger particles.
In the case of CuNPs/TiO 2 catalyst, the increase of the PD value could be attributed to a lower adhesion of CuNPs to the TiO 2 nanoparticles, which is suggested by the presence of CuNPs separated from the support (See Supplementary Information Fig. S2). Further, the smaller CuNPs particle size compared with the CuNPs/Al 2 O 3 catalyst should result in a lower molecular weight due to the increased surface area. However, rather similar molecular weights were registered.
In SET-LRP mechanism, the Cu 0 concentration plays a crucial role. In order to explore this aspect, we have synthesized a CuNPs/SiO 2 with an increased quantity of copper. From Fig. 2 it can be observed that the molecular weight slightly decreases, probably due to the increase in the active centers, in accordance with the increased concentration of Cu 0 on the surface of the support. For the CuNPs/SiO 2 (36% weight -Cu) catalyst, the TEM analysis  revealed the presence, besides the small nanoparticles, of very large (~100 nm) particles (See Supplementary  Information Fig. S3). The overall size of the CuNPs is smaller compared to the first catalyst CuNPs/SiO 2 (18% Cu, weight%), but the presence of the very large particles should explain the similar GPC results. Nevertheless, the difference between the two CuNPs/SiO 2 catalysts lies in the conversion attained at the same interval. The increase in copper content afforded a slightly higher conversion. However, the modification is not proportional to the increase of the surface area. A possible explanation could be the limitation imposed the diffusion/dissolution process of copper species (Cu I and Cu II ) into the reaction medium. Another aspect that needed to be ascertained was the reactivity of the terminal bromine of the obtained poly(butyl acrylate). Thus, styrene was used to present the successful synthesis of a BA-ST block-copolymer. This choice is motivated by the possible commercial applications of the BA-ST block-copolymers 44,45 . Table 3 and Fig. 5 demonstrate the block-copolymer synthesis evidenced by the increased molecular weight obtained.
The SET-LRP mechanism involves the CuNPs, but it is also dependent on the concentrations of Cu I and Cu II through disproportionation reaction. As the experiments revealed, an increased copper concentration did not lead to significant changes in the reaction result, which constitutes an indirect evidence that the reaction follows a SET-LRP mechanism. It is interesting to follow the evolution of the reaction in the absence of complexing ligands or using different agents, because the disproportionation reactions will be affected and the reaction could involve only CuNPs as Cu 0 source if no ligand is used. Therefore, the polymerization was realized in the absence of Me 6 TREN and in the presence of BiPy. As can be observed in Table 4 and Fig. 6, the lowest molecular weight was obtained in the case of Me 6 TREN, which can be explained by the participation of Me 6 TREN both in disproportionation and in polymer chain transfer reactions 46 . In contrast, the highest molecular weight was registered in the absence of any ligand, whereas in the case of BiPy an intermediary molecular weight was registered due to its participation in disproportionation reaction with limited chain transfer reactions.
In the case of BA polymerization in the absence of any ligand agent and using BiPy, the GPC analyses revealed an increased PD value. The results are consistent with literature data which presents the importance of N-ligands 47, 48 that stabilize Cu II and facilitate the disproportionation of Cu I . In our case this behavior explains the narrower PD value for Me 6 TREN due to the increase of Cu II concentration. However, as interesting result, a PD   value < 1.4 was obtained for the amine free reaction. Although the control of the reaction is limited, the straightforwardness of the approach would facilitate a scale-up of the process. In Fig. 7, are presented the characteristic colors resulting during the polymerization process in the absence of ligand agent, using Me6TREN and using BiPy. In the first case, a brown-violet is observed, second blue and in the third case red. These colors confirm the ATRP mechanism for the polymerization, they are specific for the cooper amine complex formed (blue and red), respectively the violet color confirms the Cu 0 nanoparticles free of ligand.
The kinetics analysis of the polymerization reaction in the absence of any ligand was performed and it was compared to the two other cases. Thus, using the plot of ln([M] 0 /[M]) as a function of time (Fig. 8), the k p app value was calculated as 0.120 min −1 (no ligand), 0.0575 min −1 (Me 6 TREN) and 0.034 min −1 (BiPy), results which are consistent with other examples in the literature for polymerization reactions involving Cu 0 and Me 6 TREN ligand 34,35 . In the case of Me 6 TREN there is a linear dependence of M n on the conversion demonstrating the "living" characteristic of the polymerization process. This behavior is manifested only up to 43% conversion in the absence of any ligand. After this point, the reaction involves termination through a chain transfer process or biradical termination.
However, the reasonable 1.38 dispersity value could be explained by the aqueous media employed for the reaction, which allows limited disproportionation and formation of an equilibrium between the copper species.
The obtained results sustain the SET-LRP mechanism for BA polymerization using CuNPs/SiO 2 catalyst in the presence of Me 6 TREN for the following reasons: (i) the high reducing capacity of NaBH 4 makes possible the    presence Cu 0 species as active species at the beginning of the reaction, thus, the high reaction rate makes Cu 0 the activating agent; (ii) the polymerization rate is dependent on the concentration of Cu 0 -CuNPs (36%)/SiO2 afforded a 95% conversion at 30 minutes; (iii) the lack of complexing ligand affords a higher molecular weight confirming limited disproportioning reaction and Cu 0 as active species;

Conclusions
Novel catalysts involving CuNPs (Cu 0 ) on different supports were synthesized and used in SET-LRP polymerization of BA. The supports consisted in TiO 2 nanoparticles, SiO 2 nanoparticles and Al 2 O 3 particles. The supports afforded under the same reaction conditions different sizes of CuNPs or different adhesion characteristics. The influence of these aspects was correlated with the GPC results for the BA polymerization experiments. The best results in terms of polydispersity were obtained for SiO 2 particles with a copper concentration of around 18% weight. Further, two consecutive experiments were performed demonstrating its reusability. The reactivity of the terminal bromine was evidenced through the synthesis of BA-ST block copolymer with an increased molecular weight compared to the BA homopolymer.  The control aspect of the reaction relies on the activity of the Me 6 TREN ligand as demonstrated by the experiments using BiPy and no ligand. Interestingly, the reaction performed in the absence of any ligand agent afforded a polydispersity value < 1.4 which sustains a somewhat limited control over the reaction.
Thus, we have shown the synthesis of a reusable catalyst able to participate in controlled living radical polymerization processes (in the presence of Me 6 TREN ligand PD value 1.25) as well as in straightforward polymerization processes with acceptable PD values (PD < 1.4) in the absence of any ligand. These aspects could facilitate the scale-up of the polymerization process.