Synthesis of well-defined alkyne terminated poly(N-vinyl caprolactam) with stringent control over the LCST by RAFT

Abstract

The reversible addition–fragmentation chain transfer (RAFT) of N-vinyl caprolactam (NVCL) using two new xanthates with alkyne functionalities is reported. The kinetic data obtained for polymerization of this non-activated monomer using a protected alkyne-terminated RAFT agent (PAT-X₁) revealed a linear increase of the polymer molecular weight with the monomer conversion as well as low dispersity (Đ) during the entire course of the polymerization. The system reported here allowed us to enhance the final conversion, diminish Đ and reduce the polymerization temperature compared to the typical values reported in the scarce literature available for the RAFT polymerization of NVCL. The resulting PNVCL was fully characterized using ¹H nuclear magnetic resonance (¹H NMR), matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS), Fourier-transform infrared spectroscopy (FTIR) and gel permeation chromatography (GPC) techniques. The temperature-responsive features of PNVCL in aqueous solutions were fully investigated under different conditions using turbidimetry. The presented strategy allows the synthesis of well-defined PNVCL with sharp and reversible phase transition temperatures around 37 °C. By manipulating the polymer molecular weight, or the solution properties, it is possible to tune the PNVCL phase transition. As a proof-of concept, the alkyne functionalized PNVCL was used to afford new linear block copolymers, by reacting with an azide-terminated poly(ethylene glycol) (N₃-PEG) through the copper catalyzed azide–alkyne [3 + 2] dipolar cycloaddition (CuAAC) reaction. The results presented establish a robust system to afford the synthesis of PNCVL with fine tuned characteristics that will enable more efficient exploration of the remarkable potential of this polymer in biomedical applications.

---

Temperature-responsive polymers (also known as thermo-responsive polymers), whose solubility depends on solution temperature, have been widely investigated for biomedical applications.^1–3 Poly(N-vinylcaprolactam) (PNVCL) is a nonionic, water-soluble, temperature-responsive polymer derived from an inexpensive monomer, N-vinylcaprolactam (NVCL). PNVCL exhibits a lower critical solution temperature (LCST) close to the physiological temperature (34–37 °C).^4–7 Furthermore, PNVCL is biocompatible^8,9 and has low cytotoxicity,^7,10,11 which makes it a very attractive polymer for different applications in the biomedical field.^{5,6,9,10,12–14} In comparison with other polymers, where the LCST is close to human body temperature temperature, i.e. poly(N-isopropylacrylamide) (PNIPAM), PNVCL stands out as a temperature-responsive material for biomedical applications since the hydrolysis of the amide group of the NVCL units does not form harmful small-molecule compounds.^10,11 Recent in vitro studies suggest that NVCL is well tolerated at concentrations below 10 mg mL⁻¹ and presents a low cytotoxicity for polymers with lower molecular weight (MW).^11,13,15 Through 3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) tests, Zhang et al.¹⁶ demonstrated that PNVCL homopolymers and NVCL copolymers with N-vinylpyrrolidone (NVP), and hydrophobic N-Boc-tryptophanamide-N-methacryl thioureas are non-cytotoxic. For concentrations ranging from 0.1 to 100 μg mL⁻¹ the cell viability of the polymer was almost 100% at 37 °C, compared to negative controls. Recently, Jayakumar and co-workers⁸ developed biodegradable and thermo-responsive polymeric nanoparticles based on chitosan and PNVCL for cancer drug delivery, confirming the low-toxicity and the efficient cell uptake of the nanocarriers.

PNVCL exhibits a “classical” Flory–Huggins miscibility behaviour (type I) in water.⁴ It is known that the decrease of the MW of PNVCL increases the LCST value, mainly due to increase of hydrophilicity of the polymer.¹⁷ Indeed, the LCST value of the PNVCL is very sensitive to changes in the polymer MW, concentration and composition of the solution, which makes the transition hydrophilic–hydrophobic very tuneable without the need of other structural changes involving, for example, the addition of hydrophobic segments (e.g. poly(vinyl acetate) (PVAc)).¹⁸ Due to the attractive properties of PNVCL for biomedical applications, several reports have been published using nanoparticles based on PNVCL for drug delivery systems (DDS),^8,15,16,19 nanoparticles for photodynamic therapy,²⁰ radiotherapy of solid tumours,²¹ hydrogels,²² films²³ and other applications such as dental restorative cements.⁹

The non-activated nature of NVCL turns its controlled polymer synthesis extremely challenging. Indeed, most of the available references involving the use of PNVCL deal with free radical polymerization (FRP) using 2,2′-azobis(2-methylpropionitrile) (AIBN) as a conventional initiator.^{4,9–11,16,24} The FRP hampers any stringent control over the structure, morphology or composition. This critical limitation is even more relevant in polymers such as PNVCL that presents strong structure-performance dependence. The available literature dealing with the synthesis of well-defined PNVCL by reversible deactivation radical polymerization (RDRP) methods is very scarce, with only a few reports about reversible addition–fragmentation chain transfer (RAFT) polymerization,^17,25–27 cobalt-mediated radical polymerization,^18,28–30 and atom transfer radical polymerization.^31–33 For the RAFT polymerization of NVCL, the choice of the RAFT agent is of extremely importance to attain PNVCL with controlled features. The RAFT polymerization mediated by xanthates, generally designated as macromolecular design via interchange of xanthates (MADIX)/RAFT polymerization, has been reported as the most suitable strategy for the polymerization of non-activated monomers, and O-ethyl xanthates the most efficient RAFT agents.^25,27 Beija and co-authors used the O-ethyl-S-(1-methoxycarbonyl)ethyl dithiocarbonate (X₁) as a RAFT agent for the synthesis of PNVCL in 1,4-dioxane. PNVCL, with a range of MW from 18 000 to 150 000 g mol⁻¹ and narrow MW distribution (Đ ∼ 1.1) was obtained.²⁶ The same RAFT agent was used to study the (co)polymerization kinetics of NVCL with methacrylic acid N-hydroxysuccinimide ester (MNHS) using different reaction conditions (different solvents and reaction temperatures³⁴), and with N-vinylpyrrolidone (PVP).^35–37 Using a similar RAFT agent, Liu et al.³⁸ reported the synthesis of a xanthate modified poly(ethylene glycol) (PEG) to mediate the direct synthesis of well-defined PEG-b-PNVCL copolymers. Yu et al.³⁹ explored the one-pot synthesis of a PNVCL based block copolymers from a dual initiator for RAFT and ring opening polymerization (ROP). Other architectures such as star-shaped PNVCL block copolymers, have also been reported from a xanthate terminated four- or six-arms star poly(ε-caprolactone) macro chain transfer agent.⁴⁰ However, the aforementioned strategies were used for the synthesis of PNVCL based block copolymers that deal with the sequential addition of monomers, because both segments can be synthesized using a single RAFT agent. For monomers that involve the use of different agents, the synthesis of block copolymers having PNVCL require either the use of specific (macro)RAFT agents or a post-polymerization reaction to attach the desired segments.

The concept of “click” chemistry, introduced by Sharpless and co-workers,⁴¹ is a very convenient synthetic approach in polymer chemistry to obtain polymers with diverse compositions, topologies or functionalities. Among several “click” coupling reactions reported in the literature, the copper-catalyzed azide–alkyne [3 + 2] dipolar cycloaddition (CuAAc) reaction is the most popular.^42–44 It is a high selective reaction with almost complete conversion obtained under mild reaction conditions.^42,43,45 The conjugation of different molecules by this post polymerization strategy, is possible by using initiators, monomers or polymers functionalized with the “click” moieties.⁴⁶ The primary goal of this work was the development of suitable polymerization conditions to afford well-controlled PNVCL via RAFT polymerization with a specific polymer chain end-functionality, the alkyne functionality. The use of such functionalized RAFT agents presents several advantages: eliminates the need of post-polymerization reactions to functionalize the polymer and facilitates the synthesis of complex structures from a simple CuAAc reaction and expands the range of applications of this temperature-responsive polymer.

Experimental

Materials

N-Vinylcaprolactam (98%, Sigma-Aldrich) was dissolved in hexane and purified by passing through a short alumina column and then recrystallized. 2,2′-Azobis(2-methylpropionitrile) (AIBN) (98%, Fluka) was purified by recrystallization from methanol before use. 1,4-Dioxane (99.8%, Acros Organics) was dried over calcium hydride and distilled under reduced pressure prior to use. Poly(ethylene glycol) methyl ether (mPEG₁₁₃, MW = 5000 Da) (Aldrich) was dried by azeotropic distillation with toluene (M_n,GPC = 5.06 × 10³; Đ = 1.03). Tetrabutylammonium fluoride trihydrate (TBAF·3H₂O) (99%, Acros Organics), 2-(2-chloroethoxy)ethanol (99%, Sigma-Aldrich), sodium azide (NaN₃) (≥99.5%, Sigma-Aldrich), hexadecyltrimethyl ammonium bromide (≥99%, Sigma-Aldrich), succinic anhydride (≥99%, Sigma-Aldrich), 4-(dimethylamino)pyridine (DMAP) (≥99%, TCI Europe), N,N′-diisopropylcarbodiimide (DIC) (99%, Sigma-Aldrich), sodium ascorbate (NaAsc) (≥98%, Sigma), copper(II) sulfate pentahydrate (CuSO₄·5H₂O) (≥98%, Aldrich), anhydrous sodium sulfate (≥98%, Fisher Chemical), dithranol (DT) (98%, Sigma-Aldrich), phosphate buffered saline (PBS) tablets (Sigma), deuterated chloroform (CDCl₃) (Euriso-top, +1% TMS), hexane (≥98.5%, Fisher Scientific), ethyl acetate (Fisher Chemical), dichloromethane (DCM) (99.99%, Fisher Scientific), diethyl ether (>99%, Fisher Scientific), petroleum ether (Fisher Scientific), and tetrahydrofuran (THF) (Fisher Scientific), were used as received. Purified water (MilliQ®, Millipore, resistivity > 18 MΩ cm) was obtained by reverse osmosis. For gel permeation chromatography (GPC), poly(methyl methacrylate) (PMMA) standards (Polymer Laboratories) (99%, ∼70 mesh, Acros) and high performance liquid chromatography (HPLC) dimethylformamide (DMF) (HPLC grade, Panreac) were used as received.

The RAFT agents used in the RAFT polymerization of NVCL, O-ethyl-S-(1-methoxycarbonyl)ethyl-dithiocarbonate (X₁), alkyne-terminated RAFT agent (AT-X₁) and protected alkyne-terminated RAFT agent (PAT-X₁) were synthesized according to the methods previously described in the literature.⁴⁷

2-(2-Azidoethoxy)ethan-1-ol,⁴⁸ carboxylic acid-terminated poly(ethylene glycol) methyl ether (mPEG-COOH) and azide terminated poly(ethylene glycol) methyl ether (mPEG-N₃)⁴⁹ were synthesized based on the procedures described in the literature. The detailed synthesis is presented in the ESI.†

Methods

The chromatographic parameters of the samples were determined by gel permeation chromatography (GPC), with refractive index (RI) (Knauer K-2301), differential viscometer (DV) and right angle light scattering (Viscotek 270 Dual Detector) detectors. The column set consisted of a PL 10 μL guard column (50 × 7.5 mm²), followed by two MIXED-B PL columns (300 × 7.5 mm², 10 μL). HPLC pump was set with a flow rate of 1 mL min⁻¹ and the analyses were carried out at 60 °C using an Elder CH-150 heater. The eluent was DMF, containing 0.3% of LiBr. The samples were filtered through a polytetrafluoroethylene (PTFE) membrane with 0.2 μm pore size before injection (100 μL). The system was calibrated with narrow PMMA standards. The M_n,GPC and Đ of the synthesized polymers were determined by using a multidetector calibration system (OmniSEC software version: 4.6.1.354). The refractive index increment (dn/dc) of the PNVCL in DMF at 60 °C was determined using an automatic refractometer for λ = 670 nm (Rudolph Research, J357 NDS-670-CC).

¹H and ¹³C nuclear magnetic resonance (NMR) spectra were recorded on a Bruker Avance III 400 MHz spectrometer, with a 5 mm TXI triple resonance detection probe, in CDCl₃ with tetramethylsilane (TMS) as an internal standard. Conversion of monomers was determined by integration of monomer and polymer peaks using MestReNova software version: 6.0.2-5475.

For matrix-assisted laser desorption ionization time-of-flight mass spectroscopy (MALDI-TOF-MS) analysis, the PNVCL samples were dissolved in THF at a concentration of 10 mg mL⁻¹ and dithranol (DT) (25 mg mL⁻¹ in THF) was used as a matrix and ionization agent. 10 μL of the PNVCL solution was mixed with 10 μL of the DT solution in a 1 mL Eppendorf tube. The dried-droplet sample preparation technique was used: 2 μL of the mixture was directly spotted on the MTP TF MALDI target, Bruker Daltonik (Bremen, Germany) and allowed to dry at room temperature, to allow matrix crystallization. The analysis was done in triplicate for each sample. External mass calibration was performed with a peptide calibration standard (Bruker Starter Kit) for 250 mass calibration points. Mass spectra were recorded using an Ultraflex III TOF/TOF MALDI-TOF mass spectrometer Bruker Daltonik (Bremen, Germany) operating in the linear positive ion mode, working under flexControl software (version 3.4, Bruker Daltonik, Bremen, Germany). Ions were formed upon irradiation by a smart beam laser using a frequency of 200 Hz. Each mass spectrum was produced by averaging 2000 laser shots collected across the whole sample spot surface by screening in the range m/z 5–20 × 10³ g mol⁻¹. The laser irradiance was set to 50% (relative scale 0–100) arbitrary units according to the corresponding threshold required for the applied matrix systems. Data analysis was done with flexAnalysis software (version 3.4, Bruker Daltonik, Bremen, Germany).

Elemental composition of the PNVCL synthesized by RAFT was determined using an elemental analyzer EA1108 CHNS-O from Fisons Instruments.

Fourier-transform infrared spectroscopy (FTIR) was performed at room temperature, at 128 scans and with a 4 cm⁻¹ resolution between 500 and 4000 cm⁻¹, using a JASCO 4200 FTIR spectrophotometer (MKII GoldenGate™ Single Reflexion ATR System).

The thermogravimetric analysis (TGA) was carried out on a TGA Q 500 machine (TA Instruments), from room temperature up to 600 °C, at a 10 °C min⁻¹ heating rate, under a dry nitrogen atmosphere (at 40 mL min⁻¹) using approximately 5 mg of the sample.

Turbidimetry analysis. The determination of the cloud point temperatures (T_CP) was achieved by turbidimetry analysis. The PNVCL polymers were dissolved in MilliQ® water at 1.0 mg mL⁻¹ concentration. The solutions were maintained at room temperature for at least 6 h to reach equilibrium. The optical transmittance at 630 nm of each PNVCL solution was monitored as a function of temperature using a modular UV-vis spectrometer equipped with a UV-vis light source (DH-2000, Ocean optics) and a temperature controlled cuvette holder (qpod®, Quantum Northwest, Inc.) with magnetic stirrer, run by a water circulating temperature controller (TC 125, Quantum Northwest, Inc.). The solution transmittance was assessed with time using SpectraScan software version: 1.0. The heating/cooling rate was 1 °C min⁻¹, 0.5 °C min⁻¹ or 0.1 °C min⁻¹. The T_CP was defined as the intercept of the tangents at the onset of turbidity. MilliQ® water was used as reference.

Procedures

Typical procedure for the RAFT polymerization of NVCL. The NVCL (4.00 g, 28.74 mmol), PAT-X₁ (15.93 mg, 0.05 mmol) and AIBN (3.93 mg, 0.02 mmol) were dissolved in 3.87 mL of 1,4-dioxane and placed in a 25 mL Schlenk reactor. The mixture was bubbled with N₂ for about 40 minutes at room temperature. The Schlenk reactor was then sealed, flushed with N₂ and placed in an oil bath at 60 °C under stirring (500 rpm). For the kinetic studies, different reaction mixture samples were collected during the polymerization by using an airtight syringe and purging the side arm of the Schlenk tube reactor with N₂. The samples were analyzed by ¹H NMR spectroscopy in order to calculate the monomer conversion and by GPC to determine the M_n,GPC and Đ. The final reaction mixture was precipitated in cold petroleum ether, dissolved in THF and precipitated again in cold petroleum ether (2×) and dried under vacuum to obtain PNVCL as white powder.

Typical procedure for the PNVCL deprotection. 2.02 g (0.06 mmol) of pure PNVCL obtained from a RAFT polymerization mediated by PAT-X₁ (M_n,th = 31.33 × 10³, M_n,GPC = 35.35 × 10³, Đ = 1.25) was dissolved in 20 mL of THF and bubbled with N₂ for about 10 minutes. The PNVCL solution was cooled down to −20 °C and 700 μL of a 0.2 M solution of TBAF·3H₂O (0.14 mmol) was slowly added to the polymer solution. After stirring for 30 minutes at low temperature, the reaction proceeded overnight at room temperature. The reaction mixture was passed through a silica column to remove the excess of TBAF and the alkyne-terminated PNVCL (AT-PNVCL) was recovered by precipitation in cold petroleum ether, dried under vacuum and analyzed by ¹H NMR and GPC (M_n,GPC = 38.90 × 10³, Đ = 1.26).

Coupling reaction between AT-PNVCL and N₃-PEG. 60 mg of the AT-PNVCL obtained after the deprotection of the protected alkyne-terminated PNVCL (M_n,GPC = 16.13 × 10³, Đ = 1.38) (3.72 μmol) and 29.3 mg of N₃-PEG (5.58 μmol) were dissolved into 3 mL of THF. The mixture was placed in a round-bottom flask equipped with a magnetic stir bar and sealed with a rubber septum. A stock solution of NaAsc (22.3 mM; 500 μL) in deionized water was added to the solution and the mixture was bubbled with N₂ for 20 min to remove oxygen. Lastly, a degassed stock solution of CuSO₄·5H₂O (11.1 mM; 500 μL) in deionized water was injected into the flask under N₂ atmosphere. The reaction was allowed to proceed under stirring at 30 °C for 24 h. The final reaction mixture was dialyzed (MW cut-off 12 400 Da) against deionized water and the block copolymer was obtained after freeze drying. The product was analyzed by GPC in order to confirm the success of the “click” coupling reaction.

Results and discussion

NVCL is a hygroscopic solid at room temperature (with a T_m ∼ 35–38 °C). This monomer is very sensitive to hydrolysis^50,51 and its purity was found to vary substantial among suppliers and even among batches from the same supplier. The presence of monomer inhibitors, namely the N,N′-di-s-butyl-p-phenylenediamine and contaminants, contributes to a yellowish color of the monomer. The detailed procedures and discussion for the monomer purification are placed in the ESI† (NVCL handling and purification). This issue is of particular importance for the following kinetic studies.

Polymerization of NVCL

NVCL is a classic non-activated monomer, as vinyl acetate (VAc), N-vinylpyrrolidone (NVP)⁵² and vinyl chloride,⁵³ since electron density around the double bond is very low. This feature turns radicals very reactive and therefore difficult to control.⁵⁴ Some authors reported the use of cobalt mediated radical polymerization of NVCL.²⁸ However, the protocol proposed is very laborious. The synthesis of low MW NVCL with narrow MW distributions is possible but for higher ratio monomer/initiator, the Đ increases and the monomer conversions remains low (∼30%). It has been reported that in order to avoid intermolecular coupling reactions, the monomer to initiator ratio should be high and the polymerization has to be stopped at low monomer conversions (<20%).³³

The RAFT polymerization mediated by xanthates has been described as the most suitable approach for the synthesis of well-controlled PNVCL with high yields and chain-end fidelity. RAFT agents such as trithiocarbonates and dithioesters can efficiently control the polymerization of more activated monomers, such as methacrylates or styrene, but are not suitable for the polymerization of non-activated vinyl monomers such as VAc or N-vinylactams (Fig. 2).^55,56 In the first report of the RAFT homopolymerization of NVCL, in 2005, by Devasia and co-authors,²⁷ the methyl 2-(ethoxy carbono thioyl thio)xanthate (X₁) was used as a RAFT agent. The reaction occurs at 80 °C and PNVCL with MW up to 33 000 g mol⁻¹ (Đ = 1.65) was obtained but no kinetic data were reported. Kinetic studies of the RAFT polymerization of low MW PNVCL mediated by a xanthate (B₁, Fig. 2) or dithiocarbamate (D₁, Fig. 2)²⁵ suggested that xanthates are the most suitable RAFT agents for the polymerization of PNVCL. Similar studies using a trithiocarbonate (A₁, Fig. 2) or a dithiocarbamate (C₁, Fig. 2) as RAFT agents¹⁷ revealed slow reactions, with MW up to 20 000 g mol⁻¹ (Đ = 1.33) after 60 h.

Fig. 1 Schematic representation of the RAFT polymerization of NVCL mediated by O-ethyl-S-(1-methoxycarbonyl)ethyl dithiocarbonate (X₁), alkyne-terminated RAFT agent (AT-X₁) or protected alkyne terminated RAFT agent (PAT-X₁).

Fig. 2 Schematic representation and examples of the different classes of RAFT agents.

Initially, in this study, the RAFT agent X₁ (Fig. 2) was evaluated for the polymerization of NVCL. The polymerization was conducted at 60 °C in 1,4-dioxane, using a monomer/solvent ratio = 1/2 (v/v), [NVCL] = 2.47 M, [X₁] = 16.47 mM and [AIBN] = 3.29 mM. The kinetic data (ESI, Fig. S6†) reveals a first-order kinetic with respect to monomer, in accordance with the literature.²⁶ The control over the polymer MW was very good taking into account the non-activated nature of monomer (Đ ∼ 1.2). Nevertheless, the polymerization was too slow, reaching only 60% of monomer conversion after 24 h.

Different reports in the literature dealing with the RAFT polymerization of NVCL, show that low concentrations of RAFT agents lead to faster reactions, and reduce the polymerization induction periods.^25,26 The degree of polymerization (DP), i.e., the ratio monomer to RAFT agent ([NVCL]₀ : [RAFT agent]₀), was increased from 140 to 600 and the xanthates with the alkyne moieties were evaluated for the RAFT polymerization of NVCL. The kinetic studies were carried out under the same reaction conditions: 60 °C in 1,4-dioxane, using a monomer/solvent ratio = 1/2 (v/v), [NVCL] = 2.47 M, [RAFT agent] = 4.12 mM and [AIBN] = 2.06 mM. The results presented in Fig. 3 show that both RAFT agents were able to mediate a successful RAFT polymerization of NVCL because both kinetics show a linear increase of MW with conversion, a good agreement between M_n,th and M_n,GPC, while maintaining narrow MW distributions (Đ ∼ 1.4). An induction period was observed in both cases; nevertheless, the results suggest that the protection of the alkyne functionality of the RAFT agent could lead to a significant reduction of the induction period, from 4 h to 1 h. This effect has already been reported for the RAFT polymerization of VAc under similar reaction conditions, which may probably be related with slower reinitiation of the initial RAFT agent.^47,57 Both reactions show a low polymerization rate, but the data also reveal that the protection of the alkyne functionality of RAFT agents leads to a faster and more controlled NVCL polymerization. Under the same experimental conditions, the PAT-X₁ exhibit higher chain transfer activity compared to the unprotected analogue. The comparison of the kinetics (Fig. 3) shows that PAT-X₁ mediates a faster reactions with higher monomer conversion. For PAT-X₁ the conversion reached 40% after 24 h, while, for the same reaction time, is only 23% for AT-X₁ (M_n,GPC = 13.7 × 10³; Đ = 1.37). Considering that both RAFT agents have the same Z group, this result may be attributed to the different R group and its influence in the RAFT agent efficiency.⁵⁶ In fact, it is well known that the leaving ability of the R group, as well as its reinitiation ability interferes with the behavior of the RAFT agent.

Fig. 3 RAFT polymerization of NVCL mediated by AT-X₁ (circle symbols) or PAT-X₁ (triangle symbols), in 1,4-dioxane at 60 °C. (a) First-order kinetic plot, (b) evolution of MW and Đ with conversion (the dashed line represents theoretical MW at a given conversion). Reaction conditions: [NVCL]₀/[RAFT agent]₀/[AIBN]₀ = 600/1/0.5 (molar), [NVCL]₀ = 2.47 M.

The chromatographic parameters of the PNVCL samples were determined using a GPC equipment coupled with a triple detector system in DMF containing 0.3% of LiBr as the elution solvent. From all the solvents reported in the literature, DMF appears to be the most suitable GPC eluent for NVCL-based polymers.³² The polymer GPC traces from the RAFT polymerization of NVCL mediated by PAT-X₁ kinetics points are plotted in Fig. 4. The GPC curves are unimodal and symmetric. It is possible to observe the movement of the peaks towards low retention volume during the course of the polymerization, indicating the controlled polymer chain grow. The MW increases linearly along the reaction time while the Đ remains low throughout the polymerization (Đ < 1.37).

Fig. 4 GPC traces with retention time for the RAFT of NVCL at 60 °C in 1,4-dioxane using PAT-X₁. Reaction conditions: [NVCL]₀/[PAT-X₁]₀/[AIBN]₀ = 600/1/0.5 (molar); [NVCL]₀ = 2.47 M.

The refractive index increments (dn/dc) of both PNVCL synthesized using AT-X₁ and PAT-X₁ were determined (ESI, Fig. S7†). There is a slight change in the dn/dc values of the PNVCL that could be related with the presence of different polymer end-groups.

Aiming to achieve fast polymerizations, the RAFT of NVCL was carried out at higher monomer/solvent ratio (1/1 (w/w)). The reaction was conducted under the same conditions: 60 °C, 1,4-dioxane, with [NVCL] = 3.70 M, [RAFT agent] = 6.16 mM and [AIBN] = 3.06 mM. The kinetic results shown in Fig. 5 reveal an increase of reaction rate maintaining a similar control over the polymer MW and Đ. The evolution of MWs is linear with conversion and the Đ was close to 1.3 throughout the polymerization. However, due to the high target degree of polymerization (DP = 600), after 50% of conversion, the solution viscosity increases considerably, leading to a reduction in radical mobility. This fact precludes the addition of new monomer units to the active chain-ends, and most probably is responsible for the limited monomer conversion around 70% and the increase of Đ after 12 h of reaction.

Fig. 5 RAFT polymerization of NVCL mediated by PAT-X₁ in 1,4-dioxane at 60 °C. (a) First-order kinetic plot, (b) evolution of MW and Đ with conversion (the dashed line represents theoretical MW at a given conversion) and (c) GPC traces with retention time. Reaction conditions: [NVCL]₀/[PAT-X₁]₀/[AIBN]₀ = 600/1/0.5 (molar), [NVCL]₀ = 3.70 M.

A series of well-defined homopolymers were prepared by RAFT polymerization, using the reaction conditions described previously, by stopping the polymerization at different monomer conversions (Table 1). The polymer obtained presented fairly low Đ, concerning the challenging nature of the monomer and the obtained M_n,GPC are in agreement with M_n,th. Using the same reaction conditions, the polymerization of NVCL in the absence of RAFT agent (FRP) (Table 1, entry 6) is faster, leading to high MW polymers with broad MW distribution (ESI, Fig. S8,† comparison of the GPC traces of PNVCL synthesized via FRP and RAFT polymerization).

Table 1 RAFT polymerization of NVCL with PAT-X₁, in 1,4-dioxane at 60 °C. Reaction conditions: monomer/solvent ratio: 1/1 (w/w), [NVCL]₀/[PAT-X₁]₀/[AIBN]₀ = 600/1/0.5

Entry	Time	Conv, %	Mn,th × 10^3	Mn,GPC × 10^3	Đ	TCP^b, °C
a FRP of NVCL, in 1,4-dioxane at 60 °C. Reaction conditions: (monomer/solvent ratio: 1/1 (w/w)), [AIBN]0 = 0.015 [NVCL]0.b Determined by turbidimetry analysis.
1	3 h 30 min	18	15.1	16.13	1.38	37.59
2	4 h 00 min	21	17.8	22.88	1.38	37.05
3	4 h 30 min	33	27.4	32.82	1.26	36.73
4	8 h 00 min	50	41.8	41.35	1.27	36.56
5	14 h 00 min	71	59.8	56.05	1.37	35.9
6^a	30 min	—	—	1650.00	>2	—

---

Chemical structure and retention of chain-end functionality

The chemical structure of the PNVCL synthesized through RAFT polymerization mediated by the PAT-X₁ was confirmed by ¹H NMR spectroscopy. The spectrum of a purified PNVCL sample is presented in Fig. 6 and reveals the resonances of the NVCL repeat unit at 4.42 ppm (m, –CH₂–CH–) and all other structural moieties: 3.20 ppm (r, –NCO–CH₂–), 2.49 ppm (n, –CH₂–NCO–) and 1.74–1.44 (o, p, q, –(CH₂)₃– (lactam ring) and l, –CH₂– (polymer backbone)).²⁵ The presence of the characteristic signal of the trimethylsilyl protons around 0 ppm (k, –Si(CH₃)₃–) in Fig. 6a evidences the success of the reaction and the retention of the RAFT functionality in the polymer α-chain end. This characteristic signal disappears after the deprotection reaction with TBAF, resulting in the alkyne-functionalized PNVCL (Fig. 6b). The ratio between the integrals of m, the proton from the polymer backbone at 4.42 ppm with the integrals of the RAFT fragment k, the trimethylsilyl protons at 0.13 ppm corresponds to the degree of polymerization (DP) and allows the calculation of the average-number MW (M_n,NMR) of the PNVCL by ¹H NMR spectroscopy, using the equation: M_n,NMR = M_CTA + DP·M_NVCL. The M_n,NMR of PNVCL was calculated to be 36.63 × 10³ and is in good agreement with the value obtained by GPC for the same polymer. The ¹H NMR spectrum of the pure polymer does not allow the assessment of the other characteristic proton peaks of the RAFT agent, i.e., the –CH₂–CH₃ near the xanthate moiety (a, b) and the ones near the alkyne terminal (c, d, h, i, j and s) due to the overlapping with other proton signals from the polymeric chain or due to solvent impurities.

Fig. 6 ¹H NMR spectra in CDCl₃ of (a) the protected alkyne-terminated PNVCL obtained by RAFT polymerization mediated by PAT-X₁ and (b) alkyne-terminated PNVCL after the deprotection reaction. Reaction conditions: [NVCL]₀/[PAT-X₁]₀/[AIBN]₀ = 600/1/0.5 (molar) in 1,4-dioxane at 60 °C, [NVCL]₀ = 3.71 M (M_n,NMR = 36.63 × 10³, M_n,GPC = 38.40 × 10³, Đ = 1.28).

The chemical structure of the synthesized PNVCL was also investigated by FTIR (ESI, Fig. S9†). A strong characteristic peak of PNVCL backbone and caprolactam ring appears at 2924 cm⁻¹ (aliphatic CH stretching peaks). The peak of C=O stretching vibration the PNVCL is present at 1625 cm⁻¹. The peaks at 1480 cm⁻¹ and 1441 cm⁻¹ are ascribed to the C–N and –CH₂– stretching vibration, respectively. The broad peak at 3500–3300 cm⁻¹ is ascribed to –OH group resultant from the hygroscopic nature of the PNVCL.⁵⁸ The PNVCL after the deprotection reaction (AT-PNVCL) maintains its structure.

In RAFT, just like in other RDRP methods, the resultant polymer can be telechelic, with distinct α and ω chain-end functionalities. The α-chain end moiety, the protected alkyne functionality, was obtained from the modification of the R group of the RAFT agent. After the polymerization, the characteristic RAFT thiocarbonylthio group remains in the opposing polymer chain-end⁵⁹ and it is responsible for its “living” character. The presence of these specific characteristic RAFT end-groups is usually assessed by a chain extension experiment. Characterization techniques such as FTIR and UV-vis analysis⁶⁰ have been described for the evaluation of the –S–C=S chain end functionality. Herein, those techniques were applied unsuccessfully, probably due to the high MW of the polymers. Therefore, the presence of the sulfur moiety in the PNVCL homopolymer, synthesized by RAFT polymerization, was confirmed by elemental analysis (ESI, Table S1†) and its chemical structure was confirmed by MALDI-TOF analysis. It should be mentioned that there is only one report in the literature concerning the MALDI-TOF analysis of PNVCL synthesized by RAFT mediated by a dithiocarbamate,²⁵ however the authors suggested that in that case polymerization did not occur in a strictly living manner.

The MALDI-TOF-MS of the protected alkyne-terminated PNVCL synthesized by RAFT polymerization, in the linear mode with m/z ranging from 5000 to 18 000 is shown in Fig. 7a and the enlargement of the 6200–7100 range is presented in Fig. 7b. Five series of peaks are separated by an interval corresponding to a NVCL repeat unit (139.20 mass units). Each group of peaks consists of the molecular ion of the polymer chain of R–PNVCL–Z where R and Z are the groups from the RAFT agent, R = Si(CH₃)₃–C≡C–(CH₃)₂–O–C–O–C(H)CH₃ and Z = S–C(S)–O–CH₂CH₃, plus the matrix (DT) (main series) or impurities in the matrix (series of less intensive peaks).⁶¹ The assignment of all the peaks shown in is presented in Table 2. The molecular structure, molecular weight of DT and the suggested DT impurities (DT_imp) reported in the literature⁶¹ and other structures identified by us are presented in the ESI (Fig. S10†). Moreover, one of the series of smaller peaks, B_m, is attributed to a polymer chain R–PNVCL, resulted from the elimination of the S–C(S)–O–CH₂CH₃ group, from the polymer chain-end under MALDI conditions, as the difference between the B_m peak and the expected MW is 122 mass units (ESI, Fig. S11†).^62,63

Fig. 7 MALDI-TOF-MS (a) in the linear mode (using DT as a matrix) from m/z 5000 to 18 000 and (b) enlargement of the spectrum from m/z 6200 to 7000 of protected alkyne-terminate PNVCL (M_n,GPC = 16.13 × 10³, and Đ = 1.38). Reaction conditions: [NVCL]₀/[CTA]₀/[AIBN]₀ = 600/1/0.5 (molar); [NVCL]₀ = 3.71 M, 1,4-dioxane, T = 60 °C.

Table 2 MALDI-TOF-MS peak assignment for protected alkyne-terminated PNVCL (Fig. 7)^a

n	A set			B set			C set			D set			E set
		MWca^b	MWexp		MWca^c	MWexp		MWca^d	MWexp		MWca^e	MWexp		MWca^f	MWexp
a n – the number of repeat units (NVCL), where PAT-X1 = 332.55 g mol^−1, NVCL = 139.20 g mol^−1, DT = 226.23 g mol^−1, SHCSOCH2CH3 = 122.21 g mol^−1; H = 1.01 g mol^−1; DTimp(I) = 240.21 g mol^−1; DTimp(II) = 184.28 g mol^−1; DTimp(III) = 198.27 g mol^−1.b MWca = PAT-X1 + nPNVCL + D + T.c MWca = PAT-X1 + nPNVCL–SHCSOCH2CH3–H.d MWca = PAT-X1 + nPNVCL + DTimp(I).e MWca = PAT-X1 + nPNVCL + DTimp(II).f MWca = PAT-X1 + nPNVCL + DTimp(III).
41	A1	6265.9	6265.2				C1	6279.7	6280.3
42	A2	6405.1	6404.7				C2	6419.1	6419.1	D1	6363.1	6362.4	E1	6377.1	6377.2
43	A3	6544.3	6543.7				C3	6558.3	6560.0	D2	6502.3	6501.9	E2	6516.3	6516.2
44	A4	6683.5	6682.8	B1	6334.1	6332.6				D3	6641.5	6640.6	E3	6655.5	6654.8
45				B2	6473.2	6473.3
46				B3	6612.4	6612.3

---

The success of the PNVCL deprotection reaction was also confirmed by the MALDI-TOF technique. The MALDI-TOF-MS spectrum of PNVCL after the deprotection reaction, in the linear mode, with m/z ranging from 5000 to 18 000 and its enlargement in the m/z range 6200–7200 are shown in Fig. S12 (ESI†). As expected, the series of peaks are separated by an interval corresponding to the NVCL repeating unit, revealing the integrity of the monomer units after the deprotection reaction. The peak series is attributed to a polymer chain R′–PVC–Z, where R group corresponds to the HC≡C–(CH₃)₂–O–C–O–C(H)CH₃ moiety in the PNVCL chain-end, plus the DT or DT_imp in the matrix. The detail information about the assignment of the peaks from the MALDI-TOF analysis of the deprotected PNVCL is presented in ESI, Fig. S13 and Table S2.†

Thermal analysis

The thermogravimetric analysis of the NVCL and the PNVCL are presented in ESI (Fig. S14 and Table S3†). The polymer shows a single degradation stage around 400 °C. Comparing the profiles of the PNVCL samples analyzed, it is impossible to establish a relationship between the synthesis strategy and the degradation profile. In this line, the differences between the TGA traces of the uncontrolled PNVCL synthesized through FRP and the well-defined PNVCL synthesized by RAFT polymerization were not relevant and are in agreement with the literature.⁵⁸ Furthermore, and more importantly, it is shown that the deprotection reactions have no deleterious effects on the polymer thermal stability.

Temperature-responsive behavior of the PNVCL

The NVCL monomer contains a seven-membered lactam ring with one hydrophilic amide group directly connected to the hydrophobic carbon–carbon backbone chain (Fig. 1). Due to the presence of hydrophilic and hydrophobic groups in its structure, the resulting polymer is soluble in both aqueous and organic media.⁶⁴ Temperature-responsive polymers can undergo reversible solvation and desolvation due to a small variation in solution temperature.^65,66 For solution temperatures above the cloud point temperature (T_CP), the PNVCL exhibits a phase transition behavior, resulting in conformational changes in the polymer that lead to the formation of hydrophobic domains and consequent change in polymer solubility.^2,67 Above T_CP, the intra- and intermolecular attractions between the hydrophobic parts of the polymer molecules are favoured compared to the interactions with water molecules, and the polymer undergoes an abrupt change in the macromolecule specific volume.²²

The phase transition temperature of PNVCL in aqueous solutions, decreases with increasing polymer concentration.¹⁷ Contrarily to other temperature-responsive polymers (e.g. PNIPAM), the PNVCL solution behaviour with temperature is strongly affected by other factors such as the polymer MW and end-groups.^4,17,26 The dependence of the LCST of the PNVCL with MW was evaluated along with the RAFT reaction kinetics by turbidimetry analysis. The transmittance of light through aqueous solutions of the polymer was measured at 630 nm using a UV-vis spectrometer and a heating or cooling rate of 1.0 °C min⁻¹. The T_CP of homopolymer aqueous solutions (1.0 mg mL⁻¹), the temperature at which the polymer becomes insoluble, was determined, from the inflection point of the curve in the transmittance vs. temperature plots.⁶⁸ As expected, the T_CP of PNVCL decreases along with increasing the MW (Fig. 8a). With the course of the reaction, the T_CP varies from 39.7 °C to 37.3 °C. Increasing the solution temperature above the T_CP, causes an abrupt change in the polymer conformation, the polymer becomes insoluble and precipitates. As long as the solution temperature is higher than T_CP, the polymer remains insoluble and the solution is cloudy (Fig. 8b).

Fig. 8 (a) Transmittance measurements as a function of temperature for 1.0 mg mL⁻¹ aqueous solution of different kinetic points of the RAFT polymerization of NVCL mediated by PAT-X₁. Reaction conditions: [NVCL]₀/solvent = 1/2 (v/v), [NVCL]₀/[PAT-X₁]₀/[AIBN]₀ = 600/1/0.5 (molar) in 1,4-dioxane, T = 60 °C and (b) pictures of the PNVCL homopolymer solution (1.0 mg mL⁻¹) at room temperature (left) and above the T_CP (right).

In an attempt to match the LCST of PNVCL with physiological temperature, some authors reported strategies involving the copolymerization of other monomers such as N-vinylpyrrolidone.^35,37,39 or MNHS³⁴ or increase the PNVCL concentration in the turbidimetry assays. Herein, the solution behavior of the pure homopolymers presented in Table 1 was also studied in order to create a library of polymers with different T_CP values. The temperature-responsive behavior of PNVCL aqueous solutions was evaluated for both heating and cooling cycles for a fixed polymer concentration of 1.0 mg mL⁻¹ and a heating/cooling rate of 1.0 °C min⁻¹. The results are shown in Fig. 9. All solutions exhibited a very sharp transition in the solution transmittance, close to the human body temperatures, and their temperature-responsive behavior was reversible. The effect of the PNVCL MW on the LCST had been studied by several authors.⁵² Nevertheless, the polymerization method presented here allows a much tighter correlation between the polymer MW and T_CP, especially for high MWs. Contrary to what would be expected, a large hysteresis in the transition temperature between the heating and cooling cycles was observed in the different PNVCL samples analyzed. The difference between the transitions temperatures increases along with PNVCL MW (Fig. 9b). It should be noted that most of the literature reports concerning the solution behavior of PNVCL only present the heating cycle. As a general practice, the authors tend to adjust the solution properties (polymer concentration) or the conditions of the turbidimetry tests, including the heating rates, to fit a desired T_CP. In the case of the PNVCL synthesized by RAFT, the temperature-responsive phenomenon of PNVCL is reversible and highly reproducible. The aqueous solution of PNVCL at 1.0 mg mL⁻¹ shows reversible changes in the transmittance between 30 and 40 °C with no changes observed in the transmittance transitions for the three consecutive heating/cooling cycles (ESI, Fig. S15†).

Fig. 9 Dependence of with polymer MW. (a) Plots of percentage transmittance vs. temperature for various PNVCL samples in an aqueous solution (1.0 mg mL⁻¹) (the solid black lines represent the heating cycle and the dashed grey lines represent the cooling cycle). (b) Variation of the cloud point temperature (T_CP) with the PNVCL molecular weight in the heating cycle (closed symbols) and in the cooling cycles (open symbols).

Due to the uncontrolled behavior of the FRP polymerization, it is not possible to tailor the PNVCL MW and consequently the T_CP. In the case of PNVCL synthesized via FRP, it exhibits a sharp transition in the solution transmittance during a heating cycle of 1.0 °C min⁻¹, for two consecutive cycles (ESI, Fig. S16†). However, in the cooling cycle, the PNVCL behaves differently. The transition in solution turbidimetry was soft and gradual, started from the beginning of the cooling cycle (45 °C). The recovery in the solution transmittance is inconstant, as evidenced by the differences between the two cooling cycles. Despite the sharp transition in the solution turbidimetry for the PNVCL synthesized by FRP, its solution behavior above the T_CP is unpredictable.

Furthermore, the T_CP is dependent on the solution heating rate, as evidenced in Fig. 10a. Slowing down the heating rate from 1.0 °C min⁻¹ to 0.1 °C min⁻¹ significantly reduces the T_CP of the sample from 36.93 °C to 35.20 °C. Moreover, the hydrophilic/hydrophobic transition of the PNVCL samples was lowered in the presence of salts, i.e. for the same heating rate, the decay in the transmittance percentage for PNVCL in PBS solution or NaCl saline solution occurs earlier than in water solutions (Fig. 10b). For the same PNVCL sample, the T_CP was shifted from 37.36 °C in water to 34.70 °C in PBS solution. These results could be related with the salting-out effect.²¹

Fig. 10 Dependence of the heating rate (a) and the presence of salts (b) in the T_CP of PNVCL samples. Percentage transmittance vs. temperature plots for the same PNVCL sample for (a) different heating rates (1.0, 0.5 and 0.1 °C min⁻¹), in aqueous solution and (b) PBS solution (pH 7.4) and physiological saline solution (1.0 mg mL⁻¹) (heating rate 1.0 °C min⁻¹).

For solution temperatures below the T_CP, there is strong interaction between the PNVCL and water molecules, causing the complete polymer solvation. At these temperatures the polymer is completely soluble. By introducing the salts into the system, the correspondent ions will preferably interact with the hydrophilic segment of the PNVCL (N–C=O bound), reducing the degree of polymer solvation. Therefore, the interactions between the hydrophobic part of the PNVCL dominates over the interactions between hydrophilic part, promoting an earlier phase separation of the polymer.¹¹

PNVCL based copolymers via a “click” coupling approach

Most of the authors aiming to obtain a specific functionality in the PNVCL chain-end, use a traditional chain transfer agent (CTA), such as 3-mercaptopropanoic acid.⁵² An uncontrolled polymerization of PNVCL occurs in the presence of the conventional initiator AIBN. The final polymer retains the acid functionality but it is obtained in low conversions and low MW. The desired functional group is further introduced in the polymer chain-end via a carbodiimide-mediated coupling reaction. This strategy was adopted by Zhang and co-workers to introduce an alkyne functionality in the chain-end of low MW PNVCL for further conjugation with an azide terminated poly(ε-caprolactone).⁶⁹ Likewise, this approach was also reported for the preparation of radiolabeled PNVCL.²¹

Herein, the synthesis of alkyne functionalized PNVCL allows the further straightforward synthesis of PNVCL block copolymers by the CuAAC reaction. After the RAFT polymerization of PNVCL using the PAT-X₁, the protective trimethylsilyl group was removed using TBAF (Fig. 11(i)). The success of the coupling reaction was confirmed by the disappearance of the characteristic trimethylsilyl signal (k) at 0.14 ppm in the ¹H NMR spectrum of the unprotected PNVCL (Fig. 6). As a proof-of-concept, an azide functionalized PEG molecule, mPEG-N₃, was conjugated to the alkyne terminated PNVCL to afford PEG-b-PNVCL copolymers (Fig. 11(ii)). The mPEG-N₃ was synthesized from the conjugation of the azido ethoxy ethanol with a carboxylic acid terminated PEG (mPEG-COOH) (the schematic representation of the synthesis of mPEG-N₃ and the correspondent FTIR-ATR spectrum are present in ESI, Fig. S1 and S2†).

Fig. 11 Schematic representation of the (i) deprotection reaction of PNVCL synthesized by RAFT mediated by PAT-X₁ and (ii) CuAAC reaction between mPEG-N₃ and AT-PNVCL to afford PEG-b-PDPA copolymers.

The CuAAC reaction, between the AT-PNVCL and mPEG-N₃, occurred in a mixture THF/water at 30 °C (Fig. 11(ii)). The success of the coupling reaction was evaluated by GPC, with a shift of the PNVCL precursor towards high MW values (low retention volume) (Fig. 12). Despite the efforts to purify the resultant PNVCL-b-PEG copolymer, the low MW tail observed in the GPC trace of the precursor mPEG-N₃ (probably resultant from a low MW impurity), was present in the final copolymer GPC trace and does not allow the determination of the PNVCL chain-end functionality.

Fig. 12 GPC traces of the alkyne terminated PNVCL (AT-PNVCL), azide terminated PEG (mPEG-N₃) and the PEG-b-PNVCL copolymer obtained after the CuAAC reaction.

Conclusions

Well-defined PNVCL with well controlled MW and narrow MW distributions was synthesized from RAFT polymerization, mediated by an alkyne functionalized RAFT agents. PNVCL presents a very sharp and reversible response to temperature in a range close to human physiological temperature. The T_CP of the PNVCL can be tuned by varying the polymer MW, solutions characteristics or the heating rate. The alkyne moiety at the PNVCL chain-end allows further conjugation to other molecules or polymers and facilitates the development of new PNVCL-based block copolymers.

Acknowledgements

JRG acknowledges FCT-MCTES for her PhD scholarship (SFRH/BD/69635/2010). The NMR data was collected at the UC-NMR facility which is supported in part by FEDER – European Regional Development Fund through the COMPETE Programme (Operational Programme for Competitiveness) and by National Funds through FCT – Fundação para a Ciência e a Tecnologia (Portuguese Foundation for Science and Technology) through grants REEQ/481/QUI/2006, RECI/QEQ-QFI/0168/2012, CENTRO-07-CT62-FEDER-002012, and Rede Nacional de Ressonância Magnética Nuclear (RNRMN). AVP's research reported in this publication was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under award number UL1TR000003.

Notes and references

1. S. Ganta, H. Devalapally, A. Shahiwala and M. Amiji, J. Controlled Release, 2008, 126, 187–204.
2. D. Schmaljohann, Adv. Drug Delivery Rev., 2006, 58, 1655–1670.
3. A. K. Bajpai, J. Bajpai, R. Saini and R. Gupta, Polym. Rev., 2011, 51, 53–97.
4. F. Meeussen, E. Nies, H. Berghmans, S. Verbrugghe, E. Goethals and F. Du Prez, Polymer, 2000, 41, 8597–8602.
5. Y. Maeda, T. Nakamura and I. Ikeda, Macromolecules, 2002, 35, 217–222.
6. A. C. W. Lau and C. Wu, Macromolecules, 1999, 32, 581–584.
7. C. K. Chee, S. Rimmer, I. Soutar and L. Swanson, React. Funct. Polym., 2006, 66, 1–11.
8. N. S. Rejinold, K. P. Chennazhi, S. V. Nair, H. Tamura and R. Jayakumar, Carbohydr. Polym., 2011, 83, 776–786.
9. A. Moshaverinia, N. Roohpour, J. A. Darr and I. U. Rehman, Acta Biomater., 2009, 5, 2101–2108.
10. N. I. Shtanko, W. Lequieu, E. J. Goethals and F. E. Du Prez, Polym. Int., 2003, 52, 1605–1610.
11. H. Vihola, A. Laukkanen, L. Valtola, H. Tenhu and J. Hirvonen, Biomaterials, 2005, 26, 3055–3064.
12. I. Negru, M. Teodorescu, P. O. Stanescu, C. Draghici, A. Lungu and A. Sarbu, Mater. Plast., 2010, 47, 35–41.
13. H. Vihola, A. K. Marttila, J. S. Pakkanen, M. Andersson, A. Laukkanen, A. M. Kaukonen, H. Tenhu and J. Hirvonen, Int. J. Pharm., 2007, 343, 238–246.
14. J. Liu, A. Debuigne, C. Detrembleur and C. Jerome, Adv. Healthcare Mater., 2014, 3, 1941–1968.
15. H. Vihola, A. Laukkanen, J. Hirvonen and H. Tenhu, Eur. J. Pharm. Sci., 2002, 16, 69–74.
16. L. F. Zhang, Y. Liang and L. Z. Meng, Polym. Adv. Technol., 2010, 21, 720–725.
17. L. Shao, M. Hu, L. Chen, L. Xu and Y. Bi, React. Funct. Polym., 2012, 72, 407–413.
18. A. Kermagoret, C.-A. Fustin, M. Bourguignon, C. Detrembleur, C. Jerome and A. Debuigne, Polym. Chem., 2013, 4, 2575–2583.
19. S. Shah, A. Pal, R. Gude and S. Devi, Eur. Polym. J., 2010, 46, 958–967.
20. C.-L. Peng, L.-Y. Yang, T.-Y. Luo, P.-S. Lai, S.-J. Yang, W.-J. Lin and M.-J. Shieh, Nanotechnology, 2010, 21, 155103–155115.
21. P. Černoch, Z. Černochová, J. Kučka, M. Hrubý, S. Petrova and P. Štěpánek, Appl. Radiat. Isot., 2015, 98, 7–12.
22. H. Vihola, A. Laukkanen, H. Tenhu and J. Hirvonen, J. Pharm. Sci., 2008, 97, 4783–4793.
23. E. Kharlampieva, V. Kozlovskaya, J. Tyutina and S. A. Sukhishvili, Macromolecules, 2005, 38, 10523–10531.
24. S. Verbrugghe, K. Bernaerts and F. E. Du Prez, Macromol. Chem. Phys., 2003, 204, 1217–1225.
25. D. Wan, Q. Zhou, H. Pu and G. Yang, J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 3756–3765.
26. M. Beija, J.-D. Marty and M. Destarac, Chem. Commun., 2011, 47, 2826–2828.
27. R. Devasia, R. Borsali, S. Lecommandoux, R. L. Bindu, N. Mougin and Y. Gnanou, Abstr. Pap. Am. Chem. Soc., 2005, 230, U4231–U4232.
28. M. Hurtgen, J. Liu, A. Debuigne, C. Jerome and C. Detrembleur, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 400–408.
29. A. Kermagoret, K. Mathieu, J.-M. Thomassin, C.-A. Fustin, R. Duchene, C. Jerome, C. Detrembleur and A. Debuigne, Polym. Chem., 2014, 5, 6534–6544.
30. A. Kermagoret, K. Mathieu, J.-M. Thomassin, C.-A. Fustin, R. Duchene, C. Jerome, C. Detrembleur and A. Debuigne, Polym. Chem., 2014, 5, 6534–6544.
31. Y. Z. Yang, J. Li, M. Q. Hu, L. Chen and Y. M. Bi, J. Polym. Res., 2014, 21, 549.
32. P. Singh, A. Srivastava and R. Kumar, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 1503–1514.
33. X. Jiang, Y. Li, G. Lu and X. Huang, Polym. Chem., 2013, 4, 1402–1411.
34. H. Peng, M. Kather, K. Rübsam, F. Jakob, U. Schwaneberg and A. Pich, Macromolecules, 2015, 48, 4256–4268.
35. X. Zhao, O. Coutelier, H. H. Nguyen, C. Delmas, M. Destarac and J.-D. Marty, Polym. Chem., 2015, 6, 5233–5243.
36. S. Maji, Z. Zhang, L. Voorhaar, S. Pieters, B. Stubbe, S. Van Vlierberghe, P. Dubruel, B. G. De Geest and R. Hoogenboom, RSC Adv., 2015, 5, 42388–42398.
37. X. Liang, V. Kozlovskaya, C. P. Cox, Y. Wang, M. Saeed and E. Kharlampieva, J. Polym. Sci., Part A: Polym. Chem., 2014, 52, 2725–2737.
38. J. Liu, C. Detrembleur, M. C. De Pauw-Gillet, S. Mornet, E. Duguet and C. Jerome, Polym. Chem., 2014, 5, 799–813.
39. Y. C. Yu, G. Li, J. Kim and J. H. Youk, Polymer, 2013, 54, 6119–6124.
40. G. D. García-Olaiz, K. A. Montoya-Villegas, A. Licea-Claverie and N. A. Cortez-Lemus, React. Funct. Polym., 2015, 88, 16–23.
41. H. C. Kolb, M. G. Finn and K. B. Sharpless, Angew. Chem., Int. Ed., 2001, 40, 2004–2021.
42. J. F. Lutz, Angew. Chem., Int. Ed., 2007, 46, 1018–1025.
43. M. Meldal and C. W. Tornoe, Chem. Rev., 2008, 108, 2952–3015.
44. V. V. Rostovtsev, L. G. Green, V. V. Fokin and K. B. Sharpless, Angew. Chem., Int. Ed., 2002, 41, 2596–2599.
45. W. H. Binder and R. Sachsenhofer, Macromol. Rapid Commun., 2008, 29, 952–981.
46. U. Mansfeld, C. Pietsch, R. Hoogenboom, C. R. Becer and U. S. Schubert, Polym. Chem., 2010, 1, 1560–1598.
47. J. R. Góis, A. V. Popov, T. Guliashvili, A. C. Serra and J. F. J. Coelho, RSC Adv., 2015, 5, 91225–91234.
48. L. Mespouille, M. Vachaudez, F. Suriano, P. Gerbaux, O. Coulembier, P. Degee, R. Flammang and P. Dubois, Macromol. Rapid Commun., 2007, 28, 2151–2158.
49. X. Ma, Z. Zhou, E. Jin, Q. Sun, B. Zhang, J. Tang and Y. Shen, Macromolecules, 2013, 46, 37–42.
50. M. F. Shostakovsky, N. A. Medzykhovskaya and M. G. Zelenskaya, Russ. Chem. Bull., 1952, 1, 627–632.
51. M. F. Shostakovsky, F. P. Sidelkovskaya and M. G. Zelenskaya, Russ. Chem. Bull., 1954, 3, 589–592.
52. S. F. Medeiros, J. C. S. Barboza, M. I. Re, R. Giudici and A. M. Santos, J. Appl. Polym. Sci., 2010, 118, 229–240.
53. C. M. R. Abreu, P. V. Mendonça, A. C. Serra, J. F. J. Coelho, A. V. Popov, G. Gryn'ova, M. L. Coote and T. Guliashvili, Macromolecules, 2012, 45, 2200–2208.
54. K. Nakabayashi and H. Mori, Eur. Polym. J., 2013, 49, 2808–2838.
55. N. S. Ieong, M. Redhead, C. Bosquillon, C. Alexander, M. Kelland and R. K. O'Reilly, Macromolecules, 2011, 44, 886–893.
56. S. Perrier and P. Takolpuckdee, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 5347–5393.
57. G. Pound, J. B. McLeary, J. M. McKenzie, R. F. M. Lange and B. Klumperman, Macromolecules, 2006, 39, 7796–7797.
58. S. Kozanoǧlu, T. Özdemir and A. Usanmaz, J. Macromol. Sci., Part A: Pure Appl. Chem., 2011, 48, 467–477.
59. J. Chiefari and E. Rizzardo, in Handbook of Radical Polymerization, John Wiley & Sons, Inc., 2003, pp. 629–690.
60. J. Chiefari, Y. K. Chong, F. Ercole, J. Krstina, J. Jeffery, T. P. T. Le, R. T. A. Mayadunne, G. F. Meijs, C. L. Moad, G. Moad, E. Rizzardo and S. H. Thang, Macromolecules, 1998, 31, 5559–5562.
61. C. Europe, Council of Europe, 2004, vol. 2, Monographs D–E, p. 1468.
62. G. Moad, E. Rizzardo and S. H. Thang, Polym. Int., 2011, 60, 9–25.
63. H. Willcock and R. K. O'Reilly, Polym. Chem., 2010, 1, 149–157.
64. I. R. Nasimova, E. E. Makhaeva and A. R. Khokhlov, J. Appl. Polym. Sci., 2001, 81, 3238–3243.
65. J. F. Coelho, P. C. Ferreira, P. Alves, R. Cordeiro, A. C. Fonseca, J. R. Góis and M. H. Gil, The EPMA Journal, 2010, 1, 164–209.
66. J. Hu and S. Liu, Macromolecules, 2010, 43, 8315–8330.
67. S. Sun and P. Wu, J. Phys. Chem. B, 2011, 115, 11609–11618.
68. D. Schmaljohann, J. Oswald, B. Jørgensen, M. Nitschke, D. Beyerlein and C. Werner, Biomacromolecules, 2003, 4, 1733–1739.
69. Q. H. Wu, J. Yi, Z. L. Yin, S. Y. Wang, Q. Yang, S. Y. Wu, X. M. Song and G. L. Zhang, J. Polym. Res., 2013, 20, 262.

---

Footnote

† Electronic supplementary information (ESI) available: Methods and characterization details; Fig. S1–S16. See DOI: 10.1039/c6ra01014h

---