Tailor-Made , Linear , and “ Comb-Like ” Polyester-Based Copolymers : Synthesis , Characterization , and Thermal Behavior of Potential 3 D-Printing / Electrospinning Candidates

Planta Piloto de Ingeniería Química, PLAPIQUI (UNS-CONICET), CP: 8000, Camino “La Carrindanga” Km 7, Bahía Blanca, Buenos Aires, Argentina Departamento de Ingeniería Química, Universidad Nacional del Sur (UNS), CP: 8000, Av. Alem 1253, Bahía Blanca, Buenos Aires, Argentina Departamento de Ingeniería Química, Facultad de Ciencias Aplicadas a la Industria (FCAI), Universidad Nacional de Cuyo (UNCuyo), CP: 5600, Bernardo de Irigoyen 375, San Rafael, Mendoza, Argentina Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Godoy Cruz, 2290 Ciudad Autónoma de Buenos Aires, Argentina Universidade Federal do Pampa, Unipampa (97650-000), Itaqui, RS, Brazil Departamento de Química, Universidade Federal de Santa María, 97105-900 Santa María, RS, Brazil


Introduction
During the last decades, the development of new polymerization techniques (which involve less sophisticated procedures from a practical point of view) has noticeably increased the potential of polymer materials to expand their uses [1].Complex macromolecular structures are now easily achieved, and a broad range of applications for new polymeric materials can be proposed at levels scarcely envisioned a few decades ago [2,3].Reversible deactivation radical polymerization (RDRP) techniques, such as atom transfer polymerization (ATRP), nitroxide-mediated polymerization (NMP), and reversible addition-fragmentation chain-transfer (RAFT) polymerization, are now well-established, and their uses in the synthesis of macromolecules are widely employed in synthetic laboratories worldwide.
For a long time, graft and block copolymers have received the attention of polymer scientists due to their attractive macromolecular characteristics and properties.This interest lies on the combination of thermodynamically incompatible macromolecular species linked together, which imposes restrictions regarding their properties and final applications.Moreover, the processability of these materials is a crucial point that will justify their further uses.In such a sense, many processing parameters such as viscosity-temperature dependence, glass-transition temperature, thermal stability, and melting temperature of semicrystalline blocks or segments are critical to decide whether or not using them [4].
Recent developments regarding biocompatible materials have enabled 3D bioprinting of functional living tissues, mainly applied to regenerative medicine to address the need for organ transplantation [5][6][7].In this aspect, 3D bioprinting materials are limited to biocompatible synthetic and natural polymers, such as targeted synthetic diblock copolymers, sodium alginate, chitosan, and acrylate-based polymers, among others [6].In fact, many ink formulations or 3Dprintable materials for biomedical purposes are made from copolymers or from the combination of complex macromolecular structures, which can be obtained by adjusting wellknown polymerization pathways [8,9].
Another area of interest in which biocompatible materials deserve special interest is electrospinning.By using this technique, the preparation of nanofibers from different materials (synthetic or natural polymers, composites, ceramics, and even inorganic oxides) can be strategically targeted in order to obtain specific properties [10].Medical studies have used this technique to design nanofiber for drug delivery systems as well as scaffolds for tissue engineering [11].The processability of the polymeric materials, their solubility behavior, and their concentration are some of the parameters of interest to produce targeted nanofiber mats or meshes.
Poly(ε-caprolactone) (PCL) is an aliphatic and hydrophobic polyester with good biodegradable and mechanical properties.This semicrystalline polymer has a relative low melting temperature (T m ~60 °C) and an appropriate glass-transition temperature for technical applications (T g ~− 60 °C).In addition, the aliphatic nature of the main chain explains its compatibility with a large variety of polymers [12], whereas its lack of toxicity is a strong reason to use it as an excellent candidate in a wide range of biomedical applications such as long-term implantable devices, scaffolds for tissue growth, and drug delivery systems, among others [13][14][15].PCL can be easily synthesized by the ring-opening polymerization (ROP) of ε-caprolactone (ε-CL) monomer promoted by suitable catalysts, such as tin(II) octanoate and a hydroxyl initiator [12].Regarding its use as 3Dprinting material, it has been reported that PCL nanoparticles have been used as ink that can be printed on a bioadhesive film for the local treatment of cervical cancer, corroborating that the printing process does not disrupt the structure of nanoparticles nor inclusion complexes [16].
Poly(2-hydryoxyethyl methacrylate) (PHEMA) is another biocompatible polymer, which is widely employed to produce hydrogel materials for ocular devices and soft contact lens [17].These applications are possible due to the hydrophilicity of the hydroxyl moiety, which imposes to PHEMA-based hydrogel physical properties that resemble those of living tissue: high water content, soft and rubbery consistency, and low interfacial tension [18].HEMA monomer can be copolymerized with other biocompatible monomers (such as ε-CL), by employing conventional freeradical polymerization mechanisms or RDRP techniques.Particularly, PCL chains grafted onto the PHEMA backbone can be obtained by RAFT polymerization [19][20][21][22], since this method allows tuning the molar masses distribution and the polydispersity indexes of the resulting complex macromolecular architectures like stars, blocks, or graft copolymers without needing sophisticated experimental procedures [23].
Poly(siloxanes) belong to the family of organic/inorganic materials, such as poly(phosphazenes), and they can be obtained from cyclic siloxane monomers.Particularly, poly(dimethylsiloxane) (PDMS) is the homopolymer made from siloxane cycles whose chemical structure is [-(CH 3 ) 2 SiO-] n , in which n denotes 3 or 4 repeating units.PDMS has unique properties such as high thermal and electric stability, and low glass-transition temperature (T g ~− 123 °C).Besides, PDMS has no chemical reactivity and displays a marked hydrophobicity and excellent physiological compatibility [24,25].The conventional synthesis of PDMS is carried out by hydrolysis and condensation reactions of dimethyldichlorosilanes in alkaline or acid media, by producing cyclic and linear polymers without a good control of the homogeneity [26].A better control of PDMS molecular parameters can be achieved by using anionic ROP of cyclic siloxanes [27][28][29].Although PDMS and its derivatives are widely employed in technical applications, their use as 3D-printing materials has been proved challenging due to poor mechanical properties and need for supporting during the process [30,31].Consequently, the searching of likely alternatives for these materials constitutes an area of broad interest.
In previous papers, we have reported the synthesis of PCL-based copolymers by employing controlled polymerization techniques [27,28,32].In an effort to expand the understanding of their physicochemical characteristics, the aim of this work was focused on the thermal behavior (at moderate temperatures) of two block copolymers series based on ε-CL monomer.In this sense, poly(dimethylsiloxane-block-εcaprolactone), P (DMS-b-ε-CL), poly(hydroxyethylmethacrylate-co-(hydroxyethylmethacrylate-graft-ε-caprolactone)block-ε-caprolactone), and P (HEMA-co-(HEMA-g-ε-CL)b-ε-CL) copolymers were synthesized by combining anionic, ROP, and RAFT polymerizations.Besides synthesis, the work is focused on the crystallization and thermal behavior of these polyester-based copolymers, both containing biocompatible blocks.
PCL has been employed as 3D bioprinting material, and it is expected that biocompatible block copolymers based 2 International Journal of Polymer Science on PCL might expand its applications in the field.As it is known, in material extrusion techniques (such as 3Dprinting or electrospinning devices), polymers are directly pushed out through a nozzle when a constant pressure is applied.The extruded material is deposited at a constant speed and fully solidifies on the substrate after coming out from the nozzle [6].In this sense, the understanding of the crystallization behavior of branched and linear copolymers based on PCL at temperatures lower than that corresponding to the linear homopolymer could lead to changes in processing parameters, especially if they are targeted for specific purposes.In such a sense, the thermal stabilities of the studied polymers were analyzed, and Avrami's isothermal crystallization model was employed to calculate kinetic parameters from the crystallization process [32][33][34][35][36].It is inferred that data obtained from this analysis will provide useful information for further applications, particularly if these materials or combinations of them are planned for using as potential inks in 3D-printing or electrospinning devices [5][6][7]16].

Synthesis of P (DMS-b-ε-CL) Block Copolymers (SCL#).
Polymerizations were carried out in two steps.First, a model poly(dimethylsiloxane) hydroxyl-ended homopolymer (PDMS-OH) was synthesized by employing anionic polymerization and hand-made polymerization apparatuses (glass-blowing techniques), by a vacuum manifold (VM), and by following the methodology reported in our previous works [3,39,40].Briefly, the reactor was attached to the VM and checked for pinholes.Then, the stopcock was closed and approximately 5 mL of commercial n-butyllithium solution in hexane (n-Bu − Li + 2 M, Aldrich) was injected into the purge section flask.After that, the stopcock of the VM was opened and the apparatus was pumped during 30 minutes in order to remove traces of hexane and air inside.Then, the reaction solvent was distilled from a collection flask already attached to the VM.After two freezing-thawing cycles, the apparatus was removed from the VM by heat-sealing, and it was washed with the diluted n-Bu − Li + solution inside.After washing, the solvent was distilled from purge section to the reactor, and once distillation was completed, the purge section was removed by heat-sealing procedures, leaving the clean reactor ready for polymerization.Then, the D 3 monomer was poured into the reactor flask, followed by the addition of sec-Bu − Li + .The reaction was left to proceed at room temperature during ~20 hours [39].Finally, THF was added and D 3 polymerization was left to proceed, at room temperature, during 20 hours.The reaction was finished by the addition of well-degassed methanol.The resulting PDMS was precipitated in cold methanol, traces of solvents were removed by employing a roto-evaporator (~60 °C), and the final product was kept in vacuum until constant weight.
Copolymerization of PDMS-OH with ε-CL monomer was carried out in a specially designed Schlenk-type reactor by following the methodology already published [42][43][44].In this sense, copolymers with a chemical composition ranging from low to intermediate PCL content were obtained by ROP of cyclic monomer.Copolymerization was performed under nitrogen atmosphere, employing degassed toluene as reaction solvent and a catalyst/PDMS-OH ratio of 0.5.Reactions were carried out at ~100-110 °C during 24 h.The resulting copolymers were precipitated in cold methanol, filtered, and vacuum-dried, at 20 °C, until constant weight.Since in this synthetic procedure the amount of ε-CL monomer was adjusted in each reaction, P (DMS-b-ε-CL) copolymers were labelled as SCL#, in which # refers to PCL content from low (1) to high (4) wt%.

Synthesis of P (HEMA-co-(HEMA-g-ε-CL)-b-ε-CL)
Copolymers (HEMACL#).Copolymerizations were carried out in a one-pot procedure through a combination of simultaneous RAFT and ROP polymerizations, by following a methodology already published [28].BSTSE, HEMA and ε-CL monomers, Vazo-88 radical initiator, dry toluene, and DMF were placed in a dry Schlenk flask equipped with a stirrer.The tube was closed, subjected to three freezepump-thaw cycles in the VM, and subsequently backfilled with nitrogen gas.The ROP catalyst DPP was then added under gentle nitrogen flow.Subsequently, the flask was closed and immediately immersed in an oil bath settled at 100 °C to promote the simultaneous copolymerization reaction.The reaction was left to proceed during 24, 2, or 1 h, and the final product was precipitated in cold methanol and dried in vacuum until constant weight [27,28].Since in this synthetic procedure the amount of grafted chains depends on the reaction time, P (HEMA-co-(HEMA-g-ε-CL)-b-ε-CL) copolymers were labelled as HEMACL#, in which # denotes low (1) to high (3) reaction times.Fourier transform infrared (FTIR) spectroscopy spectra of selected samples were obtained on a Nicolet® FTIR 520 spectrometer.Cast films from diluted solutions of the samples (1 wt% in THF) were obtained onto NaCl windows.FTIR spectra were recorded at 4 cm −1 resolution, over the 4000-400 cm −1 range, 64 scans, and dried air as background.

Nuclear Magnetic
Resonance ( 1 H-NMR). 1 H-NMR spectra of SCL# copolymers were recorded on a Bruker® 300 MHz instrument, by using deuterated chloroform (CDCl 3 , Aldrich) as solvent.About 5-10 mg of sample was diluted directly in the NMR tubes, at room temperature.The chemical composition of synthesized copolymers was obtained by using the values of integrated areas from characteristic 1 H signals of each block.For HEMACL# copolymers, 1 H-NMR spectra were acquired using an Avance DPX-400 spectrometer (400 MHz for H, 100 MHz for C).The spectra were obtained by dissolving a small quantity of the polymer sample (~10 mg), at room temperature, by using CDCl 3 as solvent.

Size Exclusion Chromatography (SEC)
. SCL# samples were characterized by SEC on a system built with a Waters® 515 HPLC pump and a Waters® model 410 differential refractometer, equipped with three mixed-bed PLgel linear columns and a precolumn with 5 μm bead size (PLgel).Elution solvent was toluene (Aldrich), flowing at a rate of 1 mL min −1 at room temperature.Injection volume was 200 μL, and PS standards were used for calibration.Mark-Houwink calibration constants used were K PS = 0 012 mL g −1 , α PS = 0 71 and K PDMS = 0 0136 mL g −1 , α PDMS = 0 69 for PDMS [32,45].SCL# were also analyzed by employing the same equipment by using THF (Ciccarelli) as solvent in order to distinguish PCL blocks in the resulting copolymers.In this case, the Mark-Houwink calibration constants used for PCL were K PCL = 1 395 × 10 −4 mL g −1 and α PCL = 0 786 [46].For HEMACL# samples, toluene was used as solvent employing the aforementioned equipment.Mark-Houwink constants employed for PCL were K PCL = 0 01298 mL g −1 , α PCL = 0 828 [47].

Differential Scanning Calorimetry (DSC).
Thermal analysis was performed by employing DSC Pyris 1 Perkin-Elmer® equipment.Samples were measured under He atmosphere, by using approximately 10 mg of each copolymer.For SCL# copolymers, samples were heated from −90 to 100 °C.Then, they were kept at 100 °C during 5 minutes in order to avoid the influence of the previous thermal history.After cooling at 10 °C min −1 , they were heated again from −90 to 100 °C at a heating rate of 10 °C min −1 .From the second heating, glass-transition temperature (T g ) of PCL block and melting temperature (T m ) of PDMS and PCL blocks were determined.For HEMACL# copolymers, samples were heated from 15 to 100 °C.Then, they were kept at 100 °C during 5 minutes in order to avoid the influence of the previous thermal history.After cooling at 10 °C min −1 , they were heated again from 15 to 100 °C at a heating rate of 10 °C min −1 .From the second heating, T m of PCL blocks and T g of PHEMA blocks were determined.
2.4.5.Isothermal Crystallization Tests.Isothermal crystallization tests were carried out by using the above-mentioned equipment and by employing sequential steps described as follows: first, samples were heated from 15 to 90 °C at 10 °C min −1 .Then, they were kept at 90 °C during 5 min and rapidly cooled (60 °C min −1 ) to selected crystallization temperatures (T c ) according to the methodology reported in previous works [32,34,48,49].Isothermal crystallization tests were carried out for temperatures between 16 and 24 °C, by using 2 °C steps.Finally, after isothermal crystallization, samples were heated up to 90 °C (10 °C min −1 ) in order to determine the heat of fusion (ΔH f ).In all cases, the heat of crystallization (ΔH c ) as a function of time (t) was recorded at different isothermal T c .

Modulated Thermogravimetric Analysis (MTGA).
Thermal stability was analyzed by using Discovery TA Instruments TGA 5500 equipment.Tests were carried out under nitrogen atmosphere, with a flow rate of 25 mL min −1 and 2 °C min −1 heating rate, in the 40-700 °C range.Weight loss curves as a function of temperature were registered.Besides, modulated thermogravimetry assays were performed to obtain continuous kinetic information regarding decomposition reactions.MTGA tests were made by using an oscillatory temperature program.Samples were analyzed by triplicate, and the resulting data (three values average) are informed with their associated standard deviation.The results obtained were compared with those for linear PCL in order to analyze the influence of the neighbor block and random branches on the thermal stability of the resulting copolymers.

Results and Discussion
SCL# copolymers were synthesized by combining classical anionic polymerization followed by ROP.PDMS-OH homopolymer was employed as macroinitiator for ROP of ε-CL monomer.Using this synthetic pathway, four different copolymers from low to intermediate PCL content were obtained.Scheme 1 shows the reaction pathways employed for this purpose.
Regarding HEMACL#, these copolymers were synthesized through a combination of RAFT and ROP polymerizations in a one-pot reaction, as it was reported in our previous work [28].Scheme 2 shows the reaction pathway employed for this purpose, in which ROP of ε-CL simultaneously takes place from the hydroxyl moieties from HEMA monomer and BSTSE RAFT agent, respectively.
The reaction proposed in Scheme 2 produces graft PCL branches in the resulting HEMACL# copolymers, whose distribution will strongly depend on reaction time (2, 1, and 24 hours for HEMACL1, HEMACL2, and HEMACL3, respectively).The macromolecular architecture obtained can be thought as a "densed and/or un-densed comblike copolymers," in which the main skeleton is a block 4 International Journal of Polymer Science copolymer based on HEMA and ε-CL, and teeth are random PCL blocks pending from HEMA monomer units.Consequently, HEMACL# copolymers can be thought as "pseudo" branched PCLs.
According to chemical characterization, Figure 1 shows selected FTIR spectra from PDMS-OH and PCL homopolymers, as well as SCL1 and HEMACL3 copolymers.PDMS-OH and PCL spectra were shifted in the transmittance axis in order to display the common absorption bands that appear in both copolymers.The broad absorption band at 3900-3200 cm −1 is associated with the stretching of the -OH group in PDMS-OH homopolymer.The spectrum of this polymer also shows an absorption band at 2963 cm −1 associated to the vibration bonds of C-H attached to Si atoms [50].At 1261 cm −1 , the out-of-phase vibrations of Si-(CH 3 ) 2 and O-Si-OR are also detected [1].Besides, at 1093 cm −1 and 1024 cm −1 , absorption bands associated to the vibration of Si-O-Si and C-Si-C bonds, respectively, can be observed [3,32].The PCL spectrum shows an absorption band at 1724 cm −1 associated to the stretching vibrations from carbonyl >C=O groups, as well as at 1240 cm −1 (vibrations of O-C=O) [33,42].On the other hand, FTIR spectra from SCL1 and HEMACL3 copolymers exhibit the typical absorption bands from their constitutive blocks, as it was reported by several authors [32,42,51].In this sense, PCL absorption signals were detected at 1724 cm −1 and 1240 cm −1 .In addition, for HEMACL# copolymer a strong stretching band at 3500-3100 cm −1 is observed, which should be assigned to -OH vibration as it was already reported by Sahiner and Demirci [51].Besides, aliphatic C-H stretching bands at 2985 cm −1 and 2854 cm −1 are also detected [51], together   [27,28].Figure 2 shows typical 1 H-NMR spectra from SCL3 and HEMACL2 copolymers.The characteristic signals from constitutive blocks and the corresponding δ (ppm) assignments are displayed in the spectra.According to the NMR spectrum provided in Figure 2 for HEMACL2, the difference between grafted and ungrafted HEMA units is the d and f signals that belong to methylene units attached to free -OH groups (d) or reacted -OH groups by esterification (f).
For SEC analysis, Figure 3 displays SEC chromatograms for PDMS-OH macroinitiator and SCL4 copolymer (toluene as solvent).The PDMS homopolymer was a model polymer, with a low polydispersity index (M n = 13,000 g mol −1 , M w / M n = 1 07, Table 1).On the other hand, the SCL4 copolymer shows an increase in M n value (M n = 65,000 g mol −1 ), with M w /M n = 1 54 (Table 1).For the other copolymers, M n values show an increasing trend as soon as w PCL increases, and M w /M n indexes display values between 1.36 (SCL1) and 1.45 (SCL3).Although M w /M n values for SCL# seem to be far from model samples, they are quite good by taking into account that ε-CL monomer was employed as received in order to simplify synthetic procedures.As it can be seen from Table 1 and Figure 3, SCL4 displays a broad molar masses distribution (when compared to PDMS-OH) and a small shoulder at higher molar masses values.These facts might be assigned to the presence of water or hydroxyl impurities from commercial ε-CL monomer as well as inaccuracies in the refractive index detector for PCL blocks for the solvent employed (toluene).The presence of water or hydroxyl impurities might lead to uncontrolled ROP due to the formation of stable tin alkoxides [52,53], whereas the use of toluene as solvent results in a less defined chromatogram when compared to that from the initial  6 International Journal of Polymer Science PDMS-OH macroinitiator.It is inferred that, with an additional purification step of commercial ε-CL monomer (for example, a previous treatment with calcium hydride), the presence of water or hydroxyl impurities will be diminished or suppressed.HEMACL# copolymers exhibited M n values between 32,300 g mol −1 and 46,000 g mol −1 , and M w /M n lower than 1.2, indicating a good control of the process over the chemical characteristics of the resulting macromolecules (Table 1).This fact proves the efficiency of the one-pot procedure to produce macromolecular architectures by employing RAFT and ROP simultaneously.As an additional fact, the expected molar masses can be achieved at reasonable reaction times (at least, 1 h) which is another feature that supports the efficiency of the proposed procedure.At this point, it is worth mentioning that HEMACL# copolymers deserve an additional purification process.In such a sense, Figure 4 displays SEC chromatograms from the crude reaction product and the resulting HEMACL3 copolymer after the purification process.As it can be seen, the crude reaction product contains residual PCL homopolymer (from BSTSE as well as from HEMA monomer) that disappears by successive precipitation employing controlled temperature and nonsolvent (methanol) fractionation.DSC measurements were performed in order to evaluate thermal transitions from linear and "graft-like" PCL-based copolymers when compared to linear PCL (at this point, it should be mentioned that the temperature range under study was selected by taking into account the T m of the PCL block from HEMACL# copolymers, which is lower than that corresponding to the PCL homopolymer employed for comparison purposes; for more details, please refer to Table 1).For SCL# samples, three transitions are detected (Figures 5(a On the other hand, for linear PCL, T mPCL is detected at ~56 °C (Figure 5(c)), whereas for HEMACL# copolymers, T mPCL from PCL grafts is observed at ~40 °C (Figure 5(d)).This reduction on the T m of PCL could be explained by taking into account the presence of random grafted PCL chains along the PHEMA backbone [27,28,54].It is reasonable to suppose that longer reaction times allow higher density of PCL grafts (HEMACL3), whereas lower PCL grafts are expected for shorter reaction times (HEMACL2).This is quite reasonable since BSTSE acts as both a RAFT and ROP initiator.For short reaction times, few PCL grafts are attached from the PHEMA block because reaction time is not enough to achieve a good RAFT conversion.On the contrary, at higher reaction times, the probability to have randomly distributed grafts along PHEMA blocks increases.In addition, for these copolymers, a final thermal event around 68 °C (associated to T gPHEMA from PHEMA block) was also observed (Figure 5(d) and Table 1) [55].
Isothermal crystallization curves of linear PCL, SCL#, and HEMACL# copolymers are displayed in Figure 6, in which the evolution of endothermic heat flow was recorded as a function of time.In Figure 6(a), the isothermal crystallization curves of linear PCL are plotted at different T c values.As it can be seen, crystallization time is increased at higher T c values, indicating that nucleation is the limiting factor for crystallization in this temperature range [56].A similar behavior was observed for SCL# and HEMACL# copolymers.As examples, Figure 6(b) shows the crystallization evolution of linear PCL, SCL2, and HEMACL2 at T c = 20 °C.It is important to note that the completed crystallization processes of SCL2 and HEMACL2 copolymers occur at higher times when compared to linear PCL.For SCL2, this rise can be associated with the noncrystallizable PDMS block (Table 2).In this sense, Lovinger et al. [57] reported that the PDMS block avoids the regular crystallization of PCL chains because of the higher mobility of the viscoelastic, amorphous PDMS block throughout the crystallization process.On the other hand, HEMACL# copolymers show a wider difference, because the presence of random, grafted branches hinders the formation of PCL crystallites by increasing crystallization time values (Table 2) [12].
As it is reported in the literature [35,58], isothermal crystallization kinetics can be better visualized by considering the relative degree of crystalline conversion as a function of time at different crystallization temperatures (Figure 7).Then, relative crystallinity, related to the crystalline volume (V c ), vs. different crystallization times (t), can be calculated from [1,32,35,49] where ΔH t and ΔH ∞ are the heat generated at time t and at long time periods, respectively, and dHc/dt is heat evolution rate.V c is obtained from the area of the exothermic peak on isothermal crystallization and is plotted vs. time (Figure 7).The slope of the graphs appears as an S-like shape.The Slike shape in the graphs with higher slopes reveals higher As examples, at a given T c value, SCL# and HEMACL# copolymers crystallize at a lower rate when compared to linear PCL as it can be deduced by exploring time values at a given relative crystallinity % (Figure 7(d)).
The above-described results allow employing Avrami's theory to obtain kinetics parameters of interest such as the half-life time for crystallization (t 1/2 ), the bulk crystallization constant (k), and Avrami's exponent (n), by following [2,15,49] This equation can be rewritten in a double logarithmic form as follows: where V c is related to the relative crystallinity from polymers at different temperatures or times, and t − t 0 is the crystallization period involved in the process.The induction time (which includes the stabilization time) has been named t 0 , and it was calculated by drawing a horizontal line from a point after the finishing crystallization process to a point at the beginning of the crystallization curve, by following the methodology reported by Lorenzo et al. [34].
In addition, n represents crystallization types regarding dimensions (one-, two-, or three-dimensional) and k is a growth rate constant that accounts for both nucleation and growth rate parameters.Plots of log −ln 1 − V c vs. log t − t 0 for each crystallization temperature (T c ) (( 3)) are shown in Figure 8. Avrami's exponent n and crystallization rate k are obtained from the slope and intercept of the curves, respectively.From these curves, t 1/2 is defined as the time required for developing 50% of final crystallinity.Usually, lower t 1/2 values are expected at higher crystallization rates (Table 2).
As example, for HEMACL3, all lines were almost straight and parallel to each other, shifting to a higher time when the crystallization temperature increases (Figure 8(a)).In addition, at T c = 20 °C, the corrected Avrami's fit was evidenced for both linear PCL and copolymers (Figure 8(b)).Moreover, for SCL# copolymers, n values were similar to those obtained for linear PCL, with calculated values between 2.4 and 3.4 (see Table 2).It is important to highlight that these values refer to a three-dimensional heterogeneous nucleation and they are in good agreement with values already reported [33,36,56].On the other hand, HEMACL# copolymers show values of n exponent from 1.8 to 2.6,  M n = 46.000g mol −1 M n = 3.000 g mol −1

HEMACL3
Molar mass/g mol −1 10 4 10 5 M n = 46.000g mol −1   Figure 4: SEC chromatograms from the crude reaction product and the resulting HEMACL3 copolymer after the purification process. 8 International Journal of Polymer Science revealing that these "comb-like" copolymers adopt a twodimensional crystallization growth with heterogeneous nucleation [32,35,36,59].The effect is strongly dependent on the number of branches, as it can be easily deduced by comparing the resulting curves for HEMACL1 and HEMACL3.It is reasonable to suppose that the presence of branches imposes dimensional restrictions over the resulting crystallites, by changing from three (linear) to two (comb-like) dimensions.In Avrami's equation, k values are related with nucleation rate and growth processes.For SCL# and HEMACL# copolymers, a decrease in this parameter with an increase in T c was observed.The presence of the PDMS block in SCL# or grafted, random PCL branches in HEMACL# lead to a reduction in k values in all the studied samples (see Table 2).This reduction could be explained by taking into account the spatial configuration of PDMS block or grafted PCL branches.During the folding and growth processes of crystals, the viscoelastic silicon block as well as the random PCL branches hinder the normal crystallization event, leading to lower k values.The effect can be easily appreciated in Figure 9, in which the evolution of t 1/2 at each T c value is shown.
For SCL# copolymers, t 1/2 values exhibited an increase of 1.6 times when compared to linear PCL (Figure 9(a)).On the other hand, for HEMACL# copolymers, a difference in t 1/2 values appears at crystallization temperatures closer to T mPCL .In this sense, HEMACL2 and HEMACL3 show t 1/2 values of 18 and 42 times higher when compared to linear PCL, respectively (as it can be appreciated for temperatures between 20 and 24 °C, Figure 9(b)).From Figure 9, it is clear that "flat" curves can be obtained for SCL# copolymers whereas marked changes were observed for HEMACL# copolymers when compared to linear PCL.The "comb-like" structure of these copolymers noticeably affects the crystallization behavior of the semicrystalline PCL block since random grafts require more time for crystallization.
Finally, activation energy for isothermal crystallization was evaluated by using the Arrhenius equation [35,56,59]:  where K 0 is a temperature-independent preexponential factor, ΔE a is the total activation energy, and R is the universal gas constant.
Arrhenius plots of n −1 ln K versus T c −1 for linear PCL, SCL#, and HEMACL# copolymers were linear, and ΔE a was determined from the slope of the curves (Figure 10).Values are summarized in Table 2.For all samples, the calculated ΔE a values were negatives.In this sense, Dhanvijay et al. [36] refer this behavior to the released energy during melt to crystalline state transformation.By comparing linear PCL with SCL# and HEMACL# copolymers, the ΔE a values of HEMACL# copolymers are the highest.ΔE a values are strongly dependent on the comonomer presence, whose chemical nature and macromolecular structure affect PCL chains.Therefore, collisions and entanglements of PDMS and PCL blocks are important when the amount of PCL increases in SCL# copolymers.On the other hand, the random grafted PCL chains hinder movement and rearrangement of the semicrystalline PCL block by reducing their crystallization ability.These results According to the thermogravimetric analysis, weight loss events associated to linear and "comb-like" copolymers as function of temperature were registered (Figure 11).In addition, values of temperature at 5% of thermal degradation (T 0 05 ) were included in  Figure 11 shows MTGA curves corresponding to linear PCL homopolymer and SCL4 and HEMACL3 copolymers, respectively.As it can be seen, linear PCL presents a typical weight loss step which is associated to PCL decomposition at 341 °C [59,60].By analyzing the first derivative curve from the MTGA thermogram (not shown), for the SCL4 copolymer three thermal degradation events were detected.These events were observed at 193 °C (rupture of polyester chains via ester pyrolysis reaction producing H 2 O, CO 2 , and 5-hexenoic acid); at 289 °C (PCL block decomposition), and at 366 °C (PDMS block degradation) [32,61].On the other hand, HEMACL# copolymers showed the thermal decomposition associated to PHEMA and PCL homopolymers already reported [28,62].
All the studied materials exhibited a reduction in the maximum degradation temperature when compared to linear PCL.In this sense, thermal degradation was faster for SCL# than HEMACL# copolymers, denoting that the PCL block is less thermally stable in the 150-450 °C range.A plausible explanation might be due to the hydroxyl-end group (-OH).For block copolymers, it is reasonable to assume one -OH group per molecule (PCL block), whereas the amount increases for comb-like copolymers (-OH groups from PCL grafts and PCL block).Since the chemical nature of end-groups strongly influences the thermal behavior of PCL [63], the differences observed can be reasonable explained.The calculated activation energy values (E ad ) for    SCL# and HEMACL# copolymers were lower than the corresponding value for linear PCL (Table 1).These results agree well with previous works regarding the thermal degradation of PCL-based copolymers reported by Ninago et al. [32].

Conclusions
Linear and block copolymers based on ε-CL were successfully synthesized through a combination of anionic, RAFT, and ROP methods.Samples showed M w /M n < 1 54, which reveal a good control over macromolecular architectures.Isothermal DSC experiments (combined with Avrami's theory) were performed in order to obtain kinetics parameters regarding crystallization behavior.For the temperature range studied (from T c = 16 °C to T c = 24 °C), k values were reduced whereas t 1/2 values were increased when compared to the linear PCL homopolymer.The effect was more notorious for HEMACL# copolymers, in which branch density might affect crystallization processes due to less chain mobility and spatial hindrance.Regarding thermal stabilities, the presence of -OH groups PCL branches reduces the energy barrier of the thermal degradation process for the corresponding HEMACL# copolymers.

( a )
Number average molar mass (M n ) and polydispersity (M w /M n ) determined by SEC and 1 H-NMR;(b) weight fraction of PCL in the block copolymer determined by 1 H-NMR;(c) glass-transition temperature (T g ) and melting temperature (T m ) determined by DSC;(d) 5% thermal degradation temperature (T 0.05 ) and activation energy for degradation (E ad ) determined by MTGA.

Table 1 :
Chemical and thermal characterization of synthesized linear PCL and PDMS-OH homopolymers, and PCL-based copolymers.
Table 1 for each sample.It is important to note that T 0 05 values denote the beginning of the thermal degradation process, which is a characteristic