Structurally-defined semi-interpenetrating amphiphilic polymer networks with tunable and predictable mechanical response

Abstract

The synthesis of structurally-defined semi-interpenetrating amphiphilic networks is realized by employing an easy and versatile synthetic concept based on the encapsulation of well-defined hydrophobic linear polymer chains within a structurally-defined 1,2-bis-(2-iodoethoxy)ethane (BIEE)-crosslinked hydrophilic polymer network. Both, the hydrophobic (poly(n-butyl methacrylate)) and hydrophilic poly(2-(dimethylamino)ethyl methacrylate) linear chain precursors to the networks have been synthesized by Reversible Addition Fragmentation chain Transfer (RAFT) polymerization. A series of BIEE-crosslinked amphiphilic semi-IPN networks was prepared by retaining the network hydrophilic content and varying only the hydrophobic content from 0–50% wt. The mechanical properties of the resulting networks containing different loadings of the encapsulated hydrophobic linear chains were tested under compressive loading conditions in their aqueous swollen state. A nonlinear hyperelastic constitutive equation was used to predict the elastic response of all network structures demonstrating that for low poly(n-BuMA) loading (i.e. up to 10% wt), no change in the materials' mechanical response is observed whereas for greater loading percentages (i.e., 35% and 50%) the networks become stiffer. The present work creates new prospects in the development of amphiphilic semi-IPN polymer networks with controllable compositional and structural characteristics and predictable mechanical behaviour realized via mathematical modeling.

---

Introduction

Amphiphilic polymer networks (APN), combining hydrophilic and hydrophobic moieties within a 3-dimensional architecture represent a highly emerging research area in polymer science. Their 3D-structure and amphiphilicity combined with the existing versatility in material design and tunable chemical composition render these materials highly promising especially in the biomedical field.

APN are classified into amphiphilic polymer conetworks^1–3 and amphiphilic (semi)interpenetrating (IPN) networks.^4–7 Their main structural difference lies on the fact that in the former the hydrophilic and hydrophobic segments are linked together via covalent bonding, whereas in the latter either a secondary polymer network (in the case of the IPN) or secondary linear polymer chains (semi-IPN) are interlaced (but not covalently bonded) with a primary polymer network.

During the last years there has been an increased interest in developing new synthetic pathways targeting to the generation of APN with structurally-defined characteristics, thus allowing the structure-to-property correlation. Several reports exist so far, dealing with the synthesis of well-defined APNs (model or quasi-model) by employing different controlled polymerization methodologies such as Reversible Addition-Fragmentation chain Transfer (RAFT) polymerization,^8–10 Atom Transfer Radical Polymerization (ATRP),¹¹ Group Transfer Polymerization^12,13 etc. or combinations of the aforementioned.^14,15 However, the above-mentioned polymerization methodologies have been mainly employed for the preparation of amphiphilic polymer conetworks whereas only a limited number of reports exist on the synthesis of (semi)interpenetrating (IPN) networks with pre-defined and controlled molecular structures.^16,17 In a first example, Li and co-workers have prepared a series of pH-responsive semi-IPN structural hydrogels consisting of PVA and well-defined poly(2-dimethylamino)-ethyl methacrylate (poly(DMAEMA)) star polymers.¹⁶ The authors have demonstrated that with the optimum combination of PVA with the PDMAEMA star polymer of an appropriate molecular weight, the pH-triggered controlled drug delivery of encapsulated pharmaceutical molecules could be realized.

In the second example, the authors describe the preparation of well-defined sliding-graft semi-IPN of PEG and poly(2-hydroxyethyl methacrylate) (PHEMA).¹⁷ The synthetic methodology involved the simultaneous occurrence of two reaction processes: (a) the copper(I)-catalyzed azide–alkyne cycloaddition (i.e. click chemistry reaction) of a diazide-functionalized PEG chain/Br-functionalized α-cyclodextrin (CD) inclusion complex with a dialkyne organic molecule namely tetrakis(2-propynyloxymethyl)methane (TPOM) to generate the network and (b) the ATRP polymerization of 2-hydroxyethyl methacrylate using the aforementioned Br-functionalized inclusion complex as the polymerization initiator.

Very recently we have reported on the synthesis of polymer (co)networks, with pre-defined and controlled architectures involving the crosslinking of well-defined poly(DMAEMA) homopolymers and polyDMAEMA-containing block copolymers prepared by RAFT, using 1,2-bis-(2-iodoethoxy)-ethane (BIEE) as a crosslinker.¹⁸ In an effort to correlate the structural characteristics of these materials to their mechanical response, we have studied the mechanical properties in compression whereas their accurate prediction was feasible by mathematical modeling using an exponential constitutive equation.¹⁹

In the present study we expand the BIEE-based synthetic concept beyond APN, targeting towards the preparation of structurally-defined semi-interpenetrating amphiphilic networks based on well-defined hydrophilic and ionisable polyDMAEMA homopolymer chains (generating the covalent hydrophilic network) and well-defined hydrophobic poly(n-butyl methacrylate) (PBuMA) homopolymer chains encapsulated within the network upon gelation in the presence of BIEE. Both, polyDMAEMA and polyBuMA linear precursors have been prepared by RAFT controlled radical polymerization.

The mechanical properties of the resulting structurally-defined amphiphilic semi-IPN networks preswollen in aqueous media containing different poly(n-BuMA) loadings were tested under compressive loading conditions. A common nonlinear hyperelastic constitutive equation was used to predict the elastic response of all network structures. By fitting this equation to the experimental data, the corresponding material constants were derived.

The present work creates new perspectives in regards to the design and development of structurally-defined amphiphilic semi-IPN polymer networks since it could be universally applied as an alternative to more synthetically-demanding and multi-step controlled polymerization processes for the generation of APN with variable functions (deriving from the selection of the linear hydrophobic component) and with tunability in regards to their mechanical performance.

Results and discussion

Synthesis of semi-interpenetrating amphiphilic networks

Initially, polyDMAEMA and polyBuMA linear homopolymers were prepared via RAFT polymerization. Unimodal polymers with controlled molecular weights (MWs) and narrow molecular weight distributions (MWDs) were obtained in both cases as demonstrated by Size Exclusion Chromatography (SEC), whereas their chemical structures were verified by ¹H NMR spectroscopy. These well-defined linear homopolymers were then employed as precursors to construct amphiphilic semi-interpenetrating polymer networks with structurally-defined characteristics. The presented synthetic concept involves the encapsulation of the hydrophobic polyBuMA linear chains within a polyDMAEMA hydrophilic network generated via the interconnection of the linear polyDMAEMA chains in the presence of BIEE. The latter reagent is capable of reacting with two adjacent tertiary amine residues that are present on the DMAEMA moieties causing their interconnection (as shown in Fig. 1) resulting to the generation of a covalent polymer network.^20–22

Fig. 1 Chemical structures of the BIEE crosslinking agent and the linear poly(DMAEMA) and poly(BuMA) homopolymers and schematic of the one-step synthetic methodology followed for the preparation of semi-interpenetrating APN with structurally-defined characteristics.

In all cases gelation was reached within 6 days under ambient conditions in the presence of air and without mechanical stirring. The sol fraction percentages (extractables) ranged between 6 and 14% which is comparable to those reported in our previous study.¹⁸

Swelling behaviour

The degrees of swelling (DSs) of the semi-interpenetrating APN networks and of the BIEE-crosslinked polyDMAEMA network (control sample) were determined in water and the obtained values together with 95% confidence intervals are provided in Table 1. From the obtained data it can be seen that the DS decreases upon increasing the content of the linear poly(n-BuMA) hydrophobic chains as expected, since the insolubility of the latter in water, renders the whole material less hydrophilic. It is noteworthy to mention that at a 50% wt hydrophobic content the DS decreases by almost 50% in comparison to the control (BIEE-crosslinked polyDMAEMA network) sample.

Table 1 Degrees of swelling with 95% confidence intervals of the water-swollen BIEE-crosslinked networks with different linear poly(n-BuMA) hydrophobic content

% wt content of polyBuMA	Degrees of swelling
0%	41.2 ± 0.6
5%	34.7 ± 0.5
10%	25.8 ± 0.9
35%	22.1 ± 0.3
50%	19.3 ± 0.6

---

Mechanical response

A comparison of the stress–strain experimental curves under unconfined compression (Fig. 2a) as well as the measured mechanical properties (Fig. 2b) of the structurally-defined amphiphilic semi-IPN networks containing different poly(n-BuMA) loadings shows that for low concentrations of poly(n-BuMA) loading (i.e. 0%, 5% and 10%), no change in the mechanical response of the materials is observed. However, for greater loading percentages (i.e., 35% and 50%) the networks become stiffer, as the stresses are higher for the same strain resulting in higher values of Young's modulus. This is attributed to the fact that when the networks are immersed in water, the hydrophobic poly(BuMA) encapsulated linear chains are characterized by a higher stiffness and consequently a higher modulus in comparison to the highly hydrated poly(DMAEMA) chains. Moreover, as discussed above, upon increasing the BuMA content the DS decreases, resulting to further increase in the materials' stiffness.

Fig. 2 (a) Stress–strain experimental curves. The figure depicts the stress–strain response under unconfined compression of the amphiphilic semi-IPN networks containing different poly(n-BuMA) loadings. Higher values of loading with poly(n-BuMA) increases the applied stress on the specimens making the polymers stiffer. (b) Young's modulus derived from the experimental stress–strain curves. The Young's modulus is larger for higher values of loading.

The obtained results are in line with a previous report by M. Vamvakaki and C. S. Patrickios on amphiphilic model co-networks synthesised by Group Transfer Polymerization consisting of 2-(dimethylamino)ethyl methacrylate and n-butyl methacrylate units.²³ More precisely, the authors reported that the elastic modulus of the water-swollen networks determined under compression was found to increase linearly with the polyBuMA content only when the latter was above 50% mol whereas at lower hydrophobic contents the networks exhibited poor mechanical properties. Moreover, when the networks were left to equilibrate in aqueous media of high pH prior to the mechanical measurements, it was found that the Young's modulus of the networks was not significantly influenced upon altering the BuMA content up to 50% mol suggesting that the mechanical properties of the co-networks were mainly governed by the poly(DMAEMA) blocks.²³ Similarly, it is suggested that the mechanical properties of the semi-IPN amphiphilic networks presented herein, are controlled by the BIEE-crosslinked polyDMAEMA network and only above a certain weight percentage, the encapsulated polyBuMA linear chains begin to contribute to the materials' mechanical response. Besides the effect of the weight percentage of the encapsulated polyBuMA linear chains on the materials' mechanical behaviour, it is likely that the molecular weight (MW) of the encapsulated hydrophobic linear chain would play a significant role to the overall mechanical response. Further experimental work will be carried out in the future, in order to verify whether a systematic increase in the MW of the encapsulated polyBuMA chains (by keeping the length of the polyDMAEMA chains constant) would result to a systematic decrease in the required polyBuMA weight % in order to observe polyBuMA-induced sensitivity of the elastic modulus.

Constitutive modeling of the networks mechanical behaviour

As shown in Fig. 3, the mechanical response of the networks is nonlinear. Based on this observation, the Mooney–Rivlin hyperelastic constitutive equation (eqn (1)) was employed to simulate the stress–strain experiments and fit the constitutive equation to the experimental data. Representative fits for all five cases tested are presented in Fig. 3. The values of the mechanical properties derived by fitting the model to the experimental results are summarized in Table 2.

Fig. 3 Representative fits of the Mooney–Rivlin constitutive equation (eqn (1)) for all network types. Dash lines present the model predictions and diamonds the experimental data.

Table 2 Values of the mechanical properties derived by fitting the model to the experimental stress–strain curves. C₁, C₂ are the material constants of the Mooney–Rivlin model, χ² the sum of the squared errors, μ the shear modulus and ν the Poisson's ratio

% wt content of polyBuMA	C1 (Pa)	C2 (Pa)	χ^2	μ (Pa)	ν
0%	2528.57 ± 215.93	51.43 ± 10.80	0.010 ± 0.003	5258.86 ± 453.46	0.43 ± 0.2
5%	2266.67 ± 348.54	58.33 ± 7.61	0.011 ± 0.003	4649.94 ± 712.31	0.45 ± 0.1
10%	2316.67 ± 240.85	51.67 ± 10.65	0.013 ± 0.002	4736.68 ± 503.12	0.43 ± 0.1
35%	3920.00 ± 404.55	62.00 ± 12.96	0.012 ± 0.002	7964.00 ± 835.02	0.40 ± 0.1
50%	5280.00 ± 355.98	52.00 ± 12.77	0.007 ± 0.002	10 664.00 ± 737.50	0.30 ± 0.1

---

The Mooney–Rivlin equation can reproduce very well the experimental data, and thus, it can be employed to predict the mechanical response of these networks – using the corresponding values of the material properties – under compressive loading conditions.

From the values of the mechanical properties provided in Table 2 we observe that the material constant C₁ increases for higher values of loading with poly(n-BuMA) and the parameter C₂ remains almost the same. For uniaxial tension/compression experiments, the term 2(C₁ + C₂/λ), where λ the stretch ratio calculated from the strain ε, as λ = 1 + ε, corresponds to the shear modulus, G. Furthermore, for homogeneous, isotropic materials the relationship between the shear and the Young's modulus is E = 2G(1 + ν), with ν the Poisson's ratio. From Table 2, we find that the predicted by the Mooney–Rivlin model shear modulus has values of 4.7–5.3 kPa for 0–10% wt polyBuMA content, 7.9 kPa for 35% wt polyBuMA content and 10.6 kPa for 50% wt polyBuMA content. From the experimentally derived Young's modulus (Fig. 2b) and the predicted shear modulus the Poisson's ratio of the networks is found to decrease from ∼0.45 for the first three concentrations to ∼0.40 for the 35% wt polyBuMA content and ∼0.30 for the 50% concentration of polyBuMA. Interestingly, the higher the concentration of polyBuMA the more compressible the networks become.

Conclusions

We have demonstrated a simple and synthetically undemanding methodology that leads to the generation of a series of semi-interpenetrating amphiphilic polymer networks with variable hydrophobic content (ranging from 0–50% wt), possessing structurally-defined characteristics. The latter derive from the fact that all structural components of these materials i.e. the hydrophilic polyDMAEMA linear segments constructing the covalent 3D network and the encapsulated hydrophobic polyBuMA linear chains, originate from RAFT controlled radical polymerization. The network formation step is realized in air, at room temperature and in the absence of mechanical stirring by employing 1,2-bis-(2-iodoethoxy)ethane (BIEE) as the crosslinking agent. Most importantly, the well-defined structural characteristics of the resulting amphiphilic semi-IPN networks containing different polyBuMA loadings, allowed for the mathematical prediction of their mechanical response under compressive loading conditions performed on water-swollen specimens, by using a nonlinear hyperelastic constitutive equation. The obtained results demonstrated that the networks become stiffer only at high poly(n-BuMA) loading percentages whereas no statistically significant effect on the mechanical behaviour is observed at low poly(n-BuMA) loadings. Through the presented synthetic approach, the expansion of amphiphilic semi-IPN polymer networks with controllable compositional and structural characteristics can be easily realized, thus enabling the development of novel materials with predictable mechanical properties.

Experimental

Synthetic procedures

Materials and reagents. Benzene (Merck, 99.7%) was stored over CaH₂ (Merck, >95%) and was distilled under reduced pressure prior to the polymerization reactions. 2-(Dimethylamino)ethyl methacrylate (DMAEMA, Merck, 99%) and n-butyl methacrylate (n-BuMA, Merck, 99%) were passed through a basic alumina column (Aldrich, activated, basic, Brockmann I, ∼150 mesh, pore size 58 Å), stored over CaH₂ and was distilled under reduced pressure immediately prior to the polymerization reactions. 2,2′-Azobis(isobutyronitrile) (AIBN) (Sigma Aldrich, 95%) was recrystallized twice from EtOH and dried under vacuum at ca. 20 °C for 3 days. Cumyl dithiobenzoate (CDTB, Aldrich, 99%), 2-cyano-2-propyldithiobenzoate (CPDTB, Sigma-Aldrich, 97%), 1,2-bis-(2-iodothoxy)ethane (BIEE, Sigma-Aldrich, 96%), n-hexane (Aldrich, 99%), methanol (Analytical grade, ACS reagent) and tetrahydrofuran (THF, Scharlau, HPLC grade) were used as received.

Synthesis of polyDMAEMA homopolymer by RAFT polymerization. The synthesis of the polyDMAEMA homopolymer was carried out by Reversible Addition-Fragmentation chain Transfer (RAFT) controlled radical polymerization as follows: in a round bottom flask (100 ml) equipped with an egg-shaped PTFE coated magnetic stirring bar and maintained under a nitrogen atmosphere, DMAEMA (20.0 g, 127 mmol) was added with a syringe. CPDTB (0.31 g, 1.4 mmol) and AIBN (0.14 g, 0.84 mmol) were dissolved in benzene (21 ml) and were added to the flask with the aid of a syringe. The reaction mixture was stirred rapidly at ca. 20 °C, degassed by three freeze–pump–thaw cycles and heated at 65 °C for 20 h. The polymerization was terminated by cooling the reaction mixture down to ca. 20 °C. The produced polyDMAEMA was retrieved by precipitation in n-hexane (160 ml) and dried at ca. 20 °C under vacuum.

Yield: 95%, number average molecular weight, M_n: 19 000 g mol⁻¹, molecular weight distribution, MWD: 1.17 (ESI, Fig. S1†).

δ_H (300 MHz; CDCl₃; Me₄Si): 0.80–1.10 (3H, m, –CH₃), 1.50–2.00 (2H, m, –CH₂), 2.34 (3H, s, –NCH₃), 2.56 (2H, s, –NCH₂), and 4.05 (2H, s, –OCH₂) (ESI, Fig. S2(a)†).

Synthesis of linear poly(n-BuMA) by RAFT polymerization. The synthesis of the polyBuMA homopolymer was carried out by RAFT as follows: in a round bottom flask (100 ml) equipped with an egg-shaped PTFE coated magnetic stirring bar and maintained under a nitrogen atmosphere, BuMA (22.4 g, 157 mmol) was added with a syringe. CDTB (0.30 g, 1.1 mmol) and AIBN (0.095 g, 0.58 mmol) were dissolved in benzene (9 ml) and were added to the flask with the aid of a syringe. The reaction mixture was stirred rapidly at ca. 20 °C, degassed by three freeze–pump–thaw cycles and heated at 65 °C for 20 h. The polymerization was terminated by cooling the reaction mixture down to ca. 20 °C. The produced polyBuMA was retrieved by precipitation in methanol (100 ml) and dried at ca. 20 °C under vacuum.

Yield: 97%, number average molecular weight, M_n: 21 000 g mol⁻¹, molecular weight distribution, MWD: 1.11 (ESI, Fig. S1†).

Poly(n-BuMA): δ_H (300 MHz; CDCl₃; Me₄Si): 2.0–0.70 (6H, m, –CH₂ and 6H, m, –CH₃), 3.91 (2H, –OCH₂) (ESI, Fig. S2(b)†).

Synthesis of the BIEE-crosslinked polyDMAEMA network. In a glass vial, polyDMAEMA (M_n = 19000, 0.10 g, 0.0052 mmol of macroCTA, 0.64 mmol per DMAEMA unit) was dissolved in THF (1.25 ml). To the solution, BIEE (58 μl, 118 mg, 0.32 mmol) was added using a micropipette. The resulting solution was stirred rapidly and it was then left at ca. 20 °C without stirring until gelation was observed (6 days). The resulting network was then placed in excess MeOH to remove the sol fraction (14% extractables).

Synthesis of semi-interpenetrating poly(n-BuMA)/polyDMAEMA/BIEE amphiphilic networks. A series of semi-interpenetrating poly(n-BuMA)/polyDMAEMA/BIEE amphiphilic networks was prepared via the encapsulation of different amounts of the polyBuMA homopolymer (5, 10, 35, 50% wt in respect to the total polymer mass) within the BIEE-crosslinked polyDMAEMA network. An example of the experimental protocol followed is provided below: in a glass vial, polyBuMA (5 mg) were dissolved in THF (1.3 ml, 8% w/v solution concentration). To the solution polyDMAEMA (100 mg, 0.0052 mmol of macroCTA, 0.64 mmol per DMAEMA unit) was added and the mixture was stirred rapidly until polyDMAEMA was fully dissolved. To the solution, BIEE (58 μl, 118 mg, 0.32 mmol) was added using a micropipette. The resulting solution was stirred rapidly and it was then left at ca. 20 °C without stirring until gelation was observed (6 days). The resulting network was then placed in excess methanol to remove the sol fraction (12% extractables).

Determination of sol-fraction. After gelation was reached, the obtained networks were placed in excess methanol (100 ml) and were left there to equilibrate for 7 days to remove the sol fraction (extractables). Subsequently, the solvent was recovered by filtration and evaporated off. The recovered extractables were dried under vacuum at ca. 20 °C for 24 h. The sol fraction was determined gravimetrically and it was calculated from the ratio of the dried mass of the extractables to the theoretical mass of all components in the network (i.e., polymers plus crosslinker). The sol fraction ranged from 6–14%.

Determination of the degrees of swelling. The washed networks were cut into small pieces and their MeOH swollen mass was determined gravimetrically before drying in a vacuum oven at ca. 20 °C for 24 h. The dry network mass was then determined, and the degrees of swelling (DSs) in MeOH were calculated as the ratio of the swollen mass divided by the dry mass. Afterwards, the dried networks were placed in deionized water for 2 weeks and the swollen network masses were measured.

Molecular characterization

¹H NMR spectra were recorded in CDCl₃ with tetramethylsilane (TMS) used as an internal standard using an Avance Brucker 300 MHz spectrometer equipped with an Ultrashield magnet. The MWs and MWDs of the polymers were determined by SEC using equipment supplied by Polymer Standards Service (PSS). All measurements were carried out at room temperature using Styragel HR 3 and Styragel HR 4 columns. The mobile phase was THF, delivered at a flow rate of 1 ml min⁻¹ using a Waters 515 isocratic pump. The refractive index was measured with a Waters 2414 refractive index detector supplied by PSS. The instrumentation was calibrated using poly(methyl methacrylate) (PMMA) standards with narrow MWD supplied by PSS.

Mechanical testing measurements

Unconfined compression experiments were performed using a high precision mechanical testing system (Instron 5944, Norwood, MA, USA). The specimens were cut in an orthogonal shape with dimensions 6 × 6 × 3 mm (length × width × thickness). According to the unconfined compression experiment the specimen was placed between two parallel platens and stress–strain experiment was performed to test the elastic response of the materials. The specimens were compressed to 30% strain with a strain rate of 0.05 mm min⁻¹, the lowest strain rate that the system can apply, in order to avoid any transient, poroelastic effects. The stress was calculated as the force measured on the load cell divided by the initial area of the specimen (i.e., 1^st Piola-Kirchhoff stress) and the strain was calculated as the displacement Δl divided by the initial length of the specimen. The Young's modulus was calculated from the slope of the linear part of the stress–strain curves. Six specimens from each material and condition were tested (n = 6).

Mathematical analysis

Mathematical modeling was employed to further analyze the mechanical properties of the semi-IPN amphiphilic polymer networks with a finite elements model. The Mooney–Rivlin, hyperelastic constitutive equation was employed with a strain energy density function of the form:where the parameters C₁, C₂ are material constants and I₁, I₂ are the first and the second invariant of the left Cauchy–Green deformation tensor and the 1^st Piola-Kirchhoff stress tensor was calculated as . A three-dimensional finite elements model of orthogonal geometry, same in size as that of the real specimens, was constructed. The model was compressed in one direction and was free to deform in the other two according to the unconfined compression experimental protocol. The material properties of the constitutive equation (i.e., C₁ and C₂) were varied so that the sum of the squared errors, reached a minimum. P^exp, P^model are the experimentally measured and predicted by the model 1^st Piola-Kirchhoff stresses, respectively and n the number of experimental data.

Statistical analysis

The data are presented as means with standard errors. Experimental groups were compared using an unpaired Student's t-test. Statistical significant difference was determined when p < 0.05.

Acknowledgements

This work is part of the Project TEXNOΛOΓIA/YΛIKA/0311(BIE)/03 that is co-funded by the European Regional Development Fund and the Republic of Cyprus through the Cyprus Research Promotion Foundation.

References

1. C. S. Patrickios and T. K. Georgiou, Curr. Opin. Colloid Interface Sci., 2003, 8, 76.
2. G. Erdodi and J. P. Kennedy, Prog. Polym. Sci., 2006, 31, 1.
3. B. Ivan, J. Feldthusen and A. H. E. Muller, Macromol. Symp., 1996, 102, 81.
4. E. S. Dragan, Chem. Eng. J., 2014, 243, 572.
5. L. K. Kostanski, R. Huang and C. D. M. Filipe, et al., J. Biomater. Sci., Polym. Ed., 2009, 20, 271.
6. T. M. Aminabhavi, M. N. Nadagouda and U. A. More, et al., Expert Opin. Drug Delivery, 2015, 12, 669.
7. L. Mespouille, J. L. Hedrick and P. Dubois, Soft Matter, 2009, 5, 4878.
8. C. Zhou, S. Qian and A. Zhang, et al., RSC Adv., 2014, 4, 8144.
9. T. C. Krasia and C. S. Patrickios, Macromolecules, 2006, 39, 2467.
10. M. Achilleos, T. Krasia-Christoforou and C. S. Patrickios, Macromolecules, 2007, 40, 5575.
11. M. D. Rikkou, M. Kolokasi, K. Matyjaszewski and C. S. Patrickios, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 1878.
12. M. Vamvakaki, C. S. Patrickios and P. Lindner, et al., Langmuir, 2007, 23, 10433.
13. M. D. Rikkou and C. S. Patrickios, Macromolecules, 2008, 41, 5957.
14. L. Mespouille, O. Coulembier and D. Paneva, et al., J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 4997.
15. T. K. Georgiou, C. S. Patrickios and P. W. Groh, et al., Macromolecules, 2007, 40, 2335.
16. W. Wu, J. Liu, S. Cao, H. Tan, J. Li, F. Xu and X. Zhang, Int. J. Pharm., 2011, 416, 104.
17. F. Yao, L. Xu, G.-D. Fu and B. Lin, Macromolecules, 2010, 43, 9761.
18. M. Achilleos, M. Demetriou, O. Marinica, L. Vekas and T. Krasia-Christoforou, Polym. Chem., 2014, 5, 4365.
19. F. Mpekris, M. Achilleos, E. Vasile, E. Vasile, T. Krasia-Christoforou and T. Stylianopoulos, RSC Adv., 2015, 5, 20011.
20. Z. G. Zhang, G. J. Lin and S. Bell, Macromolecules, 2000, 33, 7877.
21. S. Y. Liu, J. V. M. Weaver, Y. Q. Tang, N. C. Billingham, S. P. Armes and K. Tribe, Macromolecules, 2002, 35, 6121.
22. V. Butun, N. C. Billingham and S. P. Armes, J. Am. Chem. Soc., 1998, 120, 12135.
23. M. Vamvakaki and C. S. Patrickios, Soft Matter, 2008, 4, 268.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra07376j

---