Synthesis and characterization of block copolymers comprising a polyrotaxane middle block flanked by two brush-like PCL blocks

Abstract

A polyrotaxane was prepared via atom transfer radical polymerization (ATRP) of 2-hydroxyethyl methacrylate (HEMA) initiated by a polypseudorotaxane self-assembled from a distal 2-bromoisobutyryl end-capped poly(ethylene glycol) (BriB–PEG–iBBr) with α-cyclodextrins (α-CDs). Then it was used to initiate bulk ring-opening polymerization (ROP) of ε-caprolactone (CL) to give block copolymers comprising a polyrotaxane middle block flanked by two brush-like polycaprolactone (PCL) blocks. The structure of both the polyrotaxane and its CL derivatives was characterized by ¹H NMR, GPC, FT-IR, XRD, DSC and TG analyses. These polyrotaxane-derived block brush-like copolymers can be cast into self-standing films from CH₂Cl₂ and self-assemble into nano-sized particles in THF as well. Furthermore they can melt at round 60 °C, showing the potential for melt-processing.

---

Introduction

In recent years, polyrotaxanes (PRs) and polypseudorotaxanes (PPRs), obtained from the self-assembly of PEG with α-CDs, have been broadly investigated, due not only to their basic scientific importance but also their potential applications as smart materials, such as hydrogels,¹ slide-ring gels,^2,3 biosensors,⁴ carriers for drug delivery⁵ and fiber spinning,⁶etc. For avoiding the dissociation of α-CDs from the PPR axle chains, bulky groups are necessarily introduced as end stoppers of PRs. To this end a variety of bulky groups and end-capping reactions have been exploited. For example, 2,4-dinitrofluorobenzene was applied to end-cap amine-terminated PEG.⁷ Some intriguing synthetic strategies have been also developed, such as the reaction of PPR comprising permethylated α-CDs and poly(tetrahydrofuran) with an electrophile in the solid state,⁸ the Huisgen cyclization of azido- and alkynyl-modified precursor in the presence of Cu(I), etc. However, little attention has been paid to functionalizing the end-capping groups of PRs as a novel means to design and prepare functional smart materials and nanometre-scale devices for biomedical applications. In our previous work, poly(N-isopropylacrylamide) (NIPAAm) was attached as a functional end-capping agent to PPR to provide a kind of main-chain PRvia radical telomerization of NIPAAm under UV radiation in aqueous medium at room temperature.^9,10 Although they showed an LCST at round 38 °C, the degree of polymerization (DP) of the two outer PNIPAAm blocks was ill-controlled.

ATRP is such a versatile method that it can synthesize not only homopolymers with predictable molecular weight and narrow molecular weight distribution, but can also prepare copolymers with well-defined architectures, such as block copolymers,^11,12graft copolymers, star-like copolymers,¹³ and so on. Hence, it is the choice of interest in preparing PRs holding lengthily tunable and functional end stoppers. As far as we know, there are only three reports to date involving the preparation of main-chain PRs viaATRP including two studies from our group.^14–16

To explore attachment of lengthily adjustable and functional polymer blocks not only as end stoppers of PRs, but also as a synthetic strategy to functionalize these PRs, in the present study ATRP of HEMA was carried out in aqueous solution at room temperature by using PPRs self-assembled from a distal 2-bromoisobutyryl end-capped PEG with α-CDs as macro-initiator and Cu(I)Br–N,N,N′,N′,N′-pentamethyl ethylenetriamine (PMDETA) as the catalyst. Pendent hydroxyl groups in the PHEMA blocks were then used to initiate bulk ROP of ε-CL to graft PCL onto the PHEMA backbones to give novel copolymers comprising a polyrotaxane middle block flanked by two PCL molecular brushes.

To the best of our knowledge, this is the first report focused on the modification of PRs through lengthily tunable polymeric end stoppers instead of α-CDs threaded onto the polymer axle chain. The grafted PCL side chains would impart these PR-derived block brush-like copolymers not only biocompartibility and biodegradability, but also film castable and melt-processable abilities, which remain major challenges for typical PR processing due to their extraordinary crystalline capability impeding their ability to dissolving in ordinary solvents, such as THF and CH₂Cl₂, and melt-processing at relatively low temperature as well. This work may open a new avenue in preparing CD-based functional smart materials.

Experimental

Materials

α-CD was purchased from TCI, Japan and used as received. PEG (M_n = 4000, PEG 4 K) was imported from Japan and distributed domestically. Both HEMA and ε-CL were available from Acros, Belgium, vacuum distilled and stored over CaH₂ before using. 2-Bromoisobutyryl bromide, 4-dimethylaminopyridine (DMAP) and stannous octoate (Sn(Oct)₂) were provided by Alfa-Aesar, USA, and PMDETA was available from Aldrich, USA and they were used without further purification. All other reagents used were of analytical grade.

Synthesis of macroinitiator (BriB–PEG–iBBr)

For a typical preparation, 20.0 g (5 mmol) PEG 4 K was dissolved in 100 mL dry CH₂Cl₂, and then 1.01 g (10 mmol) triethylamine and 0.61 g (5 mmol) DMAP were added. This reaction mixture was cooled to 0 °C and then 2.87 g (12.5 mmol) 2-bromoisobutyryl bromide dissolved in 20 mL dry CH₂Cl₂ was added dropwise under continuous stirring over a 1 h period under nitrogen atmosphere. The reaction was continued for another 24 h at room temperature. After the reaction stopped, the resultant triethylamine hydrobromide was filtered. The solution was concentrated and precipitated in n-hexane, filtered and dried under vacuum. Finally, the crude product was dissolved in water and extracted using chloroform. After drying over magnesium chloride, the solution was again precipitated in n-hexane, filtered and dried under vacuum to give the purified product, yield 80%. FTIR/cm⁻¹: 2866 (CH₂, CH₃), 1736 (C=O), 1109 (C–O). ¹H NMR (400 MHz; DMSO-d₆)/ppm: δ 4.242 (4H, –CH₂–CH^*₂–O–C(=O)–), 3.636 (4H, –CH^*₂–CH₂–O–C(=O)–), 3.587–3.446 (360H, –OCH₂–CH₂–O–), 1.893 (12H, –CH₃).

Synthesis of PHEMA–PEG–PHEMA triblock copolymer

For comparison, a PHEMA–PEG–PHEMA triblock copolymer was prepared. 0.10 g (0.023 mmol) BriB–PEG–iBBr, 0.13 g (0.92 mmol) HEMA and 0.005 g (0.028 mmol) PMDETA were dissolved in triply distilled water and then quenched into liquid nitrogen. After 4 mg (0.028 mmol) CuBr were added, the reactor was deoxygenated using three vacuum–nitrogen-filling cycles and sealed. The reaction was then carried out at 25 °C for 6 h under stirring. Upon exposure to air, the polymerization stopped. The product was obtained by dialysis against distilled water using a cellulose membrane (MWCO 7000) for 2 days and then freeze-dried.

Synthesis of polyrotaxane (PR)

In a typical experiment, 0.46 g (0.046 mmol) α-CD was dissolved in 1.5 mL water to give a saturated solution and then added to a solution of 0.10 g (0.023 mmol) BriB–PEG–iBBr in 0.5 mL water, followed by vigorously stirring at room temperature for 2 days. Then 0.13 g (0.92 mmol) HEMA and 0.005 g (0.028 mmol) PMDETA were added to the resultant PPR suspension. After quenching with liquid nitrogen, 4.0 mg (0.028 mmol) CuBr were added. The reactor was deoxygenated using three vacuum–nitrogen-filling cycles and sealed. The polymerization was carried out at 25 °C for 6 h. Upon exposure to air, the reaction stopped and the crude product was centrifuged, washed several times using distilled water to remove the Cu²⁺ salts and finally freeze-dried. For a thorough purification, the product was again incubated in an excess of DMSO at 60 °C for 12 h, followed by precipitating in acetone and vacuum drying.

For the sake of expression, the obtained PRs were designated as PEG-mCD-n, where m stands for the feeding molar ratio of α-CDs to PEG and n for that of HEMA to PEG.

Synthesis of PR-derived block brush-like copolymer (PR-g-PCL)

In a typical synthesis, 0.10 g PR sample (only PEG-20CD-40 was used here, containing 0.14 mmol HEMA) and 0.96 g (8.4 mmol) ε-CL were added into the reactor and stirred. After 5.0 mg Sn(Oct)₂ were added, the reactor was sealed after three vacuum–nitrogen-filling cycles. The reaction was carried out at 130 °C for 12 h. After cooling to room temperature, the resulting copolymer was dissolved in chloroform, precipitated in methanol and vacuum dried. As a control sample, PCL-g-PHEMA–PEG–PHEMA-g-PCL was prepared from PEG-0CD-40 without adding α-CDs according to the same procedure.

For the convenience of expression, the as-prepared PCL-grafted block copolymers were designated as PR-xCL, where PR is referred to PEG-20CD-40, x stands for the feeding molar ratio of ε-CL to each HEMA unit. Furthermore, PCL-g-PHEMA–PEG–PHEMA-g-PCL was redesignated as PH-yCL, where PH means PEG-0CD-40 and y stands for the feeding molar ratio of ε-CL to each HEMA unit.

Measurements

¹H NMR and 2D HMBC NMR spectra were recorded on a Bruker ARX 400 NMR instrument at room temperature using DMSO-d₆ as solvent and tetramethylsilane (TMS) as internal standard. The molecular weight and polydispersity index of the copolymers were determined using gel permeation chromatography (GPC) conducted with a Waters 1515 pump, three Waters Styragel columns (HT3, HT4, HT5) and a Waters 2414 differential refractive index detector. DMF or DMF containing 0.06 mol L⁻¹LiBr was used as eluent at a flow rate of 1.0 mL min⁻¹. Polystyrene standards were utilized for calibration. FTIR spectra were measured using a Shimadzu IR Prestige-21 FTIR spectrometer at room temperature in the range between 4000 and 500 cm⁻¹ with a resolution of 2 cm⁻¹ and 20 scans. Powder samples were prepared by dispersing the samples in KBr and compressing them to form disks. Wide-angle X-ray diffraction (WXRD) measurements were performed with powder samples using a Panaltic X'pert PRO X-ray diffractometer. The radiation source used was Ni-filtered, CuKα radiation with a wavelength of 0.154 nm. The voltage was set to 40 kV and the current to 40 mA. Samples were placed on a sample holder and scanned from 10 to 50° in 2θ at a speed of 10° min⁻¹. Differential scanning calorimetry (DSC) analysis was carried out using a NETZSCH DSC 204 differential scanning calorimeter. The DSC thermograms covered a temperature range from −30 to 150 °C at a scanning rate of 10 °C min⁻¹. Thermogravimetric (TG) analysis was made using a NETZSCH TG 209 F1 thermogravimetric analyzer at a heating rate of 10 °C min⁻¹ with nitrogen used as purge gas. The transmission electron microscopy (TEM) micrographs were taken with a Hitachi-700 transmission electron microscope operated at 100 kV accelerating voltage. The crystalline morphologies of the cast copolymers were observed using XPN-203Z polarized optical microscope. The samples were prepared by dropping the sample solutions in chloroform onto a glass plates at room temperature for 24 h for complete evaporation of the solvent.

Results and discussion

Preparation and characterization of PR-derived block brush-like copolymers

As a distal macroinitiator BriB–PEG–iBBr was prepared from the reaction of HO–PEG–OH with 2-bromoisobutyryl bromide according to a previous report.¹⁷ As shown in Scheme 1, the self-assembly of this macroinitiator with α-CDs gave a PPR suspension after an aqueous solution of α-CDs was added to BriB–PEG–iBBr in water. The suspension became increasingly viscous upon increasing the amount of α-CDs. An opaque physical gel is generally formed if left standing for a long time due to the non-covalent interactions among PPRs.^18,19 Taking account its pendent hydroxyl functionality and solubility in water, HEMA was used to prepare PRs in situviaATRP. At the same time, the reaction was controlled to proceed at room temperature due to the fact that the self-assembly of the macroinitiator with α-CDs in aqueous solution is in favor of keeping as many as possible α-CDs entrapped on the polymeric backbone and that a higher temperature would cause severe dethreading of α-CDs from the polymeric axle.^20,21 Furthermore it appeared that here Cu(I)Br–PMDETA worked well, although Cu(I)Br–2,2′-bipyridine was reported to successfully catalyze ATRP of HEMA in methanol or methanol–water.¹⁷ The theoretical feed/found compositions and corresponding yields of PEG-0CD-40 and PEG-20CD-40 are summarized in Table 1.

Scheme 1 Preparation strategy of polyrotaxanes viaATRP of HEMA and brush-like PR copolymersviaROP of ε-CL.

Table 1 Molecular weight, polydispersity index and T_m of the as-synthesized PRs and PR-g-PCL products

Sample	Feed molar composition			M n,NMR ^a	M w,GPC ^b	M w/Mn^b	T m/°C^c	Yield (%)^d
	[CD]/[PEG]	[HEMA]/[PEG]	[CL]/[HEMA]
a M n,NMR was determined by ^1H NMR. b M n,GPC and Mw/Mn were determined by GPC analysis with polystyrene standards. DMF containing 0.06 mol L^−1LiBr was used as eluent. c Measured by DSC. d Yield based on final product to total feed masses. e The Mn of PR-g-PCL copolymers is difficulty to determined by ^1H NMR because of the very strong CL peaks with very weak illegible EG peak.
PEG-0CD-40	0	40	0	9 × 10^3	3.93 × 10^4	1.31	40.8	86
PH-150CL	0	40	150	—^d	2.38 × 10^5	1.32	61.8	93
PEG-20CD-40	20	40	0	2.6 × 10^4	8.42 × 10^4	1.83	—	38
PR-5CL	20	40	5	—^e	1.38 × 10^5	1.47	—	35
PR-40CL	20	40	40	—^e	1.67 × 10^5	1.38	58.2	92
PR-150CL	20	40	150	—^e	2.42 × 10^5	1.26	64.2	87
PR-300CL	20	40	300	—^e	4.53 × 10^5	1.13	64.6	85

---

The PR product was characterized by ¹H NMR analysis. The broadening of resonance peaks and slightly different chemical shifts of α-CDs are discerned in the spectrum because of the chemical environmental change after α-CDs are threaded onto the polymer chains as previously accounted.²¹ As can be seen in Fig. 1, all the resonance peaks of α-CDs in PEG-20CD-40 are evidently broadened and the corresponding resonance peaks of the hydroxyl groups (O(2)H, O(3)H, O(6)H) are also different from those of the pure α-CDs. The peak at δ 1.891 ppm corresponds to the –CH₃ contiguity to the Br end group. However, the peaks of –OH groups of HEMA are superposed with H(1) of α-CDs at δ 4.795 ppm and –CH₂–O– superposed with H(5) of α-CDs at δ 3.586 ppm.

Fig. 1 ¹H NMR spectra of pure α-CD (A), PEG-20CD-40 (B) and PR-5CL (C) in DMSO-d₆.

The other resonance peaks in PEG-20CD-40 were assigned as follows: O(2)H and O(3)H of α-CD appear at δ 5.442–5.619, O(6)H at 4.440–4.496, –CH₂– of PHEMA at 1.787 and CH₃– at 0.768–0.940 ppm, respectively. Multiple resonances at δ 3.586–3.768 ppm are ascribed to the chemical shifts of H(3), H(6) and H(5), while ethylene glycol (EG) unit protons of PEG appear at δ 3.506 ppm. This clearly indicates that α-CDs are threaded onto the PEG backbone and end stopped by the PHEMA blocks as well.

The average in-chain number of α-CDs was readily calculated from the integration of the proton resonance peak of α-CD (O(6)H) to that of the methylene protons in the central PEG axle, while the DP of PHEMA was easily determined from the integration of the proton resonance peaks of the methyl groups of HEMA to that of the methylene protons in PEG. The multiple resonances at δ 3.586–3.768 comprise the resonance peaks H(3), H(6), H(5) of α-CDs, –C(=O)–O–CH₂– of HEMA and the methylene protons of PEG. However, it is difficult to separate PEG from these broadened peaks to calculate the integration. Alternatively, we subtracted 1× integration of –C(=O)–O–CH₂– group (identified as d in Fig. 1B) of HEMA (which is equal to the integration of –C(=O)–O–CH₂– of HEMA) and 3× integration of O(6)H peak in α-CDs (which is equal to the sum of integration of H(3), H(6), H(5) of α-CDs) from the total integration of the multiple resonances at δ 3.586–3.768 to get the integration of the methylene protons of PEG. The DP of the incorporated PHEMA was found to be well in accordance with the feed molar ratio. This implied that a high initiation efficiency is achieved, and an addition of α-CDs seems not to weaken the activity of PPRs to initiate ATRP of HEMA.

As shown in Scheme 1, all the PR-derived block brush-like copolymers were prepared via bulk ROP of ε-CL initiated by pendent hydroxyl groups of PHEMA blocks in the presence of Sn(Oct)₂. The procedure was similar to that of brush-like PHEMA-g-PCL²² and PHEMA-g-(PCL-b-PEG) copolymers.²³ The starting reaction mixture was a little cloudy, and after 2 h it got sticky and more transparent. The mixture became hard to stir after about 6 h due to its high viscosity. Upon cooling to room temperature, a solid mass of product was obtained. Though DMSO is the only good solvent to dissolve PRs, these copolymers dissolved in THF, DMF, CH₂Cl₂ and CHCl₃ to form transparent solutions.

As reported,^3,24 α-CDs do not initiate ROP of δ-valerolactone (δ-VL) and ε-CL to produce any polymers while a small lactone of β-butyrolactone (β-BL) can be initiated by α-CDs to give polyester in higher yield. A proposed mechanism was that the lactone monomer first forms a 1 : 1 inclusion complex with α-CDs and then undergoes ROP. In fact ε-CL cannot form a 1 : 1 inclusion complex with α-CDs due to the bulky volume so that its bulk ROP does not proceed with α-CDs. To further confirm whether or not hydroxyl groups of α-CDs participate in bulk ROP of ε-CL here, a copolymer with a lower DP of PCL was prepared by shortening the reaction time of the PR-40CL sample to 2 h rather than 12 h, followed by ¹H NMR analysis. As shown in Fig. 1C, it clearly showed that PCL side chain was grafted to the PHEMA blocks rather than to α-CDs. As calculated using the integration ratio of the methylene group in CL unit to ethylene glycol unit in PEG, the DP of PCL per HEMA unit was found to be 5, and consequently this copolymer was referred to as PR-5CL in the following discussion. Moreover, the corresponding hydroxyl groups O(2)H, O(3)H and O(6)H of α-CDs appeared to remain unchangeable in bulk ROP as determined from the integration ratio of hydroxyl groups to ethylene glycol unit in PEG. It meant that hydroxyl groups of α-CDs indeed do not initiate ROP of CL in the reaction, in accordance with previous reports.^3,24

Further characterization was carried out using 2D HMBC NMR analysis. As shown in Fig. 2, the spectrum of PR-5CL showed that the C signals of C(O)CH₂CH₂O of HEMA unit are correlated with the resonance of protons in CL unit, marked as A, B and C in the figure. However, none of the signals of C6, C3 and C2 in α-CDs are correlated with these CL protons. This result revealed that the hydroxy groups of α-CDs do not participate in bulk ROP of ε-CL. Besides, the resonance peak at δ 4.795 ppm, may have arisen from the superposition of hydroxy groups of HEMA and H(1) of α-CDs, and is correlated with the carbon signals of α-CDs (marked as D in the figure), but not with any of the carbon signals of HEMA. It clearly indicated that the hydroxy groups of HEMA are almost completely converted.

Fig. 2 2D HMBC NMR spectrum of PR-5CL in DMSO-d₆.

When the GPC measurements were conducted in pure DMF as eluent, all the traces showed a smaller peak corresponding to a higher molecular weight besides the main peak as shown by trace (1) in Fig. 3A, this suggested that these PR-derived block brush-like copolymers enable self-aggregation even in DMF, likely due to the high crystalline character of the middle PR blocks. Given that Li⁺ ion effectively depress the hydrogen bonding strength of the self-aggregates, which is one of the most significant non-covalent interactions leading to the high PR crystalline capability and insolubility, mixing an eluent of DMF with 0.06 mol L⁻¹LiBr was used to perform the GPC measurements. As shown by trace (2) in Fig. 3A, those extra peaks were significantly reduced with this mixing eluent. The nearly symmetrical and unimodal peak of the molecular weight distributions revealed that the copolymers formed via bulk ROP of ε-CL are the PR-derived block brush-like copolymers rather than a mixture of PR and PCL. Furthermore, their molecular weights were increased with increasing the feeding molar ratio of ε-CL, as illustrated in Fig. 3B, and their molecular weight distribution also became obviously lower than their precursor, as shown in Table 1. It definitely meant that bulk ROP of ε-CL successfully proceeds and the uneven in-chained CD number in the middle PR block may no longer dominate the molecular weight distribution.

Fig. 3 (A) GPC traces of PR-300CL in DMF (1) and DMF + LiBr (2); (B) GPC traces of PR-150CL (1), PH-150CL (2), PR-40CL (3) and PR-5CL (4) in DMF + LiBr.

The FTIR spectra of the PR and PR-g-PCL copolymers are shown in Fig. 4. The peak for the carbonyl associated with the PHEMA ester appears at 1728 cm⁻¹, and the C–O–C stretching vibration of carbonyl group in ester linkages emerges at 1109 cm⁻¹. The broad bands at 3100–3700 cm⁻¹ correspond to the hydroxyl stretching vibration. C–H stretching vibration of PEG appears at 2883 cm⁻¹ while that of α-CD appears at 2926 cm⁻¹. From the spectrum of PEG-20CD-40 (B), it is easily discerned that the C–H vibration tends to be merged and shifts to higher frequency at 2940 cm⁻¹, caused by the non-covalent interaction between C–H of α-CD and the PEG backbone. The strong absorption at 1728 cm⁻¹ in PR-40CL (C) and PR-150CL (D) corresponds to the stretching mode of C=O from PCL segments, and the peak at 2943 cm⁻¹ belongs to the absorption of C–H stretch of CH₂ from PCL blocks, while the peak at 2865 cm⁻¹ is ascribed to the C–H stretching band of PEG. Otherwise, the absorption peak in the range from 3100–3700 cm⁻¹ is not clear because there are very few hydroxyl groups in the PR-derived block brush-like copolymers.

Fig. 4 FTIR spectra of BriB–PEG–iBBr (A), PEG-20CD-40 (B), PR-40CL (C), PR-150CL (D) and pure α-CD (E).

WAXD is a useful method to investigate the structure of the PR and PR-derived block brush-like copolymers. As shown in Fig. 5, the peak appears at 2θ = 19.69° (d = 4.32 Å) is the characteristic diffraction peak of the channel-type crystalline structure of PRs.²⁵ It is quite different from either the pure α-CD with major peaks at 12.04°, 14.36°, 18.24° and 21.76° or from PEG with major peaks at 19.15° and 23.35°. PCL itself exhibits intensive peaks at 21.6° and 23.9°, corresponding to the (110) and (200) planes of the orthorhombic crystal form.²² These peaks are easily discerned in the patterns of PR-40CL (B) and PR-150CL (A). The intensity of the characteristic peaks of PCL was increased with the feeding molar ratio of the CL monomer. Importantly, both the characteristic peaks of PR and PCL were clearly evidenced in PR-5CL(C), and clearly confirmed that the PCL side chains are grafted to the PHEMA backbones rather than to α-CDs.

Fig. 5 X-Ray diffraction patterns of PR-150CL (A), PR-40CL (B), PR-5CL (C), PEG-20CD-40 (D), PEG (E) and pure α-CD (F).

Further characterizations were carried out using DSC and TG analyses. As depicted in Fig. 6, the DSC curves of PR-g-PCL (F–H) and PH-150CL (E) display only a clear endothermic peak at around 60 °C. For the F–H samples, their melting points (T_m) are slightly increased with the feeding molar ratio of CL per HEMA unit. A higher T_m of PR-150CL (G) at 64.2 °C than PH-150CL (E) at 61.8 °C is maybe due to the non-covalent interactions between the α-CD containing PR middle block and pendent PCL side chains. Just like the pure α-CD (D), there is almost no endothermic peak in PEG-20CD-40 (C), because PEG is completely included into the channel of the host α-CD lattices, which restrict it from aggregating to form the crystalline phase again. A lower temperature of broad endothermic peak in PEG-0CD-40 (B) as compared with PEG (A) is certainly caused by the interference of the PHEMA blocks attached to the two terminals of PEG main chain.

Fig. 6 DSC curves of PEG 4K (A), PEG-0CD-40 (B), PEG-20CD-40 (C), α-CD (D), PH-150CL (E), PR-40CL (F), PR-150CL (G) and PR-300CL (H).

TG/DTG curves are illustrated in Fig. 7. As reported,^9,10PEG-20CD-40 (D) should undergo a two-step thermal decomposition process, the low temperature ones corresponding to the decomposition of α-CDs threaded and the high temperature ones to that of the axle chain comprising PEG as a central block and PHEMA as outer stoppers. As can be seen, it indeed exhibits two DTG peaks arising from the decomposition of α-CDs threaded at 342 °C and PEG–PHEMA at 404 °C, while the pure α-CDs (A) presents only one DTG peak at 302 °C, substantially lower than that of α-CDs threaded onto the polymeric chain. However, PH-150CL (B) displays one DTG peak at 319 °C. As a result, PR-40CL (C) experiences a three-step thermal decomposition process corresponding to the DTG peak of pendent brush-like PCL chain at 318 °C, that of α-CDs threaded at 351 °C and that of PEG-PHEMA at 399 °C, respectively. It implied that the introduction of PCL would lower the thermal stability of the polyrotaxanes.

Fig. 7 TG/DTG curves of α-CD (A), PH-150CL (B), PR-40CL (C), PEG-20CD-40 (D).

Preparation of the films

The flexible tough films of the PR-g-PCL copolymers were prepared by dissolving them in CH₂Cl₂, casting onto a Teflon panel (50 mm in diameter) and drying overnight at room temperature. As shown in Fig. 8, when the feeding molar ratio of CL per HEMA unit is lower than 200, the copolymer cannot cast into a self-standing film. This may be due to the fact that a higher content of the rigid and crystalline middle PR block makes it too brittle. If the length of the incorporated PCL side chains is long enough, such as its DP equal to 300 here, the brush-like blocks of the PR-derived copolymer begin to dominate the physical properties to give a semi-transparent and flexible tough film.

Fig. 8 Appearances of the casting films of PR-150CL (A) and PR-300CL (B).

The microstructure of the films were then analyzed by using a polarizing microscope, as shown in Fig. 9. From the micrographs, the PR-g-PCL copolymer films all display the separated dark and bright domains, representing a heterogeneous condensed phase structure. In comparison, the PH-150CL film exhibits a regular crystalline morphology. In agreement with the DSC results, these casted films were found to melt at around 60 °C, where amorphous transparent morphologies were obtained. However, after cooling to room temperature, their morphologies became very different from the samples directly casted from CH₂Cl_2, as evidenced in Fig. 9. All of them present regular crystalline morphologies after melting.

Fig. 9 Polarized optical micrographs of the casting films of PH-150CL (A), PR-150CL (B) and PR-300CL (C) from CH₂Cl₂ and the samples (A′), (B′) and (C′) obtained by melting and then cooling to room temperature.

In general, low temperature melt-processing is superior to solution-processing for biomedical applications. Regarding investigations as discussed above, one of these PR-derived block brush-like copolymers, PR-300CL, was tested using melt-processing in this study. It was smoothly extruded on melting index apparatus at about 60 °C into a spring-like specimen with some flexibility as shown in Fig. 10.

Fig. 10 Spring-like sample of PR-300CL produced using melt-processing.

Self-assembly of copolymer PR-40CL in THF

The PCL brush-like blocks of the copolymers have excellent solubility in many organic solvents, such as THF, CH₂Cl₂ and CHCl₃, while the middle PR block is almost insoluble in them due to its crystalline structure. This intriguing feature imparts these PR-derived block brush-like copolymers their amphiphilic properties. This enables them to self-assemble into the unique core–shell structure in selective organic solvents. The self-assembly of PR-40CL was investigated. It was dissolved in THF at a concentration of 1.0 mg mL⁻¹ in which the self-assembled nanoparticles were evidenced by TEM observation. Fig. 11 shows a typical TEM image of the self-assembly morphologies. A clear core–shell structure is presented with the particle size in a range of 100∼150 nm. Some very smaller ones observed may correspond to the unimolecular particles.

Fig. 11 TEM micrograph of nanoparticles formed from 0.1 mg mL⁻¹PR-40CL solution in THF.

As described above, the middle PR blocks belonging to the block brush-like copolymers prefer to self-assemble into rigid crystalline domains via intermolecular hydrogen bonding, while the brush-like PCL blocks are apt to create a loose lyophilic coating around the resulting rigid crystalline domains yielding the unique core–shell structure in THF.^20,27 It can be inferred that the rigid PR blocks pose as the dark core of the nano-sized particle, while the flexible and soluble PCL brush-like blocks present the bright shell. A proposed sketch of the self-assembly of the PR-derived block brush-like copolymers in THF is illustrated in Scheme 2.

Scheme 2 Proposed sketch of the self-assembly of the copolymer in THF.

It is interesting that there are only few reports describing the self-assembly of main-chain PRs to date,^20,27,28 especially those conducted in organic solvents. Since these CD-based polyrotaxanes possess a lot of hydroxyl functional groups, they are suitable as scaffolds to develop polymeric prodrugs, due to the fact that a variety of drugs can be covalently attached to the CD rings.²⁶ Moreover, the unique core–shell structure nanoparticles also show potential in sustained drug release, as the drug can release along with the swelling and degradation of the shell formed by the brush-like PCL blocks. Moreover, drugs which are not suitable in drug release systems formed in aqueous solution are available due to the organic solvent formation environment for nanoparticles here.

Conclusion

A novel main-chain polyrotaxane was prepared in situviaATRP of HEMA initiated by polypseudorotaxane made from the self-assembly of a distal 2-bromoisobutyryl end-capped PEG with α-CDs and catalyzed by Cu(I)Br–PMDETA in aqueous solution at room temperature. It was then grafted with PCL side chains viaROP of ε-CL to give the PR-derived block brush-like copolymers. When the length of grafted PCL chains is long enough, they are enabled to be cast into the semi-transparent and self-standing films and self-assemble into the unique core–shell structure nanoparticles in THF. Furthermore they can melt at around 60 °C, showing the prospect for melt-processing. It is expected that this study could offer a new approach to prepare CD-based functional smart biomaterials, such as polymeric prodrugs, scaffolds for tissue engineering, carriers for controlled drug release, sliding gels, etc.

Acknowledgements

We acknowledge support from the Natural Science Foundation of China (no. 20374008, 20674006 and 20711140361).

References

1. J. Li, X. Li, Z. Zhou, X. Ni and K. W. Leong, Macromolecules, 2001, 34, 7236.
2. T. Karino, M. Shibayama and K. Ito, Physica B, 2006, 385–386, 692.
3. K. Ito, J. Araki and T. Kataoka, Soft Matter, 2008, 4, 245.
4. T. Ooya, M. Eguchi and N. Yui, J. Am. Chem. Soc., 2003, 125, 13016.
5. J. Li and X. J. Loh, Adv. Drug Delivery Rev., 2008, 60, 1000.
6. J. Arakia, T. Kataoka, N.K., A. Teramoto, K. Ito and K. Abe, Polymer, 2006, 47, 8241.
7. A. Harada, J. Li, T. Nakamitsu and M. Kamachi, J. Org. Chem., 1993, 58, 7524.
8. N. Kihara, K. Hinoue and T. Takata, Macromolecules, 2005, 38, 223.
9. H. Yu, Z. Feng, A. Zhang, D. Hou and L. Sun, Polymer, 2006, 47, 6066.
10. H. Yu, Z. Feng and A. Zhang, J. Polym. Sci., Part A: Polym. Chem., 2006, 44, 3717.
11. W. Batsberg, X. Chen, B. Gao and J. Kops, Polymer, 1998, 39, 911.
12. R. Krishnan and K. S. Srinivasan, Eur. Polym. J., 2003, 39, 205.
13. Y. Higaki, H. Otsuka and A. Takahara, Polymer, 2006, 47, 3784.
14. X. Zhang, X. Zhu, X. Tong, L. Ye, A. Zhang and Z. Feng, J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 5283.
15. L. Ren, F. Ke, Y. Chen, D. Liang and J. Huang, Macromolecules, 2008, 41, 5295.
16. X. Tong, X. Zhang, L. Ye, A. Zhang and Z. Feng, Polymer, 2008, 49, 4489.
17. K. L. Robinson, M. A. Khan, P. B. de, X. S. Wang and S. P. Armes, Macromolecules, 2001, 34, 3155.
18. M. Ceccato, P. L. Nostro, C. Rossi, C. Bonechi, A. Donati and P. Baglioni, J. Phys. Chem., 1997, 13, 2436.
19. H. Wei, H. Yu, A. Zhang, L. Sun, D. Hou and Z. Feng, Macromolecules, 2005, 38, 8833.
20. K. C. Tam, J. Liu and H. R. Sondjaja, Langmuir, 2007, 23, 5106.
21. T. Zhao and H. W. Beckham, Macromolecules, 2003, 36, 9859.
22. W. Yuan, J. Yuan, F. Zhang, X. Xie and C. Pan, Macromolecules, 2007, 40, 9094.
23. H. Lee, W. Jakubowski, K. Matyjaszewski, S. Yu and S. S. Sheiko, Macromolecules, 2006, 39, 4983.
24. A. Harada, M. Osaki, Y. Takashima and H. Yamaguchi, Acc. Chem. Res., 2008, 41, 1143.
25. A. Harada and M. Kamachi, Macromolecules, 1990, 23, 2821.
26. G. Wenz, B. H. Han and A. Müller, Chem. Rev., 2006, 106, 782.
27. C. Travelet, G. Schlatter, P. Hébraud, C. Brochon, A. Lapp, D. V. Anokhin, D. A. Ivanov, C. Gaillard and G. Hadziioannou, Soft Matter, 2008, 4, 1855.
28. J. W. Chung, T. J. Kang and S. Y. Kwak, Macromolecules, 2007, 40, 4225.

---