Shape-shifting thermoreversible diblock copolymer nano-objects via RAFT aqueous dispersion polymerization of 4-hydroxybutyl acrylate

2-Hydroxypropyl methacrylate (HPMA) is a useful model monomer for understanding aqueous dispersion polymerization. 4-Hydroxybutyl acrylate (HBA) is an isomer of HPMA: it has appreciably higher aqueous solubility so its homopolymer is more weakly hydrophobic. Moreover, PHBA possesses a significantly lower glass transition temperature than PHPMA, which ensures greater chain mobility. The reversible addition–fragmentation chain transfer (RAFT) aqueous dispersion polymerization of HBA using a poly(ethylene glycol) (PEG113) precursor at 30 °C produces PEG113–PHBA200–700 diblock copolymer nano-objects. Using glutaraldehyde to crosslink the PHBA chains allows TEM studies, which reveal the formation of spheres, worms or vesicles under appropriate conditions. Interestingly, the partially hydrated highly mobile PHBA block enabled linear PEG113–PHBAx spheres, worms or vesicles to be reconstituted from freeze-dried powders on addition of water at 20 °C. Moreover, variable temperature 1H NMR studies indicated that the apparent degree of hydration of the PHBA block increases from 5% to 80% on heating from 0 °C to 60 °C indicating uniform plasticization. In contrast, the PHPMAx chains within PEG113–PHPMAx nano-objects become dehydrated on raising the temperature: this qualitative difference is highly counter-intuitive given that PHBA and PHPMA are isomers. The greater (partial) hydration of the PHBA block at higher temperature drives the morphological evolution of PEG113–PHBA260 spheres to form worms or vesicles, as judged by oscillatory rheology, dynamic light scattering, small-angle X-ray scattering and TEM studies. Finally, a variable temperature phase diagram is constructed for 15% w/w aqueous dispersions of eight PEG113–PHBA200–700 diblock copolymers. Notably, PEG113–PHBA350 can switch reversibly from spheres to worms to vesicles to lamellae during a thermal cycle.


Covalent Stabilization of Diblock Copolymer Nano-objects Using Glutaraldehyde
A typical protocol used for crosslinking PEG 113 −PHBA 260 spheres is as follows. A 15% w/w acidic aqueous dispersion of PEG 113 −PHBA 260 spheres (0.2 g) was diluted to 5% w/w using water (0.4 g; PHBA = 455 µmol), adjusted to pH 7 (1 M NaOH) and stirred for 4 h in a 7 ml reaction vial. Glutaraldehyde (GA; 34.1 mg, 341 µmol, GA/PHBA molar ratio = 0.66) was then added and stirred at 20 °C for 16 h. Then an aliquot (0.01 g) was extracted and diluted with water (4.99 g; final target solids concentration = 0.05% w/w) and stirred for 24 h prior to preparation of the corresponding TEM grid (see below for further details). Crosslinking of PEG 113 −PHBA x worms and vesicles was also performed at 5% w/w. In order to crosslink PEG 113 -PHBA x diblock copolymer nano-objects at a particular temperature, 5% w/w dispersions were either equilibrated at the desired reaction temperature for 1 h prior to GA addition and then allowed to crosslink for 16 h before dilution with water (final target solids concentration = 0.05% w/w) that had been pre-equilibrated at the same temperature.

Copolymer Characterization
1 H NMR Spectroscopy. Spectra were recorded in either CD 3 OD or D 2 O using a 400 MHz Bruker AVANCE-400 spectrometer with 64 scans being averaged per spectrum. The relative degree of (partial) hydration of the PHBA chains was calculated by comparing the S4 integrated signals assigned to the two COO-CH 2 protons labeled c' and the four CH 2 CH 2 protons labeled b' relative to the four oxyethylene protons labeled a' assigned to the PEG 113 stabilizer block at each temperature (see Figure 6). Thus, a degree of hydration of 100% is calculated in such experiments if the apparent diblock composition corresponds to that observed by 1 H NMR spectroscopy when employing a good solvent for the PEG and PHBA blocks (e.g. CD 3 OD).

Gel Permeation Chromatography. Copolymer molecular weights and dispersities
were determined using an Agilent 1260 Infinity GPC system equipped with both refractive index and UV−visible detectors. Two Agilent PLgel 5 μm Mixed-C columns and a guard column were connected in series and maintained at 60 °C. HPLC-grade DMF containing 10 mM LiBr was used as the eluent and the flow rate was set at 1.0 mL min −1 . Refractive index detection was used for calculation of molecular weights and dispersities by calibration against a series of ten near-monodisperse poly(methyl methacrylate) standards (with M n values ranging from 370 to 2,520,000 g mol −1 ).

Differential Scanning Calorimetry. DSC studies were performed using a TA
Instruments Discovery DSC 25 instrument equipped with TZero low-mass aluminum pans and vented lids. Copolymers (and homopolymers) were equilibrated above their T g for 10 min before performing two consecutive thermal cycles at 10 °C min −1 . Two cycles were performed to minimize the thermal history of each sample and ensure removal of any residual solvents. Only data obtained during the second thermal cycle are presented.
Fourier Transform Infrared Spectroscopy. FTIR spectra were recorded using a Thermo-Scientific Nicolet IS10 FT-IR spectrometer equipped with a Golden Gate Diamond ATR accessory. Spectra were recorded for freeze-dried homopolymers and copolymers after drying in a vacuum oven at 30 °C for 3 days. Each spectrum was averaged over 256 scans.
Transmission Electron Microscopy. Unless stated otherwise, as-prepared copolymer dispersions were diluted at 20 °C using acidified deionized water (pH 3) to generate 0.05% w/w aqueous dispersions. Copper/palladium TEM grids (Agar Scientific, UK) were coated in-house to produce thin films of amorphous carbon. These grids were then treated with a plasma glow discharge for 30 s to create a hydrophilic surface. One droplet of an aqueous copolymer dispersion (20 μL; 0.05% w/w) was placed on a freshly-treated grid for 1 min and then blotted with a filter paper to remove excess solution. To stain the deposited nanoparticles, an aqueous solution of uranyl formate (10 μL; 0.75% w/w) was placed on the sample-loaded grid via micropipet for 45 s and then carefully blotted to remove excess stain. Each grid was then dried using a vacuum hose. Imaging was performed using a Philips CM100 instrument operating at 100 kV and equipped with a Gatan 1k CCD camera. Rheology. An AR-G2 rheometer equipped with a variable temperature Peltier plate and a 40 mL 2° aluminum cone was used for all experiments. The dispersion viscosity, loss modulus and storage modulus were measured as a function of applied strain, angular frequency, and temperature to assess the gel strength, gel viscosity and critical gelation temperature. Temperature sweeps were conducted using 20% w/w copolymer dispersions at an applied strain of 1.0% and an angular frequency of 1.0 rad s -1 . In these latter experiments, the copolymer dispersion was subjected to a single thermal cycle 4 (heating up to 60°C, followed by cooling to 1°C), and then equilibrated at 1°C for 15 min prior to measurements. To obtain these data, the dispersion was initially cooled to 0°C and held for 60 s prior to heating at 2°C intervals, allowing a thermal equilibration time of 60 s between each measurement. SAXS Studies. SAXS patterns were recorded for 1.0% w/w aqueous copolymer dispersions at Diamond Light Source (station I22, Didcot, UK) using monochromatic synchrotron X-ray radiation (λ = 0.124 nm, with q ranging from 0.015 to 1.300 nm -1 , where q = 4π sin θ/λ is the length of the scattering vector and θ is one-half of the scattering angle) and a 2D Pilatus 2M pixel detector (Dectris, Switzerland). Alternatively, a Xeuss 2.0 (Xenocs) SAXS instrument equipped with a FOX 3D multilayered X-ray mirror, two sets of scatterless slits for collimation, a hybrid pixel area detector (Pilatus 1M, Dectris) and a liquid gallium MetalJet X-ray source (Excillum, λ = 1.34 Å) was used. In the latter case, SAXS patterns were recorded at a sample-to-detector distance of approximately 1.20 m (calibrated using a silver behenate standard). Glass capillaries of 2.0 mm diameter were used as sample holders. SAXS data were reduced (integration, normalization and absolute intensity calibration using SAXS patterns recorded for deionized water assuming that the differential scattering cross-section of water is 0.0162 cm -1 ) using Dawn software supplied by Diamond Light Source. 3 For the variable temperature experiments, the sample holders were placed in a HFSX350-CAP temperature-controlled stage (Linkam Scientific, Tadworth, UK) and 10 min was allowed between each measurement to ensure thermal equilibration.

Small Angel X-ray Scattering (SAXS) Models
In general, the intensity of X-rays scattered by a dispersion of nano-objects [usually represented by the scattering cross section per unit sample volume, ] can be Σ Ω ( ) expressed as: where ( , 1 ,… , ) is the form factor, 1 ,…, is a set of k parameters describing the structural morphology, ( 1 ,… , ) is the distribution function, S(q) is the structure factor and N is the nano-object number density per unit volume expressed as: where ( 1 ,… , ) is volume of the nano-object and φ is their volume fraction in the dispersion. For all SAXS experiments conducted herein, a dilute copolymer concentration of 1.0 % w/w was utilised. As such, for all analysis and modelling it was assumed that s(q) = 1.
Sphere model. The spherical micelle form factor equation for Equation S1 is given by 4 : Where R s is the core radius of the spherical micelle, R g , is the radius of gyration of the PEG 113 corona block. The core block and the corona block X-ray scattering length contrast is given by and , respectively. Here , and are the X-ray scattering length densities of the core-forming block (ξ PHBA = 10.65  10 10 cm -2 ), the coronal stabilizer block (ξ PEG113 = 11.37  10 10 cm -2 ) and the solvent (ξ sol = 9.42  10 10 cm -2 ). V s and V c are the volumes of the core-forming block and the coronal stabilizer block, respectively. Using the molecular weights of the PHBA and PEG 113 blocks and their respective mass densities (ρ PHBA = 1.16 g cm -3 and ρ PEG113 = 1.23 g cm -3 ), the individual block volumes can be calculated from , where M n,pol corresponds to the number-average molecular weight = , of the block determined by 1 H NMR spectroscopy. The sphere form factor amplitude is used for the amplitude of the core self-term: Where . A sigmoidal interface between the two blocks Φ( was assumed for the spherical micelle form factor (equation S4). This is described by the exponent term with a width σ accounting for a decaying scattering length density at the micellar interface. This σ value was fixed at 2.2 during fitting. The form factor amplitude of the spherical micelle corona is: The radial profile, , can be expressed by a linear combination of two cubic b splines, ( ) with two fitting parameters s and a corresponding to the width of the profile and the weight coefficient, respectively. This information can be found elsewhere, 5,6 as can the approximate integrated form of Equation S5. The self-correlation term for the corona block is given by the Debye function:

S6
Where R g is the radius of gyration of the PEG 113 coronal block. The aggregation number of the spherical micelle is: Where x sol is the volume fraction of solvent in the PHBA micelle core. An effective structure factor expression proposed for interacting spherical micelles 5 has been used in equation S1: Herein the form factor of the average radial scattering length density distribution of micelles is used as and is a hard-sphere interaction structure factor based on the Percus-Yevick approximation, 7 where R PY is the interaction radius and f PY is the hard-sphere volume fraction. A polydispersity for one parameter (R s ) is assumed for the micelle model which is described by a Gaussian distribution. Thus, the polydispersity function in Equation S1 can be replaced as: Where is the standard deviation for R s . In accordance with equation S2, the number density per unit volume for the micelle model is expressed as: where is the total volume fraction of copolymer in the spherical micelles and is the ( 1 ) total volume of copolymer in a spherical micelle .

Worm-like micelle model
The worm-like micelle form factor in Equation S1 is expressed as 4 : S11 ( ) = 2 2 2 ( ) + 2 ( , ) + ( -1) 2 ( ) + 2 2 ( ) where all the parameters are the same as in the spherical micelles model (Equation S3). The self-correlation time for the worm-like micelle core or radius is: where S13 2 ( , ) = [ 2 1 ( ) ] 2 and J 1 is the first-order Bessel function of the first kind, and a form factor F worm (q, L w , b w ) for self-avoiding semi-flexible chains represent the worm-like micelle, where b w is the worm Kuhn length and L w is the mean worm contour length. A complete expression for the chain form factor can be found elsewhere. 8 The self-correlation term for the corona block is given by the Debye function shown in Equation S6. The mean aggregation number of the wormlike micelle is given by: where x sol is the volume fraction of solvent within the worm-like micelle core. Possible semispherical caps at the ends of each worm are ignored in this form factor. The R g obtained for the PEG 113 coronal block of 2.7 nm is comparable to the estimated value of 2.6 nm.

Vesicle model
The vesicle form factor in Equation (S1) is expressed as: The X-ray scattering length contrast for the membrane-forming block (PHBA) and the coronal stabilizer block (PEG 113 ) is given by = ( − ) and = ( − ), respectively, where ξ m , ξ vc and ξ sol are the X-ray scattering length densities of the membrane-forming block (ξ PHBA = 10.65  10 10 cm -2 ), the coronal stabilizer block (ξ PEG113 = 11.37  10 10 cm -2 ) and the solvent (ξ sol = 9.42 10 10 cm -2 ). V m and V vc are the volumes of the membrane-forming block and the coronal stabilizer block, respectively. Using the molecular weights of the PHBA and PEG 113 blocks and their respective mass densities (ρ PHBA = 1.16 g cm -3 and ρ PEG113 = 1.23 g cm -3 ), the individual block volumes can be calculated from , where M n,pol corresponds to the number-average molecular = , weight of the block determined by 1 H NMR spectroscopy. The amplitude of the membrane self-term is: where is the inner radius of the membrane, is the outer radius of the membrane, , . It should be noted that Equation S16 differs from that reported in the original work. More specifically, the exponent term in Equation S16 represents a sigmoidal interface between the blocks, with a width σ in accounting for a decaying scattering length density at the membrane surface. The numerical value of σ in was fixed at 2.2. The mean vesicle aggregation number, N v , is given by: is the solvent (i.e. water) volume fraction within the vesicle membrane. A simpler expression for the corona self-term of the vesicle model than that used for the spherical micelle corona self-term was preferred because the contribution to the scattering intensity from the corona block is much less than that from the membrane block in this case. Assuming that there is no penetration of the solvophilic coronal blocks into the solvophobic membrane, the amplitude of the vesicle corona self-term is expressed as: where the term outside the square brackets is the factor amplitude of the corona block copolymer chain such that: The mean experimental R g value of 2.7 nm for the PEG 113 coronal block is close to the theoretical value (2.6 nm). The latter can be calculated from the contour length of the PEG 113 block, L PEG113 = 113  0.37 nm = 41.8 nm (the projected contour length of an ethylene glycol repeat unit (0.37 nm) was based on a known literature value obtained for the crystal structure of PEG homopolymer). 9 Assuming a PEG Kuhn length of 1 nm, 10 an approximate R g of (41.8  1/6) 0.5 = 2.6 nm was calculated.
For the vesicle model, it was assumed that two parameters are polydisperse: the overall radius of the vesicles and the membrane thickness (R m and T m , respectively). Each is assumed to have a Gaussian distribution, so the polydispersity function in Equation (S1) can be expressed as: where σ Rm and σ Tm are the standard deviations for R m and T m , respectively. Following Equation S2, the number density per unit volume for the vesicle model is expressed as: where φ is the total volume fraction of copolymer in the vesicles and ( 1 , 2 ) is the total volume of copolymers in a vesicle [ ( 1 , 2 ) = ( + ) ( 1 , 2 )].
diblock copolymer used for a direct comparison of the thermoresponsive behavior of these two systems.   15 The broad band at around 3360 cm -1 is assigned to the OH stretch associated with the pendent hydroxyl groups in the HBA repeat units of the linear PEG 113 -PHBA 600 diblock copolymer (see red trace). [16][17][18][19][20] Notably, this feature is much less intense in the FT-IR spectrum recorded for GA-crosslinked PEG 113 -PHBA 600 (see green trace). Literature precedent suggests that the new band at 1110 cm -1 (and most likely also the 1345 cm -1 band) is the result of acetal formation (see green trace). 21 Finally, the characteristic strong carbonyl stretch for GA at 1635 cm -1 (not shown above) is not observed in the FT-IR spectrum recorded for the final GA-crosslinked PEG 113 -PHBA 600 vesicles, indicating minimal contamination by residual GA.   Table S1. Summary of the target diblock copolymer compositions, DLS particle diameters and polydispersities, and TEM morphology assignments for the PEG 113 -PHBA x diblock copolymer nano-objects used to construct the phase diagram shown in Figure 4. [N.B. All TEM studies were conducted on glutaraldehyde-crosslinked nano-objects].  . DLS studies were conducted on 0.1% w/w aqueous copolymer dispersions (originally synthesized at 20% w/w). SAXS studies were performed on 1.0% w/w aqueous dispersions using well-known sphere, worm or vesicle models for data fits (see the SAXS models section in the supporting information for more information). TEM analysis was conducted on 0.05% w/w aqueous dispersions of glutaraldehyde-crosslinked PEG 113 -PHBA x nano-objects. Crosslinking conditions: GA/HBA molar ratio = 0.66, [copolymer] = 5% w/w, 20°C, pH 7.
SAXS analysis of linear PEG 113 -PHBA 200 spheres using a well-established spherical micelle model 4 indicated a volume-average diameter of 35 ± 4 nm for the PHBA cores, which is consistent with an overall hydrodynamic z-average diameter of 44 ± 6 nm reported by DLS (and an estimated TEM number-average diameter of 38 ± 5 nm for the corresponding GAcrosslinked spheres). The SAXS pattern recorded for PEG 113 -PHBA 350 nano-objects was fitted using Pedersen's worm-like micelle model. 4 The worm core cross-sectional diameter was calculated to be 41 ± 5 nm, which is slightly lower than estimated from TEM studies of the GA-crosslinked worms (40 ± 8 nm). However, this is not unreasonable given the additional mass conferred by the GA crosslinker. Moreover, although covalent stabilization is essential for visualization of the copolymer morphology, some degree of worm deformation (flattening) may occur on drying, which would lead to a slightly larger apparent worm core diameter. Unfortunately, the q range used for these SAXS studies was too short to enable determination of the overall worm contour length. TEM studies indicated highly linear PEG 113 -PHBA 300-400 worms with a broad distribution of worm lengths (0.5-10 µm). The SAXS pattern obtained for the PEG 113 -PHBA 500 nano-objects could be satisfactorily fitted using a well-known vesicle model. 8 This indicated relatively small vesicles with an overall volumeaverage diameter of 143 ± 21 nm and a mean vesicle membrane thickness of 23 ± 3 nm. This membrane thickness is smaller than both the sphere diameter and worm thickness, which suggests strong interdigitation of the PHBA chains. 22,23 As expected, the above vesicle diameter is lower than the z-average diameter of 153 ± 48 nm reported by DLS. However, TEM studies of the corresponding GA-crosslinked vesicles indicated a slightly higher number-average diameter (172 ± 56 nm), which suggests that some degree of deformation occurs during grid preparation. Figure S7. Temperature-dependent rheological data obtained for a 15% w/w aqueous dispersion of PEG 113 -PHPMA 260 short worms at an applied strain of 1.0% and an angular frequency of 1.0 rad s -1 . This dispersion was held at 5 °C for 30 min prior to heating to ensure thermal equilibration. (a) G' (black diamonds) and G'' (black triangles) and (b) complex viscosity during initial heating (red circles) and subsequent cooling (blue squares) runs. The higher hydrophobicity of the PHPMA-cores inhibited any thermoresponsive behavior. PEG 113 -PHPMA 260 short worms indicated no thermoresponsive behavior: this dispersion always remained a viscous fluid (G' = 2.0 Pa and G'' = 2.5 Pa) and never formed a freestanding gel. This is consistent with observations made by Lovett et al. 24 and Warren et al., 25 who found that PHPMA-based nano-objects exhibited little or no thermoresponsive behavior if the PHPMA DP exceeded 200.  Table S3. Summary of z-average DLS diameters, structural parameters calculated from SAXS analysis and number-average TEM diameters obtained for PEG 113 -PHBA 260 nanoobjects at 10, 36 or 50°C. DLS studies were conducted using 0.1% w/w aqueous dispersions. SAXS studies were performed on 1.0% w/w aqueous dispersions using wellknown sphere, worm or vesicle models for data fits (see the SAXS models section in the supporting information for more information). TEM analysis was conducted using 0.05% w/w aqueous dispersions of GA-crosslinked PEG 113 -PHBA x nano-objects. Crosslinking conditions: GA/HBA molar ratio = 0.66, [copolymer] = 5% w/w, pH 7. Figure S9. Variation in the sphere-equivalent hydrodynamic diameter (blue diamonds) and DLS polydispersity (red triangles) obtained for GA-crosslinked PEG 113 -PHBA 300 worms prepared at 20 °C. The minimal change in each parameter confirms effective covalent stabilization. In contrast, the linear PEG113-PHBA300 nano-objects undergo interconversion between spheres, worms and vesicles over the same temperature change, as indicated in Fig. 7.     (c) SAXS patterns recorded for a 1.0% w/w aqueous dispersion of linear PEG 113 -PHBA 500 diblock copolymer lamellae at 50°C. The mean inter-sheet stacking distance was 115 nm was calculated from the diffraction peak labeled q* using the equation shown in the inset. 26