Polymerisation‐Induced Self‐Assembly of Graft Copolymers

Abstract We report the polymerisation‐induced self‐assembly of poly(lauryl methacrylate)‐graft‐poly(benzyl methacrylate) copolymers during reversible addition‐fragmentation chain transfer (RAFT) grafting from polymerisation in a backbone‐selective solvent. Electron microscopy images suggest the phase separation of grafts to result in a network of spherical particles, due to the ability of the branched architecture to freeze chain entanglements and to bridge core domains. Small‐angle X‐ray scattering data suggest the architecture promotes the formation of multicore micelles, the core morphology of which transitions from spheres to worms, vesicles, and inverted micelles with increasing volume fraction of the grafts. A time‐resolved SAXS study is presented to illustrate the formation of the inverted phase during a polymerisation. The grafted architecture gives access to unusual morphologies and provides exciting new handles for controlling the polymer structure and material properties.

1 Instrumental Methods

Nuclear Magnetic Resonance Spectroscopy
1 H Nuclear Magnetic Resonance (NMR) spectra and 1 H-13 C Heteronuclear Single Quantum Coherence (HSQC) spectra were recorded in deuterated chloroform (CDCl3) on Bruker Avance III HD (300 MHz or 400 MHz) spectrometer at 300 K. Chemical shift values (δ) are reported in ppm. Tetramethylsilane (TMS) was used as the internal standard.

Size Exclusion Chromatography
Size Exclusion Chromatography (SEC) was carried out using an Agilent Infinity II MDS instrument equipped with differential refractive index (DRI), viscometry (VS), dual angle light scattering (DALS) and multiple wavelength UV detectors. The system was equipped with 2×PLgel Mixed C columns (300×7.5 mm, 200 to 200,000 g/mol operating range) and a PLgel 5 µm guard column. Analyte samples were prepared in CHCl3 and filtered through a Fisherbrand PTFE syringe filter with 0.2 μm pore size before injection. Samples were run in CHCl3 at 1 ml/min at 30 o C. Experimental number-average molar mass (Mn,SEC) and dispersity (Đ) values of synthesised polymers were determined with Agilent GPC/SEC software by using Agilent EasyVial poly(methyl methacrylate) (PMMA) calibration.

Small-Angle X-Ray Scattering
Small-Angle X-ray Scattering (SAXS) measurements were performed using a Xenocs Xeuss 2.0 equipped with a micro-focus Cu Kα source collimated with scatterless slits providing a 0.8 mm diameter beam. SAXS patterns were recorded using a Pilatus 300K detector with a pixel size of 0.172 mm×0.172 mm. The sample to detector distance was calibrated using silver behenate (AgC22H43O2) providing a value of 2.481(5) m, providing an effective scattering vector Q range of 0.005-0.24 Å -1 , where Q is defined as where 2θ is the scattering angle and λ is the X-ray wavelength. Reaction mixtures were mounted at 20 wt% without dilution in a 1 mm (Ø) borosilicate glass capillaries or a Perspex holder with Kapton tape if too viscous for the capillaries. Data were collected for 20 min at 25 °C unless otherwise stated. A radial integration of the 2D scattering profile was performed using FOXTROT software and the resulting data corrected for the absorption, sample thickness and background. 1 Finally, the scattering intensity was then rescaled to absolute intensity using glassy carbon as a standard. 2 SAXS data were analysed using model-dependent analysis implemented in SasView software. 3 The scattering length density (SLD) defining the "scattering power" of a material, is defined as the sum of X-ray scattering lengths, bi, of N atoms within a given molecular or particle volume, Vm, as given by (3. 3) The SLD of a material can also be calculated using the bulk density ρ, atomic molar mass Mi and Avogadro's constant NA, 4 where (3.4) In this study, the SLD of dodecane, LMA and BzMA were calculated as 7.41×10 -6 Å -2 , 8.22×10 -6 Å -2 , and 9.52×10 -6 Å -2 , respectively, and fixed for the fitting procedure.

Graft length series
Graft DP 1 to 15 were analysed using a spherical form factor with a Gaussian radial polydispersity applied and a sticky hard sphere structure factor (Table S4). 5,6 Graft DP 18 was analysed using a cylindrical form factor with a Gaussian radial polydispersity applied. 7 Graft DPs 24 and 37 were analysed using a flexible cylinder form factor, describing a cylinder with a total persistence length which can be split into shorter segments which can be considered rigid, described by the Kuhn length (Table S5). 8,9 Similarly to the cylindrical form factor described above, a Gaussian radial polydispersity was also applied to the flexible cylinder model. Graft DP 53 was analysed using a vesicular form factor with parameters describing the wall thickness and vesicle radius (Table S6) 6 Similarly to the spherical form factors, a Gaussian polydispersity was also applied to the wall thickness and vesicle radius. Finally, graft DP 100 was analysed using a "raspberry" form factor, 10 describing small spheres within a larger spherical structure (Table S7). In this case, the SLD of the small spheres was fixed to that for LMA, and the SLD of the larger spheres fixed to that for BzMA. The fractional penetration depth of small spheres within the larger spheres was set to 1, representing small spheres distributed throughout the larger sphere. Similar to models above, a Gaussian radial polydispersity was applied.

Time-resolved SAXS experiment
The reaction carried out in the time-resolved SAXS experiment was equivalent to (6.12) conducted at 20 wt% solids targeting a graft length of 106 repeating units. Reaction mixture was prepared in a septum-capped vial by dissolving pLMA915-CTA10% (6) (10.72 mg, 3.6 μmol side chain CTAs) and BzMA (70.88 mg, 386 μmol) in n-dodecane (412 μl). V-601 initiator stock solution (23 μl, 0.10 μmol, 1.0 mg/ml in n-dodecane) was added. The reaction mixture was deoxygenated by bubbling N2 into the solution for 15 min and transferred into a 1 mm (Ø) borosilicate glass capillary under an argon blanket. The capillary was kept at room temperature, in the dark and under argon until mounting (<2 h). The capillary was mounted at room temperature, aligned, and then heated up to 70 °C at the rate of 5 °C/min to start the reaction using a Linkam HFSX 350 temperature stage. Data were collected at 70 °C continuously over 320 min and binned to a 5 min time resolution.
Data collected from 5 to 50 min were fitted using a Gaussian coil form factor describing individual polymer chains in terms of the zero-angle intensity, I0, proportional to the volume fraction of polymer chains in solution, the volume of an individual chain and the SLD contrast between polymer and solvent, and the radius of gyration, Rg. 11 Data from 55 to 85 min were fitted using an ellipsoidal form factor with parameters describing the radii along the polar and equatorial axes. 12 Data collected from 140 to 320 min were fitted using the raspberry form factor as described above.

Scanning Electron Microscopy
Scanning electron microscopy (SEM) was carried out using a Zeiss SUPRA 55-VP instrument operating at 2-12 kV accelerating voltage. Polymer samples were spin-coated or merely deposited onto silicon wafers directly from reaction mixtures and placed under vacuum overnight. Samples were coated with carbon using an Emitech K950X turbo-pumped evaporator prior to imaging.
Cryo-SEM images were taken on a Zeiss Supra 55VP fitted with a Gatan Alto 2500 cryo transfer system.

Transmission Electron Microscopy
Transmission electron microscopy (TEM) was conducted using a JEOL 2100Plus instrument operating at 80-200 kV and equipped with a Gatan Orius 11-megapixel digital camera. Samples were deposited directly from reaction mixtures unless otherwise stated onto 300 mesh carbon-coated copper grids. Excess sample was blotted off with a filter paper.
4-Cyano-4-(((dodecylthio)carbonothioyl)thio)pentanoic acid (CPADTC) RAFT agent was prepared by adapting a previously reported protocol (Scheme S1). 14 Ground NaOH (6.23 g, 0.16 mol) was suspended in 500 ml diethyl ether using an overhead stirrer. 1-Dodecanethiol (30.0 g, 0.15 mol) was added dropwise and stirring was continued for 10 min. Carbon disulfide (11.9 g, 0.16 mol) was added in one shot and stirring was continued for 1 h. Reaction mixture was cooled down in the freezer and solids were filtered, washed with cold diethyl ether and dried, giving sodium dodecyl carbonotrithioate as a yellow solid (25.2 g, 57%). The product was used in the next step without purification.

Preparation of Poly(lauryl methacrylate-stat-hydroxyethyl methacrylate) Copolymers
Scheme S2 RAFT solution polymerisation of LMA and HEMA used in this work to construct graft copolymer main chains.
A series of poly(lauryl methacrylate-stat-hydroxyethyl methacrylate) (p(LMA-s-HEMA)) copolymers (1-5) were prepared using RAFT polymerisation (Scheme S2) with the following general procedure.  Figure S3). Reaction was stopped at 70-98% conversion after ≥14 h by cooling down to room temperature. The copolymers were not isolated prior to the functionalisation step. A portion of polymer (1) was precipitated in MeOH thrice to obtain a 1 H NMR spectrum of an isolated product ( Figure S4).
Monomer conversion was determined using 1 H NMR spectroscopy by setting the area δ = 3.5-4.5 ppm as a constant and quantifying the disappearance of the vinyl signals, δLMA = 5.5 ppm and δHEMA = 5.6 ppm. SEC samples were sampled directly from reaction mixtures.
Theoretical number-average molar masses (Mn,th) were calculated as All main chains and their characteristics are listed in Table S1.

Table S1
Main chain copolymers employed in this study.

Functionalisation of Poly(lauryl methacrylate-stat-hydroxyethyl methacrylate) Copolymers
Scheme S3 Synthetic route used in this work for main chain functionalisation.
The Steglich esterification (Scheme S3) was used with the following general procedure to couple CPADTC onto copolymer main chains (1-5) to yield functionalised main chains (6-10) ( Table S2). An aliquot of the polymer solution from the copolymerisation step (4.00 g p(LMA424-s-HEMA50) (2), 2.1 mmol HEMA) was added to a dry 250 ml round-bottom flask and toluene was removed by rotary evaporation at room temperature. CPADTC (1.26 g, 3.1 mmol) and a stir bar were added, and the flask was sealed with a septum. Anhydrous DCM (125 ml) was cannulated into the flask and the mixture was stirred at room temperature until the polymer and CPADTC had fully dissolved. DMAP (25 mg, 0.2 mmol) was added as a solid and the flask was immersed in an ice bath. Solid DCC (0.612 g, 3.0 mmol) was added to start the reaction and stirring was continued for 15 minutes over ice, during which dicyclohexylurea started to precipitate out of solution. The flask was taken off ice and stirring was continued overnight. The reaction mixture was filtered twice through cotton wool and concentrated using a rotary evaporator at room temperature. The product was precipitated into methanol, centrifuged, isolated, and redissolved in DCM. Precipitations (n ≥ 3) were carried out until no free CTA remained in the product.
The polymers were characterised with 1 H NMR spectroscopy ( Figure S6) and SEC ( Figure S5).
The extent of functionalisation, or the maximum number of grafting points per main chain (nCTA), was calculated from the 1 H NMR spectra of the isolated polymers as where nHEMA and nLMA are the number of HEMA and LMA units, respectively, as given by conversion.
The degree of functionalisation (nCTA%) of the main chain was defined as Theoretical number-average molar masses (Mn,th) of the functionalised main chains, pLMADP,tot-CTAnCTA%, were calculated as ',)* = BC9,)* + +,-+3-4,+ where Mpre,th is the theoretical molar mass of the precursor copolymer, p(LMA-s-HEMA) ( Table S1), and MCPADTC is the molar mass of CPADTC. A The x and y in the formula pLMAx-CTAy indicate the DP of the main chain (DPtot) and the degree of functionalisation (nCTA% , Eq S3), respectively. B Degree of functionalisation of HEMA repeating units based on the 1 H NMR spectrum of the product and the theoretical structure of the precursor copolymer. C Number of CPADTC units per polymer (Eq S2) D Calculated based on the structure of the precursor copolymer (1-5) and number of CPADTC units (Eq S4). E Experimental number-average molar mass and dispersity as given by SEC analysis in CHCl3 with DRI detection and PMMA calibration.

"Grafting From" Dispersion Polymerisation of Benzyl Methacrylate
Scheme S4 Dispersion polymerisation conditions used in this work for studying PISA of graft copolymers. The graft copolymers were characterised using 1 H NMR spectroscopy and SEC by sampling directly from the reaction mixture. Theoretical number-average molar masses (Mn,th) of the graft copolymers, pLMA-g-pBzMA, were calculated as where Mmc,th is the theoretical molar mass of the functionalised main chain, pLMADP,tot-CTAnCTA% (Table  S2), and MBzMA is the molar mass of BzMA.
Reinitiation efficiency of the CPADTC units (Iinit) was calculated as The apparent number of grafts (ng) was calculated by taking into account the maximum number of grafting points (Eq S2) and the reinitiation efficiency of CPADTC as given by 1 H NMR (Eq S7). The apparent number of grafts was calculated as The apparent grafting density was defined as where DPtot is the length of the main chain. The resulting apparent DP (DPapp) of the pBzMA grafts was calculated by correcting for the apparent number of grafts as where DPp is the graft length given by conversion assuming 100% reinitiation efficiency, that is (S11) Figure S7 Representative 1 H NRM spectra (400 MHz, CDCl3) of a dispersion polymerisation of BzMA before and after reaction using functionalised main chain pLMA915-CTA10% (6) and targeting graft length of 5 repeating units. 42% of CPADTC units remain unfragmented after reaction as indicated by shift δ = 3.32 → 3.19 ppm

Effect of Graft Length
The effect of graft length on the resulting morphologies was studied at 20 wt% by preparing a series of graft lengths using one main chain ( Table S3). The appearances of the reaction mixtures after reaction are shown in Figure S8. SEC was carried out in CHCl3 to confirm an increase in molecular weight with increasing graft length. The resulting materials (6.1-6.12) were characterised using SAXS without modification. SEM and TEM were used to visualise the nanostructures (Figures S11 and S20).  S5). B Graft length as given by conversion assuming full reinitiation of CPADTC units (Eq S11). C Reinitiation efficiency of CPADTC units as given by 1 H NMR of end conversion sample. D,E Apparent graft length and grafting density, taking into account the reinitiation efficiency (Eq S10 and S9). F Molar ratio of BzMA and LMA in the graft copolymer. G Theoretical number-average molar mass as given by conversion (Eq S6). H Experimental number-average molar mass and dispersity as given by SEC analysis in CHCl3 with DRI detection and PMMA calibration.

Figure S8
Left: Appearance of the reaction mixtures at room temperature after the PISA of pLMA915-g-pBzMAx, derived from main chain pLMA915-CTA10% at 20 wt%. Right: Recovery of gel-like material (6.6) after agitation.

Table S7
Parameters obtained through fitting SAXS data for graft length DP 100 to a raspberry form factor describing small pLMA spheres within a larger pBzMA particle. Values marked with * were held constant throughout the fitting procedure.

Effect of Main Chain Length
The effect of main chain length on the PISA was studied at 20 wt% by polymerising a series of graft lengths with three functionalised main chains of dissimilar lengths but similar degrees of functionalisation ( Table S8). The appearances of the reaction mixtures after reaction are shown in Figure S12. SEC was carried out to confirm an increase in molecular weight with increasing graft length ( Figure S13). TEM was used to visualise some of the resulting materials.

Table S8
Graft copolymers pLMAx-g-pBzMAy prepared for studying the effect of main chain length on PISA transitions using main chains of dissimilar lengths but similar degrees of functionalisation.  S5). B Graft length as given by conversion assuming full reinitiation of CPADTC units (Eq S11). C Molar ratio of BzMA and LMA in the graft copolymer. D Theoretical number-average molar mass as given by conversion (Eq S6). E Experimental number-average molar mass and dispersity as given by SEC analysis in CHCl3 with DRI detection and PMMA calibration.

Effect of Grafting Density
The effect of grafting density on the PISA was studied at 20 wt% by polymerising a series of graft lengths with three functionalised main chains with dissimilar degrees of functionalisation but similar lengths ( Table S9). The appearances of the reaction mixtures after reaction are shown in (Figure S14). SEC was carried out to confirm an increase in molecular weight with increasing graft length ( Figure  S15).  S5). B Graft length as given by conversion assuming full reinitiation of CPADTC units (Eq S11). C Molar ratio of BzMA and LMA in the graft copolymer. D Theoretical number-average molar mass as given by conversion (Eq S6). E Experimental number-average molar mass and dispersity as given by SEC analysis in CHCl3 with DRI detection and PMMA calibration.
Figure S15 SEC profiles of pLMA-g-pBzMA graft copolymers prepared targeting various graft lengths and grafting densities. Analysis was conducted in CHCl3 with DRI detection and PMMA calibration.

Effect of Concentration
The effect of total mass concentration on the PISA was studied at 10 wt% and 20 wt% by polymerising two series of graft lengths with each of three functionalised main chains of dissimilar lengths but similar degrees of functionalisation (Table S10). The appearances of the reaction mixtures after reaction are shown in (Figure S16). SEC was carried out to confirm an increase in molecular weight with increasing graft length ( Figure S17).  S5). B Graft length as given by conversion assuming full reinitiation of CPADTC units (Eq S11). C Molar ratio of BzMA and LMA in the graft copolymer. D Theoretical number-average molar mass as given by conversion (Eq S6). E Experimental number-average molar mass and dispersity as given by SEC analysis in CHCl3 with DRI detection and PMMA calibration. F Sampling unreliable due to heterogeneity of reaction mixture. G Targeted graft length.

Figure S17
Size exclusion chromatograms of pLMAx-g-pBzMAy graft copolymers prepared at 10 wt% (A-C) and 20 wt% (D-F) in the concentration dependence study. Analysis was carried out using CHCl3 as the eluent, DRI detection and PMMA calibration. Graft lengths indicated in brackets could not be reliably quantified with 1 H NMR and correspond to targeted DP.

Figure S19
Destructive effects of the electron beam. Images show loss of colloidal shape and internal morphology of nanostructures prepared at 20 wt% solids.