Multi-responsive (diethylene glycol)methyl ether methacrylate (DEGMA)-based copolymer systems

Abstract

Multi-responsive (diethylene glycol) methyl ether methacrylate–methacrylic acid (DEGMA–MAA) copolymers were synthesised for the first time via RAFT polymerisation, and the structure–activity interplay of statistical and block copolymers in solution was compared. In addition to MAA, DEGMA was also copolymerised with 2-(dimethylamino)ethyl methacrylate (DMAEMA) to generate two sets of well-defined weakly acidic or weakly basic copolymers, respectively. The temperature, pH and salt-responsive properties of all polymers was determined via UV-visible spectroscopy and dynamic light scattering, and their solution structures as a function of temperature were visualised via electron microscopy. Results from this study indicate that the DEGMA–MAA statistical and block copolymers display similar stimuli-responsive property trends to the well-characterised DEGMA–DMAEMA copolymers, and show similar lower critical solution temperature (LCST) modulations and aggregate structures in water.

---

Introduction

Polymers that respond reversibly to changes to their solution environment have been extensively studied for a number of applications such as sensors, chromatography, drug encapsulation and drug delivery.^1–5 Depending on the polymer system employed, small, externally applied changes in stimuli such as temperature, pH, and light can cause dramatic changes in a polymer's solution conformation, three-dimensional structure, function, and solubility. One of the most studied stimuli-responsive polymer systems is the thermoresponsive polymer poly(N-isopropylacrylamide) (PNIPAM), which exhibits a characteristic phase transition at approximately 32 °C in water. Numerous studies have investigated the ability to fine-tune the lower critical solution temperature (LCST) of PNIPAM systems based on the incorporation of different comonomers, for use in applications such as chromatography for the production of stimuli-responsive packing materials, and for tissue culture surface modification, to enable reversible detachment of cells from a surface.^6–11 Whilst the PNIPAM system is exemplary for certain applications, other thermoresponsive polymer systems are gaining traction as viable alternatives with thermoresponsive properties that can be equally well controlled.^10,12–14

One such system is the thermoresponsive (oligoethylene glycol)methacrylate (OEGMA)-based polymers that exhibit temperature-responsive properties comparable to NIPAM-based polymers. In a point-by-point comparison between these two sets of polymers, the LCSTs of OEGMA polymers were found to be as equally sensitive to temperature, but considerably less sensitive to differences in polymer chain length, compared to NIPAM-based polymers.¹³ In addition, as OEGMA-based polymer systems contain ethylene glycol side chains of varying lengths (between ∼1–9 ethylene glycol units), they exhibit additional properties that are desirable for biological applications, including high biocompatibility. Furthermore, the polymers are uncharged and present with little to no toxicity in vivo.¹⁵

The number of ethylene glycol units in the monomer side chain is known to cause dramatic differences in the LCST of OEGMA homopolymers.¹⁶ Therefore for this study, polymers were designed to contain (diethylene glycol)methyl ether methacrylate (DEGMA) as the predominant comonomer, due to the fact that the LCST of the homopolymer is close to room temperature. To date, a number of DEGMA-based thermoresponsive (co)polymers have been synthesised for a variety of applications, with many studies focussing on modulating the transition temperature of polymers by incorporating a number of different comonomers, such as OEGMA.^17–19 Depending on the desired properties and applications, numerous reports exist on the synthesis of DEGMA and/or OEGMA-containing polymers as either statistical²⁰ or block^21–23 copolymers and/or grafted brushes.²⁴ Polymeric block structures are known to promote the self-assembly of core–shell structures such as micelles, vesicles or polymersomes, and have been studied for numerous applications that exploit this property.^19,21 In addition to OEGMA monomers, the copolymerisation of DEGMA with 2-(dimethylamino)ethyl methacrylate (DMAEMA) has also been well studied.^2,21 Due to its hydrophilicity, homopolymers of this monomer also display intrinsic thermo- and pH-responsive properties, and many studies have investigated the application of DEGMA–DMAEMA copolymers for drug delivery applications.^2,21,23–25

Whilst OEGMA, DMAEMA and many other monomers have been copolymerised with DEGMA, it is somewhat surprising to note that the copolymerisation of DEGMA and MAA via reversible addition–fragmentation chain-transfer (RAFT) polymerisation,^26–29 to the authors' knowledge, has not been reported; nor have the stimuli-responsive properties of these polymers been studied. The only reports of DEGMA–MAA are (i) Yamamoto et al.²⁴ and this polymerisation was achieved using tert-butyl methacrylate and atom-transfer radical polymerisation (ATRP); and (ii) Laloyaux et al.³⁰ where this polymer was grown from a silicon surface via ATRP. Furthermore, whilst MAA has been copolymerised via RAFT with other PEG-based monomers in the past,^31–34 DEGMA–MAA copolymers have not been reported.

Therefore, this study presents the synthesis of DEGMA–MAA statistical and block copolymers, and investigates the temperature, salt and pH-responsive properties of these polymers, and DEGMA–DMAEMA copolymers, in parallel. Few studies directly compare the stimuli-responsive properties of similar sets of statistical and block copolymers,³⁵ and as such we synthesised the polymers via RAFT polymerisation to ensure that the polymer compositions were well-defined, such that a general comparison could be completed. Proton (¹H) NMR spectroscopy was used to monitor polymerisation kinetics, provide an estimate of the various reactivity ratios, and to analyse the final composition of synthesised polymers. Thermo-, salt- and pH-responsive properties were determined via UV-Vis spectroscopy and dynamic light scattering (DLS), with self-assembled structures for the DEGMA–MAA copolymers visualised using cryogenic transmission electron microscopy (cryoTEM).

Experimental

Materials

(Diethylene glycol) methyl ether methacrylate (DEGMA), methacrylic acid (MAA), 2-(dimethylamino)ethyl methacrylate (DMAEMA) were purchased from Sigma-Aldrich. Each monomer was purified by running it through a short column packed with inhibitor removers (Sigma-Aldrich). All monomers were purified using hydroquinone monomethyl ether inhibitor remover, and DEGMA was further purified using tert-butylcatechol inhibitor remover. 4-Cyano-4-[(phenylthiocarbonyl)sulfanyl]pentanoic acid (4-CPDB) was obtained from Sigma-Aldrich and was used as received. The initiators 1,1′-azobis(cyclohexanecarbonitrile) (VAZO-88) and 2,2′-azobis(2-methylpropionitrile) (AIBN) were obtained from Wako Chemicals (Japan) and were recrystallised from methanol prior to use. All analytical grade solvents were purchased from Sigma-Aldrich or Merck and were used as received. NMR solvents were purchased from Cambridge Isotope Laboratories and used as received.

Instrumentation

Gel Permeation Chromatography (GPC) for all polymers was carried out on a Shimadzu system equipped with a CMB-20A controller system, an SIL-20A HT autosampler, an LC-20AT tandem pump system, a DGU-20A degasser unit, a CTO-20AC column oven, an RDI-10A refractive index detector, and four Waters Styragel columns (HT2, HT3, HT4, and HT5). Each column was 300 mm × 7.8 mm², providing an effective molar mass range of 100 to 4 × 10⁶. N,N-Dimethylacetamide (DMAc; with 4.34 g L⁻¹ of lithium bromide (LiBr)) was used as an eluent with a flow rate of 1 mL min⁻¹ at 80 °C. Number average (M_n) and weight average (M_w) molar masses and dispersity (Đ) values were evaluated using Shimadzu LC Solution software. The GPC columns were calibrated with low dispersity poly(methyl methacrylate) (PMMA) standards purchased from Polymer Laboratories. Reported polymer molar masses are relative to these PMMA standards. A 3rd-order polynomial was used to fit the log M_p vs. time calibration curve, which was near linear across the molar mass ranges. This GPC system was chosen for molecular weight analysis due to the ability of DMAc to solubilise all reaction components and purified products, and the compatibility of the column setup and subsequent analysis with the charged nature of the polymers. Proton nuclear magnetic resonance (¹H NMR) spectra were recorded on a Bruker BioSpin 400 MHz NMR spectrometer (Billerica, MA) at 298 K and using a 10 s delay between acquisitions. This was found to be a sufficient relaxation time to allow accurate end group analysis for these polymers. Data was analysed using TopSpin software. Chemical shifts (δ_H) are reported in parts per million (ppm) and referenced to residual solvent signals (CH₂Cl₂: δ 5.32 ppm). Deuterated dichloromethane (CD₂Cl₂) was used for ¹H NMR characterisation as the solvent signal did not overlap with the proton peaks of the benzoate group, allowing end group analysis to be conducted on all polymers. In situ NMR experiments were performed on a Bruker BioSpin 500 MHz with a 5 mm inverse-detection ¹H–¹³C–¹⁵N PATCI auto-tuning cryoprobe equipped with z-gradient. Samples were loaded into the spectrometer at 298 K and heated to 368 K for 24 h to allow for high conversion. Deuterated DMF (DMF-d₇) was used for the in situ experiments to match the solvent used in the batch polymerisation experiments (DMF). Spectra were acquired every 30 minutes during the time course of the reaction and were processed using TopSpin software. All spectra were referenced to the residual solvent signal at DMF-d₇: δ 2.75 ppm. Reactivity ratios were calculated via non-linear least square fitting of equations as described in the ESI† using Microsoft Excel software. Dynamic light scattering (DLS) measurements were obtained on a Malvern Zetasizer NanoZS operating with a laser beam at 633 nm and a scattering angle of 173°. Samples were dissolved in Milli-Q water and filtered through a 0.2 μM PTFE filter and allowed to equilibrate at room temperature for 24 h. Samples were autosampled using a Malvern Zetasizer NanoSampler with an attached quartz flow cell (ZEN0023). Samples were heated (and then cooled) from 5 to 60 °C in 5 °C intervals, with 120 s equilibration time at each temperature, and measured six times. The intensity and the volume distribution of the particle sizes were calculated applying the NNLS mode. Lower critical solution temperature (LCST) measurements were performed by heating a solution of the polymer (water or buffer, as specified) at different concentrations (1.0, 2.5, 5.0, 10.0 mg mL⁻¹) from 5 to 70 °C, with a heating/cooling rate of 1.0 °C min⁻¹, and 30 second equilibration time. Three heating/cooling cycles were conducted, with the reported LCST values calculated from absorbance measurements obtained in the second cycle. Absorbance (A) values were converted to percent transmittance (% T) according to eqn (1):

% T = 100(10^−A) (1)

Absorbance of the solutions during these cycles was recorded using a Shimadzu UV-1650 UV/Vis spectrometer with an 8-micro multi-cell temperature controlled cuvette chamber (path length 10 mm), operated using a Shimadzu temperature controller and Tm analysis software. Reported LCST values correspond to the temperature at which 50% transmittance of the solution occurs. Cryogenic transmission electron microscopy (cryoTEM) measurements were performed on a Tecnai 12 Transmission Electron Microscope (FEI, Eindhoven, The Netherlands) at an operating voltage of 120 kV. Images were recorded using either a Megaview III CCD camera with AnalySIS camera control software (Olympus), or a FEI Eagle 4k × 4k CCD camera. Measurements were made using ImageJ software. The samples for cryoTEM investigations were dissolved in Milli-Q water (resistivity of 18.0 MΩ cm⁻¹) at a concentration of 5 mg mL⁻¹, filtered (0.22 μm PTFE filter) and stored at 4 °C for 24 h prior to conducting the measurements. Each sample was preheated for at least 30 min in a heating block at a temperature 5 °C above the LCST of the polymer recorded at 5 mg mL⁻¹ (see Fig. 4 or Table 3). Using a preheated tip attached to a micropipette, 5 μL of the polymer solution was then placed on carbon-coated 300-mesh copper grids that had been glow-discharged to render the carbon film hydrophilic. The samples were deposited, and then rapidly blotted and plunged into a liquid ethane reservoir. The samples were then transferred to a liquid nitrogen reservoir and stored at −176 °C to avoid the formation of crystalline ice layers.

General procedure for the preparation of statistical copolymers via RAFT

DEGMA and either MAA or DMAEMA, 4-CPDB, VAZO-88 and DMF were added to a Schlenk flask and oxygen was removed via three freeze–pump–thaw cycles. The flask was then placed in an oil bath and heated at 90 °C for 4 h. The resultant polymers were isolated by precipitation from cold pentane three times and dried under high vacuum to remove residual solvent, producing 1 or 2, respectively. Polymer conversions were calculated from ¹H NMR spectra by comparison of the residual vinylic protons to newly formed polymer peaks present in the spectra, as described in the Results and discussion section.

General procedure for the preparation of block copolymers via RAFT

DEGMA, 4-CPDB, VAZO-88 and DMF were added to a Schlenk flask and oxygen was removed via three freeze–pump–thaw cycles. The flask was then placed in an oil bath and heated at 90 °C for 4 h. The resultant polymer was isolated by precipitation from cold pentane three times and dried under high vacuum to remove residual solvent, producing 3.

The macroRAFT-CTA 3 was then chain extended in the presence of MAA or DMAEMA under the above conditions for 20 h. The resultant polymers were isolated by precipitation from cold pentane three times and dried under high vacuum to remove residual solvent, producing 4 or 5, respectively. Polymer conversions were calculated from ¹H NMR spectra by comparison of the residual vinylic protons to newly formed polymer peaks present in the spectra, as described in the Results and discussion section.

General procedure for in situ ¹H NMR polymerisations

A portion of a stock solution containing DEGMA, 4-CPDB, AIBN and either MAA or DMAEMA, was mixed with deuterated N,N-dimethylformamide (DMF-d₇) in a gas-tight Schlenk NMR tube. The vessel was degassed using three freeze–pump–thaw cycles, before the sample was loaded into the NMR spectrophotometer. A ¹H spectra was obtained at 298 K (25 °C) before the sample was removed and the cryoprobe heated to 348 K (75 °C). The sample was reloaded into the spectrophotometer, and spectra were recorded at 30 minute intervals over the course of 24 h. Serial processing of the ¹H spectra was done using the TopSpin software.

Full experimental details of all polymerisations can be found in the ESI.†

Results and discussion

Batch synthesis of the polymer sets

Polymers 1–5 were synthesised via RAFT polymerisation using 4-cyano-4-[(phenylthiocarbonyl)sulfanyl]pentanoic acid (4-CPDB) as the chain transfer agent and VAZO-88 as the initiator, in DMF at 90 °C (Scheme 1).

Scheme 1 (Top) Synthesis of statistical copolymers 1 and 2; and (bottom) synthesis of the P(DEGMA) macroRAFT-CTA 3, and corresponding block copolymers 4 and 5, via RAFT polymerisation.

Statistical copolymers 1 and 2 were synthesised with a starting monomer feed ratio of 0.75 : 0.25 DEGMA : MAA or DEGMA : DMAEMA, respectively. The corresponding block copolymers 4 and 5 were synthesised from the purified P(DEGMA) macroRAFT-CTA 3, which was chain extended with either MAA or DMAEMA. The monomer feed ratio of MAA or DMAEMA to the macroRAFT-CTA 3 was adjusted so that the corresponding block copolymers 4 and 5 contained a similar 0.75 : 0.25 ratio of monomer units in the final products, analogous to the compositions of the matching statistical copolymers 1 and 2.

Experimental results from NMR and GPC analysis of the pure polymers is shown in Table 1. GPC analysis showed the molar mass distributions to be unimodal for all polymer except for the block copolymer 5 which had a slight high molar mass shoulder (see ESI†). Furthermore, all polymers had dispersity (Đ) values of less than 1.23 which is indicative of a living polymerisation process.

Table 1 Experimental results and analysis of polymers 1–5 by ¹H NMR and GPC

Code	Polymer	Mn,theo (g mol^−1)	Conv.^a (%)	Mn,NMR (g mol^−1)	DPNMR	M1 : M2 (%)^b	Mn^c	Mw^c	Đ^c
a Percent (%) overall conversion of monomer to polymer as calculated from ^1H NMR.b Calculated from the relative integration of side-group and backbone signals in the ^1H NMR spectra of purified polymers.c From GPC (DMAc). Mn and Mw are relative to PMMA standards.
1	P(DEGMA-stat-MAA)	13 800	67	12 900	77	79 : 21	11 800	13 100	1.11
2	P(DEGMA-stat-DMAEMA)	14 000	61	17 200	96	74 : 26	21 200	26 000	1.23
3	P(DEGMA)	10 900	57	12 900	67	100 : 0	10 900	12 000	1.10
4	P(DEGMA-b-MAA)	15 700	73	15 300	95	71 : 29	11 300	13 200	1.16
5	P(DEGMA-b-DMAEMA)	19 400	91	18 700	104	64 : 36	17 100	20 500	1.20

---

The number-average molar mass (M_n,NMR) of polymers 1–5 (Table 1) were all determined via the integration of the five phenyl–CH proton signals corresponding to the aromatic Z-group from 4-CPDB, which appeared at 7.0–8.5 ppm, and the –CO–O–CH₂– proton signals of the DEGMA (and DMAEMA in the case of 2 and 5) side chains that appeared between 3.9 and 4.4 ppm. Small differences in the peak shapes for the phenyl protons were observed between each sample (see ESI†) which arose from the interaction of the polymer chains with each other in solution, as well as the monomer closest to the dithiobenzoate moiety, which differed between each of the polymers.

Calculated M_n,NMR values were mostly in agreement with theoretical values (M_n,theo; see Table 1). In the case of 2 however, the theoretical value (M_n,theo) was lower than the M_n,NMR value, which indicated that the overall conversion was slightly higher than that observed in the ¹H NMR spectra of the crude reaction mixture (calculated to be 61%).

Overall, the ¹H NMR spectra of all polymers, as shown in the ESI,† display typical peaks corresponding to the backbone and different side chain groups present. In particular, the spectra of polymers 2 and 5 contained signals identical to those previously reported for DEGMA–DMAEMA copolymers.²¹

The mole ratio of monomers in the final polymer (% M₁ : M₂; see Table 1) was additionally determined by ¹H NMR. To calculate the relative amount of DEGMA in 1 and 4, a signal corresponding to a portion of the side chain of DEGMA (–CO–O–CH₂–) was integrated, and compared to the integral corresponding to the backbone –CH₃ groups. By calculating the integral corresponding to the proportion of –CH₃ groups from the DEGMA monomers, and subtracting this from the integral of the total backbone –CH₃ groups, the amount of MAA in each copolymer was determined. By comparing these sets of integrals it was found that the ratio of DEGMA to MAA in 1 was 79 : 21, and 71 : 29 in 4. A similar analysis was conducted on the polymers 2 and 5. Based on the ratio of integrals of relevant side chain signals (–CO–O–CH₂– for DEGMA and –N–CH₂– for DMAEMA), the % M₁ : M₂ was found to be 74 : 26 and 64 : 36 (DEGMA : DMAEMA) for 2 and 5, respectively.

In situ ¹H NMR experiments – polymerisation kinetics

In parallel to the batch reactions, in situ ¹H NMR polymerisations were also conducted to monitor polymerisation kinetics. Individual and total monomer conversions for DEGMA, MAA and DMAEMA over time are shown in Fig. 1. Since the in situ ¹H NMR reaction conditions varied slightly from the batch polymerisation conditions, the resulting polymers are referred to as 1* P(DEGMA-stat-MAA) and 2* P(DEGMA-stat-DMAEMA) in this section. All conversions were calculated from integration of the vinylic region of the spectra with respect to polymer formed, and were 87% for 1* and 95% for 2* after 24 hours (total monomer conversions). From Fig. 1, it is clear that whilst DEGMA and DMAEMA polymerise at nearly identical rates (Fig. 1, right), the rate of MAA polymerisation is slightly slower than that of DEGMA (Fig. 1, left). This figure clearly shows that the rate of MAA polymerisation is similar to DEGMA in the first hour of the reaction, and that this rate decreases significantly over the course of the 24 hour polymerisation.

Fig. 1 Individual (red/blue data points) and total (black line) monomer conversions (%) over time (h) for the polymerisation of P(DEGMA-stat-MAA) (1*) (left) and P(DEGMA-stat-DMAEMA) (2*) (right) at 75 °C, in deuterated DMF (DMF-d₇), followed by ¹H NMR spectroscopy.

The mole fraction of each monomer during the course of the two polymerisations is shown in Fig. 2. In both cases, the mole fraction of monomers in the starting feed was 0.75 : 0.25 (DEGMA : MAA or DMAEMA), and this ratio remained constant during the synthesis of 2*. This result is consistent with the mole fraction of monomers observed in 2 (0.76 : 0.24), obtained during the batch polymerisation of DEGMA with DMAEMA.

Fig. 2 Mole fraction of monomer as a function of total monomer conversion (%) during the synthesis of 1* and 2* as measured by ¹H NMR spectroscopy.

In the case of 1*, significant composition drift is observed during the polymerisation. Fig. 2 clearly shows that the feed ratio deviated significantly from the initial 0.75 : 0.25 ratio (DEGMA : MAA) over the course of the polymerisation due to the relatively slow polymerisation rate of MAA. Again, this result is consistent with the observed mole fraction of monomers obtained from the batch polymerisation to synthesise P(DEGMA-stat-MAA) (1), which was 0.79 : 0.21 (DEGMA : MAA) following a 4 hour reaction time.

This result, in combination with the observed difference in polymerisation rate suggests that the reactivity ratios between DEGMA and MAA differ significantly. In turn, this means that the resultant copolymer may not be truly statistical, which is important to know when analysing and interpreting results from studies on the solution properties of these polymers. Therefore, monomer reactivity ratios were estimated from the above data and were found to be r₁ = 1.46 and r₂ = 0.54 for DEGMA : MAA, respectively, and r₁ = 1.02 and r₂ = 0.98 for DEGMA : DMAEMA, respectively. Further details on how these values were calculated can be found in the ESI.† In the case of DEGMA : DMAEMA, the reactivity ratios are both approximately equal to 1, which indicates that each of the monomers does not preference either itself or another (propagating) monomer and the polymers of these two monomers should have a random structure. This is not the case with DEGMA : MAA however, as r₁ > 1 > r₂ which indicates a gradient, or more block-like structure in the polymer chain. Whilst these reactivity ratios have only been estimated for polymers 1* and 2*, for the purposes of this study the general outcomes can still be applied to batch synthesised copolymers 1 and 2. Therefore, the following sections discussing the stimuli-responsive properties of these polymers will be done in light of the fact that the composition of 1 is most likely a gradient copolymer, whilst 2 is most likely a statistical copolymer.

Thermoresponsive properties of the polymer sets

The turbidity of all polymers (1–5) in solution was studied using UV-Vis spectroscopy (500 nm) to determine the temperature at which the polymer chains transition from a coil to globule state in solution. Polymers 1–5 were dissolved in Milli-Q grade water at a concentration of 1.0 mg mL⁻¹ and were analysed to determine the absorbance of solutions between 5 and 80 °C, at a heating rate of 1.0 °C min⁻¹. Three heating and cooling cycles were performed on each sample, and the second heating cycle was analysed. Absorbance values were then converted to percent transmittance (% T) values, and the reported LCST values are the temperature at which 50% normalised percent transmittance occurs.

The heating curves of polymers 1–5 as function of temperature are shown in Fig. 3. The thermal transitions of all polymers were reversible, with a hysteresis and LCST variation of approximately 1 °C observed between the heating and cooling cycles in all samples.

Fig. 3 Temperature dependence of the transmittance (normalised, % T) of 1 mg mL⁻¹ solutions of polymers 1–5 in Milli-Q water. Data points are coloured according to the polymer composition: black: P(DEGMA) (3); red: P(DEGMA–MAA) (1 and 4), blue: P(DEGMA–DMAEMA) (2 and 5). Data points are shaped according to polymer composition: circle: statistical copolymer; square: macroRAFT-CTA and block copolymers.

It is evident from Fig. 3 that the LCSTs of the MAA containing copolymers 1 (23 °C) and 4 (28 °C) were lower than the LCSTs of the DMAEMA-containing copolymers 2 (34 °C) and 5 (45 °C). This effect is due to the greater hydrophobicity of the methacrylic acid side chains in unbuffered water that causes a decrease in the number of hydrogen bonds that need to break between water and the copolymer. As a result, the coil to globule transition, observed as the LCST, is shifted to a lower temperature. Likewise, DMAEMA is a known hydrophilic monomer, and thus its incorporation in the polymer chain is the reason that 2 and 5 have higher LCSTs overall. Falling between these two sets of LCST transitions is the DEGMA homopolymer 3, which exhibits a phase transition at 31 °C. This value is slightly higher than the average LCST of 26 °C which has been previously reported for similar polymers at the same concentration in water.^{13,17–19,36–39} Of these reported polymers previously synthesised via RAFT, two were approximately the same size polymer, with similar dispersities; and the LCSTs were measured at the same concentration in water.^17,18 However, both of these polymers used different RAFT agents during their synthesis, namely 2-cyano-2-propyl benzodithioate and benzyl 4-cyano-4-(ethylthiocarbonothioylthio)pentanoate, which gave two different end groups that were more hydrophobic than the 4-cyano pentanoic acid end group in polymer 3.^17,18 The effect of the polymer end group on the LCST has been previously reported and is the most likely reason for the 5 °C difference in LCST observed for this DEGMA homopolymer.^15,40

Another interesting observation from Fig. 3 is that the LCSTs of the statistical copolymers are both slightly lower than the LCSTs of the corresponding block copolymers, and that this difference is smaller for the MAA-containing copolymers 1 and 4 (5 °C) compared to the DMAEMA containing copolymers 2 and 5 (11 °C). This effect is attributed to the differences in structure (block versus statistical/gradient) and thus the varying distribution of the charged groups along the polymer chain, rather than the slight differences in chemical chain composition (i.e. % M₁ : M₂), as discussed below. In addition, given that the influence of chain length on the phase transition of DEGMA-based copolymers is known to be negligible,¹³ we do not think that the slight differences in chain length are playing a role in these LCST modulations.

From Table 1, the relative amounts of comonomers for 1 and 4 are similar, with the amount of MAA slightly higher in the block copolymer 4 (29%) than it is in the gradient copolymer 1 (21%). Given that MAA imparts a hydrophobicity to the polymer chain, one would expect the polymer with the higher amount of MAA to have a lower LCST. However, what is observed is actually a 5 °C increase in the LCST of the block copolymer compared with that of the gradient copolymer. A similar trend is also observed in the case of the basic copolymers, where the ratio of DEGMA : DMAEMA is 74 : 26 for the statistical copolymer 2 and 64 : 36 for the block copolymer 5. Since DMAEMA is a hydrophilic monomer, the LCST of the block copolymer would be expected to be slightly higher than the statistical copolymer, given the additional 10% of DMAEMA units in the polymer chain of 5. What is observed is an 11 °C difference in LCST, which arises from both the difference in copolymer chain structure and mole ratio of monomers between the two polymers.

The final trend that was observed from this data is that the statistical copolymers 1 and 2 and the macroRAFT-CTA 3 all displayed much sharper phase transitions (between 7 and 10 °C) than 4 and 5, where transitions occurred over a temperature window of 20–25 °C. This is most likely because polymers 1–3 display more uniform hydrogen bonding with water and therefore display one sharp transition. In the case of the block copolymers 4 and 5 however, they are more likely to form micelles and aggregate into self-assembled structures (globules) which would happen more slowly, leading to an observed transition that is less sharp.

Concentration and salt effects on the thermoresponsive properties

LCSTs of polymer solutions as a function of concentration were also studied with the results shown in Fig. 4. Cloud points of polymer samples 1–5 in water at various concentrations (1.0, 2.5, 5.0 and 10.0 mg mL⁻¹) were again determined by UV-Vis spectroscopy. For each polymer, the LCST decreased as the concentration increased, with this trend generally plateauing at concentrations above 5 mg mL⁻¹. The main reason for the observed decrease in LCST as the concentration is increased, is the fact that at higher concentrations the polymer chains are more likely to aggregate as there are more polymer chains in the same solution volume, and the polymer chains may experience less water solvation overall. Therefore, these conditions ultimately result in a polymer transitioning at a lower temperature.⁴¹ This result is in good agreement with previous studies that have identified similar links between LCSTs and solution concentration of different thermoresponsive polymers.^13,17,19,41

Fig. 4 LCST (°C) of polymers 1–5 in water as function of concentration (mg mL⁻¹) in Milli-Q grade water. Data points are coloured according to the polymer composition: black: P(DEGMA) (3); red: P(DEGMA–MAA) (1 and 4), blue: P(DEGMA–DMAEMA) (2 and 5). Data points are shaped according to polymer composition: circle: statistical copolymer; square: macroRAFT-CTA and block copolymers. NB: polymers 3 and 4 showed same LCST at 2.5 mg mL⁻¹.

Polymers 1–5 also exhibited a decrease in their LCST as the salt concentration increased, as shown in Table 2. This ‘salting out’ effect is well known and has been studied in detail for NIPAM polymers^42,43 though this effect is less pronounced for OEGMA-type polymers.^13,18 This effect occurs due to the increase in surface tension at the polymer/water interface causing partial dehydration of the polymer chains in the presence of salts, such as sodium chloride (NaCl). As a consequence of this, less hydrogen bonds are formed between the polymers and water, and the observed result is a decrease in the LCST of the polymer.

Table 2 The LCSTs of 1.0 mg mL⁻¹ solutions of 1–5 in either Milli-Q grade water, or 1.0 M NaCl solution

NaCl (M)	Polymer LCST (°C)
	1	2	3	4	5
0.0	23	34	31	28	45
1.0	14	24	17	19	25

---

pH-Responsive properties of the polymer sets

The pH responsive properties of these polymers were also observed via turbidity, as shown in Fig. 5a–c. Polymers 1–5 were dissolved in different buffer solutions (1 mg mL⁻¹) that were pH adjusted to 3.0, 7.0 and 10.8. The MAA-containing copolymers 1 and 4 only displayed thermoresponsivity at pH 3, with no LCSTs observed at higher pH (Fig. 5a). In both cases, the LCST of 1 and 4 decreased by 3–5 °C from the LCST obtained in unbuffered water where the pH was measured to be 4.1 and 4.0, respectively. This decrease in LCST is most likely a combination of a slight increase in hydrophobicity of the protonated carboxylic acid pendant moieties and carboxylate of the RAFT end group (R-group) as the pH was reduced from ∼4 to pH 3 (thus the pH < pK_a of these acid groups), combined with the increased ionic strength of the buffered solution, which causes a reduction in the solvation of the polymer chain in solution and thus the LCST. Whilst the MAA-copolymers were additionally tested for thermoresponsivity in buffers at pH 7 and 10.8, the polymers did not elicit an LCST (they remained soluble throughout the entire temperature cycles), which has been previously observed in other studies on different MAA-containing thermoresponsive polymers.²⁴

Fig. 5 pH-responsive properties of polymers 1–5 in different buffer solutions: (a) MAA-containing copolymers 1 and 4 at pH 3 (pink), as compared to unbuffered Milli-Q water (black); (b) P(DEGMA) macroRAFT-CTA 3 at pH 3 (pink) and 7 (green), as compared to unbuffered Milli-Q water (black); (c) DMAEMA-containing copolymers 2 and 5 at pH 7 (green) and 10.8 (blue), as compared to unbuffered Milli-Q water (black). Buffer solutions: pH 3: citric acid–sodium phosphate dibasic, pH adjusted with orthophosphoric acid; pH 7: sodium phosphate–sodium phosphate dibasic, pH adjusted with orthophosphoric acid; pH 10.8: sodium bicarbonate–sodium carbonate buffer, pH adjusted with 1.0 M NaOH.

A small shift (−2 °C) in LCST is also observed for the macroRAFT-CTA 3 in pH 3 buffer, as compared to its LCST in water (Fig. 5b). Furthermore, the LCST of this polymer shifts to a higher value (37 °C; +6 °C), as the pH of the solution was increased to 7. No LCST was observed for 3 at pH 10.8. The reason for these shifts may be to do with the protonation/deprotonation of the carboxylic acid end group of the polymer as the pH and ionic strength of the solution changes. Furthermore, it is possible that a very small proportion of the DEG side chains may have hydrolysed to carboxylic acids during the measurement of the LCST which would further exacerbate these LCST shifts. This is unlikely, however, given that these solutions were made fresh from dried polymers immediately prior to conducting LCST measurements.

In the case of the DMAEMA-containing copolymers (Fig. 5c), there is a clear pattern to the observed LCSTs as the pH of the solutions is changed. The LCST of the polymers at pH 10.8 is very close to the LCST of the polymer in unbuffered Milli-Q grade water where the measured pH was 8.8 and 8.4 for 2 and 5, respectively. This is due to the fact that the pK_a of the tertiary amine in the DMAEMA side chain is approximately pH 6,⁴⁴ and the pH of these solution are well above that. Again, the LCSTs of the statistical copolymers are much lower than the LCSTs of the corresponding block copolymers. As expected, decreasing the pH to 7 causes the polymers to become more hydrophilic (as pH < pK_a for the DMAEMA tertiary amine), and the higher LCSTs observed are in agreement with reported values.^21,44

Investigation of polymeric self-assembly

The thermoresponsivity of all copolymers was further investigated by dynamic light scattering (DLS) and cryogenic transmission electron microscopy (cryoTEM; for DEGMA-MAA polymers only) in order to determine the size of the self-assembled aggregates above the polymer LCSTs. For analysis by DLS, copolymer samples in water (2.5 mg mL⁻¹) were analysed on a Malvern Zetasizer NanoS at temperatures well below the polymer LCST (between 5 and 15 °C) and at every subsequent temperature (in 5 °C intervals) up to 60 °C (above the LCST of all polymers). One heating and one cooling cycle was performed on each sample. Fig. 6 shows the volume distribution of particle sizes in each copolymer (1, 2, 4, and 5) at temperatures above and below the individual polymer LCSTs (see Table 3). Below 5–15 °C (temperature assigned based on LCST), all polymers displayed a unimodal peak of less than 10 nm, corresponding to fully soluble hydrated polymer chains in random coil conformations. An increase in the temperature of each solution to a value higher than the polymer LCST then caused collapse of the copolymer chains into globular states, and subsequent aggregation into larger particles as evidenced by an even bimodal distribution of peaks (peak 1: ∼10 nm; and peak 2: ∼600–700 nm on average) during this transition period (data not shown). This bimodal distribution transitioned fully to a unimodal distribution of the larger peak upon further increases in temperature, with the exception of 4. Interestingly, this block copolymer exhibited two phase transitions: one at approximately 35 °C (∼200 nm particles) and a second at approximately 50 °C (∼500 nm particles; Fig. 6). This clearly demonstrates heat-induced process of polymer globule assembly into larger aggregate structures in solution.

Fig. 6 DLS measurements at selected specified temperatures of (left) 1 P(DEGMA-stat-MAA) and 4 P(DEGMA-b-MAA), and (right) 2 P(DEGMA-stat-DMAEMA) and 5 P(DEGMA-b-DMAEMA). Polymer solutions were 2.5 mg mL⁻¹ in water, with measurements collected at 5 °C temperature intervals, over a 5–60 °C range. Specific values reported here correspond to temperatures below and above the specific polymer LCSTs for 2.5 mg mL⁻¹ solutions, as shown in Table 3. The data shown is only a small selection of the DLS data collected at 5 °C intervals between 5 and 60 °C.

Table 3 LCSTs and transition windows of 2.5 and 5.0 mg mL⁻¹ solutions of copolymers 1, 2, 4, and 5 in Milli-Q water as measured by UV-Vis spectroscopy and DLS, respectively

Polymer	UV-Vis^a	UV-Vis^a	DLS^b
	5.0 mg mL^−1	2.5 mg mL^−1	2.5 mg mL^−1
a 5–70 °C range, 1 °C min^−1, 3 heating/cooling cycles.b 15–60 °C range for 1, 2, and 5 and 5–60 °C range for 4, 5 °C min^−1, 1 heating/cooling cycle.
1	17	20	30–35
2	29	31	25–30
4	22	27	10–15 & 35–40
5	33	40	25–30

---

It is interesting to note from Fig. 6 that in both sets of polymers, the statistical and block copolymers appear to form spherical polymer aggregates in solution above the polymer LCST. Self-assembly for block copolymers is a well characterised phenomenon, with similar DEGMA–DMAEMA block copolymers previously shown to form multi- and unilamellar vesicles.²¹ For random copolymers however, whilst examples of flower-like micelle formation have been reported for statistical copolymers,^45,46 very few systematic studies have directly compared the structures obtained from statistical and block copolymers of the same two monomer pairs.⁴⁷ Therefore, to investigate further, the self-assembled structures of the gradient and block copolymers of DEGMA–MAA were both visualised and compared through the use of cryogenic transmission electron microscopy (cryoTEM). Polymer solutions at 5 mg mL⁻¹ were prepared and heated to temperatures 5 °C above each polymer's LCST and allowed to equilibrate for at least 30 minutes before preparation of the specific microscopy sample. The images obtained from cryoTEM of P(DEGMA-stat-MAA) 1 and P(DEGMA-b-MAA) 4 are shown in Fig. 7 (left and right, respectively), alongside histograms (inset) showing the frequency of aggregate sizes in each cryoTEM sample.

Fig. 7 CryoTEM images of 1 (left) and 4 (right) solutions in Milli-Q water (5.0 mg mL⁻¹, preheated) showing the formation of polymer aggregates. (Inset) Histograms showing the frequency of aggregate sizes in each cryoTEM sample.

For P(DEGMA-stat-MAA) 1, small aggregates of approximately 40–70 nm were visualised using cryoTEM (Fig. 7, left image), and these were much smaller than those observed via DLS. The most likely reason for this is due to the unconventional method used to obtain this image. During visualisation of this sample, it was apparent that the refractive index of the aggregates of 1 was about the same as that of the thin ice sheets within the sample grid network. Therefore, in order to visualise the polymer aggregates, the electron beam had to be focussed on particular points in the lacy carbon grid in order to disrupt the crystalline ice layers and allow the polymer aggregates to be imaged. Given that this technique is non-standard and destructive to the sample, it is not known what effect this is having on the aggregate structures and whether or not the image is actually representative of the actual aggregate sizes and structures present in the sample.

For P(DEGMA-b-MAA) 4 however, the method of obtaining the image for this sample (Fig. 7, right) was much more straightforward. Large particles of approximately 350–450 nm were observed for this sample, which matched particle sizes observed in DLS analysis.

Whilst initial cryoTEM experiments were undertaken on the same samples that were used for DLS (i.e. at 2.5 mg mL⁻¹), no aggregates could be visualised due to the low concentration (and thus low frequency) of aggregates in solution. Therefore, samples at a concentration of 5.0 mg mL⁻¹ had to be used in this instance, and in both cases, spherical particles with good dispersion and an array of particle sizes was observed, as summarised in the inset histograms in Fig. 7.

Overall, approximate size and shape of the polymer aggregates has been determined via a combination of DLS and cryoTEM, and work is ongoing to determine whether polymers 1 and 4 are forming ordered self-assembled structures such as micelles or vesicles. Given that both polymers display aggregation above their LCSTs, it would be interesting to know more information about the integrity of these structures and how they differ between the statistical and block copolymers. Therefore this data, in combination with the above results, would have significant implications for researchers interested in using these particular stimuli-responsive polymers in the future.

Conclusion

The synthesis and stimuli-responsive properties of DEGMA–MAA statistical and block copolymers is reported for the first time, and these results were compared to an equivalent set of DEGMA–DMAEMA copolymers. RAFT polymerisation was used for the synthesis of all polymers, which were then characterised via ¹H NMR and GPC. The stimuli-responsive properties of these polymer sets were examined via UV-Vis spectroscopy and dynamic light scattering under various solution conditions, including different polymer concentrations, sodium chloride concentrations, and solution pHs. Small differences were observed between the LCSTs of the statistical and block copolymers in water, with the statistical copolymers generally transitioning at lower temperatures than corresponding block copolymers. All polymers exhibited concentration dependence of the LCST in water, and displayed both pH and salt sensitivities. Self-assembly of each polymer was investigated and minor differences were observed between the particles sizes of statistical and block copolymers as observed via DLS. Images obtained via electron microscopy for P(DEGMA–MAA) gradient and block copolymers showed that both polymers self-assembled to form medium to large aggregates above the individual polymer's LCST.

Acknowledgements

The authors wish to acknowledge funding from the CSIRO Office of the Chief Executive (OCE). Many thanks to Jo Cosgriff, Roger Mulder and Katherine Locock for their help with NMR polymerisations and data analysis. Thank you also to Adam Costin and Georg Ramm (Monash University) and Lynne Waddington, Mark Greaves, and Chi Huynh (CSIRO) for their help with cryogenic electron microscopy.

References

1. I. N. Savina, I. Y. Galaev and S. V. Mikhalovsky, in Smart Polymers and their Applications, ed. M. R. Aguilar and J. S. Román, Woodhead Publishing, 2014, pp. 408–436,  DOI:10.1533/9780857097026.2.408.
2. M. A. Kostiainen, C. Pietsch, R. Hoogenboom, R. J. M. Nolte and J. J. L. M. Cornelissen, Adv. Funct. Mater., 2011, 21, 2012–2019.
3. Y. Zhao, L.-P. Lv, S. Jiang, K. Landfester and D. Crespy, Polym. Chem., 2015, 6, 4197–4205.
4. L. Ionov, Adv. Funct. Mater., 2013, 23, 4555–4570.
5. X. Li, Y. Gao and M. J. Serpe, Macromol. Rapid Commun., 2015, 36, 1382–1392.
6. E. Ayano and H. Kanazawa, Anal. Sci., 2014, 30, 167–173.
7. Z. Tang, Y. Akiyama and T. Okano, J. Polym. Sci., Part B: Polym. Phys., 2014, 52, 917–926.
8. V. Aseyev, H. Tenhu and F. Winnik, in Self Organized Nanostructures of Amphiphilic Block Copolymers II, ed. A. H. E. Müller and O. Borisov, Springer, Berlin, Heidelberg, 2011, vol. 242, ch. 57, pp. 29–89.
9. Y. Guan and Y. Zhang, Soft Matter, 2011, 7, 6375–6384.
10. J. Ramos, A. Imaz and J. Forcada, Polym. Chem., 2012, 3, 852–856.
11. M. A. Cooperstein and H. E. Canavan, Langmuir, 2010, 26, 7695–7707.
12. R. Hoogenboom, in Complex Macromolecular Architectures, John Wiley & Sons (Asia) Pte Ltd, 2011, pp. 685–715,  DOI:10.1002/9780470825150.ch22.
13. J.-F. Lutz, Ö. Akdemir and A. Hoth, J. Am. Chem. Soc., 2006, 128, 13046–13047.
14. D. Roy, W. L. A. Brooks and B. S. Sumerlin, Chem. Soc. Rev., 2013, 42, 7214–7243.
15. J. F. Lutz, J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 3459–3470.
16. Z. Hu, T. Cai and C. Chi, Soft Matter, 2010, 6, 2115–2123.
17. K. Bebis, M. W. Jones, D. M. Haddleton and M. I. Gibson, Polym. Chem., 2011, 2, 975–982.
18. M. Luzon, C. Boyer, C. Peinado, T. Corrales, M. Whittaker, L. Tao and T. P. Davis, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 2783–2792.
19. J.-F. Lutz and A. Hoth, Macromolecules, 2006, 39, 893–896.
20. T. Ishizone, A. Seki, M. Hagiwara, S. Han, H. Yokoyama, A. Oyane, A. Deffieux and S. Carlotti, Macromolecules, 2008, 41, 2963–2967.
21. C. Pietsch, U. Mansfeld, C. Guerrero-Sanchez, S. Hoeppener, A. Vollrath, M. Wagner, R. Hoogenboom, S. Saubern, S. H. Thang, C. R. Becer, J. Chiefari and U. S. Schubert, Macromolecules, 2012, 45, 9292–9302.
22. S. T. Hemp, A. E. Smith, W. C. Bunyard, M. H. Rubinstein and T. E. Long, Polymer, 2014, 55, 2325–2331.
23. A. Schmalz, H. Schmalz and A. H. E. Muller, Soft Matter, 2012, 8, 9436–9445.
24. S.-I. Yamamoto, J. Pietrasik and K. Matyjaszewski, Macromolecules, 2008, 41, 7013–7020.
25. D. Fournier, R. Hoogenboom, H. M. L. Thijs, R. M. Paulus and U. S. Schubert, Macromolecules, 2007, 40, 915–920.
26. J. Chiefari, Y. K. Chong, F. Ercole, J. Krstina, J. Jeffery, T. P. T. Le, R. T. A. Mayadunne, G. F. Meijs, C. L. Moad, G. Moad, E. Rizzardo and S. H. Thang, Macromolecules, 1998, 31, 5559–5562.
27. Y. K. Chong, T. P. T. Le, G. Moad, E. Rizzardo and S. H. Thang, Macromolecules, 1999, 32, 2071–2074.
28. G. Moad, in Controlled Radical Polymerization: Mechanisms, ed. K. Matyjaszewski, B. S. Sumerlin, N. V. Tsarevsky and J. Chiefari, American Chemical Society, Washington, DC, 2015, ch. 12,  DOI:10.1021/bk-2015-1187.ch012.
29. C. Boyer, M. H. Stenzel and T. P. Davis, J. Polym. Sci., Part A: Polym. Chem., 2011, 49, 551–595.
30. X. Laloyaux, B. Mathy, B. Nysten and A. M. Jonas, Macromolecules, 2010, 43, 7744–7751.
31. J. A. Jones, N. Novo, K. Flagler, C. D. Pagnucco, S. Carew, C. Cheong, X. Z. Kong, N. A. D. Burke and H. D. H. Stöver, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 6095–6104.
32. C. R. Becer, S. Hahn, M. W. M. Fijten, H. M. L. Thijs, R. Hoogenboom and U. S. Schubert, J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 7138–7147.
33. Y. Y. Khine, Y. Jiang, A. Dag, H. Lu and M. H. Stenzel, Macromol. Biosci., 2015, 15, 1091–1104.
34. C. Tsitsilianis, G. Gotzamanis and Z. Iatridi, Eur. Polym. J., 2011, 47, 497–510.
35. K. Li and Y. Cao, Soft Mater., 2010, 8, 226–238.
36. S. Han, M. Hagiwara and T. Ishizone, Macromolecules, 2003, 36, 8312–8319.
37. Y. Maeda, T. Kubota and H. Yamauchi, Langmuir, 2007, 23, 11259–11265.
38. T. Cai, M. Marquez and Z. B. Hu, Langmuir, 2007, 23, 8663–8666.
39. F. J. Hua, X. G. Jiang and B. Zhao, Macromolecules, 2006, 39, 3476–3479.
40. P. J. Roth, F. D. Jochum, F. R. Forst, R. Zentel and P. Theato, Macromolecules, 2010, 43, 4638–4645.
41. F. Hua, X. Jiang, D. Li and B. Zhao, J. Polym. Sci., Part A: Polym. Chem., 2006, 44, 2454–2467.
42. Y. Zhang, S. Furyk, D. E. Bergbreiter and P. S. Cremer, J. Am. Chem. Soc., 2005, 127, 14505–14510.
43. Y. J. Zhang and P. S. Cremer, Curr. Opin. Chem. Biol., 2006, 10, 658–663.
44. M. Wagner, C. Pietsch, A. Kerth, A. Traeger and U. S. Schubert, J. Polym. Sci., Part A: Polym. Chem., 2015, 53, 924–935.
45. P. Guo, W. Guan, L. Liang and P. Yao, J. Colloid Interface Sci., 2008, 323, 229–234.
46. L. Li, K. Raghupathi, C. Song, P. Prasad and S. Thayumanavan, Chem. Commun., 2014, 50, 13417–13432.
47. Q. Guo, T. Zhang, J. An, Z. Wu, Y. Zhao, X. Dai, X. Zhang and C. Li, Biomacromolecules, 2015, 16, 3345–3356.

---

Footnote

† Electronic supplementary information (ESI) available: Detailed experimental section, reactivity ratio calculations, GPC traces and ¹H NMR spectra (including in situ NMR experiments) of all polymers, and DLS data for polymer 3. See DOI: 10.1039/c6ra14425j

---