Thermoresponsive Hyperbranched Polymers with ... - ACS Publications

Nov 23, 2017 - cloud point than the linear counterpart at the low polymer concentration, whereas their cloud points were similar at the high polymer c...
0 downloads 11 Views 2MB Size
Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

Thermoresponsive Hyperbranched Polymers with Spatially Isomerized Groups: NMR Implication to Their Thermoresponsive Behaviors Bin Wang,† Ting Xiao,‡ Xiao-Bin Fu,‡ Ting-Ting Jiang,‡ Yu Chen,*,†,§ and Ye-Feng Yao*,‡ †

Tianjin Key Laboratory of Molecular Optoelectronic Science, Department of Chemistry, School of Sciences, Tianjin University, and Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin 300354, P. R. China ‡ Department of Physics & Shanghai Key Laboratory of Magnetic Resonance, School of Physics and Materials Science, East China Normal University, North Zhongshan Road 3663, Shanghai 200062, P. R. China § Tianjin Engineering Technology Center of Chemical Wastewater Source Reduction and Recycling, School of Science, Tianjin Chengjian University, Tianjin 300384, P. R. China S Supporting Information *

ABSTRACT: The influence of stereochemical difference on the phase transition of thermoresponsive polymer was studied in detail in this work. To this end, we synthesized two thermoresponsive hyperbranched polymers having the almost same chemical composition but differing in the spatial distribution of the chemical groups. These samples exhibit remarkably different low critical solution temperatures (LCSTs). A detailed NMR study on the samples revealed that before the transition the chemical groups in the two samples have very different packing arrangements. A microscopic phase separation was realized in the polymer having the densely packed structure. The origin of the different LCSTs of the polymers was well explained by the entropy change due to the dehydrate/hydrate processes in the transition.



larger and more compact aggregate during the heating process.7 Moreover, it was found that cyclic PNIPAM exhibited a higher cloud point than the linear counterpart at the low polymer concentration, whereas their cloud points were similar at the high polymer concentration.5,6 Some chiral polymers were found to exhibit quite different thermoresponsive properties compared with the optically inactive one.11−13 Two acid-labile poly(methacrylamide)s with the pendant cyclic ortho ester moieties of trans and cis configurations were reported to exhibit different thermoresponsive behavior in water.14 Two polymers with different pyrrolidone-based side groups had been addressed to show different thermoresponsive phase transition behaviors due to the different steric hindrance effect of their side groups.15 NMR is a very powerful technique for the structure analysis of materials because it provides not only information about the chemical nature of the molecules but also unique information, such as the spatial information on the groups, the inter/ intramolecular interaction, etc. NMR has long been explored for studying the phase transition behavior of the traditional thermoresponsive linear polymers and their interaction with various additives, and valuable information has been

INTRODUCTION During the past decade, polymers with stimuli-responsive properties, such as fast and reversible conformational or phase changes in response to variations in temperature or pH, have gained much interest in many aspects.1,2 One of the most appealing stimuli-responsive species is the thermoresponsive hydrophilic polymer having the lower critical solution temperature (LCST) in aqueous solution, which means that above a specific temperature its solubility in water dramatically decreases.3 With respect to thermoresponsive polymers, LCST control is a very important issue in their design and practical applications. Normally, the LCSTs of thermoresponsive polymer can be modulated through varying the ratio of hydrophilic and hydrophobic units, and the more hydrophilic one shows a higher LCST than the more hydrophobic counterpart.3 Beyond the well-known microscopically hydrophilic/hydrophobic modulation method, topology alteration4−10 and the stereochemical effect11−15 of molecular parameters were also found to have influence on the phase transition behavior recently. For instance, Steinhauer et al. found that block and gradient copolymers of 2-hydroxyethyl acrylate and 2-methoxyethyl acrylate showed not only different phase transition temperature but also different hysteresis on heating and cooling.4 Cyclic poly(N-isopropylacrylamide) (PNIPAM) was reported to show two kinetic growth stages while linear PNIPAM only form © XXXX American Chemical Society

Received: August 31, 2017 Revised: November 23, 2017

A

DOI: 10.1021/acs.macromol.7b01887 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Scheme 1. Preparation of Isomers A and B

obtained.16−21 Recently, NMR interpretation of the structure and spatial information on the compact and globular dendritic polymers gradually attracted attention. Chai et al. studied the structure and conformation of DAB dendrimers in chloroform and benzene solutions via multidimensional NMR techniques.22 Holycross and Chai reported a detailed NMR study on the structures of hyperbranched polyethylenimine (HPEI) in chloroform solution.23 Our group used high-resolution 1H NMR to study the structure and dynamics of thermoresponsive HPEI derivatives in aqueous solution before and after the phase transition.24 Thermoresponsive polymers with different stereochemical structures have the same composition; however, the stereochemical difference can render certain polymers different thermoresponsive properties.11−15 To our best knowledge, what is the origin of the different thermoresponsive properties induced by the stereochemical difference is not clear at the molecular level.



1H, CH), 3.15 (t, 2H, CH2), 1.94 (s, 3H, CH3), 1.83 (m, 2H, CH2), 1.52 (m, 2H, CH2), 1.36 (m, 2H, CH2) (Figure S1 in the Supporting Information). Synthesis of 2-Amino-6-isobutyramidohexanoic Acid. Lysine·HCl (1.46 g, 7.99 mmol) was dissolved in 16.0 mL of NaHCO3 solution (1.00 M). CuSO4·5H2O (1.05 g) in 8.00 mL of water was added, and the mixture turned dark blue. After the further addition of 1.00 g (11.9 mmol) of NaHCO3 and 3.00 mL of acetone, isobutyric anhydride (1.42 g, 8.98 mmol) was added dropwise into the solution and stirred for 24 h. The blue slurry was filtered, and the residue was dried under vacuum, yielding the copper complex as a blue solid. The copper complex was suspended in 50 mL of water, and then 8hydroxyquinoline (1.20 g, 8.27 mmol) was added. The suspension was stirred for 10 h, and the color turned from blue to yellow. After filtration, the filtrate was washed three times by 50.0 mL of ethyl acetate. The aqueous phase was freeze-dried to give the product. Yield: 64%. 1H NMR (D2O, 500M, ppm): δ = 3.69 (t, 1H, CH), 3.15 (t, 2H, CH2), 2.44 (m, 1H, CH), 1.84 (m, 2H, CH2), 1.52 (m, 2H, CH2), 1.38 (m, 2H, CH2), 1.05 (d, 6H, CH3) (Figure S2). Synthesis of Methyl 6-Acetamido-2-isobutyramidohexanoate (Isomer A) and Methyl 2-Acetamido-6-isobutyramidohexanoate (Isomer B) (see Scheme 1). 6-Acetamido-2-aminohexanoic acid or 2-amino-6-isobutyramidohexanoic acid (2.0 mmol) was dissolved in 10.0 mL of methanol, and then the solution was cooled to −10 °C. SOCl2 (0.22 mL, 3.0 mmol) was slowly added to the solution in 1 h. The reaction was kept at room temperature for 2 h and was refluxed at 70 °C for 5 h. After removing the volatile under reduced pressure, the residue was dissolved in 4.0 mL of methanol and was precipitated into diethyl ether. The precipitate was collected and dried under vacuum to give methyl 6-acetamido-2-aminohexanoate or methyl 2-amino-6-isobutyramidohexanoate. Methyl 6-acetamido-2-aminohexanoate or methyl 2-amino-6isobutyramidohexanoate (1.0 mmol) and triethylamine (0.56 mL, 4.0 mmol) were dissolved in 10 mL of chloroform. Isobutyric anhydride or acetic anhydride (3.0 mmol) was slowly added to the solution and stirred for 24 h at room temperature. After diluting with 20 mL of CHCl3, the solution was washed with 30 mL of saturated NaHCO3(aq), 30 mL of HCl (1.0 M), and 30 mL of saturated NaCl(aq). The organic layer was dried over anhydrous MgSO4 and concentrated under reduced pressure to afford a crude product. The product was further purified by recrystallization in ethyl acetate and gave a white solid. Isomer A: yield 85%; mp 96−98 °C. FT-IR: 3300 cm−1 (υ-N−H), 2968 cm−1 (υ-C−H), 1740 cm−1 (υ-CO in ester group), 1646 and 1552 cm−1 (υ-CO, in amide group). 1H NMR (D2O, 500M, ppm): δ = 6.16 (br, 1H, NH), 5.93 (br, 1H, NH), 4.57 (m, 1H, CH), 3.74 (s, 3H, CH3), 3.21 (m, 2H, CH2), 2.43 (m, 1H,

EXPERIMENTAL SECTION

Materials. Hyperbranched polyethylenimine (HPEI, Aldrich, Mn = 104 g/mol, Mw/Mn = 2.5) was dried under vacuum prior to use. Triethylamine (TEA, 99%, Tianjin Kewei Chemical Company) was dried over CaH2 and distilled before use. Isobutyric anhydride (97%) and acetic anhydride (98%) were purchased from Alfa Aesar and used without further purification. Lysine·HCl (99%) and 8-hydroxyquinoline were purchased from HEOWNS and used directly. Thionyl chloride (99.5%), CuSO4·5H2O, NaHCO3, and 3 Å molecular sieve were purchased from Tianjin Kewei Chemical Company and used directly. Benzoylated cellulose membrance (MWCO 1000 g/mol) was purchased from Sigma and used as received. Synthesis of 6-Acetamido-2-aminohexanoic Acid. Lysine·HCl (1.46 g, 7.99 mmol) was dissolved in 16.0 mL of NaHCO3 solution (1.00 M). CuSO4·5H2O (1.05 g) in 8.00 mL of water was added, and the mixture turned dark blue. After the further addition of 1.00 g (11.9 mmol) of NaHCO3, acetic anhydride (0.98 g, 9.6 mmol) was added dropwise into the solution and stirred for 24 h. The blue slurry was filtered, and the residue was dried under vacuum, yielding the copper complex as a blue solid. The copper complex was suspended in 50.0 mL of water, and then 8-hydroxyquinoline (1.20 g, 8.27 mmol) was added. The suspension was stirred for 10 h, and the color turned from blue to yellow. After filtration, the filtrate was washed three times by 50.0 mL of ethyl acetate. The aqueous phase was freeze-dried to give the product. Yield: 54%. 1H NMR (D2O, 500M, ppm): δ = 3.69 (t, B

DOI: 10.1021/acs.macromol.7b01887 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Scheme 2. Preparation of Thermoresponsive HPEIs Conjugated with Spatially Isomerized Groups

b. 2D NMR Experiments. 2D 1H−1H NOESY spectra were acquired on a Varian 700 MHz instrument by using 7000 Hz spectral windows in f1 and f 2, a recycle delay of 2 s, and a 100 ms mixing time. Eight scans were averaged for each of 512 complex t1 increments. The data were processed with Gaussian weighting function in both dimensions and zero filling to a 1024 × 2048 data matrix before Fourier transformation. 2D 1H−1H COSY spectra were acquired on the 700 MHz instrument by using 7000 Hz spectral windows in f1 and f 2 and a 2 s recycle delay. Four scans were averaged for each 256 complex t1 increments. The data were processed with Gaussian weighting functions in both dimensions and zero filling to a 1024 × 2048 data matrix before Fourier transformation. 2D 1H−13C HSQC spectra were acquired on the 700 MHz instrument by using 44 000 and 5600 Hz spectral windows in the 13 C( f1) and 1H( f 2) chemical shift dimensions, respectively, and a 2 s recycle delay. Eight scans were averaged for each 512 complex t1 increments. The data were processed with Gaussian weighting function in both dimensions and zero filling to a 2048 × 4096 data matrix before Fourier transformation. 2D 1H−13C HMBC spectra were acquired on a 700 MHz instrument by using 44 000 and 5600 Hz spectral windows in the 13 C( f1) and 1H( f 2) chemical shift dimensions, respectively, and a 2 s recycle delay. Sixteen scans were averaged for each 512 complex t1 increments. The data were processed with Gaussian weighting function in both dimensions and zero filling to a 2048 × 4096 data matrix before Fourier transformation. 2D DOSY expriments were acquired at the 700 MHz instrument by using 10 kHz spectral width in f2, and a recycle delay of 2 s. The pulse gradient durations range from 1.9 ms to 4.0 ms, and the diffusion time ranges from 150 ms to 250 ms. The pulse gradients were incremented from 2 to 95% of the maximum gradient strength in a linear ramp (16 steps). 8 scans and 4 dummy scans were used for the signal acquisition of each sample. The data were processed with exponential window function with a line broadening factor of 5 Hz and zero filling to 64k data points.

CH), 1.98 (s, 3H, CH3), 1.82 (m, 1H, CH2), 1.67 (m, 1H, CH2), 1.55 (m, 2H, CH2), 1.34 (m, 2H, CH2), 1.16 (m, 6H, CH3) (Figure S3). 13 C NMR (D2O, 500M, ppm): δ = 177.20, 173.11, 170.56, 52.38, 51.44, 39.02, 35.41, 32.17, 28.53, 23.11, 22.35, 19.61 (Figure S4). MS (ESI) m/z: [M + H]+, 272.9 (Calculated: 273.2). Isomer B: yield 87%; mp 111−114 °C. FT-IR: 3301 cm−1 (υ-N−H), 2967 cm−1 (υ-C−H), 1740 cm−1 (υ-CO in ester group), 1645 and 1553 cm−1 (υ-CO, in amide group). 1H NMR (D2O, 500M, ppm): δ = 6.35 (br, 1H, NH), 5.66 (br, 1H, NH), 4.47 (m, 1H, CH), 3.67 (s, 3H, CH3), 3.20 (m, 2H, CH2), 2.31 (m, 1H, CH), 1.99 (s, 3H, CH3), 1.78 (m, 1H, CH2), 1.67 (m, 1H, CH2), 1.45 (m, 2H, CH2), 1.28 (m, 2H, CH2), 1.08 (m, 6H, CH3) (Figure S5). 13C NMR (D2O, 500M, ppm): δ = 177.77, 172.88, 170.51, 52.29, 52.11, 38.42, 35.45, 31.37, 28.94, 22.91, 22.18, 19.57 (Figure S6). MS (ESI) m/z: [M + H]+, 273.2 (Calculated: 273.2). Synthesis of HPEI-C4-C2 and HPEI-C2-C4. A mixture of sodium methoxide (10 mg), HPEI (80 mg, 0.61 mmol of −NH2), isomer A or B (0.53 g, 1.9 mmol), 3 Å molecular sieve (40 mg), and anhydrous methanol (0.5 mL) was heated at 70 °C for 48 h under nitrogen flow conditions. The resulting mixture was diluted with 10 mL of methanol. After centrifugation, the supernatant was further filtered by 0.45 μm film and dialyzed in methanol for 48 h. The dialysis solvent was exchanged every 12 h. The product was concentrated by reduced pressure and dried under vacuum. Characterization. Light transmittance of the polymer solution was measured on a temperature-controlled Purkinje General (China) T6 UV/vis spectrophotometer at 660 nm. The heating rate was 0.1 or 0.15 °C min−1. The cloud point temperature (Tcp) was taken from the intersection of the maximal slope tangent and the initial horizontal tangent in the resulting transmittance versus temperature curve. The temperature error is ±0.2 °C. FT-IR spectra were recorded on a Nicolet 5DXC FTIR spectrometer. The measurement was done using KBr pellets, and the scanning range was 4000−400 cm−1. Electrospray ionization/mass spectrometry (ESI-MS) was performed on an LCQ Advantage MAX mass spectrometer. NMR Experiments. The NMR experiments were mainly performed on a Varian 700 MHz instrument equipped with four RF channels, a z-axis pulse field gradient (PFG) accessory, and a 5 mm Varian 1H/13C/31P/2H four-channel probe with a PFG coil. Some experiments were performed on a Bruker DRX 500 MHz instrument equipped with a 5 mm standard PFG probe. a. 1D NMR Experiments. High-resolution 1H NMR spectra were acquired on a Varian 700 MHz instrument by using 7 kHz spectral width, 8 μs (90°) pulse width, a recycle delay of 4 s, and 32 scans. For the quantitative 1H spectra, the recycle delay was set to 20 s in the experiments. The data were processed with an exponentially decaying window function with a line broadening factor of 2 Hz. 13C NMR spectra were acquired on the 700 MHz instrument by using 40 kHz spectral width, 15 μs (90°) pulse width, a recycle delay of 2 s, and 2048 scans. The data were processed with Gaussian weighting function. All the experiments had TMS as the internal reference.



RESULTS AND DISCUSSION

Two thermoresponsive hyperbranched polymers based on the HPEI skeleton were designed and synthesized in this work. In these samples, a pair of spatial isomers with two kinds of amide units (Scheme 1), i.e., the acetamide (C2) and isobutyramide (C4) groups, were designed to be located on the different layers of the globular HPEI macromolecules (Scheme 2). The two samples are named as HPEI-C4-C2 and HPEI-C2-C4. To achieve the structures, we have exploited the reaction mechanism between the amino groups of HPEI and the isomers, i.e., C4-C2 and C2-C4. It has been realized that in the amidation reaction of HPEI the reaction first takes place in the primary amines located at the periphery, and the inner primary and secondary amines are less reactive due to the steric C

DOI: 10.1021/acs.macromol.7b01887 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules hindrance effect.25 Following this mechanism, the resulted HPEI-C4-C2 and HPEI-C2-C4 molecules are expected to have the spheroidal structures as illustrated by the cartoon picture in Scheme 2. The previous work has shown that it is difficult for the bulky reactants to react with these inner primary and secondary amines of HPEI even though the reactants in the reactive acyl chloride state.25 In this work, 1H NMR measurements (Figures S7 and S8) show that maximally 24% of amino groups of HPEI react with these isomers. Both HPEI-C4-C2 and HPEI-C2-C4 are soluble in water. The aqueous solutions of these two polymers become turbid after being heated to certain temperature and become transparent again when the solutions are cooled down, indicating that they are thermoresponsive in water. For the quantitative analysis, the temperature dependence of the light transmittance of the aqueous solutions of these two polymers was studied (Figure 1). An obvious phase transition can be

simply counting of the amount of the hydrophobic groups cannot explain the remarkably different phase transition behaviors of these molecules. For HPEI-C4-C2 and HPEIC2-C4, it is the spatial distribution of the hydrophobic groups in the spheroidal macromolecules that results in the pronounced Tcp difference. What follows next is to reveal the structures and spatial distribution of the hydrophobic groups in HPEI-C4-C2 and HPEI-C2-C4. High-resolution solution NMR has been performed on the two samples in D2O to discover the molecular origin of the different transition behaviors. Figures 2a and 2b exhibit the 1H NMR spectra of HPEI-C4-C2 and HPEI-C2-C4. It is observed that the signals in the high-resolution 1H NMR spectra of HPEI-C4-C2 and HPEI-C2-C4 are very similar, indicating the very similar chemical composition of the samples. We have assigned the signals in the 1H NMR spectra of HPEI-C4-C2 and HPEI-C2-C4. The details for the signal assignment are referred to the 1H COSY, 13CDEPT-135, 1H−13C HSQC, and HMBC spectra in the Supporting Information (Figures S9− S14). The temperature-dependent 1H spectra of HPEI-C4-C2 and HPEI-C2-C4 ranging from 288 to 333 K were measured (Figure 3). In these spectra, the signals of the methylene groups of HPEI (M, N for HPEI-C4-C2; M′, N′ for HPEI-C2-C4) show relative large line width at the low temperatures. With increasing temperature, these signals show a clear narrowing tendency, indicating the increase of the local mobility of the groups. In contrast to the methylene groups, the signals from the C4-C2 and C2-C4 groups keep almost constant in the same temperature range. Generally, the molecular mobility increases upon raising temperature. Reflecting such a temperature influence, the width of 1H signal usually decrease with increasing temperature. This has been observed in the temperature dependence of the signals of the methylene groups of HPEI (M, N; M′, N′). For the signals of the C4-C2 and C2-C4 groups, the almost constant line width upon temperature alteration indicates that there are some additional factors counteracting the narrowing effect from increasing temperature. The packing arrangement of the groups, which can induce an additional restriction on the segmental mobility, has been considered as the origin in the molecular level. To reveal the packing arrangement of the groups, the 1H NOESY experiments have been performed on HPEI-C4-C2 and HPEIC2-C4. Note that in many NMR studies on the thermoresponsive polymers the signal intensities in the 1H NMR spectra

Figure 1. Typical influence of temperature on the light transmittance. The polymer concentration is 1 mg/mL; 0.1 M phosphate buffer saline with pH = 8.

observed in both the aqueous solutions; however, these two polymers have markedly different phase transition temperature that is called the cloud point temperature (Tcp) here. HPEI-C4C2 shows a Tcp of 54.3 °C, whereas that of HPEI-C2-C4 is only 30.6 °C. Generally speaking, the phase transition temperature of thermoresponsive polymer decreases with the increase of the molecular hydrophobicity. HPEI-C4-C2 and HPEI-C2-C4 have the same ratio of hydrophilic and hydrophobic units; therefore,

Figure 2. High-resolution 1H NMR spectra of (a) HPEI-C4-C2 and (b) HPEI-C2-C4. These 1H NMR spectra were recorded on the 700 MHz NMR instrument. The experimental temperature is 288 K. The solution concentration is 1 mg/mL for both samples in D2O; 0.1 M phosphate buffer saline with pH = 8. D

DOI: 10.1021/acs.macromol.7b01887 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 3. Temperature-dependent high-resolution 1H NMR spectra of (a) HPEI-C4-C2 and (b) HPEI-C2-C4. These 1H NMR spectra were recorded on the 700 MHz NMR instrument. The recycle delay is 20 s to ensure the quantification of the signals in the spectra (see Figure S15). The solution concentration is 1 mg/mL for both samples in D2O; 0.1 M phosphate buffer saline with pH = 8.

Figure 4. 1H NOESY spectra of (a) HPEI-C4-C2 and (b) HPEI-C2-C4. These 1H NMR spectra were recorded on the 700 MHz NMR instrument. The experimental temperature is 288 K. The solution concentration is 1 mg/mL for both samples in D2O; 0.1 M phosphate buffer saline with pH = 8.

for HPEI-C2-C4) and between the methyl group of the isopropyl group (A and A′) and the methylene group (N for HPEI-C4-C2 and N′ for HPEI-C2-C4). These cross-peaks are marked in the spectra using the red circles. For HPEI-C4-C2, the cross-peaks between A and C and between A and N are clearly observed in the spectrum, whereas no clear cross-peaks between A′ and C′ and between A′ and N′ are observed in the spectrum of HPEI-C2-C4. From the structural models in Scheme 2 and the branch models in Figure 4, it is clear that in the same branch the spatial distance between the groups A/A′ and the groups C/C′ or N/N′ is far enough to give rise to the NOE and thus the cross-peaks in the 2D NOESY spectra. Therefore, it is most likely that the cross-peaks, if appeared, are not from the groups of the same branch, but rather from the groups of the neighbored branches. The appearance of the cross-peaks between the groups A and C and between A and N in the NOESY spectrum of HPEI-C4-C2 indicates that the

will significantly reduce when the temperature is above the LCST.16−21 In this study, we have not observe the similar phenomena. We analyzed the temperature dependence of the integrated signal intensities and found that the signal intensity reduces only 10−15% for HPEI-C4-C2 and HPEI-C2-C4 when the temperature is above the LCST transition (see Figure S16). From Scheme 2, it is realized that the main structural difference between HPEI-C4-C2 and HPEI-C2-C4 lies on the spatial distribution of the chemical groups in the HPEI molecular skeleton. To unveil this difference, 2D 1H NOESY NMR has been applied to the samples. Figures 4a and 4b show the 1H NOESY spectra of HPEI-C4-C2 and HPEI-C2-C4 acquired at 288 K (below Tcp). The difference in the spectra of HPEI-C4-C2 and HPEI-C2-C4 is clearly realized in the crosspeaks between the methyl group of the isopropyl group (A for HPEI-C4-C2 and A′ for HPEI-C2-C4) and the methyl group connected to the carbonyl group (C for HPEI-C4-C2 and C′ E

DOI: 10.1021/acs.macromol.7b01887 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 5. 1H NOESY spectra of (a) HPEI-C4-C2 and (b) HPEI-C2-C4, acquired at 333 K. These spectra were recorded on the 700 MHz NMR instrument. The solution concentration is 1 mg/mL for both samples in D2O; 0.1 M phosphate buffer saline with pH = 8.

333 K (Figure 5b). This is probably due to the interplay between the 1H Larmor frequency, ω, and the correlation time of the segmental motion (τc). Similar phenomena have been observed by us24 and the other group.22 HPEI-C4-C2 and HPEI-C2-C4 are synthesized from the same batch of HPEI. The 1H NMR spectra show that these samples contain the same amount of branches. Therefore, a similar molecular size is expected for these molecules, if the molecules are in the ideally “freely stretched” state. Because of the plenty of the hydrophilic groups in the molecules, one may expect that both HPEI-C4-C2 and HPEI-C2-C4 can have a well-hydrated state in the dilute aqueous solutions before the phase transition. But the above NOESY spectra have shown that the groups of HPEI-C4-C2 have a denser packing arrangement than HPEI-C2-C4 before the phase transition, which indicates that HPEI-C4-C2 is less hydrated than HPEIC2-C4 before the phase transition. Reflecting the different packing arrangement of the groups, HPEI-C4-C2 thus is expected to have a smaller size than HPEI-C2-C4. This has been confirmed by the observations in the 1H PFG diffusion NMR of HPEI-C4-C2 and HPEI-C2-C4. The 1H PFG diffusion NMR is a standard NMR method to measure the diffusion coefficients of the molecules in solution. The diffusion coefficients obtained from this method can be directly correlated to the molecular size. The signal decay in the PFG-NMR diffusion measurement can be described by eq 1

branches in HPEI-C4-C2 are not in a good hydration state, but rather have a close packing arrangement due to the hydrophobic interactions among A, C, and N groups. In contrast, for HPEI-C2-C4, the missing cross-peaks indicate that the branches are likely in a good hydration state and thus are not closely packed together. Therefore, from the NOESY spectra in Figures 4a and 4b, it is reasonable to deduce that the branches in HPEI-C4-C2 have a denser packing arrangement than HPEIC2-C4. Note that in Figures 4a and 4b the cross-peaks between A and B or A′ and B′ are observed in both of the spectra. The origin of these cross-peaks most likely is from the groups of the same isopropyl group. Figure 5 shows the 1H NOESY spectra of HPEI-C4-C2 and HPEI-C2-C4 acquired at 333 K (above Tcp). In this temperature, the two molecules have undergone the phase transition. A dense packing of the branches in the molecules thus is expected. In the 1H NOESY spectra of HPEI-C4-C2 and HPEI-C2-C4 (Figure 5), a new set of cross-peaks appears between C/C′ and N/N′, indicating the dense packing of the groups in these molecules. Note that for HPEI-C4-C2 the group C is the end of the branch, which is expected to be spatially far away from the HPEI skeleton. The appearance of the cross-peaks between C and N indicates that some portions of the branches in HPEI-C4-C2 have achieved a very dense packing arrangement, manifesting the shrinkage of HPEI-C4C2 molecule due to the phase transition. In the 1H NOESY spectrum of HPEI-C2-C4 at 333 K, besides the cross-peaks between C′ and N′, the cross-peaks between A′ and C′ between A′ and N′ appeared, indicating some portions of the branches in HPEI-C2-C4 are densely packed in the molecules due to the phase transition. A similar phenomenon is observed in HPEI-C4-C2 before transition (the NOESY spectrum acquired at 288 K, Figure 4a). The similarity between the NOESY spectrum of HPEI-C2-C4 after transition and the NOESY spectrum of HPEI-C4-C2 before transition further confirms the presence of the densely packing structures in HPEI-C4-C2 before transition. It is observed that the crosspeaks between A′ and C′ between A′ and N′ in Figure 5b are much weaker than those in Figure 4a. This indicates that after transition the groups of HPEI-C4-C2 pack much denser than HPEI-C2-C4, which is well consistent with the observation in the 1H PFG NMR later. Note that there are no cross-peaks between A′ and B′ in the NOESY spectrum of HPEI-C2-C4 at

ln(I /I0) = − [Dg 2γ 2δ 2(Δ − δ /3)]β

(1)

where I is the signal intensity measured under the gradient strength of g, I0 is the signal intensity without gradient, D is the self-diffusion coefficient, γ is the gyromagnetic ratio of the spin, δ is the duration of gradient pulse, Δ is the diffusion time, and β is called the stretch parameter which will be the focus of the following study. In the literature, the β parameter has been used to quantify the influence of the polydispersity of the moving particles on the signal decay.26 β is equal to 1 for a monodisperse sample and ranges from 0.5 to 1 for a polydisperse sample. The value of β decreases when the polydispersity increases. Figure 6 shows the plots of CH3 signal intensity of C2 group versus [g2γ2δ2(Δ − δ/3)] of the dilute HPEI-C4-C2 and HPEIC2-C4 solutions (1 mg/mL) at 288 K. A clear bending-up F

DOI: 10.1021/acs.macromol.7b01887 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

different packing arrangements of the groups. The dense packing arrangement of the groups in HPEI-C4-C2 can be considered as a type of “shrinkage” of the macromolecules. But due to the chemical nature of HPEI-C4-C2, such a shrinkage should not be a uniform one and also will not be identical in the different HPEI-C4-C2 molecules. This very likely will result in the increase of the polydispersity of the molecular size. This well explains the smaller β value of HPEI-C4-C2 than HPEIC2-C4. After transition, HPEI-C4-C2 shows a bigger diffusion coefficient than HPEI-C2-C4, indicating a denser packing in HPEI-C4-C2 than HPEI-C2-C4, which is consistent with the observation in the 1H NOESY spectra at 333 K. The β values are similar for HPEI-C4-C2 and HPEI-C2-C4. This indicates that after transition the HPEI-C4-C2 and HPEI-C2-C4 molecules have very similar polydispersity in the size. But because HPEI-C4-C2 most likely has a much denser packing than HPEI-C2-C4, the similar β values of HPEI-C4-C2 and HPEI-C2-C4 might be just incidentally identical. Based on the above observations, a deep understanding of the phase transitions of HPEI-C4-C2 and HPEI-C2-C4 can be achieved at a molecular level. The main structural difference of the two isomers is the different C2 and C4 arrangements of the chain ends. Chemically known C4 is longer and more hydrophobic than C2. When C4 is arranged outside and C2 is inside in the chain end for HPEI-C2-C4, this leads to less hydrophobic interaction among the polymer chains and more exposure of hydrophilic carbonyl groups to the surface of the “globular” polymer in aqueous solution. In this way, the polymer behaves more hydrated or has more interaction with water (see Figure S18) and vice versa for HPEI-C4-C2 (see Figure S19). Apparently, the polymer chains of HPEI-C2-C4 are more “extended” than those of HPEI-C4-C2 in water. This has been clearly demonstrated by the diffusion NMR measurements before Tcp. These two-carbon space is rather small in structure for these two polymers. So both have very slight difference in diffusion coefficient values but much distinct difference in β values. More “extended” chain ends of HPEIC2-C4 (less hydrophobic interaction among chains) make the polymer more uniform in size, while more “folded” chain ends (more hydrophobic interaction among chains) make the polymer various in size (various folding styles). Of course, at high temperature (after the phase transition), the polymer chain motion increases and more chain interaction occurs. Thus, both polymers are “shrunk” with dense packing, which leads to smaller size and more variation or pseudopolydispersity in size. This is well in line with the observations in the NOESY spectra (the appearance of the NOE interaction between C/C′ and N/N′) and the PFG NMR (the larger diffusion coefficient and the bigger difference in the β values). The different starting states of HPEI-C4-C2 and HPEI-C2C4 are considered as the origin of the different transition temperatures of these molecules. The thermal dynamics of the phase transition has been taken into account for a deep understanding of the different transition temperatures. In the literature, the phase transition of thermoresponsive polymer in water is an endothermal process.27 Entropy increase has been considered as the dominant driving force for the phase transition. The more the entropy change, the lower the transition temperature. For the origin of the entropy change in the transition process, the water molecules are considered to play a crucial role. Below the phase transition temperature, the water-soluble units are likely to be wrapped by the hydration

Figure 6. Plots of CH3 signal intensity of C2 group versus [g2γ2δ2(Δ − δ/3)] of the dilute HPEI-C4-C2 and HPEI-C2-C4 solutions. The 1H PFG NMR data were recorded at the 700 MHz NMR instrument. The solution concentration is 1 mg/mL for both samples in D2O; 0.1 M phosphate buffer saline with pH = 8. The experimental temperatures are 288 K (before transition) and 333 K (after transition).

curvature is observed in all the decay curves, indicating the presence of the polydispersity of the moving particles. Fitting the curves yields the self-diffusion coefficient D and β values (Figure S17). The obtained D and β values of HPEI-C4-C2 and HPEI-C2-C4 before and after the phase transition are listed in Table 1. Table 1. β and diffusion coefficients for the HPEI-C4-C2 and HPEI-C2-C4 aqueous solutions at 288 and 333 K sample HPEI-C4-C2 HPEI-C2-C4

temp (K) 288 333 288 333

diffusion coefficient (10−11 m2/s) 5.2 13.8 4.4 8.4

± ± ± ±

0.1 0.2 0.1 0.1

β 0.76 0.74 0.90 0.75

± ± ± ±

0.02 0.02 0.03 0.02

As mentioned above, HPEI-C4-C2 and HPEI-C2-C4 are from the same batch of HPEI, and the 1H NMR shows that these samples have the almost same chemical compositions (including the almost same amount of branches). Therefore, one can expect that HPEI-C4-C2 and HPEI-C2-C4 have similar molecular size and polydispersity. If so, the 1H PFG NMR of HPEI-C4-C2 and HPEI-C2-C4 before transition should give similar diffusion coefficient D and similar β value. However, in Figure 6 it is clear that the signal of HPEI-C4-C2 decays faster than HPEI-C2-C4, indicating that HPEI-C4-C2 has a higher diffusion coefficient than HPEI-C2-C4. The diffusion coefficient from the curve fitting (Table 1) confirms this point. Moreover, the β value of HPEI-C4-C2 is much smaller than HPEI-C2-C4. This indicates that HPEI-C4-C2 has a smaller molecular weight and a higher polydispersity than HPEI-C2-C4. These observations apparently contradict the above point that HPEI-C4-C2 and HPEI-C2-C4 have similar molecular weight and polydispersity. This contradiction can be explained by considering the different packing arrangement of the groups in HPEI-C4-C2 and HPEI-C2-C4: The groups of HPEI-C4-C2 are more densely packed than HPEI-C2-C4. This indicates that in the solution the size of HPEI-C4-C2 most likely is smaller than HPEI-C2-C4. The smaller HPEI-C4-C2 diffuses faster than HPEI-C2-C4. The difference in the molecular size well explains the different diffusion coefficients of HPEI-C4-C2 and HPEI-C2-C4. The different β values of HPEI-C4-C2 and HPEI-C2-C4 can be explained by the G

DOI: 10.1021/acs.macromol.7b01887 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Notes

water molecules. The temperature increase results in the increase of the thermal motion energy of the hydration water molecules and thereby the possible transition of the water molecules from the hydration state to the “free” state. This process can dramatically increase the entropy of the system and thus has been considered as the dominant driving force of the phase transition. For HPEI-C4-C2 and HPEI-C2-C4, these two macromolecules have the different hydration states before transition. The highly hydrated state of HPEI-C2-C4 before transition indicates that significant amount of water molecules are interacted with HPEI-C2-C4. Compared with of HPEI-C2C4, HPEI-C4-C2 is much less hydrated, indicating that less water molecules are interacted with HPEI-C2-C4. This means that the entropy increase of the phase transition of HPEI-C4C2 is less than that of HPEI-C2-C4; thus, higher temperature is required for HPEI-C4-C2 to ensure the phase transition.

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was financially supported by the National Natural Science Foundation of China (21174039, 21274106, and 21574043). The authors are grateful to one of the reviewers for proposing the structural models of the two molecules and stimulating us to think over the NOESY data and the correlations to the packing structures in the two molecules.





CONCLUSIONS In summary, in this work we synthesized two thermoresponsive hyperbranched polymers. In these hyperbranched polymers, a pair of spatial isomers with two kinds of amide units, i.e., the acetamide and isobutyramide groups, were “planted” in the different layers. The remarkably different transition behaviors of these polymers have been related to the different spatial distribution of the acetamide and isobutyramide groups. The detailed NMR study reveals the mechanism of the thermal transition on the molecular level: The different spatial distribution of the acetamide and isobutyramide groups in the hyperbranched polymers induces the different packing arrangement of the chemical groups in the hyperbranched polymers. The hyperbranched polymer having a densely packed structure exhibits the characteristics of a microscopic phase separation in the molecule before the phase transition. Reflecting the densely packed structures, the groups in this polymer are less hydrated than those in the polymer having the loosely packed structures. The entropy change due to the dehydration/hydration processes is considered as the driving force of the phase transition. The different hydration states of the two hyperbranched polymers can cause the different entropy change in the phase transition. This explains the different transition temperatures of the two hyperbranched polymers.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.7b01887. Details for experiments and the signal assignments with the 1H COSY, 13CDEPT-135, 1H−13C HSQC, and HMBC spectra (PDF)



REFERENCES

(1) Galaev, I. Y.; Mattiasson, B. ’Smart’ Polymers and What They Could Do in Biotechnology and Medicine. Trends Biotechnol. 1999, 17, 335−340. (2) Moon, H. J.; Ko, D. Y.; Park, M. H.; Joo, M. K.; Jeong, B. Temperature-Responsive Compounds as in Situ Gelling Biomedical Materials. Chem. Soc. Rev. 2012, 41, 4860−4883. (3) Weber, C.; Hoogenboom, R.; Schubert, U. S. Temperature Responsive Bio-compatible Polymers Based on Poly(ethylene oxide) and Poly(2-oxazoline)s. Prog. Polym. Sci. 2012, 37 (5), 686−714. (4) Steinhauer, W.; Hoogenboom, R.; KeuL, H.; Moeller, M. Block and Gradient Copolymers of 2-Hydroxyethyl Acrylate and 2Methoxyethyl Acrylate via RAFT: Polymerization Kinetics, Thermoresponsive Properties, and Micellization. Macromolecules 2013, 46, 1447−1460. (5) Qiu, X. P.; Tanaka, F.; Winnik, F. M. Temperature-Induced Phase Transition of Well-Defined Cyclic Poly(Nisopropylacrylamide)s in Aqueous Solution. Macromolecules 2007, 40, 7069−7071. (6) Xu, J.; Ye, J.; Liu, S. Y. Synthesis of Well-Defined Cyclic Poly(Nisopropylacrylamide) via Click Chemistry and its Unique Thermal Phase Transition Behavior. Macromolecules 2007, 40, 9103−9110. (7) Ye, J.; Xu, J.; Hu, J.; Wang, X.; Zhang, G.; Liu, S.; Wu, C. Comparative Study of Temperature-Induced Association of Cyclic and Linear Poly(N-isopropylacrylamide) Chains in Dilute Solutions by Laser Light Scattering and Stopped-Flow Temperature Jump. Macromolecules 2008, 41, 4416−4422. (8) Lahasky, S. H.; Hu, X.; Zhang, D. Thermoresponsive Poly(αpeptoid)s: Tuning the Cloud Point Temperatures by Composition and Architecture. ACS Macro Lett. 2012, 1 (5), 580−584. (9) Li, Q.; Gao, C.; Li, S.; Huo, F.; Zhang, W. Doubly ThermoResponsive ABC Triblock Copolymer Nanoparticles Prepared through Dispersion RAFT Polymerization. Polym. Chem. 2014, 5 (8), 2961− 2972. (10) Ma, L.; Wang, G.; Sun, S.; Wu, P. The Influence of a Thermoresponsive Polymer on the Microdynamic Phase Transition Mechanisms of Distinctly Structured Thermoresponsive Ionic Liquids. Phys. Chem. Chem. Phys. 2017, 19 (33), 22263−22271. (11) Aoki, T.; Muramatsu, M.; Torii, T.; Sanui, K.; Ogata, N. Thermosensitive Phase Transition of an Optically Active Polymer in Aqueous Milieu. Macromolecules 2001, 34, 3118−3119. (12) Seto, Y.; Aoki, T.; Kunugi, S. Temperature- and PressureResponsive Properties of L- and DL-forms of Poly(N-(1hydroxymethyl)propylmethacrylamide) in Aqueous Solutions. Colloid Polym. Sci. 2005, 283 (10), 1137−1142. (13) Fang, W.-C.; Zhang, R.; Yao, Y.-F.; Liu, H.-J.; Chen, Y. Specific Thermoresponsive Behaviours Exhibited by Optically Active and Inactive Phenylalanine Modified Hyperbranched Polyethylenimines in Water. Chin. J. Polym. Sci. 2017, 35 (8), 1035−1042. (14) Huang, X.; Du, F.; Liang, D.; Lin, S.-S.; Li, Z. Stereochemical Effect of Trans/Cis Isomers on the Aqueous Solution Properties of Acid-Labile Thermoresponsive Polymers. Macromolecules 2008, 41, 5433−5440. (15) Liu, P.; Xiang, L.; Tan, Q.; Tang, H.; Zhang, H. Steric Hindrance Effect on Thermoresponsive Behaviors of PyrrolidoneBased Polymers. Polym. Chem. 2013, 4 (4), 1068−1076.

AUTHOR INFORMATION

Corresponding Authors

*(Y.C.) E-mail: [email protected]. *(Y.Y.) E-mail: [email protected]. ORCID

Yu Chen: 0000-0002-0559-6136 Ye-Feng Yao: 0000-0002-6274-2048 Author Contributions

B.W. and T.X. contributed to this work equally. H

DOI: 10.1021/acs.macromol.7b01887 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (16) Hanyková, L.; Spěvácěk, J.; Ilavsky, M. 1 H NMR study of Thermotropic Phase Transition of Linear and Crosslinked Poly(vinyl methyl ether) in D2O. Polymer 2001, 42, 8607−8612. (17) Spěvácě k, J.; Geschke, D.; Ilavsky, M. 1H NMR Study of Temperature Collapse of Linear and Crosslinked Poly(N,Ndiethylacrylamide) in D2O. Polymer 2001, 42, 463−468. (18) Spěvácě k, J.; Hanyková, L.; Starovoytova, L. 1H NMR Relaxation Study of Thermotropic Phase Transition in Poly(vinyl methyl ether)/D2O solutions. Macromolecules 2004, 37, 7710−7718. (19) Spěvácě k, J. I.; Hanyková, L.; Labuta, J. Behavior of Water during Temperature-Induced Phase Separation in Poly(vinyl methyl ether) Aqueous Solutions. NMR and Optical Microscopy Study. Macromolecules 2011, 44, 2149−2153. (20) Chen, J.; Gong, X.; Yang, H.; Yao, Y.; Xu, M.; Chen, Q.; Cheng, R. NMR Study on the Effects of Sodium n-Dodecyl Sulfate on the Coil-to-Globule Transition of Poly(N-isopropylacrylamide) in Aqueous Solutions. Macromolecules 2011, 44, 6227−6231. (21) Spěváček, J.; Konefał, R.; Č adová, E. NMR Study of Thermoresponsive Block Copolymer in Aqueous Solution. Macromol. Chem. Phys. 2016, 217 (12), 1370−1375. (22) Chai, M.; Niu, Y.; Youngs, W. J.; Rinaldi, P. L. Structure and Conformation of DAB Dendrimers in Solution via Multidimensional NMR Techniques. J. Am. Chem. Soc. 2001, 123, 4670−4678. (23) Holycross, D. R.; Chai, M. Comprehensive NMR Studies of the Structures and Properties of PEI Polymers. Macromolecules 2013, 46, 6891−6897. (24) Jiang, S.; Yao, Y.; Chen, Q.; Chen, Y. NMR Study of Thermoresponsive Hyperbranched Polymer in Aqueous Solution with Implication on the Phase Transition. Macromolecules 2013, 46 (24), 9688−9697. (25) Liu, Y.; Fan, Y.; Liu, X.-Y.; Jiang, S.-Z.; Yuan, Y.; Chen, Y.; Cheng, F.; Jiang, S.-C. Amphiphilic Hyperbranched Copolymers Bearing a Hyperbranched Core and Dendritic Shell: Synthesis, Characterization and Guest Encapsulation Performance. Soft Matter 2012, 8 (32), 8361−8369. (26) Gong, X.; Hansen, E. W.; Chen, Q. Molecular Weight Distribution Characteristics (of a Polymer) Derived from a Stretched-Exponential PGSTE NMR Response Function-Simulation. Macromol. Chem. Phys. 2012, 213, 278−284. (27) Schild, H. G.; Tirrell, D. A. Microcalorimetric Detection of Lower Critical Solution Temperatures in Aqueous Polymer Solutions. J. Phys. Chem. 1990, 94, 4352−4356.

I

DOI: 10.1021/acs.macromol.7b01887 Macromolecules XXXX, XXX, XXX−XXX