Cloud points in aqueous solutions of poly(N-isopropylacrylamide) synthesized by aqueous redox polymerization

The mean-square radius of gyration, second virial coefficient and intrinsic viscosity were determined in methanol at 25.0 °C for poly(N-isopropylacrylamide) (PNIPA) samples in the range of

weight-average molecular weight Mw from 1.04 × 105 to 2.65 × 106, which were synthesized by aqueous redox polymerization using a redox catalyst consisting of ammonium persulfate and sodium

metabisulfite. Analyses of the quantities showed that the samples have almost the same degree of branching as that of PNIPA samples previously synthesized by radical polymerization using

azobis(isobutyronitrile) as an initiator in tert-butanol. The cloud point was also determined in aqueous solutions of two PNIPA samples with Mw=1.04 × 105 and 1.86 × 105 and compared with

the results of the previous samples. It was found that the cloud point is definitely higher for the present samples than for the previous ones, and, moreover, the Mw dependences of the cloud

point for the two kinds of samples are opposite to each other. As the two kinds of samples have almost the same stereochemical composition and degree of branching, such difference in the

behavior of the cloud point is considered to arise from the difference in their chain-end groups.

In a series of experimental studies recently made of aqueous poly(N-isopropylacrylamide) (PNIPA) solutions,1, 2, 3, 4 it was shown that the behavior of their cloud-point curves is

considerably affected by the kind of end group of the PNIPA sample used1, 3 and also by primary structure, that is, the degree of branching of the sample.2 Furthermore, it was pointed out

that the cloud-point curve does not seem to correspond directly to a liquid–liquid binodal.4

The PNIPA samples used previously1, 2, 3, 4 were synthesized by radical polymerization using azobis(isobutyronitrile) (AIBN) as an initiator and by living anionic polymerization using

diphenylmethylpotassium as an initiator, so that they had hydrophobic groups at their chain ends, that is, the isobutyronitrile group for the former and the diphenylmethyl group for the

latter. The above-mentioned aqueous solution behavior of PNIPA may arise from the effect of the hydrophobic chain-end groups, which is enhanced with decreasing molecular weight. To verify

more clearly the effect of the chain-end groups on the behavior of the cloud-point curve, it is desirable to pursue further investigations of aqueous solution behavior of PNIPA samples with

a hydrophilic chain-end groups.

In this study, therefore, we prepared such PNIPA samples by aqueous redox polymerization using a redox catalyst consisting of ammonium persulfate and sodium metabisulfite, whose chain-end

groups are hydrophilic and/or ionic ones. We first characterized the PNIPA samples so prepared by analyzing dilute solution properties such as the mean-square radius of gyration 〈S2〉, second

virial coefficient A2 and intrinsic viscosity [η], in methanol at 25.0 °C, then determined the cloud point in the aqueous solutions, and finally made a comparison of present results with

the previous ones for the PNIPA samples having the hydrophobic chain-end groups.1

Original PNIPA samples were synthesized by aqueous redox polymerization using a redox catalyst consisting of ammonium persulfate (NH4)2S2O8 (Wako Pure Chemical Industries, Ltd, Osaka, Japan)

as an oxidative reagent and sodium metabisulfite Na2S2O5 (Wako Pure Chemical Industries, Ltd) as a reductive reagent, following the procedure reported by Wooten et al.5 The monomer

N-isopropylacrylamide (Tokyo Chemical Industry Co., Ltd, Tokyo, Japan; ∼20 g), which had been recrystallized three times from a 9:1 mixture of n-hexane and benzene and then dried in a vacuum

for 12 h, was dissolved in 250 ml of phosphate-buffered saline (PBS) with pH=7.3–7.5 (Lonza Walkersville, Inc., Walkersville, MD, USA). Subsequently, 0.09 g of Na2S2O5 in 1 ml of PBS was

added to the monomer solution, followed by 0.03 g of (NH4)2S2O8 in 1 ml of PBS. The reaction mixture was then stirred under dry nitrogen at 25 °C for 24 h. Then, the polymerization mixture

was poured into 250 ml of methanol, to precipitate an original polymer sample.6 The resulting polymer was then dissolved in pure water and dialyzed seven times against pure water for 24 h

using a cellulose tube.

The original samples so prepared were separated into fractions of narrow molecular weight distribution by fractional precipitation using acetone as a solvent and n-hexane as a precipitant,

or by a column elution method with a 6:4 mixture of n-hexane and acetone as an eluent. We obtained nine test samples named Rx (x=10, 19, 21, 48, 66, 96, 130, 220 and 270), which are

generally called R samples. Each of the samples was dissolved in 1,4-dioxane, then filtered through a Teflon membrane Fluoropore (Sumitomo Electric Industries, Ltd, Osaka, Japan) with a pore

size of 1.0 μm, and finally freeze-dried from their 1,4-dioxane solutions. They were dried in a vacuum at ∼80 °C for 12 h just before use. The ratio Mw/Mn of the weight-average molecular

weight Mw to the number-average molecular weight Mn was determined for the 9 samples from analytical gel permeation chromatography in the same manner as before1, 3 using tetrahydrofuran as

an eluent and 12 standard polystyrene (PS) samples as reference standards.

To specify the chain-end group of the present PNIPA samples, we synthesized an extra PNIPA sample with a very small Mw by mixing 10 g of N-isopropylacrylamide in 250 ml of PBS with 1.50 g of

Na2S2O5 in 25 ml of PBS and 0.50 g of (NH4)2S2O8 in 25 ml of PBS, in the same manner as above. The reaction mixture was stirred under dry nitrogen at 25 °C for 30 min. The resulting polymer

was dialyzed and then dried in the above-mentioned manner. The extra sample was named R0.5, whose values of Mw, Mn and Mw/Mn were determined to be 4.77 × 103, 3.34 × 103 and 1.43,

respectively, from analytical gel permeation chromatography with a column Shodex SB-804 HQ (Showa Denko KK, Tokyo, Japan) connected to a solvent delivery pump PU-980 (JASCO Corporation,

Tokyo, Japan) and a refractive index detector RI-8000 (Tosoh Corporation, Tokyo, Japan); a 0.2 M aqueous NaNO3 solution was used as an eluent and nine standard poly(oxyethylene) samples as

reference standards.

Tetrahydrofuran used in analytical gel permeation chromatography was of reagent grade with no stabilizer. Deuterated dimethyl sulfoxide used for 1H- and 13C-nuclear magnetic resonance (NMR)

spectroscopy was of reagent grade. Methanol used for static light scattering (LS) and viscosity measurements was purified by distillation after refluxing over calcium hydride for ∼6 h. Water

used for the determinations of the cloud point was highly purified through a water purification system Simpli Lab (Millipore Corporate, Billerica, MA, USA); its resistivity was 18.2 MΩ cm.

To determine the values of the fraction fr of racemo diads for all the nine samples, R10 through R270, 1H NMR spectra for the samples in deuterated dimethyl sulfoxide at 170 °C at ∼1 wt%

were recorded on a spectrometer EX-400 (JEOL Ltd, Tokyo, Japan) at 399.8 MHz using an rf pulse angle of 90° and a pulse repetition time of 8 s. To specify the chain-end group of the present

PNIPA samples, the 13C NMR spectrum of the extra sample R0.5 in deuterated dimethyl sulfoxide at 150 °C at ∼20 wt% was recorded on the same spectrometer, operated at 100.5 MHz with a pulse

repetition time of 3 s. Tetramethylsilane was added to each solution as an internal standard.

LS measurements were carried out to determine Mw and A2 for all nine samples, R10 through R270, and 〈S2〉 for the eight samples, R19 through R270, in methanol at 25.0 °C. A light-scattering

photometer Fica 50 (Fica, Saint Denis, France) was used for all measurements with vertically polarized incident light of wavelength λ0=436 nm. To calibrate the apparatus, the intensity of

the light scattered from pure benzene was measured at 25.0 °C at a scattering angle of 90°, where the Rayleigh ratio RUu(90°) of pure benzene was taken as 46.5 × 10−6 cm−1.7 The

depolarization ratio ρu of pure benzene at 25.0 °C was determined to be 0.41±0.01. The scattered intensity was measured at eight different concentrations and at scattering angles ranging

from 22.5 to 142.5°, and then converted to the excess unpolarized components ΔRUv of the reduced scattered intensity using the scattered intensity from the solvent methanol. The data

obtained were treated by using the Berry square-root plot.8 For all samples, corrections for the optical anisotropy were unnecessary as the degree of depolarization was negligibly small.

The most concentrated solution of each sample was prepared gravimetrically and made homogeneous by continuous stirring at room temperature for 1 or 2 days. It was optically purified by

filtration through a Teflon membrane Fluoropore with a pore size of 0.45 or 0.10 μm. The solutions of lower concentration were obtained by successive dilution. The weight concentrations of

the test solutions were converted to polymer mass concentrations c using the densities of the respective solutions calculated from the partial specific volumes v2 of the samples and the

density ρ0 of the solvent methanol. The quantity v2 was measured using an oscillating U-tube density meter DMA5000 (Anton-Paar, Graz, Austria). The values of v2 so determined in methanol at

25.0 °C were 0.869, 0.876, 0.878 and 0.883 cm3 g−1 for the samples R10, R19, R21 and R48, respectively, and 0.884 cm3 g−1 for the samples with Mw≳6 × 105, independent of Mw. For ρ0 of

methanol at 25.0 °C, we used the literature value 0.7866 g cm−3.9

The refractive index increment ∂n/∂c was measured at a wavelength of 436 nm using a differential refractometer DR-1 (Shimadzu Corporation, Kyoto, Japan). The values of ∂n/∂c in methanol at

25.0 °C were determined to be 0.1865, 0.1864, 0.1855 and 0.1846 cm3 g−1 for samples R10, R19, R21 and R48, respectively, and 0.1851 cm3 g−1 for samples with Mw≳6 × 105, independent of Mw.

For the refractive index n0 of methanol at 25.0 °C at a wavelength of 436 nm, we used the literature value 1.3337.9

Viscosity measurements were carried out for all nine samples, R10 through R270, in methanol at 25.0 °C using a conventional capillary viscometer of the Ubbelohde type. The flow time was

measured to a precision of 0.1 s, keeping the difference in the flow time between the solvent and solution longer than 20 s. The test solutions were maintained at 25.0 °C within ±0.005 °C,

during the measurements.

The most concentrated solution of each sample was prepared in the same manner as in the case of the LS measurements. The solutions of lower concentration were obtained by successive

dilution. The polymer mass concentrations c were calculated from the weight fractions and the densities of the solutions. Density corrections were also made in the calculations of the

relative viscosity ηr from the flow times of the solution and solvent. The data obtained for the specific viscosity ηsp and ηr in the range of ηr