Introduction

Copolymers of [2-(acryloyloxy)ethyl]trimethylammonium chloride (AETAC) and acrylamide (AAm) (AETAC-co-AAm) (Fig. 1) are important in various industries. For example, they are used as flocculants in paper manufacturing [1, 2], for the enrichment and recovery of minerals from mining [3], and in wastewater treatment [4,5,6]. Because it comprises polyelectrolytes, AETAC-co-AAm provides better flocculation than nonionic polyacrylamide (PAAm) [7,8,9,10]. In recent years, more than two kinds of acrylamide-based polymer flocculants have been used because sludge dewatering has become difficult owing to increasing organic matter content in wastewater [11,12,13].

Fig. 1
figure 1

The chemical structure of AETAC-co-AAm

Full size image

The properties of AETAC-co-AAm in aqueous media have been studied by light scattering and viscometry [4]. The relationships between the weight average molecular weight (Mw), root mean square radius of gyration (Rg in nm) and intrinsic viscosity ([η] in dL/g) were determined for AETAC-co-AAm with 30% AETAC: Rg = 0.033 Mw0.54, [η] = 1.05 × 10−4 Mw0.73 (1 M NaCl 25 °C, 4.5 × 105 ≤ Mw ≤ 2.7 × 106). The effect of ionic strength on [η] was also investigated for NaCl concentrations (Cs) between 10−2 and 1 mol/L, with a semirigid configuration of the copolymer chains at low Cs values being suggested owing to the absence of a linear relationship between [η] and Cs−1/2 [4]. The interaction between AETAC-co-AAm and an anionic perfluorinated surfactant was studied by NMR spectroscopy to elucidate the mechanism of flocculation [6]. Full spectral assignment of 1H and 13C resonances was achieved with 1D and 2D experiments. The 1H NMR spectra of AETAC-co-AAm with 14% and 23% AETAC in D2O (1% w/v) showed a higher multiplicity than the corresponding spectra of AETAC-co-AAm with 42% and 54% AETAC, and a stiff rod-like structure of the copolymer chains at higher AETAC compositions was suggested. This interpretation was consistent with the 13C NMR spectra of AETAC-co-AAm, in which the number of resonances with higher AETAC compositions was smaller than that with lower AETAC compositions [6].

The aim of the present study was to contribute to the characterization of AETAC-co-AAm in aqueous media by applying diffusion-ordered two-dimensional NMR spectroscopy (DOSY). DOSY is an NMR method [14,15,16,17] that reports diffusion coefficients (D) for individual resonances in NMR spectra. DOSY and the related pulsed field gradient (PFG) NMR diffusometry have been used to determine the Mw and molecular weight distribution of synthetic polymers [15, 18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34] and to characterize mixtures of polymers [35,36,37] and block copolymers [33, 38,39,40,41,42,43]. A useful tutorial on DOSY experiments for polymers was recently published [44]. Nevertheless, high-molecular-weight AETAC-co-AAm in aqueous media is challenging to characterize using DOSY experiments because their D values are as small as 10−11.5 m2/s or less and their 1H resonances suffer from line broadening owing to restricted segmental mobility of the polyelectrolyte chains.

One of the most commonly used pulse sequences for DOSY experiments is the bipolar pulse pair stimulated echo sequence with a longitudinal eddy current delay (bpp-ste-led) (Fig. 2) [45]. The observed echo intensity f(G) at a given PFG strength G (in G/cm) is given by the Stejskal–Tanner expression:

$$f(G) = f(G_1)\exp \left( { - \frac{{2\tau _1}}{{T_2}} - \frac{{\tau _2}}{{T_1}}} \right)\exp \left[ { - (\gamma \delta G)^2D\left( {\Delta - \frac{\delta }{3}} \right)} \right],$$
(1)

where G1 represents the first (the smallest) value of the gradient ramp steps; τ1, τ2, δ, and Δ are delays, which are shown in Fig. 2; γ is the magnetogyric ratio of the observed nuclide; and T1 and T2 are the longitudinal and transverse relaxation times, respectively. Modern NMR spectrometers are generally equipped with a PFG system (probe and amplifier) generating a maximum G of several tens to 60 G/cm and a maximum δ of several milliseconds. To obtain 5% echo attenuation f(G)/f(G1), which is required for reliable DOSY experiments, a delay, Δ, of approximately 460 ms (Eq. 1) is required to determine a D of 10−11 m2/s under the following experimental parameters: a maximum G of 60 G/cm and a δ of 5 ms. A larger Δ is necessary as D decreases and as Mw of the solute polymer increases because the relationship between D and Mw roughly obeys D ∝ Mw−0.6. Applying a large Δ for DOSY experiments of AETAC-co-AAm severely reduces the f(G) intensity because T2 of the copolymer resonances is expected to be less than several tens of milliseconds (see Eq. 1 and Supplemental Material) owing to low segmental mobility, as described above. In our previous studies [37, 46], DOSY experiments of AETAC-co-AAm with Mw ranging from 5.1 × 106 to 1.5 × 107 were carried out using a diffusion probe system with a maximum G of 1200 G/cm. The D values determined by DOSY agreed with those determined by dynamic light scattering (DLS). However, the DOSY experiments gave broad distributions of D owing to the low signal-to-noise ratio (S/N) and poor resolution of the spectra. Recently, an inverse-geometry diffusion probe system (with a maximum G of 1800 G/cm) optimized to perform 1H NMR diffusometry was devised. In contrast to conventional (normal geometry) diffusion probes, which are useful for NMR diffusometry of low γ nuclei, the inverse-geometry diffusion probe is optimized for 1H DOSY experiments of high molecular weight polymers. For example, Lopez and coworkers reported 1H NMR diffusometry experiments for polystyrene sulfonates with Mw up to 3.4 × 105 in D2O using this inverse-geometry diffusion probe [47]. The Mw values of AETAC-co-AAm investigated in our present study are larger than those of polystyrene sulfonates by one order of magnitude. This inverse-geometry diffusion probe enabled us to obtain 1H DOSY spectra with greater resolution than those obtained in our previous studies.

Fig. 2
figure 2

The bpp-ste-led pulse sequence [45]. The encoding gradients of a stimulated echo (ste) sequence were applied as symmetrical bipolar pulse pairs (bpp) of the total duration, δ. The longitudinal eddy-current delay (led) sequence was extended with a delay period, Te

Full size image

Experimental procedure

Materials

AETAC was purchased from MT Aquapolymer Inc. (Tokyo). AAm and 2,2ʹ-azobis(2-amidinopropane) dihydrochloride (V-50) were purchased from Sigma‒Aldrich.

Preparation of polyacrylamide and AETAC-AAm copolymers

Radical polymerization was performed in water at 40 °C for 15 h with V-50 as the initiator. Polyacrylamide (sample A) and AETAC-co-AAm (samples B, C, D) with different monomer compositions and molecular weights were synthesized by changing the concentration of V-50 from 0.01 to 1.06 mol% in the monomer mixture. The reactions were stopped by cooling, and the polymers were isolated by repeated precipitation in acetone. Polyacrylamide and AETAC-co-AAm powders were obtained by vacuum drying at room temperature.

NMR measurements

The polyacrylamide and AETAC-co-AAm powders were dissolved in D2O (99.9 atom% D), 0.1 M NaCl in D2O or 1 M NaCl in D2O containing 0.05% trimethylsilylpropanoic acid to prepare samples of 0.2 wt% polymer or copolymer. The powders of sample A (or C) and sample D were mixed, and then the solvent was added. These sample solutions were transferred to a sample tube with a 5-mm outer diameter, and the liquid height was adjusted to 30 mm. DOSY measurements were performed using a Bruker AVANCE NEO 400 spectrometer equipped with the Diff BBI probe and the Diff-5-3 program implemented in TopSpin 4.1.0 software. For the measurements of the polymer samples in D2O containing 1 M NaCl at 30 °C, data were acquired using the bpp-ste-led sequence [45] (PULPROG: diffSteBp) with a gradient pulse length (δ) of 3.2 ms (gradient pulse shape: smsq), a diffusion delay (Δ) of 10 ms, an eddy current delay (Te) of 5 ms, and a gradient recovery delay (tg) of 0.5 μs. The parameters for the measurements under other conditions are given in the figure legends. All experiments were recorded using 128 scans for each gradient strength in 128 linear G lamp steps. The acquisition time was 0.84 s with a relaxation delay of 4.16 s using 8192 data points for the F2 dimension covering 4850 Hz. The transmitter frequency offset was adjusted to the frequency of the N(CH3)3+ resonance. An exponential window with a broadening factor of 3 Hz was applied prior to Fourier transform along the F2 dimension. Careful phase adjustment and baseline correction were performed. The CONTIN method [48, 49] implemented in the TopSpin 4.1.4 software was adopted for inversion of the Laplace transform along the F1 dimension.

SEC-MALS measurements

Size exclusion chromatography coupled with multiangle light scattering (SEC–MALS) measurements were performed in 1 M NaCl at 25 °C to determine the Mw, Mw/Mn, and root mean square radius of gyration (Rg) of the polyacrylamide and AETAC-co-AAm samples. The powders of these samples were dissolved in 1 M NaCl to give 0.2 wt% solutions. An MALS detector (WYATT DAWN HELEOS-II) and two SEC columns (Shodex OHpak SB-806 M HQ with a maximum porosity of 2 million and OHpak SB-807 HQ with a maximum porosity of 50 million) were used for the analysis. Aqueous NaCl (1 M) was used as the eluent, and the flow rate was 0.6 mL/min. A Wyatt Optilab T-rEx differential refractometer was calibrated using a series of 1 M NaCl solutions of known concentrations [50].

Dynamic light scattering measurements

DLS measurements of the polymer samples were carried out for 0.2 wt% solutions in 1 M NaCl at 30 °C using a Malvern Panalytical Zetasizer Nano ZSP system and a 10-mm polystyrene cell. The powders of polyacrylamide and AETAC-co-AAm were dissolved in 1 M NaCl to prepare samples of 0.2 wt% polymer or copolymer. The powders of sample A and sample D were mixed, and then the solvent was added. Inversion of the Laplace transform was conducted using the CONTIN method implemented in Zatasizer software (version 8.02).

Intrinsic viscosity measurements

The [η] of the polymers was measured in 1 M NaCl at 25 °C using a Cannon–Fenske viscometer (size 75). The intrinsic viscosity values were extrapolated from the linear plots of the reduced specific viscosity as a function of concentration [50].

Results and discussion

Table 1 summarizes the average AETAC composition, Mw, Mw/Mn, Rg, and [η] of samples A to D. It should be noted that the [η] values increased with increasing Mw. According to the work of Mabire et al., in a 1 M NaCl solution, [η] is affected by the ionic strength of the solution as well as the composition and Mw of the polymer, and Rg is affected by Mw but not by the composition of AETAC-co-AAm [4].

Table 1 Homopolymer of AAm and copolymers of AETAC and AAm prepared by radical polymerization at 40 °Ca
Full size table

Figure 3b shows the 1D NMR spectrum extracted from the DOSY data of sample B measured in D2O at 40 °C. The spectrum was acquired at the first gradient step, G1. The parameters for the DOSY experiment are shown in Table 2 as Run 2. According to the assignments by Proietti et al. [6], the broad resonances at 1.35–2.41 and 2.41–3.04 ppm can be assigned to the CH2 (1) and CH (2) groups of the copolymer main chain, respectively. We assigned the strong resonance at 3.24 ppm to the (CH3)3N+ group (5) and the resonances at 3.77 and 4.59 ppm to the CH2 groups of the AETAC side chain (3, 4). A comparison of this spectrum with that of Fig. 3a in our previous study [37, 46] (the parameters are shown in Table 2 as Run 1) reveals that the intensities of resonances 1–4 were higher than that of resonance 5. We mainly attribute this increase to the reduction of the diffusion time (Δ) from 100 ms (Run 1) to 10 ms (Run 2), which was made possible by the use of the Diff BBI probe. The T2 values for resonances 1–4 were smaller than that of resonance 5 (see Supplementary Material). Thus, using a large Δ leads to severe loss of the echo intensity for small T2 resonances (see Eq. 1).

Fig. 3
figure 3

1D NMR spectra extracted from the DOSY data of sample B in D2O at 40 °C. The spectra were acquired at the first gradient step, G1. The DOSY experiments were performed using the GR probe (previous study, Run 1) (a) and the Diff BBI probe (present study, Run 2) (b). See Table 2 for the parameters of these experiments

Full size image
Table 2 Parameters for the DOSY experiments of AETAC-co-AAm (sample B) used in our previous studies (Run 1) [37, 46] and in the present study (Run 2)a
Full size table

Figure 4 shows the DOSY contour plots with F1 (the D dimension) and F2 (the chemical shift dimension) projections obtained from the CONTIN analysis of samples A and D and a 1:1 mixture of A and D in D2O at 40 °C. The plots for sample A (Fig. 4a) show the resonances due to the CH2 (6) and CH (7) groups of the polyacrylamide main chain and the log–Gaussian distribution of D with the peak maximum (Dp) at a log D of −11.52 m2/s. In contrast, almost no resonances other than that due to the (CH3)3N+ group (5) of the AETAC side chain were observed in the plots for sample D (Fig. 4c). The log Dp was found at −12.33 m2/s. The plots for the mixture of samples A and D (Fig. 4b) showed resonances 6 and 7 of polyacrylamide and resonance 5 of AETAC-co-AAm. The distributions of D for samples A and D were well separated, with log Dp values at −12.09 and −12.27 m2/s, respectively. Polyacrylamide (sample A) is uncharged and has a flexible chain structure, whereas AETAC-co-AAm (sample D) is a polyelectrolyte, which we assume has a relatively expanded chain structure in D2O owing to electrostatic repulsion between the charged AETAC units. We consider that the expansion of AETAC-co-AAm chains reduces the segmental mobility of the main chain and makes T2 of resonances 1–4 very small. Even with a small Δ of 12 ms, which was achieved by the use of the Diff BBI probe, it seems difficult to observe resonances 1–4 from DOSY experiments in pure D2O.

Fig. 4
figure 4

DOSY-CONTIN contour plots with F1 (the D dimension) and F2 (the chemical shift dimension) projections of sample A (a), sample D (c), and a 1:1 mixture of samples A and D (b). The DOSY experiments were performed in D2O at 40 °C. Sample concentration = 0.2 wt%; Δ = 12 ms; δ = 5.4 ms; number of linear G steps = 64 (a) or 128 (b, c); G64 = 700 G/cm (a); and G128 = 1400 and 1420 G/cm in (b) and (c), respectively

Full size image

It should be noted that the log Dp for sample A in Fig. 4b (−12.09 m2/s) was significantly smaller than that in Fig. 4a (−11.52 m2/s). We attribute the decrease in Dp when sample A was mixed with sample D to the concentration of 0.2 wt% in D2O far exceeding the overlap concentration (c*). The individual polymer chains start to overlap as the concentration approaches c*, which is defined as the concentration at which a given dilute conformation’s pervaded volume is equal to the solution concentration [51]. The need to use sufficiently low-concentration samples to determine the correct D by diffusometry has been noted in the literature [22, 32, 52, 53]. However, DOSY experiments for AETAC-co-AAm samples at concentrations below 0.2 wt% were impractical in terms of sensitivity of the spectrometer in the present study.

Figure 5 shows the DOSY contour plots with F1 and F2 projections of samples A and D and the mixture of samples A and D in 1 M NaCl in D2O at 30 °C. In contrast with Fig. 4b, c, resonances 1–5 of AETAC-co-AAm were observed in Fig. 5b, c, and resonances 6 and 7 of polyacrylamide were observed in Fig. 5a, b. It was reported that the intramolecular electrostatic repulsion between charged segments by (CH3)3N+ in AETAC-co-AAm is screened by Cl− and that polymer chain expansion is not appreciable in D2O solution with salt [9]. As a result, the log Dp for sample D increased markedly to −11.78 m2/s (Fig. 5c). The distributions of D for samples A and D in Fig. 5b separated well, with log Dp values at −11.51 and −11.64 m2/s, respectively. These values agreed reasonably well with those determined by DLS-CONTIN analysis for sample A (−11.26 m2/s) and sample D (−11.58 m2/s) in 1 M NaCl in D2O at 30 °C (Fig. 6). However, we note that the D determined by DLS is a mutual-diffusion coefficient in contrast with the self-diffusion coefficient determined by DOSY. Separations of samples A and D along the diffusion axis determined by DOSY (Fig. 5b) and DLS (Fig. 6) compare unfavorably with the SEC separation, which showed a bimodal distribution of molecular weight corresponding to samples A and D (Fig. 7). We attribute the insufficient separation in the SEC curve for the mixture of samples A and D to overloading of the sample solution in the SEC-MALS instrument.

Fig. 5
figure 5

DOSY-CONTIN contour plots with F1 and F2 projections of sample A (a), sample D (c), and a 1:1 mixture of samples A and D (b). The DOSY experiments were performed in 1 M NaCl in D2O at 30 °C. Sample concentration = 0.2 wt%; Δ = 10 ms; δ = 3.4 ms; number of linear G steps = 128; and G128 = 1100, 1400, and 1650 G/cm in (a–c), respectively

Full size image
Fig. 6
figure 6

Distribution of D determined by DLS-CONTIN analysis of sample A, sample D, and a 1:1 mixture of samples A and D in 1M NaCl in D2O at 30 °C. Sample concentration = 0.2 wt%

Full size image
Fig. 7
figure 7

Molecular weight distributions of sample A, sample D, and a 1:1 mixture of samples A and D measured by SEC–MALS using 1 M NaCl in water as the eluent at 25 °C

Full size image

Figure 8b shows the DOSY contour plots with F1 and F2 projections of a 1:1 mixture of samples C and D in 0.1 M NaCl in D2O at 30 °C. The Mw of sample D (3.9 × 106) was twice as large as that of sample C (1.9 × 106), and the AETAC composition of sample D (80.9 mol%) was larger than that of sample C (59.4 mol%). However, the F1 (the diffusion dimension) projection showed no sign of separation or broadening arising from the two copolymers. Thus, it appears difficult to separate the diffusion profiles of these samples by DOSY analysis under the conditions of this study.

Fig. 8
figure 8

DOSY-CONTIN contour plots with F1 and F2 projections of sample C (a), sample D (c), and a 1:1 mixture of samples C and D (b). The DOSY experiments were performed in 0.1 M NaCl in D2O at 30 °C. Sample concentration = 0.2 wt%; Δ = 11 ms; δ = 3.8 ms; number of linear G steps = 128; and G128 = 1250, 1550, and 1500 G/cm in (a–c), respectively

Full size image

Our results demonstrate that the inverse-geometry diffusion probe system is useful for the DOSY-CONTIN analysis of polyelectrolytes with Mw over one million in D2O containing 0.1 to 1 M salts. This is particularly relevant to the study of sludge solutions. For example, the electrical conductivity of common sludge is approximately the same as that of 0.1 M NaCl aqueous solution (22 μS/m) [54]. Many studies have been conducted to elucidate the mechanism of sludge dehydration using polymer flocculants, although few have focused on the structure of these polymers [55, 56]. We expect that the application of the DOSY-CONTIN method will provide information on the structure and mobility of polymer flocculants in stimulated sludge solutions.