Position-Specific Substitution in Cellulose Ethers Studied by DNP Enhanced Solid-State NMR Spectroscopy

Abstract

Ethyl hydroxyethyl cellulose (EHEC) and methyl ethyl hydroxyethyl cellulose (MEHEC) are hydrophilic cellulose ethers commonly employed as rheology modifiers in diverse industrial applications. The performance of these polymers, and their resistance to degradation by various cellulase enzymes, depends on their intricate molecular structure. Distribution of the etherifying groups, within the anhydroglucose units and along the polymer chain, is the key property to control. However, characterizing such structural properties is challenging, necessitating the development of novel analysis methods. In this study, we demonstrate the application of solid-state nuclear magnetic resonance (NMR) spectroscopy, enhanced by dynamic nuclear polarization (DNP), for this purpose. We prove that the hydrophilic EHEC and MEHEC samples are homogenously swelled in D₂O/H₂O-based radical solutions, a necessity to ensure uniform DNP enhancement throughout the material. And we illustrate how the high sensitivity enhancements obtained can be used to perform selective, J-coupling-based C1 to C2 transfer experiments to measure the fraction of substituted C2 positions in these cellulose ethers. Moreover, with further refinement, the methodology outlined in this work holds promise for elucidating C3-specific substitution patterns.

Keywords: DNP; EHEC; MEHEC; cellulose ethers; solid‐state NMR; substitution.

FIGURE 1

Molecular structures of possible repeating units of ethyl hydroxy ethyl cellulose (EHEC) and methyl ethyl hydroxy ethyl cellulose (MEHEC). In MEHEC, additional methyl groups are introduced, making molecular structure more complex.

FIGURE 2

DNP enhanced ¹³C CP/MAS and ¹³C detected ¹H saturation recovery data for the EHEC2 sample swelled in 12‐mM AMUPol D₂O:H₂O (9:1) radical solution. (a) ¹³C CP/MAS spectra illustrating the large sensitivity enhancement when the microwaves are turned on/off. (b) ¹³C detected ¹H saturation recovery data acquired with microwaves turned on, sodium formate was added to the radical solution, and the curves illustrate the practically identical buildup time detected on the carbonyl signal (CO) of the formate ion and the C1 carbon (C1) of EHEC. (c) ¹³C detected ¹H saturation recovery data acquired with microwaves turned off, the buildup times differ from the ones observed with the microwaves turned on but are also here, practically the same for CO and C1.

FIGURE 3

¹³C–¹³C DNP‐enhanced refocused INADEQUATE spectrum of EHEC2. (a) 2D spectrum, it is easy to spot the C1/C2 correlations and chemical shift change of the substituted and free C2 positions, located around 75 and 84 ppm in the direct (horizontal) single quantum (SQ) dimension and around 180 and 189 ppm in the indirect (vertical) double quantum (DQ) dimension. (b) 1D slices of the rows in the 2D spectrum that correspond to the C1/C2 correlations, where peak integrals are indicative of substitution at the C2 position.

FIGURE 4

1D, J‐coupling‐based correlation experiment for quantification of C2 substitution. (a) Spectrum from selective transfer experiment of the EHEC2 sample at MAS = 10 kHz, T = 104 K, and 12‐mM AMUPol D₂O:H₂O (9:1) radical solution and 2048 scans/2 h experimental time. The pulse sequence with selection step and perfect echo is shown. (b) Spectrum from the control experiment in which the π/2 transfer pulse has been removed from the perfect echo. The spectrum thus only contains T′ ₂ decayed artifact signal indicated by the small black arrow. (c) Difference spectrum (transfer experiment (a)—control experiment (b)), here, the signals are assumed to come only from the J‐coupling transfer and are thus integrated. The signal in spectral region 90.1–80.4 ppm is considered to originate from substituted C2 and the signal in the 80.4–71.9‐ppm region from free, nonsubstituted C2. The percentage of the integral relative to the integral of the whole region is shown. A dashed line indicates the baseline.

FIGURE 5

DNP‐enhanced spectra from the 1D transfer experiment (no control spectra subtracted) acquired on the MEHEC2 sample with different τ delay in the spin‐echoes, showing a clear buildup of signal for the 4‐ms delay. The buildup of signal and small difference in proportion of the C2_sub correlation signal supports the claim that differences in J _C1C2 and T′ ₂ are not a major problem affecting the results. The inner panel shows a zoomed‐out view of the spectra, illustrating the T′ ₂ decay of the large C1 signal.