Analysis of calcium salt of carboxymethyl cellulose: size distributions of parent carboxymethyl cellulose by size-exclusion chromatography with dual light-scattering and refractometric detection

Abstract

Generally, the insoluble calcium salt of carboxymethyl cellulose can be quantitatively transformed in a batch process into its sodium form using chelating resin Lewatit TP-208. The resulting sodium carboxymethyl cellulose, having a fairly low degree of substitution close to 0.6, was analysed using size exclusion chromatography with multiangle light scattering/refractometric detection. Commercial samples of sodium carboxymethyl cellulose having degree of substitution 0.7 and 1.2 were used to highlight the relation between the degree of substitution and the presence of aggregated structures. It was found that all calcium/sodium transformed carboxymethyl cellulose samples contain a significant amount of macrogel particles larger than 5 μm which can be removed by centrifugation or filtration. Centrifuged and filtered samples were then shown to contain significant amounts of smaller aggregated structures (fringed micelles) which were found also in sodium carboxymethyl cellulose reference samples where their amount was inversely proportional to their degree of substitution. Because size exclusion chromatography separates macromolecules according to their hydrodynamic volume, linear and aggregated structures having the same hydrodynamic volume, but entirely different molar mass, are eluted together and a significant heterogeneity in terms of molar mass at a fixed elution volume is found especially at lower sample elution volumes. Hence, molar mass distributions become undefined. Fortunately, the elution was found to be homogeneous in terms of experimentally accessible radii of gyration and reliable size distributions could be evaluated.

Introduction

Calcium salt of carboxymethyl cellulose (CaCMC) finds the main use in tablet formulations where it is used as a binder, diluent and disintegrant. Although CaCMC is insoluble in water, it is an efficient disintegrant because it swells to several times of its original volume on contact with water (Wade & Weller, 1994). Commercial products are mostly characterized only by some particle size of dry product and pH value of their aqueous dispersion. Being insoluble in any solvent, direct characterization of CaCMC in terms of molecular weight and size is impossible and can be done only after its quantitative transformation to some soluble form, preferably sodium salt. More detailed knowledge of their composition like calcium content and molar mass as well as particle size of parent sodium carboxymethyl cellulose is desirable in pharmaceutical industry because there is a close relation between these parameters and the CaCMC tablet technology/application performance. The preferred CaCMC disintegrant should have a high molar mass, together with a low average degree of substitution (DS); typical DS of CaCMC was reported to be 0.6±0.1 (Doelker, 1993). There is a danger of its incomplete solubility in water even if such a CaCMC derivative is transformed to its sodium form (NaCMC) because a minimum DS which guarantees complete water solubility should be at least 0.7 (Just & Majewicz, 1985). Highly swollen but insoluble gel particles/aggregates then appear in solutions of the samples having DS below 0.7 and their amount increases with decreasing DS.

Ion-exchange is a natural choice when CaCMC should be transformed to the sodium form. This requires to perform ion-exchange in the batch mode because of CaCMC insolubility in any solvent. Fortunately, chelating ion-exchange resins should have sufficient selectivity to accomplish quantitative Ca/Na exchange in a batch experiment. (Samuelson, 1963). Having CaCMC transformed to the soluble form, size exclusion chromatography may be used to characterize its molar masses and dimensions. The use of dual light scattering/refractometric detection appears to be a must here because the low degree of substitution predetermines the presence of both macrogel and microgel particles (Just & Majewicz, 1985) together with polymer coils formed from linear chains in the transformed samples. Such particle system is characterized by a variable mass/hydrodynamic volume ratio of the various particles present, which means that both molar mass and size information are needed to interpret SEC separation experiments where separation is governed by a hydrodynamic volume of the analyzed particles preferably expressed by its cube root (Potschka, 1987, Potschka, 1988, Potschka, 1991) in terms of a radius R_SEC. In the case of SEC of polyelectrolytes using packings of the same charge, both the effective hydrodynamic volume of a solute and the interfacial pore wall effect must be accounted for. In general, separation is governed by some overall radius R

$$R=R_{\text{SEC}}+R_{\text{IF}}=R_{\text{SEC}}+{\kappa }^{-1}\overline{a}$$

where $\overline{a}$ is the average electrostatic repulsion distance at equilibrium in multiples of Debye length κ⁻¹, R_SEC is the rotationally averaged mean radius of the solute and R_IF is the interfacial contribution to the total solute radius R. Actually, R_IF is introduced to account for the reduced pore size under low-salt conditions. For spherical particles like proteins, R_SEC should be simply the sphere radius but for the case of the coiled polyelectrolyte, it can be a complex function of ionic strength. The generally accepted universal calibration approach predicting congruent polymer elution irrespective of their composition and architecture then assumes negligible R_IF, i.e. high ionic strength conditions, and

$$R_{\text{SEC}}\sim {\{[\eta ]M\}}^{1/3}$$

where [η] is the intrinsic viscosity and M is the molar mass of a macromolecule in bulk solution. The Flory–Fox theory (Flory & Fox, 1951) predicts that

$$[\eta ]M=\Phi {[{〈r_{\text{g}}^2〉}^{1/2}]}^3$$

where ${〈r_{\text{g}}^2〉}^{1/2},$ mostly abbreviated as r_g, is the root mean square radius (commonly denoted radius of gyration) of a macromolecule and Φ should be a universal proportionality constant originally assumed to be valid for flexible uncharged polymers soluble in organic solvents if their molar mass exceeds a minimum value around 50 000. Assuming validity of Eq. (3), R_SEC should be proportional to r_g as another universal calibration parameter. Nevertheless, the universality of r_g must be always verified experimentally before use, because it is highly questioned especially for semirigid polymers soluble in aqueous solvents, mostly due to observed variations of the Φ value with their architecture and/or molar mass (Radke, 2001, Ioan et al., 2001)

It will be shown in this paper that CaCMC can be quantitatively transformed into its sodium salt form, which makes possible SEC analysis. NaCMC polymers having variable DS will be used as model systems and compared to transformed CaCMC samples. An evidence of the presence of both macrogels and fringed micelles in the transformed samples will be presented. Guidelines how to interpret molar mass and size data when the samples contain linear macromolecules and non-negligible mass amounts of aggregates of the same hydrodynamic volume will be described and the impact of their low DS on the data will be evaluated.