Chroma Class 103

J.H. Kowalski, R.A. Mateer, R.A. Beck


The functional properties of many carbohydrate polymers used in food, pharmaceutical and other industries often depends on molecular branching.

Polymeric carbohydrates such as dextrans, glycogen, amylopectin, pectin and synthetic polyglycans commonly exist as highly branched or dendritic structures. Taken as individual molecules, such structures typically display one reducing sugar end and a multiplicity of non-reducing ends. Thus, molecular branching is a function of molecular sites between the reducing and non-reducing ends of the molecule that undergo branching.

While molecular weight, polydispersity and RMS Radii data help define a polymer, quantification of its unique branching attributes can be challenging. However, branching characteristics of carbohydrate polymers may provide pivotal characterization criteria for:

  • Structure and function correlation
  • Intellectual property claims
  • Specifications for a standard of identity
  • Manufacturing process control.


Size exclusion chromatography (SEC) used with high pressure liquid chromatography (HPLC) and multi-angle laser light scattering (MALLS) can detect relative evidence of high or low amounts of branching that is unique to a polymer.

The mechanism of separation for SEC is simple—larger elutes faster. An SEC column contains porous beads that will entrain smaller particles longer. But a pure SEC separation is rare, since other things can affect the elution of particles.

This creates a situation for analysts where they may be able to obtain “relative” differences between materials (holding all things equal), but absolute molar mass (MM) is unobtainable, and in the worst case, the samples are not very much alike and the information obtained from SEC is misleading.

Multi-angle laser light scattering (MALLS) is a technique that corrects for many of the errors inherent in SEC. Without delving too deeply into the mathematics involved with classical light scattering, there is a direct correlation between the molar mass of a polymer molecule and the way it interacts with electromagnetic radiation.

The intensity of scattered light is proportional to the concentration and molar mass of a particle. If one measures concentration with an additional detector (RI/UV), the molar mass is easily solved for. The calculations are complicated by the fact that the measured angle becomes important if the particles are greater than a certain size (~λlaser/20). By measuring many angles, this angular dependence must be accounted for, but it allows one to extract additional information from the light scattering data.

The comparison of the intensity of scattered light at each measured angle is used in a Zimm plot which extrapolates to zero angle (100% scattered light regardless of particle size), and zero concentration. This plot gives a direct measure of a mathematical size of the polymer (RMS Radius).

When a polymer is polydisperse, the plot of MM vs. elution time gives a profile of the molecular weight distribution. The plot of RMS Radius to elution time does the same for size. By combining these two plots, we get some very useful information.

The rate of growth for a polymer in terms of size to molar mass changes differently depending on the conformation. A rod-shaped polymer would have the greatest slope followed by a random coil, and then a sphere (e.g., globular protein).

A branched polymer of a given molar mass is expectedly smaller in size than a polymer that is linear. When the molar mass increases, the slope of the plot “size vs. MM” is different. A linear material has a steeper slope for that plot than a branched molecule would. The occurrence of parallel lines would suggest that two materials were of equivalent degrees of branching, but that their ÒdensityÓ is different. Perhaps the degree of solvation is different or there is a difference in functionality that would cause more or less electrostatic forces within the molecule.

MALLS and Branching

If one could take a linear material and plot RMS Radius vs. MM and do the same for an analogous branched polymer, one would expect a different slope. This implies that if one could determine the actual branching number independently for a given material at several different degrees of branching, there should be a way to correlate the degree of branching to the slope of the RMS Radius vs. MM plot.

Plot of RMS Radius vs. MM shows the dependence of size on molar mass.
Plot of RMS Radius vs. MM shows the dependence of size on molar mass.

The number of branches per molecule can be estimated by MALLS with the use of a linear analog. This is not possible with some natural polymers that do not have linear materials that are identical in structure.

Let us consider a process where a material is being modified. The molar mass is being changed, and we do not know if complete side-chains are being truncated from the molecule or just parts. The degree of branching may or may not be affected. If we analyzed the material during different steps or conditions during this process, it stands to reason that by using the initial sample as a benchmark, we would be able to see changes in the slope of RMS Radius vs. MM, which would indicate change or no change.

If we analyzed the samples with an additional method to absolutely characterize the number of branches, we should be able to create a shortcut by correlating the branching to the slope. That would allow a quick method (i.e., MALLS) for the determination of many properties of the processed material, namely MM distribution, RMS Radius distribution, and branching.

Carbohydrate Tertiary Structure

There are many conformations of polymers as they exist in solution. This fact creates the need for characterization by a method that is not hampered by assumptions. Conventional size-exclusion chromatography (SEC) requires the use of standards that have the same properties as the material of interest. This is very difficult when working with biopolymers and carbohydrates since the makeup of these may not be completely known. There may be varying functional groups affecting column interactions, or conformational differences which change the relative hydrodynamic size of the polymer. MALLS not only correctly calculates molar mass, but gives information about the tertiary structure and branching.

Linear and branched starch molecules
A. Linear starch molecule
B. Branched starch molecule

This diagram shows three possible configurations of a polymer of the same molar mass.

Diagram of three possible configurations of a polymer
A. Random coil — linear polymer
B. Random coil — branched polymer
C. “Rod”-like configuration — branched polymer

The polymer with the smallest hydrodynamic radius would be the random coil branched material. The largest would be the “rod”-like material. All three would have different elution times even though they have the same molar mass.

Obviously, for branched polymers it is difficult to determine the type of standards needed to produce a similar elution profile by SEC.

Linear vs. Branched Carbohydrate

This chromatogram shows the near overlay of a 410 kD dextran sample with a carbohydrate with a MM of 64 kD at the apex. This clearly shows the difference in conformation. These peaks would separate well if the conformations were similar.

Chromatogram of a 410 kD dextran sample and a carbohydrate with a MM of 64 kD

Branched Carbohydrate

Here is an example of two carbohydrates with different degrees of branching showing different slopes for RMS Radius vs. Molar Mass.

Chromatogram of two carbohydrates with different degrees of branching showing different slopes for RMS Radius vs. Molar Mass

Determination of Percent Branching

Determination of Percent Branching. Polymers constructed of monosaccharides including their respective acids, amine-derivatives, and so on, commonly result in polymeric homo- or heteroglycans with varying degrees of branching. Analytical determination of the total molar concentration of all repeating monosaccharidic subunits in a branched polymer can be achieved by classical methods involving anthrone reagent. Galacturonic acid polymers may require a comparable reagent such as carbazole.

Analytical Rationale. Using the example in Table 1 (see PDF), the polymer has a dendritic structure initiated at a single α1→4 linkage to a second glucose residue and so on. Each site of branching over the polymer requires a α1→6 glycosidic linkage. There is only one reducing sugar or reducing end (RE) displayed by the polymer because the anomeric carbon (carbon 1 or C1) is free. However in the case of every branching site over the polymer there are two non-reducing sugar residues.

Considering this, the number of α1→6 branches is one less than the number of non-reducing ends (N) in the polymer. Based on the schematic structure (Table 1), the number of branches is 15 and the number of non-reducing ends (NREs) is 16. For all practical intents where molecular weights are high, the number of α1→6 linkages approximates the number of reducing ends. Thus, the percent branching or percent α1→6 glucose residues in a polymer can be estimated from:

Branching (%) = [α1→6 linkages in sample] /
[total monosaccharide residues in sample]

Since the number of α1→6 linkages is equal to the number of non-reducing ends, this value can be determined by periodic acid oxidation (PAO). PAO specifically reacts with non-reducing ends of the polymer. This releases formic acid (HCOOH) whose concentration parallels the number of non-reducing ends. The prerequisite for PAO is the presence of vicinal groups on adjacent carbon atoms that include –OH and –OH, –OH and –NH2, =C=O and =C=O or =C=O and –CHO.

Table 1. Schematic diagram for a branched polysaccharide
Table 1. Schematic diagram for a branched polysaccharide modeled after a non-specific homopolysaccharide such as amylopectin or glycogen that is composed of a monosaccharide [♦] such as glucose. The model has one initial monosaccharide residue with a reducing-end (RE) engaged in a α1→4 link to a second residue shown in structural tier “A”. The monosaccharide in tier “A” engages in an ordinary α1→4 linkage to another sugar residue plus an α1→6 linkage to a second sugar residue in the polymer. Both of these sugar residues are indicated in tier B. Each tier of the polymeric structure where a branch occurs will feature one respective α1→4 branch and one α1→4 linkage to other glucose residues in the polymer. Tier J represents most terminal non-reducing residues of the dendritic polymer. As nonreducing saccharidic residues they are subject to stoichiometric reaction with sodium periodate and quantification to assess branching.


The combination of MALLS and periodate oxidation for branching analysis seems to be a useful method for measuring differences in similar materials. There are no simple ways to measure branching without using a linear analog as a reference.

Taking the percent branching number from the periodate test and correlating it to the MALLS data gives a simple means of calculating differences between products in a process, or between starting materials that have the same makeup.


Kratochvil P. Classical Light Scattering from Polymer Solutions, Elsevier, New York, 1987.

Provder T. Chromatographic Characterization of Polymers — Hyphenated and Multidimensional Techniques, American Chemical Society, Washington, DC, 1995.

Boureng HG and Lindberg B. Methods in structural polysaccharide chemistry. Advances in Carbohydrate Chemistry. M. L. Wolfrom. ed., 1960.

Dyer JR. Use of periodate oxidation in biochemical analysis. Methods of Biochemical Analysis. D. Glick ed., Vol. 3:1956.

Montgomery R and Smith R. End group analysis of polysaccharides. Part IV. End group determination by periodate oxidation. Methods of Biochemical Analysis. D. Glick ed., Vol. 3:1956.

Rendina G. Experimental Methods in Modern Biochemistry. W. B. Saunders Co., Philadelphia. 1971. (Pp. 161-166).

See all our Chroma Classes:

Chromaceutical Advanced Technologies, Inc.