Fractionation of Cellulose


All cellulose samples obtained from natural sources exhibit large polydis-persities with regard to their molar mass. This feature does not only impede basic research (where experimental observations cannot be attributed to a certain chain length), it may also constitute an obstacle for the industrial application of cellulose derivatives (where too short or too long chains turn out to be harmful for realization of desirable effects). In the past we have made attempts to fractionate cellulose by means of readily soluble derivatives for which the substituents can be removed quanti-tatively after fractionation. Presently we are fractionating unsubstituted cellulose either through fractional dissolution of activated "solid" cellulose or from solution.

Cellulose derivatives

We have used trimethylsilylcellulose (TMSC) (which was highly substituted to avoid fractions of polymer chains with different degree of substitution (DS)) for that purpose and a mixed solvent, composed of toluene and dimethylsulfoxide.

GPC diagram

Fig. 1. GPC diagram of starting material and of different fractions obtained by the fractionation.

The preparation of larger amounts of material by means of the Continuous Polymer Fractionation (CPF) turned out to be rather time consuming and not very efficient, so that we are investigating other ways.

Fractional dissolution

We have tried to extract the lower molecular weight fractions of unsubstituted cellulose by means of the mixed solvent DMAc+LiCl containing of insufficient amounts of LiCl.

fractional dissolution

Fig. 2. Weight percent of cellulose extracted by means of the mixed solvent DMAc+LiCl upon a stepwise augmentation of the LiCl content.

Intrinsic viscosities of  cellulose fractions

Fig. 3. Intrinsic viscosities (in alkaline aqueous solution of ferric tartaric acid complex at 25 C) of the cellulose fractions, obtained by pressing the extracts through a spinning nozzle into water.

Fractionation from solution

Special solvents:

Starting from a 3 wt% solutions of cellulose in Ni-tren (dihydroxo-tris (2-aminoethyl) amine-nickel(II)-complex) we have tried to precipitate the polymer by a dropwise addition of 0.2 N H2SO4, although there is a separation with respect to chain length, the method is not useful because of polymer degradation. For this reason Ni-tren could be used only when absolutely necessary (for example for gravimetric analysis) and kept the time the polymer is in contact with this solvent as short as possible.


Low molecular weight cellulose can be fractionated from solutions in DMAc+LiCl using acetone as a precipitant. This system is, however, not apt for long chains because of the high viscosity of the solutions.

Polymer incompatibility:

By another interesting route, suggested by the early work of Tompa, consists in fractionation of polymers by means of liquid/liquid phase equilibria triggered by a second polymer instead of a non-solvent. It works because the phase separation of a solutions of polymers A and of polymer B, both exhibiting a broad molecular weight distribution, in a common solvent is associated with an accumulation of the short chains of A in the B-rich phase, whereas the low molecular weight components of B are preferentially incorporated into the A-rich phase. We have examined the quasi-ternary system (DMAc+LiCl)/PMMA/Solucell

Phase diagram of the system DMAc+LiCl /Cellulose

Fig. 4. Part of the phase diagram of the system DMAc+LiCl /Cellulose (Solucell)/PMMA (Open circles: cloud points; full squares: overall compositions at which fractionation experiments have been performed).

From the data collected in Table 1 it can be seen that the shorter cellulose chains accumulate in PMMA-rich phase, while the longer ones remain in cellulose-rich phase. The fractionation of PMMA proceeds analogously.

Table 1. Intrinsic viscosities of the cellulose fractions (in alkaline aqueous solution of ferric tartaric acid complex at 20 C)

Over-all composition Cell. rich phase ('') PMMA rich phase (')

Phase equilibrium

wt %
wt %
[η] Cell
[η] Cell
1 1.5 0.6 480 220 350 350
2 1.5 1.2 360 300 180 340
3 1.5 1.0 400 300 180 340
4 1.0 1.0 460 300 410 410
5 1.0 1.2 540 280 480 360
6 2.0 0.6 530 270 460 350
7 0.5 1.4 410 330 ----- 390


The orienting experiments reported here demonstrate that a stepwise extraction of cellulose and fractionation of cellulose operating on the basis of polymer incompatibility in the described manner constitutes a promising method for the fractionation of unsubstituted cellulose. Presently it is unclear which of the options is more efficient. Considerable additional work is required to investigate the potentials for an optimization of the different processes and to develop suitable fractionation strategies for the production of larger amounts of cellulose with narrow molecular weight distribution.