Allomorphs of cellulose

Isogai, A. (1994). Allomorphs of cellulose and other polysaccharides. In Cellulose Polymers Blends and Composites, Gilbert, R. D. (Ed.), p. 21. Hanser Verlag, Munich. 

Yet another set of interdisciplinary studies are represented by the work of Hayashi and coworkers, wherein they attempt to shed light on the questions of reversibility, or lack thereof, in transformations between the allomorphs of cellulose and its derivatives. 

In addition to their diffractometric studies reported in prior publications, they add in their contribution to the present symposium analyses of the infrared spectra as well as analyses of the CP-MAS 13c NMR spectra. Their thesis is not inconsistent with the proposals of Atalla and coworkers concerning differences between the conformations of celluloses I and II. However, Hayashi and coworkers go beyond this by proposing that the differences in conformation can be preserved in the course of heterogeneous derivati-zation reactions, and also in the process of generating the other allomorphs of cellulose, namely celluloses III and IV, from the two primary allomorphs 1 and II. 

The crystalline allomorphs of cellulose differ from each other also by the shapes of the crystalline unit cells. The projection of the Cl-unit cell has a parallelepiped shape, CIV-unit cell has a square shape, while unit cells of Cll and CIII have a rhombic shape. 

There are two forms of cellulose, cellulose I and cellulose II. Cellulose 1 is found in nature and is composed of parallel chains [16]. Cellulose II, the more stable form, is composed of antiparallel chains. There are two distinct allomorphs of cellulose I, 1 and I [17]. The crystal structures of these molecules have been determined [18, 19]. The allomorph I has 1-chain triclinic xmit cell and 1 has 2-chain monoclinic unit [20]. Cellulose I is metastable and is readily converted to 1. The ratio of cellulose I and 1 ... 

Fig. 7 Two allomorphs of cellulose I (parallel) and cellulose II (anti-parallel) structures. Choroselective polymerization of fi-CV produced cellulose I or cellulose II...

Figure 7. Relationship between various allomorphs of cellulose.

Cellulose III. Cellulose III results from treatment of cellulose with Hquid ammonia (ammonia mercerization) or amines. Cellulose III can be made from either Cellulose I or II. When treated with water. Cellulose III can revert to its parent stmcture. Some cellulose III preparations are much more stable than other preparations. The intensities on diffraction patterns from Cellulose III differ slightly depending on whether the Cellulose III was made from Cellulose I or II, and thus these allomorphs are called IIIj or IHjj- Workers studying III concluded, based partiy on the results of I and II, that the packings of IIIj and IIIjj are parallel and antiparallel, respectively (67). IIIjj also is thought to have hydrogen bonds between the corner and center chains. 

One of the most special aspects of cellulose polymorphy is the transformation from I to II. The conversion of the parallel-packed cellulose I structures to an antiparallel cellulose II structure is interesting because it can occur without loss of the fibrous form. This transformation is widely thought to be irreversible, although there are several reports [231-233] of regenerated cellulose I. The observation that there are two different forms of cellulose III and of IV is also remarkable. The two subforms of each allomorph have essentially identical lattice dimensions and at least similar equatorial intensities. Other intensities are different, particularly the meridional intensities, depending on whether the structures were prepared initially from cellulose I or II. The formation of the III and IV structures is reversible and the preceding polymorph (I or II) results. 

Artificial cellulose showed the cellulose II allo-morph, a thermodynamically more stable form with an antiparallel structure, by X-ray diffraction study, when a crude celluase was employed for the enzymatic polymerization.123 The other allomorph cellulose I is a thermodynamically metastable form with a parallel structure, which living cells normally produce, but was believed impossible to be realized in vitro. Interestingly, however, the in vitro synthesis of cellulose I was successfully achieved by using a purified cellulase.125 The molecular packing of glucan chains in a crystal is affected by the purity of the enzyme as well as the enzymatic polymerization conditions. A novel concept choroselectivity was therefore proposed, which is concerned with the intermolecular relationship in packing of polymers having directionality in their chains.126... 

One of the discoveries growing out of the early diffractometric studies of cellulose was that it can occur in a number of allomorphic forms in the solid state, each producing distinctive X-ray diffractometric patterns. In addition to the cellulose II form, which has been discussed extensively, two other forms have been recognized these are cellulose III and cellulose IV. It is of interest to consider them briefly because they reflect the capacity of cellulose to aggregate in a wide variety of secondary and tertiary structures and because some of the higher plant celluloses produce diffraction patterns that are not unlike those of cellulose IV. Furthermore, they reflect the tendency for some of the celluloses to retain some memory of their earlier states of aggregation in a manner not yet understood. 

Quite early in the x-ray diffractometric studies of cellulose it was recognized that its crystallinity is polymorphic. It was established that native cellulose, on the one hand, and both regenrated and mercerized celluloses, on the other, represent two distinct crystallographic allomorphs (14). Little has transpired... 

In the first detailed comparison of the Raman spectra of celluloses I and II, it was concluded that the differences between the spectra, particularly in the low frequency region, could not be accounted for in terms of chains possessing the same conformation but packed differently in the different lattices (33). As noted above, that had been the general interpretation of diffrac-tometric studies of Che two most common allomorphs. The studies of the Raman spectra led Co the proposal Chat two different stable conformations of the cellulose chains occur in the different allomorphs. 

When cellulose trinitrate Ij(TNG Ij) and triacetates Ij and IIj (TAG Ij and IIj) were prepared from allomorphs of the I family in the fibrous state under low-swelling conditions, they could be saponified into cellulose I. On the other hand, TNG Ijj, TAG Ijj and TAG IIu can be saponified into cellulose II (A-6). When the esters of the I family were recrystallized by heat treatment, they were transformed irreversibly into corresponded esters (for example, TAG Ij to TAG Iji). They crossed the barrier of irreversibility and were saponified into cellulose II. 

Atalla and Van der Hart (11, 12) concluded, based on their Raman and NMR spectra, that the molecules in cellulose I and II have different conformations. Based on x-ray analyses, Sarko et al. (13i H) and Blackwell et al. (15, 16) both concluded that crystal structures of cellulose I and II were based on parallel and antiparallel packing, respectively, of chains that have similar backbone conformations. Sarko (17) concluded that the allomorphs in the I and II families were based on parallel and antiparallel chains, respectively. The irreversibility may arise from the increase in entropy when parallel packing is converted to antiparallel packing. 

In the present work, we tried to determine which proposal is better using IR and solid state 13C NMR. There are many papers on the spectra of cellulose I and II (12, 18-24.), however, there are few on the other allomorphs. Mann and Marrinan (21) found differences in the OH stretching bands of IR spectra of IIIj and IIIn and of IVj and IVjj. Chidambareswaran et al. (25) reported IR spectra for several allomorphs, but their spectra lacked enough resolution for detailed discussion of chain conformation. The 13C NMR spectra of the other allomorphs have not been reported. 

Fig. 1 X-ray diffractograms of the allomorphs in cellulose I family by the reflection method (the reflection plane is parallel to the membrane surface). A I, valonia cellulose, B IVj prepared from A through IIIi(C), C IIIi prepared from valonia.

Table I. 0-H Stretching with Parallel of Cellulose Allomorphs...

Table IV. Glucose Ring Stretching of Allomorphs in Cellulose I Family ...

Recently, we have obtained better NMR data for the allomorphs with a Bruker instrument at 200 MHz for protons. The Cl signal of IIIj seemed to be a singlet, but its profile was broadened unsymme-trically to the lower ppm side and suggested additional weak peaks. The half-width was almost the same as that of cellulose I. The common characteristics related to half-width of the Cl signal within each families were clearly evident. 

Dudley et al. (29) suggested the observed doublets of Cl and C4. in cellulose II were due to independent chains in the unit cell. Cael et al. (30) explained the 13C NMR spectrum of cellulose I with the eight-chain unit cell using Dudley s proposal. But the proposal could not explain the similarity of the Cl and C4. signals of the allomorphs in each of the families. They should show similarity between IIIi and IIIii or IVi and IVn because their chain packing should be similar to each other. When the high resolution NMR spectra of these are measured the problem will be more clear. 

Two classes of experiments were conducted. In both sets of experiments, fibers in which the cellulose chains are oriented parallel to the fiber axis were used. In the first class of experiments, the plane of polarization of the incident light was changed relative to the axis of the fibers by rotating the fibers around the optical axis of the microscope (see Figure 2a). The dependence of the band intensities on the polarization of the incident light was studied to determine the directional character of the vibrational motions. This information was used to advance the assignment of the Raman spectrum of cellulose. Spectra from Valonia, ramie, and mercerized ramie fibers, which have different allomorphic compositions, were compared to study the structural differences between the allo-morphs. 

---