The Raman spectra of two forms of native celluloses, Ia and Ip, have been discussed in detail elsewhere (13), but are presented here in support of evidence for a helical model for native biological structures. Furthermore, these forms are presented to support findings from the molecular modeling program that two different rotamers of the exocyclic group at C6 occur in the native state. Finally, spectra are presented to illustrate the effects oflateral dimensions on spectral resolution.

The spectra presented in Figure 6.9 were recorded for samples of the alga Valonia ventricosa and the tunicate Halocynthia roretzi. The first is an alga wherein the fibrils of cellulose are 65% Ia and 35% Ip. The tunicate appears to be predominantly Ip. The importance of these spectra derives because fibrils in both instances are approximately 20 by 20 nm in lateral dimension. Thus, resolution of the spectra is sufficient to allow confident discussion of their interpretation. Two features of the spectra are noteworthy. The first supports the observation based on molecular modeling that two distinct rotamers of the exocyclic group at C6 occur in the native form, which is evident from the appearance of two distinct bands above 1450 cm-1. Presently, it is not possible to associate the individual bands with the corresponding rotamer, but there is little question that two distinct rotamers occur. The relative intensities of the two bands differ in the spectra of the two forms. In the spectrum of the predominantly Ia Valonia, the lower frequency band is higher than it is in the spectrum of the predominantly Ip Halocynthia. The occurrence of two bands above 1450 cm-1 is

unique to the two forms Ia and Ip. The spectra of celluloses II and III both have a single sharp band in this region (33).

Since questions have been raised within the cellulose science community regarding sen­sitivity of Raman spectra to molecular conformation, we present here some of the con­siderations that persuade us. Raman spectroscopy is a branch of vibrational spectroscopy complementary to infrared (IR). Raman spectra are no less sensitive to perturbations of molecular structure or environment than IR and are indeed better suited to studies of bio­logical systems because of the very low scattering coefficient of water. Both IR and Raman spectroscopy involve transitions between molecular vibrational states. The key difference is that Raman spectra are more sensitive to vibrational transitions involving highly covalent

bond systems, whereas IR spectra are more sensitive to transitions involving highly polar systems of bonds.

The basis for establishing sensitivity of Raman spectra to molecular conformation was es­tablished through extensive normal-coordinate analyses of six classes of model compounds related to saccharides (34-42). The most persuasive are analyses of vibrational spectra of the inositols (34, 39). The inositols are cyclohexane hexols, differing from each other only in the distribution of hydroxyl groups between axial and equatorial orientation and their positions relative to each other. Differences in their spectra leave very little doubt that for molecules where the skeletal structure is made up of C-C and C-O bonds, for which reduced masses, bond energies, and force constants are similar and the coupling of vibrational modes is high, the individual spectra are determined by organization of atoms in space relative to each other within the molecule.

For demonstration of the sensitivity of Raman spectra to conformation particularly in the context of celluloses, we present in Figure 6.10, Raman spectra of cellulose from Cladophera glomerata, which is a fresh water alga that has fibrils very similar to those of Valonia, both with respect to cross section and the balance between Ia and Ip. Spectrum (I) is for the alga in its native state after purification with an acid chlorite treatment; it is very similar to the spectrum of Valonia presented in Figure 6.9. Spectrum (IIII) is recorded after conversion to cellulose III by treatment in anhydrous liquid ammonia at -30° C; the very high level of order is retained by allowing the ammonia to evaporate gradually at ambient temperature.

Spectrum (Ini) is after recovery of cellulose I primarily in the Ip form from cellulose III by boiling in water.

Three comparisons in Figure 6.10 are noteworthy. First are the significant differences between the spectra of native cellulose I and IIII; the key difference between celluloses I and III are differences in conformation. Second, the spectrum of IIII recovers similarity to that of the alga but is now less well resolved because of fibrillation accompanying the conversion back to cellulose I first noted by Chanzy and coworkers (43). Lateral dimensions of fibrils of cellulose IIII range between 3 and 6 nm, which are in contrast to the lateral dimensions of the original Caldophera cellulose at 20 by 20 nm. The spectrum of IIII is almost indistinguishable from that of cotton. Thus, lateral dimensions of nanofibrils are clearly very important to resolution of bands in the spectra. Most of the bands that are very sharp with relatively low bandwidth for cellulose I from the native Caldophera with 20 by 20 nm fibrils are now broadened considerably in the spectrum of IIII. Chanzy and coworkers found a similar effect with SS 13C NMR spectra. Finally, the Raman spectrum of tunicate Ip cellulose shown in Figure 6.9 is much more similar to those of the Ia algal celluloses in Figures 6.9 and 6.10, than to that of the fibrillated Ip sample of IIII.

We conclude that spectra in Figure 6.9 confirm that conformations of the Ia and Ip forms are almost identical and that results of the diffractometric studies reflect two erroneous assumptions. The first, by Atalla and VanderHart that Ia and Ip are “two distinct crystalline forms” rather than simply “two distinct forms,” the second by the authors of diffractometric studies in further assuming that Ia and Ip belong to different crystallographic space groups, which is also in contrast to the findings of Sugiyama and coworkers (12).