Modifications in solvent clusters embedded along the fibers of a cellulose polymer network cause paper degradation

Plants, algae, and their derivatives (paper, textiles, etc.) are complex systems that are chiefly composed of a web of cellulose fibers. The arrangement of solvents within the polymeric structure is of great importance since cellulose degradation is strongly influenced by water accessibility and external agents. In this paper we develop a model that is able to deconvolve the scattering contributions of both polymeric structures and solvent clusters trapped along the polymeric fibers. The surface morphology of cellulose fibers and the spatial distribution of water-filled pores and their dimensions have been recovered from small angle neutron scattering and atomic force microscopy data in function with paper degradation. In addition to providing a boost to the effort to preserve cellulose-supported material (included cultural heritage), the relevance of our model resides in the exploitation of a large number of biopolymer networks that are known to share structures similar to that of cellulose.

Figure 1

Image of the sample $M3$, without prior preparation, has been obtained by using a Zeiss EVO40 variable pressure electron microscope. A random network of many thick fibers tangled together to form a complex structure can be observed. Fibers appear to be inhomogeneous; in addition, the diameters’ size distribution does not appear to be uniform. The void spaces among the fibers are also inhomogeneous and, in some cases, can be very large, measuring up to several tens of microns.

Figure 2

Idealized structure of paper made out of cellulose fibers. The cellulose network, of mesh size $\zeta$, is sketched as a collection of cylindrical segments bonded to each other (a), with an average diameter $d$ and an average length $\ell$ (b). Each cylindrical segment, representing the cellulose fiber, is in turn composed of a collection of densely packed microfibrils. Small spherical clusters of radius $R$, representing trapped solvents, are localized in volumes of size $\xi$ that are intercalated along the axis of the cylinder (c).

Figure 3

The scattering intensity profiles, as simulated from the model Eq. (17), are reported according to the Krakty plot (a) or multiplied by $q^3$ (b). The parameters used in the simulation are $A_1=0.00154$, $\alpha =3.81$, $A_2=0.148$, $d_f=1.55$, and $\xi =20\text{nm}$, while $R$ varies from 1.4 (dashed-dotted line) to 2.1 nm (continuous line) with 0.1 nm steps. Shown in the inset is the behavior of $R$ in function with the inverse of the peak position in both cases (open square and open circles).

Figure 4

Neutrons scattering intensities profiles subtracted from the incoherent background and multiplied by $q^3$ for samples $M{2}_D$, $M{2}_H$, and $M{2}_{HD}$. The peak at $\sim 1.15{\text{nm}}^{-1}$ is very evident for $M{2}_D$ and $M{2}_H$ samples, while it almost disappears in case of sample $M{2}_{HD}$ sample.

Figure 5

Neutrons scattering intensities versus scattering vector $q$ from all samples exposed to ${\text{D}}_2\text{O}$ atmosphere (listed in Table ) are reported. Data are arbitrarily shifted for clarity. Best-fit curves are represented by continuous lines (see text for details).

Figure 6

The mean radius of the cluster is reported for all the samples investigated. Modern, artificially aged, and the ancient samples are plotted separately to make comparison easier. The cluster radius increases with the natural aging. The presence of a layer of coating (gelatin) on the fibers surface appears to protect against the increase of the cluster radius.

Figure 7

The mass fractal dimension $d_f$ for all samples exposed to the ${\text{D}}_2\text{O}$ atmosphere is reported. The dependence of $d_f$ on aging suggests that the solvent clusters within fibers rearrange their relative positions in a less compact structure.

Figure 8

Representative AFM images of $F6$, $W4$, and $N1$ samples are reported. Fibers appear to be composed of an assembly of densely packed protofibrils, which bestow a certain degree of roughness. The good agreement between neutrons scattering intensities profiles (open circles) and the intensities profiles, as obtained from different scales AFM images analysis (crosses), strongly supports the model we suggest and also allow for the calculation of the surface fractal dimension.