The structures of different celluloses (microcrystalline and native) were characterized by diffuse reflectance infrared spectroscopy (DRIFT), analyzing the bands due to O−H stretching, C−O stretching, and C−H bending modes. The respective crystallinities were determined as 73 ± 2% and 40 ± 2%, and the crystalline structure in both was characterized as type Iβ. A comparative study of adsorption of an aromatic ketone, benzophenone, on these two substrates was made, using the same technique. Through the modifications observed in the carbonyl stretching band of benzophenone, it was possible to distinguish adsorption in three different environments:  entrapped between the glycosidic chains in crystalline domains, in amorphous domains, and as crystallites adsorbed at the cellulose surface. By deconvolution of spectra and subsequent analysis of the main components' relative intensities, it has been possible to calculate the approximate amounts of entrapped molecules per gram of cellulose as a function of total benzophenone concentration. These results show that, for low benzophenone concentrations, adsorption occurs preferentially in crystalline domains of cellulose, independently from its crystallinity. On the contrary, as concentration is increased, crystallinity has a determinant role on the total amount of adsorbed benzophenone and on its molecular distribution in the different domains of cellulose.

Benzophenone and fluorenone, which have a nonrigid and a rigid structure, respectively, were used as probes to study the nature of the adsorption process onto microcrystalline cellulose. Diffuse reflectance techniques were used in the UV−vis and infrared regions. Luminescence studies revealed that whenever fluorenone or benzophenone are entrapped into the natural polymer chains and in close contact with the substrate, a strong quenching effect exists for both probe's luminescence at room temperature. For fluorenone, the fluorescence quantum yields (ΦF) determined were about 0.10 when dichloromethane, cyclohexane, and benzene (solvents which do not swell cellulose) were used for sample preparation, while for dioxane, acetone, ethanol, and methanol (solvents which efficiently swell cellulose) ΦF was approximately 0.01. These values are about 1 order of magnitude higher than those obtained in solution, showing the importance of the rigid dry matrix in reducing the nonradiative pathways of deactivation of the (π,π*) fluorenone first excited singlet state. Complementary, infrared studies showed that the carbonyl group of benzophenone is affected by entrapment (when the solvents used induce the swelling of cellulose), whereas in fluorenone the same band is insensitive to the adsorption process, not allowing the differentiation between entrapped molecules and surface crystallites of this ketone. These observations implied that benzophenone is entrapped between the chains of the polymer forming hydrogen bonds between the carbonyl and the hydroxyl groups of the glycosidic chains, while the rigidity of fluorenone apparently restrains the ketone−substrate interactions to the aromatic rings. Through the modifications observed in the carbonyl stretching band of benzophenone, it was possible to establish a swelling effect scale for the solvents, which is compared with previous results.