Reference-SiO₂ and Paracetamol-SiO₂ samples were synthesized by sol-gel process using tetraethoxysilane (TEOS) (Sigma-Aldrich 98%) as silica precursor. Molar alkoxide:water ratios (R_w) used were 1:4 and 1:8. For the Reference-SiO₂ sample synthesis was as follows, suitable amounts of water and ethanol were mixed into a rounded three neck flask (Table 1), then 18.5 mL of TEOS were added dropwise under continuous stirring at room temperature. The obtained sol was left under the same conditions until gel formation. Paracetamol-SiO₂ sample was obtained similarly, but 2.5 mg of paracetamol (Bristol-Myers Squibb) per gram of SiO₂ were mixed with deionized water and added at the beginning of the process. Once the gels were obtained, the samples were dried at room temperature and then grinded for further analysis.

Infrared absorption spectra of the nanomaterials were obtained on FTIR Affinity-1equipment. A wafer of

potassium bromide (Sigma-Aldrich) for each of the four different samples was prepared for the analysis, the sample (5 wt%) was pressed together with KBr (95 wt%) (2000 ton/in²).

Thermograms were carried out using a Simultaneous Thermal Analyzer STA i-1000. Samples were placed in a platinum pan and heated from room temperature to 800˚C at a rate of 10˚C/min in a N₂ atmosphere.

Images for the morphology analysis were obtained by means of a high-resolution Transmission Electron Microscope (TEM) JEOL JEM-2100F, operated at 200 kV and equipped with an energy dispersive spectroscopic (EDS) microanalysis system (Oxford). The micrographs were obtained using a Gatan Orius camera.

Nitrogen adsorption-desorption isotherms were obtained using a Micromeritics Belsorp II, Bell Japan Inc. The Brunauer-Emmett-Teller (BET) method was used to calculate specific surface areas (SBET). Pore volumes and pore size distributions were obtained using BJH method.

A wafer made of each Paracetamol-SiO₂ nanomaterial (1:4 and 1:8 ratios) was placed into a glass containing deionized water (50 - 75 mL). Sampling was performed at different time periods during a total of 200 hours. To assess the amount of drug released, samples were analyzed using ultraviolet spectroscopy (Cary-1 UV-visible, Varian). The increase in main absorption bands reported for paracetamol was monitored in all samples. After measurements, samples were returned to its corresponding glass to maintain a constant volume. A calibration curve was also realized and absorbance spectra were collected. In order to calculate drug concentration, Lambert-Beer law was used. Drug release curves were obtained by plotting cumulative drug concentration versus time. All the determinations were made by triplicate.

Paracetamol release kinetics from each nanomaterial were analyzed by several mathematical models. Depending on these estimations, suitable mathematical models to describe dissolution profiles were determined. The following plots were made: % of drug released versus time (zero-orderkinetic model); ln dissolution % drug remaining versus time (first-order kinetic model); dissolution % drug released versus square root of time (Higuchi model); cube root of drug % remaining in matrix versus time (Hixson-Crowell cube root law); and dissolution % drug release versus time (hyperbola).

Figure 1 and Figure 2 show the infrared spectra of the materials. In the low energy region, characteristic bands of silica were observed in all the samples, the band located around 477 cm⁻¹ corresponds to Si-O-Si vibrations, while the peaks in the interval from 900 to 980 cm⁻¹ are typical of the Si-OH bonds (Figure 1(a) and Figure 2(a)) [20] . The FTIR vibration band of paracetamol located at 3330 cm⁻¹ corresponds to N-C-H (Figure 1(b) and Figure 2(b)). If paracetamol is occluded in the silica sample; some bands appear at the interval from 800 to 1500 cm⁻¹and correspond to symmetric and asymmetric vibrations of C-H and C-N of organic groups. The band at 1251 cm⁻¹ corresponds to -H-C-N- vibrations and the bands located at 1600, 1643 and 1655 are attributed to C=O (amide) stretching vibrations (Figure 1(a) and Figure 2(a)). These bands were higher for 1:8 samples than 1:4 samples. In the high energy region, a broad band accompanied of the other small peaks between 3100 and 3700 cm⁻¹ appears in all samples, these bands are formed by stretching vibrations from H-O-H, C-H and N-H bonds [21] [22] . The vibration bands with the higher intensities were exhibited in Paracetamol-SiO₂ 1:8 sample (Figure 2(a) and Figure 2(b)), thus the difference in water concentrations slightly affects the properties of the drug adsorbed in silica matrix.

Thermograms of all the nanomaterials are shown in Figure 3(a) and Figure 3(b). Incorporation of paracetamol into silica sol-gel in both ratios (1:8 and 1:4) causes slight modifications in thermal behavior of SiO₂ being more evident in R_w = 1:4 systems (Figure 2); where a difference in total amount of weight loss differs in 3%. At the beginning, Paracetamol-SiO₂ sample lose about 13% from room temperature to 130˚C, while this lost was 13% for SiO₂ alone. This first step is due to residual solvent and adsorbed water. The minimal difference in weight loss may be related to gelation process, since Paracetamol-SiO₂ sample took longer time than SiO₂ leading to complete hydrolysis-condensation reactions. Between 130 and 400˚C both samples showed a loss of 4% for SiO₂ and 6% for Paracetamol-SiO₂. Paracetamol melting (167˚C - 169˚C) and decomposition (326˚C) took place at this temperature range [23] [24] . The oxidative decomposition of the organic chains corresponding to the organic species dispersed in the silica matrix occurred between 250˚C - 320˚C. From the mass losses observed at the range between 250˚C and 280˚C on the three thermogram curves, we noticed that the alkoxide:water ratio used for the synthesis (which determines the nature of the matrix) does not influence significantly. The EG quantity which chemically bounds in the silica network structurally bonded -OH groups, takes place around 450˚C. For the 1:8 ratio materials, thermal behavior was very similar between both samples. The first loss (ca. 14.5%) was recorded around 156˚C for SiO₂ while Paracetamol-SiO₂ sample lost 13.5% at 141˚C. At the

156˚C - 246˚C range SiO₂ showed a minimal loss of 0.6% that did not occurred in the 1:4 SiO₂ sample, this may be explained because of the difference in water content. The following event occurred from 246˚C to 442˚C with a loss of 3% and the last one is observed between 468˚C and 553˚C with a loss of almost 1%. Thermal behavior of paracetamol-SiO₂ 1:8 sample showed a small difference compared to 1:4 samples. Since the first one registered a two-step mass loss between 141˚C and 530˚C with a lower loss. Both losses, 203˚C - 329˚C and 329˚C - 424˚C were around 2%. It is well known that increased values of water promote silicon ratio hydrolysis leading to minimum alkoxide residuals, with increased gelation rates [25] [26] .

Transmission electron micrographs from Paracetamol-SiO₂ 1:4 are shown in Figure 4. From the TEM images, aggregates of particles with estimated size of 10 - 60 nm were observed.

N₂ isotherms and BET surface areas for both samples and its references were obtained (Figure 5(a) and Figure 5(b)). We observed that both 1:4 and 1:8 samples were classified as type IV isotherms according to IUPAC classification [27] the adsorption is very similar up to P/P₀ = 0.4. All the samples presented an H2 hysteresis loop typical of materials with pores in the form of ink, characteristic of the majority of inorganic oxides. The most striking features of type IV isotherm are the hysteresis loops and the plateaus at high P/P₀, the isotherms of this type are produced by adsorbents for which capillary condensation occurs in the mesopore range, such as many silica gels, the majority of inorganic oxides some other metals [28] [29] . The surface areas obtained from the BET analysis of the isotherms are summarized in Table 1, this parameter did not considerably vary between the samples.

The idea of varying the water concentration is to have a different specific surface area and mean pore diameter, in Table 1 we can observe that Paracetamol-SiO₂:water 1:4 has an specific surface area of 561 m²/g, while Paracetamol-SiO₂:water 1:8 considerably increases to 825.

Figure 6 and Figure 7 show paracetamol release curves the reservoirs monitored over a period of 0 - 3 and 0 - 200 hours, the amount of released drug was 50% and 60% respectively in both samples (1:4 and 1:8). The main purpose of paracetamol addition at the beginning of the synthesis with different amounts of water is to bond paracetamol to the silica network avoiding the degradation of the drug, allowing the formation of a gel with the drug bonded by means of Van der Waals interactions to the silica complex. In this way we ensure that paracetamol, which remains at the surface, will be released during the first hour, and the drug that remains at the internal surface will be slowly released in latter times (Figure 7(a) and Figure 7(b) and Figure 8).

The amount of drug released from SiO₂ 1:4 reservoir correspond to 60% and 80% at time periods of 0 - 3 (Figure 6(a)) and 0 - 200 hours respectively (Figure 7(b)). A slower release rate is observed in 1:8 nanomaterial (Figure 7(a) and Figure 7(b)) with respect to 1:4, this could be explained because of the higher BET area showed by paracetamol 1:8 sample since it is encapsulated in a network with major stability in the extent that is released. After first 3 hours both reservoirs followed a sustained type profile determined largely by stoichiometric ratio of the silica matrix and particle size [30] [31] . Hyperbola t was the model that best fit for all in vitro release profiles of paracetamol, with lineal correlation coefficients of R² = 0.9 for both nanomaterials.