Characterization of pine bark extracts 1. Thin layer chromatography

TLC analyses of FSFE extracts were performed using silica gel plates (20 cm×20 cm, thickness 0.2 mm) (Merck, Germany). A hexane-ethyl acetate (8:2, v/v) mixture was used as the mobile phase. To observe the CO2 soluble volatile oil compounds, an anisaldehyde solution was used as the spray reagent while a NP solution (methanol-2-aminoethyl diphenylborinate, 99:1, v/v) was used for the analysis of ethanolic extracts phenolic compounds (at 365 nm) (Wagner et al., 1984). Plates were drawn using the ACD/TLC Plate Tool for ChemSketch, Freeware version 10.02 (Advanced Chemistry Development, Inc.).

8.3.4.2. Gas chromatography

The compositions of the extracts obtained by hydrodistillation (HD) are presented in terms of relative area (%) and were determined using a GC-MS apparatus (QP-5000, GC-MS, Shimadzu, Tokyo, Japan). The system was equipped with a fused silica capillary column (30 m×0.32 mm i.d., 0.45 μm, OV-5, Ohio Valley Specialty Company, Ohio, USA). Helium was employed as the carrier gas (1.0 mL/min) and 1 μL of sample was diluted in dichloromethane and injected. The injector temperature was 513 K and that of the detector was 503 K. The oven temperature was raised from 333 to 553 K, at 3 K/min, and maintained at 553 K for 20 min. The identification of compounds was done based on the comparative analyses between the mass spectra of the substances with those present on the GC-MS system database (Nist. 62 lib.), on literature (McLafferty and Stauffer, 1989) and on the Kovats retention indexes (Adams, 1995). These were obtained by co-injection of a standard mixture of n-alkanes (C8 -C24) and applying the van den Dool and Kratz equation (Van Den Dool and Kratz, 1963). GC-MS analyses were performed at Laboratório de Produtos Naturais, Centro de P&D Recursos Genéticos Vegetais, Instituto Agronômico de Campinas, Brazil.

The chemical compositions of the CO2 SFE extracts (1^st step FSFE) and of Soxhlet extracts (SoE) are also shown in Table 8.2 (in terms of relative area (%)) and were determined using a gas chromatograph (Tremetrics 9001, Thermo Scientific, Austin, USA), equipped with a fused capillary column (15 m × 0.32 mm i.d., 0.45 μm, CP-Sil 8 CB, Varian, Middelburg, The Netherlands). The carrier gas was helium (2-3 mL/min) and 2 μL of sample

(0.1 mg/mL, in ethyl acetate) was injected. The injector temperature was 503 K and that of the detector was 523 K. The temperature program was 323 K (for 5 min), 323-553 K (at 5 K/min) and 553 K (for 5 min). Kovats indexes were determined relatively to the retention times of a series of n-alkanes and using the Kovats’ method (Adams, 1995). The identification of substances was done based on the comparison of their Kovats indexes with those of the hydrodistillation (HD) extracts and also based on their chromatogram profiles.

In order to obtain relative area standard deviation values, GC analyses were done in duplicate for HD and for SoE samples. For FSFE samples and despite the identification have been made for all the obtained kinetic point samples, only the total composition profiles are presented in Table 8.2 because it was difficult to obtain a significant value variation since there were differences between the vials compositions.

8.3.4.3. High-performance liquid chromatography

HPLC analyses were performed, at room temperature, in a C18 column (250×4 mm i.d., 5 μm, Eurospher, Berlin, Germany) equipped with a pre-column and coupled to an UV detector (WellChrom k-2500, Kanuer, Berlin, Germany) and to a HPLC pump (WellChrom Maxi-Star k-1000, Knauer, Berlin, Germany). A mobile phase, constituted by methanol/acidified water (5% formic acid, v/v) in a proportion of 8:2 (v/v), was employed in an isocratic elution (80 min), at a flow rate of 1 mL/min. Samples were microfiltered (0.20 μm) before injection. For all samples, chromatographic profiles were measured at 280 nm and the concentrations (expressed as % (w/w) on a dry basis) of catechin and epicatechin were calculated from previously determined duplicated calibration curves.

8.3.4.4. Antioxidant activity: β-carotene and linolenic acid coupled reaction assay The antioxidant activities of the obtained extracts were determined by the coupled reaction of linolenic acid and β-carotene (Hammerschmidt and Pratt, 1978). Reaction were monitored at 470 nm by UV-vis absorbance readings (of extracts and control samples - the reaction medium with no extract) at 0 h (Abs ) and after 1, 2 and 3 h (^t0 Abs ) of reaction. Antioxidant ^tn activities were expressed as oxidation inhibition percentages (Equation 8.1).

Abs 100

8.3.5. Calculation procedures

The mean geometric diameter (dmg) of pine bark particles was calculated according to the American Society of Agricultural Engineers ASAE S319.2 method (ASAE, 1993), using sieves of 18/120 mesh (through 18 mesh and on 120 mesh).

( )

In this equation, d is the geometric mean length of particles on the i_i ^th screen

5

For all employed extraction methods, total yields were calculated as the ratio between the total extract mass and the feed mass, on a dry basis (d.b.). For FSFE experiments, the total extract mass was determined summing the extracted masses and the extract masses retained in the adsorption column. Extract masses recovered from tubing line washing were also considered for the calculation of 2^nd extraction step yields.

The overall kinetics extraction curves were constructed using just the accumulated masses of extracts which were collected at a given extraction time interval. Therefore, the masses collected from the adsorption column, as well as those collected in the cleaning process, were not considered for the kinetic representations.

For the 1^st FSFE step, a linear regression analysis was used to fit experimental data while, for the 2^nd extraction step, each overall high pressure extraction curve was fitted by a curve formed by two straight lines. Second step fitting was done by minimizing the least regression error (in the least squares sense) using the fminsearch function of Matlab (R2006a), and the first line was identified as the constant extraction rate period (CER). The corresponding kinetic parameters were calculated according to Rodrigues et al. (2002) (Rodrigues et al., 2002): mass transfer rate, MCER (Kg/s); mass ratio of solute in the solvent phase at measuring-cell outlet, YCER (kg/kg) and duration of it, tCER (s). For the 1^st step FSFE, and because the CER period was not yet entirely completed, tCER was fixed at 90 minutes and, consequently, the mass transfer rates and mass ratios of solute in solvent phase were calculated just for 90 minutes of extraction (M90min and Y90min, respectively). The extract yields corresponding to these two consecutive extraction periods were denoted by R90min and RCER (kg/kg), respectively.

8.4. Results and discussion