Avocado seeds were harvested and classified at La Fortuna farm municipality of Mariquita, Tolima (5.255607 N, -74.998586 W). Size reduction was made with flake cuts of 3 mm and the material was dried at 55 °C during 14 h in a Memmert UF 55 oven, then they were ground and filtered with a 500-micron mesh. For colorant extraction, three solvents were used independently: distilled water, an aqueous solution of NaOH (0.5%) (Devia Pineda and Saldarriaga, 2005) and a mixture of distilled water and ethanol (1:1). A ratio of 0.05 milled seed-solvent, as well as the reflux system at a temperature of 45 °C and the extraction time of 120 min, remained constant. Extraction yields were calculated using a moisture determinant XM 60 HR.

The antioxidant activity was measured using the DPPH method applied by Brand-Williams et al. (1995), which is based on the absorbance reduction of the DPPH 1, 1-diphenyl-2-picrylhydrazyl radical, measured at 515 nm.

To determine the antioxidant activity, 200 μL of each extract at different concentrations and 2800 μL of DPPH solution (0.1 mM) were mixed and taken to a dark chamber during 30 minutes at room temperature, and the absorption was measured at 518 nm in a Jenway 7305 spectrophotometer. The 2-carboxylic-6-hydroxy-2, 5, 7, 8 tetramethylchroman acid (Trolox) was used as a control. Tests were replicated three times.

The percentage of antioxidant inhibition from the extracts at different concentrations (expressed in ppm) was evaluated. The pH value of the NaOH aqueous extract (pH 9.3), water extract (pH 5.1) and ethanol extract (pH 5.8) was neutralized to pH 7.0 The percentage of inhibition was calculated according to equation 1.

A: Blank absorbance

A₁: Sample absorbance

The results were expressed as the maximum concentration of the inhibitory mean (IC₅₀), defined as the amount of antioxidant necessary to reduce the initial concentration of DPPH to 50%.

The antimicrobial activity was evaluated using the Kirby Bauer disk diffusion method (Hudzicki, 2009). Strains of Staphylococcus aureus ATCC 29213 and Escherichia coli ATCC 25992 were used for this purpose. Those strains were isolated in EMB and Baird Parker agar respectively. Three to five colonies of each microorganism were placed in a saline solution (5 ml, 0.85%) and the concentration was adjusted to the 0.5 McFarland tube (1.5×10⁶ CFU mL^-1). Once the suspension was adjusted, it was massively seeded with a sterile brush in Mueller Hinton agar. The sterile filter paper discs (10 mm in diameter), used for the test, were previously impregnated with 0.1 mL of the extract at 5000 ppm and dried in a closed petri dish. The treatments were classified in A1 (aqueous extract; 5000 ppm and pH 5.4), A2 (NaOH aqueous extract; 5000 ppm and pH 8.2), A3 (water-ethanol extract; 5000 ppm and pH 5.5). A paper disc with chloramphenicol (10 mg mL^-1) was used as the positive control (A4). All experimental units were incubated at 37 °C for 24 h (Casana et al., 2016). Then inhibition halos were read.

Matured Hass avocado was reduced in size and dried in a Memmert UF 55 plus oven at 50 °C for 12 h. Subsequently, the free fat total extraction was carried out using the Soxhlet method for 4 h with hexane (grade HPLC, Scientific). The fatty acids profile was analyzed by means of FID gas chromatography in a Thermoscientific Trace 1310 chromatographer with an Rtx-5 Restek Corporation column of 30 m long, 0.32 mm internal diameter and a film thickness of 1µm under the following conditions: injection volume 1.0 L, injector temperature 230 °C, detector temperature 250 °C, column pressure 23.04 psi, hydrogen flow in the detector 45 mL min^-1, air flow in the detector 450 mL min^-1, makeup gas flow (N₂) 45 mL min^-1, split flow 70.2 mL min^-1, split ratio 40:1.

Temperature ramp: The initial column temperature was 190 °C (during 12 min) and rose up to 220 °C with a ratio of 2.0 °C per 4.0 min. The fatty acids composition was found by comparing the peaks retention times obtained with the Fatty Acids Methyl Esters (FAMEs) standards.

Physicochemical parameters such as density was determined following the NTC 336 (ICONTEC, 2002a), Iodine index following the NTC 283 (ICONTEC, 1998a), peroxide index following the NTC 236 (ICONTEC, 1998b), refractive index following the NTC 289 (ICONTEC, 2002b), acidity index following the NTC 218 (ICONTEC, 1999), and saponification index following the NTC 335 (ICONTEC, 1998c). Tests were measured in duplicate.

The extract with the highest yield was used to calculate the color parameters by using the CIELab space coordinates (L*, a*, b*) in a Konica Minolta Cr-5 colorimeter. Different concentrations of liquid dye avocado were tested (1%, 2%, and 3%) in the liquid soap matrix. The soup emulsion was stored at room temperature and exposed to light for one month. The color difference between the samples was expressed in ΔE* (Hikita et al., 2001). The color variation can be estimated through the ΔE* which is determined through the L*, a*, and b* parameters (Manayay et al., 2013) (Equation 2).

ΔE*: Color variation or alteration

Texapon 40 Sodium Lauryl Ether Sulphate (SLES) and distilled water were mixed and placed on a magnetic stirring plate at 200 rpm. Then sodium benzoate, cocamidopropyl betaine, glycerine, colorant, alcohol, and salt were homogenously added in a strict sequential order. Different formulations are shown in Table 1.

Table 1

Color behavior was evaluated by using the CIELab space coordinates (L*, a*, b*) in a Konica Minolta Cr-5 colorimeter (with liquid analysis accessory). The pH variation was measured with a Lovibond SD 300 potentiometer. Soap samples were stored at room temperature for one month and exposed to light.

An airtight vial was overflowed with a soap sample and incubated for 5 d to determine the biochemical oxygen demand. Dissolved oxygen was measured before and after the incubation phase and the BOD calculation results from the difference between the initial and final values of dissolved oxygen (Gender Cevallos and Arnao Ramirez, 2005). Commercial liquid soap and a sample soap (formula 4) were used to determine the BOD.

Statgraphics (Centurion version) was used to perform the analysis of variance of the color measurement, extraction performance, and antimicrobial activity. Simple linear regression was used to predict the percentage of antioxidant inhibition at different concentrations.

The process of drying avocado seeds reported a yield of 27.90±0.99%. Sodium hydroxide showed the most efficient extraction with a percentage of total biomass extracted (weight/weight) of 35.72±0.43 and CIELab color coordinates of L*=0.15, a*=0.05 and b*=-0.44, followed by alcohol extraction (33.16 ±0.13 and CIELab color coordinates of L*=76.89, a*=15.49 and b*=66.74) and distilled water showed the lowest extraction yield (11.61±0.89 and CIELab color coordinates of L*=85.08, a*=5.05 and b*=50.25).

Treatments with sodium hydroxide (T3 and T4, Table 2) showed the lowest percentage of inhibition. Samples extracted with NaOH at a concentration of 150 μg mL^-1 and pH 9.3 showed a percentage of inhibition of 24.72%. However, at pH 7.0 the percentage of inhibition decreased to 14.03%.

Table 2

Samples treated with water at pH 7.0 showed a decrease in the percentage of inhibition at all concentrations (reference value=150 μg mL^-1). T1 reached 51.69% (IC₅₀ value=153.87 μg mL^-1) while T2 reached 38.96% (IC₅₀ value 187.66 μg mL^-1). Both IC₅₀ values were close to the Trolox control (84.10 μg mL^-1). Samples neutralized at pH 7.0 and treated with NaOH and water, showed less antioxidant activity.

Otherwise, treatments with a mixture of water and ethanol showed an increase in antioxidant activity when neutralized to pH 7.0. The ethanol extract at pH 5.8 (150 μg mL^-1) showed a lower percentage of inhibition (24.21%) than at pH 7 (35.01%). IC₅₀ values from T5 (630.00 μg mL^-1) and T6 (265.67 μg mL^-1) were the closest to the control. Table 2 shows the results of the maximum concentration of the inhibitory mean (IC₅₀). The ANOVA (Analysis of Variance) indicates that there are statistically significant differences (P<0.05) between the treatment and the control.

Nagaraj et al. (2010) studied the antioxidant activity of methanol-water (4:1) extracts by the DPPH method and obtained a percentage of inhibition of 60.8%. In the present research, methanol was not used due to its incompatibility with the liquid soap matrix formulation; water was used instead. The aqueous extract showed 51.95% of inhibition while the Trolox control reached 80.76%. This difference is mainly due to the type of solvent used (Fu et al., 2016).

None of the treatments showed antimicrobial activity against any of the evaluated strains; except for the chloramphenicol control. The extracts were not subjected to any type of bioactive compounds isolation or fractionation; A1 (aqueous extract; 5000 ppm and pH 5.4), A2 (NaOH aqueous extract; 5000 ppm and pH 8.2), A3 (water-ethanol extract; 5000 ppm and pH 5.5) and A4 (Chloramphenicol; 10 mg mL^-1). Although the use of sodium hydroxide as a solvent, showed to be the most efficient for the extraction of the dye and achieve greater dyeing power, this is not suitable for the extraction of compounds with antimicrobial activity. Therefore, in this case, the polarity of the solvent and variables such as temperature, extraction time, the nature of the matrix, the specific characteristics of the compounds and their location within the matrix must be taken into account for the optimization of the extraction of compounds with bioactive characteristics (Osorio-Tobón and Meireles, 2013).

Some studies have analyzed the antimicrobial activity from the avocado seed extracts and have defined its seeds as a good source of phytochemical components with high bioactivity (Dabas et al., 2013; Nagaraj et al., 2010). However, the solvent used in these studies was methanol in different conditions, and terpenoids and other bioactive compounds were fractionated.

The drying performance of the avocado pulp was 47.41±1.22%. A paste with a rigid texture and dark green color was obtained afterward. The yield percentage of oil extraction from the dehydrated avocado pulp was 71.26±1.25%.

The iodine index is a measure of the degree of unsaturation of the fat components. There is a clear difference between the obtained index and what is established by the NTC 258 (ICONTEC, 2011). A value of 177.52 cg I₂ g^-1 shows that the obtained oil has a high degree of unsaturation, different from the reported by Restrepo Duque et al. (2012) (77.85 cg I₂ g^-1) and Acosta Moreno (2011) (75-94 cg I₂ g^-1).

The peroxide index measures the fresh oil oxidation or its degree of rancidity at the time of the test (Lafont et al., 2011). A value of 38.45 meq peroxide kg^-1 reflects a high degree of rancidity. Similarly, a study that used the same extraction method for the Hass cultivar oil reported a value of 31.66 meq peroxide kg^-1. However, the advanced state of ripeness and the prolonged heat treatment to which the pulp was subjected should be considered when comparing its value to olive oil with a maximum permitted limit of 20 meq peroxide kg^-1.

The saponification index of the avocado oil (190.74 mg KOH g^-1) was higher than the reported by Restrepo Duque et al. (2012) (175 mg KOH g^-1). This indicates a greater presence of low molecular weight fatty acids since the esters of this type of molecules require more KOH for saponification (Lafont et al., 2011). Soap and cosmetics industry demand a minimum value of 185 mg KOH g^-1 (Lafont et al., 2011), this suggests that oil of this type can be used in such type of factories.

The percentual sum of the oleic and linoleic fatty acids was 57.33 ± 0.33% (Table 3), a similar value was reported by Acosta Moreno (2011) (59.1%) for the Hass avocado cultivar (Acosta Moreno, 2011). This indicates that the obtained oil has a high degree of unsaturation; verified with the iodine value (177.52 cg I₂ g^-1). For palmitate and stearate, the sum was 24.27%. This value is higher than the reported by Acosta Moreno (2011) who obtained 16.99% for such saturated fatty acids. The difference is significant and may be due to the quality of the original raw material.

Table 3

The colorant obtained from the extraction with sodium hydroxide was added to the liquid soap since it showed the best performance and high colorant strength.

The color difference in all treatments can be considered small and inconspicuous since the ΔE* values are less than 1.5 (Obón et al., 2009), meaning that the color given at different concentrations by the NaOH aqueous extract and avocado seed is stable. Also, the change in pH is not affected over time. According to the CIELab coordinates, the colorant has a tendency towards yellow; behavior that is linked to the concentration and reflected in the b* coordinate (where a positive number indicates yellow, and a negative number indicates blue). When the colorant concentration increased from 1% to 3%, the value of b* increased from 41.51 to 68.07. A slight tendency towards the range of red was observed in coordinate a* (where a positive number indicates red, and a negative number indicates green), this may also be closely related to the concentration since as colorant concentration increases from 1% to 3%, the a* value increases from 9.2 to 33.23. However, the L* value (where numbers between 0 - 50 indicate black or darkness and between 51 -100 indicates whiteness or clarity) decreased from 79.8 to 54.21 as the concentration increased from 1% to 3%.

In this sense, the colorant extracted with NaOH was completely stable in the liquid soap matrix (pH of 6.2) during a month of storage. The formulation containing oil at 2% also remains stable in color and pH during the same period. The ANOVA indicated that there are no statistically significant differences between the treatments for parameter ΔE* (P>0.05).

The BOD of the natural colorant was 10.35 mg L^-1, higher than the value of the commercial soap (9.42 mg L^-1). This indicates that more oxygen is required to degrade the organic matter in the soap matrix by a microbial population.

Table 4