To the optimization tests, liquid humus samples obtained from the fermentation of agricultural wastes were used. The mixture consisted of 20% wastes (4 L poultry + 4 L of orange peel) with 80% water (32 L), in volume.

A mix (2000 µg∙L⁻¹) of reference material p,p’-DDT, p,p’-DDE, p,p’-DDD, lindane, aldrin, heptachlor, chlordane, dieldrin and HCB (hexachlorobenzene) in hexane: toluene (1:1, in volume) was used for testing. From the reference material, there was prepared a stock solution of 200 µg∙mL⁻¹ in ethyl acetate. Ethyl acetate HPLC grade was supplied by Tedia. Helium (purity: 99.9990%) was supplied by White Martins.

The chromatographic system used was a GC/MS, Shimadzu, GC2010 Plus, equipped with split/splitless injector and a mass spectrometric detector, Shimadzu, MS-QP 2010 Plus. The capillary column used was a BP-5 capillary column, 30 m × 0.25 mm i.d. and 1.0 µm film thickness (Agilent). Helium at a constant flow rate of 1 mL∙min⁻¹ was used as carrier gas. The injector and detector temperature were 280˚C. Other operational conditions are described in Table 1. The OCPs were quantified by analytical curve constructed with 6 concentration levels in the range of 10 - 100 mg∙L⁻¹ for each reference material.

The spectrometer was operated in full scan mode for determination of retention times of the compounds and in the SIM mode to analyze the samples. The operating conditions of the mass spectrometer are shown in Table 2.

The extraction procedure was performed in a microwave digestion system (SPEEDWAVE FOUR^®, with Built-in Non-Contact, Temperature and Pressure Measurement, BERGHOF, Germany) with a capacity of 12 vessels of 60 mL and 1450 W maximum power. To the optimization tests, they were used liquid humus samples obtained from the fermentation of agricultural wastes (20:80, in volume) with the extraction solvent (ethyl acetate). The extraction temperature ranged from 70˚C to 110˚C, sample and solvent volumes ranged from 5 to 10 mL, the irradiation times (10 min), ramp (5 min) and power (80% - 1160 W) were fixed based on previous experiments.

After the extraction step, the vials were cooled until room temperature and then opened. Then, the material was filtered in 45 μm filter paper, concentrated under vacuum in rotary evaporator (FIZATOM) at 80˚C until dryness, reconstituted with 1 mL of ethyl acetate and placed in a vial for the chromatographic determination. For tests with blanks, the above procedures were followed using the sample without addition of the reference material.

The extraction conditions of POPs evaluated in liquid humus samples were optimized based on the analysis of response surface methodology (RSM) of the investigated factors. In this study, the effects of three factors on the response were evaluated using a 2³ factorial design, 16 experiments in duplicate were randomly performed according to the experimental design. The area for each compound was chosen as response-variable in the optimization process.

A two-way interaction linear model was adjusted to the experimental data based on Equation (1):

where is the predicted value (area for each compound), β’s are the model coefficients, and the encoded factors: temperature (, 70˚C - 110˚C), sample volume (, 5 - 10 mL) and solvent volume (, 5 - 10 mL). The ranges used to evaluate the independent variables were selected based on preliminary experiments.

Previous experiments in our laboratory, to extract the compound p,p’-DDT in sewage sludge samples evaluated the significance of the parameters: temperature, extraction time, and percentage of acetone in the solvent mixture (ethyl acetate-acetone). Based on this, new parameters were selected for the present study. Then, the solvent chosen to extraction was ethyl acetate and the extraction time for all tests was 10 minutes.

The use of microwave energy in the extraction process generates heat of solvent and consequently of the sample which can suffer degradation or conversion of some compounds by different ways, such as exposure to high temperatures and pressure inside the microwave extraction vessel [5] . To achieve the best extraction condi-

tions, a factorial design was used. The actual and encoded levels of the independent variables (temperature, sample volume and solvent volume) for each experiment are shown in Table 3.

Table 4 shows the significance of the extraction process. Bold numbers indicates significant factors as identified by the analysis of variance at the 95% confidence level.

From the calculated results it can be seen that the factor 3 (solvent volume) was the most important for the majority of the investigated analytes (strongly positive) except for the compounds heptachlor and o,p’-DDD. Only to three compounds, the factor 2 (sample volume) influenced the extraction process, and the major significance of this factor is in the compound hexachlorobenzene. The factor 1 (temperature) was significant for most compounds, but the interaction of this one with the other two factors, i.e., the factor interactions (β₁₂ and β₁₃) had low significance in the process, and the factor interactions β₁₂ (temperature x sample volume) has significance in five compounds. The factor interactions β₁₃ (temperature x solvent volume) was not significant in the extraction process. The interaction between sample volume x solvent volume (β₂₃) is what most influences the process, eight of eleven evaluated compounds were significantly influenced. All calculated regression coefficients (R²) are in the range 0.99, indicating a good correlation to quantify and interpret the relationship between the experimental data and the factor effects studied. Equation (2) describes an example of the proposed model for the compound Aldrin.

Response surface graphics are very useful to estimate and to identify the optimal ranges for the two or more factors on the response [10] . In order to determine the optimum ranges of the factors, graphs were plotted using the z-axis (response) against two of the independent variables (temperature, sample volume and solvent volume), maintaining a third one at a fixed value. Figures 1-3 show an example of the main interactions in the process for the compound Aldrin. Similar plots but not identical were obtained for all the other investigated compounds.

Figure 1 shows the interaction between temperature and the sample volume in response. Increasing the temperature from 70˚C to 110˚C and increasing the sample volume from 5 to 10 mL, the response value increase. So, best responses are achieved when using 10 mL of sample at a temperature of 110˚C, using 10 mL of the extrac-

tion solvent. Figure 2 shows the interaction between solvent volume with the extraction temperature in response. Increasing the solvent volume from 5 to 10 mL and increasing the temperature from 70˚C to 110˚C the response value increase. Extraction temperature at 110˚C combined with a solvent volume of 10 mL provides a better response. Figure 3 shows that best responses are achieved when the maximum sample volume is equal to the maximum solvent volume. The combination of 10 mL of sample volume, 10 mL of solvent volume (ethyl acetate) at 110˚C were considered to provide the best conditions.

Considering the extraction temperature significance in the process, i.e., higher the temperature, higher the yield of the extraction, the chosen temperature was 110˚C. Elevated levels of temperature (above 110˚C) would increase the internal pressure inside extraction vessels, which could result in loss of analyte. Thus, the maximum levels of factors sample volume and solvent volume, related to recovery levels, achieved the final extraction conditions were defined, without the need of additional experiments.

After defining the best extraction conditions, the recovery of the compounds were evaluated. The recovery assays experiments were performed by adding 125 µL of stock solution of 200 µg∙mL⁻¹ to the extraction vessel, the final volume was 20 mL. Thus, inside of vessels, the concentration of POPs was 1.25 µg∙L⁻¹. After the extraction procedure, the sample was determined according to the procedure previously described.

Table 5 shows the linear regression, coefficient of determination, average recovery and relative standard deviation of the OCPs studied.

^*Means ± standard deviation (duplicate) and relative standard deviation (R.S.D).

In the analysis of blank chromatograms, no peaks related to OCPs were observed. The recoveries values (87.98% - 105.68%) showed that the methodology was suitable for determination of organochlorine pesticide in liquid humus samples.