Determination of Metals in Olive Oil by Electrothermal Atomic Absorption Spectrometry: Validation and Uncertainty Measurements

Abstract

The determination of trace metals in organic matrices is still highly demanding despite improvements in analytical instrumentation. The present study was undertaken in order to evaluate electrothermal atomic absorption spectrometry for the determination of cadmium, copper, iron, lead, and nickel in olive oil. A variety of approaches were used. The most suitable digestion procedure was heating the samples at 300°C for 24 hours and ashing in a muffle furnace at 450°C for 16 hours. The validation data were detection limits of 0.2–153 ng g⁻¹; mean trueness on certified reference materials of 81–94%; mean recovery on spikes of 90–120%; and repeatability of 12–53%. The combined relative uncertainty was 0.298–0.766. Oils processed by pressing had higher copper, iron, and lead concentrations than oils processed by centrifugation. The reported method provides an efficient way for monitoring trace metal content during olive oil production.

Keywords:

• Electrothermal atomization atomic absorption spectrometry
• Heavy metals
• Olive oil

Notes

^a Arithmetic mean of 5 parallel measurements, ±standard deviation.

^b <LOD, calculated from the calibration, values are provided in Table 3.

Note: LOD and LOQ were calculated from calibration slope and from signal variation of blank sample.

^a The complete analysis (digestion and measurement) of five aliquots of the real sample.

^b The analytes were added before the sample digestion in four parallel additions of 9.93 ng g⁻¹Cu, 0.5 ng g⁻¹ Cd, 5.14 ng g⁻¹ Ni, 4.65 ng g⁻¹ Pb, and 282 ng g⁻¹ Fe.

^c Standard deviation of blank sample mass concentration (solution mass 5 g, sample mass 5 g).

^d Standard deviation of analyte signal about the regression line.

^e Slope of the calibration line.

Note: Metal content was in the ng g⁻¹ range.

^a Amount of added analyte: 10–20 ng g⁻¹.

^b Amount of added analyte: 12–48 ng g⁻¹.

^c 3 standard deviations of blank signal.

^d mean blank signal plus 3 standard deviations of blank signal.

^e DA, dry ashing; WA, wet ashing by microwave oven.

^f PSA, potentiometric stripping analyzer; ISE, ion selective electrode.

^g  N/A, not analyzed.

Note: Total of 17 triplicate samples were analyzed. Values below limit of detection were regarded as zero in statistical analysis.

^a Different upper case superscripts indicate samples that were significantly different (p < 0.05).

^b P, percentage of samples containing metals above detection limit.

Note: The method details are given in Table 4.

^a D, mean difference of duplicate values; x _m, mean of duplicate values; D/x, difference of duplicates divided by its mean; sd, standard deviation of D/x values; (n = 2 for P/B and 4 for R); w _ss, metal content in spiked sample; w _s, metal content in sample; w _sp, spiked content.