Overview
Oliveira and colleagues bought 40 chocolate bars made in Brazil in Natal stores between April 2024 and May 2025, covering four dark cocoa levels, milk and white chocolate, and sugar-free milk and white chocolate. They measured total arsenic by the silver diethyldithiocarbamate spectrophotometric method and compared results with Brazilian limits. At least one bar of most types exceeded the limit; sugar-free milk chocolate was lowest.
Key numbers
- 40% cocoa bars: 1.025, 1.069 and 1.105 µg total arsenic (tAs)/g for three bars, but 0.029 and 0.047 µg/g for the other two (Table 2).
- Milk chocolate bars: 0.780 to 1.046 µg tAs/g; 70% cocoa bars: 0.169 to 0.213 µg tAs/g (Table 2).
- Sugar-free milk chocolate: one bar “NQ”, one printed “0.000”, others 0.042 to 0.083 µg tAs/g (Table 2).
Methods (brief)
Samples of 2.5 g were dry-ashed with magnesium oxide and magnesium nitrate, arsenic was reduced to arsine with zinc and potassium iodide, and the arsine was trapped in silver diethyldithiocarbamate and read by spectrophotometry. Each bar was analysed in duplicate against a calibration curve prepared on the day. No certified reference material or spike recovery was run, and detection and quantification limits were not determined in-house.
Evidence Fitness
Forty bars across eight product types with per-bar results, but using an older colorimetric method without in-house validation. The white chocolate results do not follow the printed calibration curve, which matches the milk chocolate curve, so those five values are flagged. Several low values fall below the quantification range the authors cite for the method. Only total arsenic is reported, without inorganic speciation. No DOI is printed in the PDF; the manifest DOI is used. Plain chocolate bars are direct evidence for chocolate and only a proxy for chocolate-nut-candy, which is not linked.