Cadmium in cacao and chocolate
Cadmium is taken up from soil by the cacao tree, so it is in the cocoa and rises with cocoa content. How much depends on the soil under a particular farm far more than on the country, and the limits, warning levels and intake figures quoted about it are different kinds of number.
Three kinds of number, often confused
- A maximum level
- A concentration in the product, in mg/kg, above which it may not be sold. Law in the European Union and in Australia and New Zealand; a voluntary standard at Codex.
- A warning threshold
- California’s figure. It triggers a notice, not a ban, and is set a thousandfold below a level at which no effect was observed.
- A reference intake
- An amount per kilogram of body weight, from all food together, regarded as tolerable over a lifetime. Stated per week, per month or per day depending on who set it.
None of the three marks the point at which a bar of chocolate becomes harmful, and this page is not medical advice.
Every cadmium level read, by product category
All in milligrams per kilogram of product. The categories are not the same from one regime to the next, so rows cannot be matched one to one: the European Union divides chocolate at 30 and 50% cocoa solids, Codex at 30, 50 and 70%, and California at 65 and 95% cacao.
| Regime | Product category | mg/kg | What kind of level |
|---|---|---|---|
| European Union | Milk chocolate with less than 30% total dry cocoa solids | 0.1 | Law |
| European Union | Chocolate with less than 50% total dry cocoa solids; milk chocolate with 30% or more | 0.3 | Law |
| European Union | Chocolate with 50% or more total dry cocoa solids | 0.8 | Law |
| European Union | Cocoa powder sold to the final consumer, or as an ingredient in sweetened cocoa powder or drinking chocolate | 0.6 | Law |
| Codex | Chocolate containing or declaring less than 30% total cocoa solids | 0.3 | Voluntary international standard |
| Codex | Chocolate with 30% to less than 50% total cocoa solids | 0.7 | Voluntary international standard |
| Codex | Chocolate with 50% to less than 70% total cocoa solids | 0.8 | Voluntary international standard |
| Codex | Chocolate with 70% or more total cocoa solids | 0.9 | Voluntary international standard |
| Codex | Cocoa powder, 100% total cocoa solids | 2.0 | Voluntary international standard |
| Australia and New Zealand | Chocolate and cocoa products, whatever the cocoa content | 0.5 | Law |
| California (2018 settlement) | Chocolate products with up to 65% cacao | 0.4 | Warning trigger under a court settlement |
| California (2018 settlement) | More than 65% and up to 95% cacao | 0.45 | Warning trigger under a court settlement |
| California (2018 settlement) | More than 95% cacao | 0.96 | Warning trigger under a court settlement |
The European levels were first set in 2014 and have applied since 1 January 2019; they were read in the 2023 regulation as published, not in a consolidated version. The regulation sets no cadmium level for cocoa beans.
The Californian figures are those of a 2018 court settlement, which bind its parties and provided for later reductions. A later modification was not read.
No level was found for the United States at federal level, for Canada or for any Latin American country in the texts read. That is a statement about what was read.
The reference intakes, each in its own unit
| Body | Kind of value | As stated | The same, per week |
|---|---|---|---|
| European Food Safety Authority | Tolerable weekly intake | 2.5 micrograms per kg of body weight per week | 2.5 µg/kg |
| Joint FAO/WHO Expert Committee on Food Additives | Provisional tolerable monthly intake | 25 micrograms per kg of body weight per month | 5.8 µg/kg |
| US Food and Drug Administration | Toxicological reference value | 0.21 to 0.36 micrograms per kg of body weight per day | 1.5 to 2.5 µg/kg |
The last column is this site’s arithmetic, for comparison only: a month is taken as 30 days. The bodies chose different periods deliberately, because cadmium accumulates over years and a single week matters little.
The arithmetic people do, done properly
How much chocolate a week would, with no other food counted, supply the whole of the European weekly intake of 2.5 micrograms per kilogram of body weight, if the chocolate contained cadmium at exactly the level shown. One milligram per kilogram is one microgram per gram, so the sum is body weight times 2.5, divided by the level.
| Body weight | At 0.10 mg/kg | At 0.30 mg/kg | At 0.80 mg/kg |
|---|---|---|---|
| 20 kg | 500 g | 167 g | 63 g |
| 60 kg | 1.5 kg | 500 g | 188 g |
| 80 kg | 2.0 kg | 667 g | 250 g |
What this leaves out: every other food. Staple foods eaten in quantity are the main contributors to dietary cadmium, so nobody has the whole allowance free for chocolate.
And in the other direction: the levels shown are the European maximum for each category, not what a typical bar contains, and laboratory digestions suggest only part of the cadmium in chocolate is available to be absorbed.
It is not a recommendation. A tolerable intake describes a lifetime average. Passing it in one week is not an injury, and staying under it is not a guarantee.
Where the cadmium comes from
Cadmium is in soil everywhere in small amounts. The cacao tree takes it up through its roots and some of it ends in the bean. It is in the plant, not on it, which is the difference from lead.
- The soils in question are mostly not polluted. In a survey of 560 sites across Ecuador total soil cadmium averaged 0.44 mg/kg, described as typical of young, unpolluted soils, and on 55 Honduran farms it averaged 0.25 mg/kg, in the range of uncontaminated soils. Beans averaged about 1 mg/kg in both. (Argüello et al. 2019; Gramlich et al. 2018)
- Geology explains much of it. In Honduras bean cadmium differed strongly by the rock beneath and was highest on river-laid soils, and no effect of fertiliser or of nearby industry was found. (Gramlich et al. 2018)
- Not every study agrees on the source. On 19 farms in southern Ecuador cadmium was concentrated near the surface and fell with depth, which the authors read as a sign of human activity. (Chavez et al. 2015)
- Official texts carry both views. The European regulation that set the limits says that in some regions of producing countries soil cadmium “can be naturally high”; the international code of practice lists human sources (ore processing, fuel burning, waste, manure, phosphate fertiliser) alongside natural ones (volcanic activity, marine shale). (source 1; source 2)
Not shown: The share of cadmium in any region's cocoa that is natural as against added by people. The studies infer it from patterns; none measured it directly.
What makes a soil give up its cadmium
- How much is there, and how acid the soil is. Across Ecuador, bean cadmium rose with total soil cadmium and fell as soil pH rose, and fell also with manganese oxides and organic carbon. Together these explained about two-thirds of the variation. (Argüello et al. 2019)
- A synthesis of published data from many growing regions found the same two leaders, total soil cadmium and soil pH, with leaf cadmium the third useful predictor. (Wade et al. 2022)
- On organic matter the studies disagree: it lowered bean cadmium in the Ecuadorian survey, helped explain variation in Costa Rica, and had a negligible direct effect in the synthesis. (Argüello et al. 2019; Carrillo et al. 2023; Wade et al. 2022)
- They disagree too on whether “available” cadmium, the fraction a soil test extracts, predicts better than the total. It was the best predictor in Honduras and negligible in the synthesis. (Gramlich et al. 2018; Wade et al. 2022)
- Older orchards had less: bean cadmium fell by a factor of 1.4 as orchard age rose from 4 to 40 years in Ecuador, and was lower in older and higher farms in Costa Rica. (Argüello et al. 2019; Carrillo et al. 2023)
Not shown: What would happen on one farm if its soil were changed. These are comparisons between sites, not experiments.
Regional variation: hotspots, not origins
Surveys are reported here as what they sampled and found. None of them supports a statement about a country's cocoa as a whole.
- Within Peru, modelled maps place elevated soil and bean cadmium mainly in four northern departments with pockets elsewhere. The authors estimate that fewer than 20% of cacao-farming households nationally might be affected by cadmium rules, and as many as 89% in one department. (Thomas et al. 2023)
- Within Costa Rica, 150 farms gave beans from 0.12 to 3.23 mg/kg: a range of more than twentyfold inside one small country, with the higher values in two regions. (Carrillo et al. 2023)
- Within Ecuador, 45% of bean samples exceeded 0.60 mg/kg, which means 55% did not. (Argüello et al. 2019)
- At retail, one American survey found cadmium higher in products labelled as from Latin America than from Africa. It was a convenience sample, with origin as stated on the pack. (Abt et al. 2018)
Not shown: That any named origin is dependably high or low. The variation inside each country surveyed is larger than a country label can carry.
Not shown: Anything about West African or Asian soils. No soil survey from either was read; the regions appear only in retail comparisons.
Inside the tree: genetics and rootstocks
- The kind of tree matters a great deal. A hundred different cacao types growing on uniform soil in Trinidad's genebank differed thirteenfold in bean cadmium. (Lewis et al. 2018)
- The difference shows in the bean and not in the rest of the tree. In a Honduran trial of 11 cultivars the wood and leaves did not differ between cultivars, and the beans did. (Engbersen et al. 2019)
- Cadmium seems to reach the bean indirectly. A tracer put into soil under two trees appeared in young leaves first and in the nibs last, reaching half its final level there after 191 days, which points to cadmium being moved on from stem and leaves. (Vanderschueren et al. 2023)
- In the Ecuadorian survey the effect of genotype, CCN-51 against Nacional, was inconsistent. (Argüello et al. 2019)
- Grafting on to a low-cadmium rootstock is the obvious application. The international code of practice says it has so far only been tried experimentally. (source 2)
Not shown: That any named variety or genetic group is reliably low in cadmium. Most of the variation in the genebank study was between individual types.
Not shown: That a low-uptake rootstock lowers cadmium in the beans of what is grafted on to it. The trees studied were not grafted for the purpose.
Mitigation: what has been tried
This is the weakest part of the evidence, and the part most often oversold.
- Liming, to make the soil less acid, is described in the international code of practice as the most effective method developed so far, where soil pH is below six. (source 2)
- In the one multi-year field trial read, in two Ecuadorian orchards, lime raised soil pH by a unit and lowered cadmium in the leaves on the acid soil. Bean cadmium was unchanged by lime, gypsum or compost after 22 to 30 months. (Argüello et al. 2023)
- A pot experiment suggests one reason. Most roots were in the top layer; liming only that layer reduced leaf cadmium less than liming all the soil, and drove the seedlings to take more from the unlimed soil below. (Argüello et al. 2020)
- Biochar and lime that worked at low rates in the laboratory needed higher and more frequent application in the greenhouse and more again in the field. The abstract reports no size of reduction. (Ramtahal et al. 2019)
- Authors of the data synthesis infer that breeding and rootstocks are more likely to succeed than treating soil. That is an inference from what predicts cadmium, not a trial result. (Wade et al. 2022)
- Zinc, microbial treatments and field trials of rootstocks: no field trial reporting cadmium in beans was found.
Not shown: That any soil treatment lowers cadmium in beans in the field. The trial that measured beans found no change in the time it ran, and an effect may take longer than two and a half years.
From bean to bar
- The shell holds more than the nib. Concentration was highest in the shell and lowest in the pulp, with the nib, the placenta and the pod husk between. (Vanderschueren et al. 2020; Lewis et al. 2018)
- Fermentation can move some cadmium outward from nib to shell, but only if it acidifies the nib to below pH 5. The average fall in nib cadmium was a factor of 1.3, and a laboratory incubation in acetic acid reproduced it. (Vanderschueren et al. 2020; Vanderschueren et al. 2023)
- On one Colombian farm no post-harvest step changed bean cadmium. What lowered it was removing the shell and then diluting the cocoa: 4.17 mg/kg in whole beans, 3.28 in nibs, 1.60 in the bar. (Bravo et al. 2022)
- Cadmium stays with the cocoa solids. Pressing liquor left virtually all of it in the powder and not in the butter. (Mounicou et al. 2003)
- So cadmium rises with cocoa percentage. It was correlated with percent cocoa in an American retail survey and, weakly, with labelled cocoa solids in a Japanese one, and it is the Commission's stated reason for setting different limits at different cocoa contents. (Abt et al. 2018; Kataoka et al. 2018; source 1)
- Along three production lines total cadmium fell by between 10.5 and 33%, with winnowing and conching the steps that mattered. (Kruszewski et al. 2018)
Not shown: The cadmium in any bar from its cocoa percentage. Percentage shifts the odds; the beans decide the level, and a low-percentage bar made from high-cadmium beans can exceed a high-percentage bar made from low ones.
Not shown: That fermentation is a reliable way to reduce cadmium. One study found a modest fall under the right conditions and another found none.
How much reaches people
- The international expert committee estimated average cadmium exposure from cocoa products at 0.02 to 1.6% of its tolerable monthly intake across 17 regional diets. For children who eat a great deal of cocoa products it put total exposure at 96% of that intake, said this was probably an overestimate, and did not consider cocoa's contribution a concern. (source 3)
- The American regulator says experts internationally have found chocolate a minor source of cadmium exposure, quoting a World Health Organization range of 0.1 to 9.4% of dietary cadmium. (source 4)
- The European position is worded differently. The Commission's summary of the European assessment puts staple foods eaten in quantity as the main contributors and lists chocolate among the foods with the highest concentrations; the same regulation calls chocolate and cocoa powder an important source of exposure, frequently eaten by children. (source 1)
- In a Belgian study chocolate supplied 7 to 9% of dietary cadmium, at an average of 28 g a day, in a sample the authors think leaned towards chocolate eaters. (Vanderschueren et al. 2023)
- Not all of it is absorbed. In laboratory digestions 10 to 50% of the cadmium in cocoa powder was potentially available, and cadmium in chocolate was less available than cadmium in wheat flour, by a factor of 2 for dark chocolate and 5 for cocoa powder. (Mounicou et al. 2003; Vanderschueren et al. 2023)
Not shown: How much cadmium a person absorbs from chocolate. The digestion studies are laboratory models and their results range widely.
Not shown: Your own exposure, which depends on everything else you eat, and on smoking, which none of these figures includes.
California's warnings, which are not limits
Much of what is written about heavy metals in chocolate traces back to one Californian law. It sets warning thresholds, and they are far lower than any safety limit by design.
- For a substance listed as a reproductive toxicant, the law requires a warning above one thousandth of the level at which no effect was observed. The resulting figure is 4.1 micrograms a day for cadmium by mouth, and 0.5 micrograms a day for lead. (source 5; source 6; source 7)
- A 2018 court settlement applies this to chocolate as concentrations that trigger a warning, rising with cacao content: for cadmium 0.400 parts per million up to 65% cacao, 0.450 from there to 95%, and 0.960 above 95%. (source 8)
- The settlement provided for those triggers to step down after six years unless an expert committee recommended otherwise. (source 8)
- The effect of the choice of benchmark is visible in one data set: of 72 American products, 35% exceeded the Californian level for cadmium per serving, while the medians were below it. (Hands et al. 2024)
Not shown: That a product over a Californian level is unsafe. The level sits a thousandfold below an observed no-effect level and governs a notice, not a sale.
Not shown: The triggers in force today. The settlement was read as made in 2018; a later modification was not read.
Sources
Studies are cited by author and year and summarised, with their limits, on the research pages. The official documents are listed here. Each was downloaded and read, and the passage relied on was checked to be in it.
- European Commission. Commission Regulation (EU) No 488/2014 of 12 May 2014 amending Regulation (EC) No 1881/2006 as regards maximum levels of cadmium in foodstuffs (2014-05-12). eur-lex.europa.eu/eli/reg/2014/488/oj
Law · EU - Codex Alimentarius Commission (FAO/WHO). Code of Practice for the Prevention and Reduction of Cadmium Contamination in Cocoa Beans (CXC 81-2022) (2022). www.fao.org/fao-who-codexalimentarius/sh-proxy/en/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites%252Fcodex%252FStandards%252FCXC%2B81-2022%252FCXC_081e.pdf
Voluntary international text · International - World Health Organization. Evaluation of certain food additives and contaminants: seventy-seventh report of the Joint FAO/WHO Expert Committee on Food Additives (WHO Technical Report Series 983) (2013). iris.who.int/handle/10665/98388
Risk assessment · International - US Food and Drug Administration. Cadmium in Food and Foodwares (page content current as of 2024-03-05). www.fda.gov/food/environmental-contaminants-food/cadmium-food-and-foodwares
Official guidance, not law · United States - California Office of Environmental Health Hazard Assessment. Proposition 65 in Plain Language (copy archived by the Internet Archive). oehha.ca.gov/proposition-65/general-info/proposition-65-plain-language
Official guidance, not law · California, United States - California Office of Environmental Health Hazard Assessment. Cadmium (Proposition 65 chemical page; copy archived by the Internet Archive). oehha.ca.gov/proposition-65/chemicals/cadmium
Law · California, United States - California Office of Environmental Health Hazard Assessment. Lead and Lead Compounds (Proposition 65 chemical page; copy archived by the Internet Archive). oehha.ca.gov/proposition-65/chemicals/lead-and-lead-compounds
Law · California, United States - Superior Court of California, County of San Francisco. Consent Judgment, As You Sow v. Trader Joe's, Inc. et al., San Francisco Superior Court, Case No. CGC-15-548791 (as published by the California Attorney General) (2018-02-15). oag.ca.gov/system/files/prop65/judgments/2014-00635J3734.PDF
Law · California, United States
Reviewed 9 October 2026.
Related: the cadmium studies, lead, nickel, mould toxins and other contaminants, cadmium and heavy metals, how cocoa and chocolate are tested and what legally counts as chocolate. All data.