🌿 Processing and Cyanide Safety

Cassava contains cyanogenic glucosides because they are the plant's chemical defence against herbivores and pathogens. That is the honest starting point for everything in this module. The toxin is not a processing accident or a sign of a bad root. It is the plant working exactly as it evolved to work.

How much is there? The sourced answer is a very wide range, and the width is itself the lesson. One synthesis reports cyanogenic glucoside content in fresh roots from about 15 to 400 ppm. More extreme reports give 10 mg per kg in the least toxic varieties up to 2,000 mg per kg in the most toxic. Leaves are reported far higher still, as high as 2,000 mg per kg fresh weight in one source and 540 to 1,450 mg per kg in another. These are different studies reporting different distributions. They are not a contradiction to be averaged into one number.

The same variety varies too, with soil, drought stress, plant age and even which part of the root is tested. So there is no fixed cyanide number for cassava, for a variety, or even for a field.

One historical classification, from 1954 and cited in an FAO review, grades fresh peeled root as innocuous below 50 mg HCN per kg, moderately poisonous at 50 to 100, and dangerously poisonous above 100. Treat that as one historical reference scale, not as today's regulatory line.

Now the disease, because this is why the module exists.

Konzo is a paralytic disease caused by chronic cyanide exposure from badly processed cassava combined with a protein-poor diet. It is real, documented, and has affected tens of thousands of people in central and southern Africa, overwhelmingly children and women of childbearing age. The paralysis is spastic and it is irreversible. There is no treatment that reverses it. A person who develops konzo does not recover.

Read that again, slowly. This is not a food-quality issue. It is not a matter of a product tasting bitter or a customer complaining. It is a permanent, crippling disability, and it falls hardest on children.

Two things drive it together: cyanide from cassava that was not processed properly, and a diet short of protein. Protein matters because the body uses sulphur-containing amino acids to detoxify cyanide, and a poor diet has less of them to spend. That is why konzo appears in poverty and in food crises, when people eat more cassava, process it faster to eat sooner, and have less of everything else. The exact moment when a household is most tempted to shorten fermentation is the moment when shortening it is most dangerous.

The most recent figures available to this course date to a 2011 review synthesising data up to 2009. Current surveillance data was not retrieved, so this course gives no current case count.

Now the honesty that this course owes you about the legal limit.

You will hear a figure of 10 ppm, or 10 mg of HCN-equivalent per kilogram, quoted as the safe limit for cassava food products. It appears repeatedly in WHO-linked and national-regulator secondary literature. But in the research behind this course it could not be traced to one specific clause of one primary Codex or WHO document. The Codex standards for cassava flour and for gari both set moisture, fibre and ash limits firmly, and neither states a numeric cyanide limit; both cross-refer to the general Codex contaminants standard, which was not retrieved.

So here is how to hold it. Ten ppm is the number every source agrees on. It is not a clause you can quote chapter and verse, and this course will not pretend otherwise. What actually governs you is your own national standards body. Nigeria's Standard Organisation is cited at a maximum of 10 mg per kg cyanide in gari, given here as an example of how such a rule is written, not as a universal number every country enforces.

One last sourced fact that should end any argument about raw roots. A Ugandan study of twelve varieties, six local and six improved, found every one of them exceeded the cited 10 ppm reference value in the raw, unprocessed state, including the sweeter ones.

This is the most important lesson in the module. If you understand it, you can judge a process you have never seen before and adapt safely to new equipment or a new variety. If you only memorise steps, you cannot.

There are two mechanisms, and only two.

Mechanism one is physical removal. Some of the toxin is simply taken away with material you throw out. Peeling removes the outer root layers, which generally carry a different, often reported as higher, cyanogenic glucoside concentration than the inner flesh. Pressing and dewatering squeeze out liquid that carries dissolved cyanogenic compounds and their breakdown products. Soaking leaches soluble compounds into the water. In every one of those cases the toxin leaves in something you discard, which is exactly why the peel, the pressed liquid and the soak water must be discarded and never returned to the food.

Mechanism two is chemical conversion, and it is the one people misunderstand.

Inside an intact cassava cell, the cyanogenic glucosides, mainly linamarin, are kept physically separate from the plant's own enzyme, linamarase. They sit in different compartments and nothing happens. Break the cells, and the two meet. Linamarase breaks down linamarin and releases hydrogen cyanide gas.

So grating or crushing is the trigger. It ruptures the cells and starts the reaction. And here is the sentence that saves lives: grating does not remove cyanide by itself. It starts the chemical reaction that allows the cyanide to be driven off as gas during the subsequent drying, frying or fermentation step.

A grated mash that has not then been given time and conditions for that reaction to run, and for the gas to escape, is not a detoxified mash. It is a mash in which the reaction has been started and interrupted. That is the single most common way cassava processing fails.

What conditions does the reaction need? Reported optimal activity for the purified enzyme is pH 5 to 8 and 30 to 40 degrees Celsius. That is consistent with why warm, moist fermenting mash, and not cold dry flour, is where the detoxification chemistry actually happens fastest. Cold slows it. Dry slows it. Warm and moist is where the enzyme works.

Hydrogen cyanide has a boiling point of roughly 26 degrees Celsius, which is below the temperature of a warm day and far below any drying, roasting or frying temperature. That is why it leaves as a gas. Drying, roasting and frying drive off residual hydrogen cyanide, and at the same time reduce moisture to a level that halts microbial spoilage.

One consequence of that gas matters for your safety as a processor. The cyanide leaves the food into the air where you are standing. Ferment, press and fry in the open air or under good ventilation, never in a closed room, and keep children out of the working area.

Now sort the steps by mechanism, because this is the classification you should be able to do from memory.

• Peeling: physical removal.
• Grating or crushing: neither, on its own. It is the trigger for conversion.
• Fermentation: chemical conversion, by continuing enzymatic breakdown, plus acidification. In one study pH fell from 6.3 plus or minus 0.2 to 4.0 plus or minus 0.3, and titratable acidity rose from 0.08 plus or minus 0.03 percent to 0.36 plus or minus 0.05 percent, over 96 hours.
• Soaking or retting: both. It leaches soluble compounds out and it gives the enzymatic breakdown time in warm, moist conditions.
• Pressing or dewatering: physical removal.
• Drying, roasting, frying: chemical conversion completed, by driving the gas off.

Here is the whole module in one sentence. Cyanide reduction in cassava processing works by physically removing some toxin with the peel and the pressed-out liquid, and by chemically converting the rest into a gas that boils off during grating through to drying. Every step in every traditional method exists to serve one or both of those. Skipping or shortening a step, especially fermentation and pressing, removes the time those mechanisms need to finish.

Gari is fermented, granulated, roasted cassava and one of West Africa's great staple foods. It is also, when made properly, a complete detoxification chain: it uses both mechanisms from Lesson 2, several times over.

The reported process is: peel, wash, grate, ferment, dewater and press, sieve and granulate, then fry, which is often called garifying, in a dry pan.

Now take it step by step with the mechanism named.

Peel and wash. Physical removal. The peel goes out, and it must go out, not into the food and not back through the grater.

Grate. The trigger. Cells rupture, linamarase meets linamarin, and the reaction starts.

Ferment. Reported at 1 to 5 days, with the low end increasingly used in commercial settings. Traditional fermentation is commonly cited as taking 4 to 6 days to achieve sufficient detoxification. During this time enzymatic breakdown continues and natural organisms acidify the mash, with pH falling from about 6.3 to about 4.0 over 96 hours in one study.

Dewater and press. Physical removal. Traditionally by cloth bag over up to a week or longer; by hydraulic press at up to 25 kg per square centimetre, moisture falls from 60 to 70 percent to about 50 percent in as little as 15 minutes in Brazilian farinha processing. That comparison is worth noticing: mechanisation compresses a week-long step into minutes without changing the underlying chemistry, because the liquid still has to be physically removed either way.

Sieve and granulate. Preparation for even frying.

Fry. Chemical conversion completed. Reported at 30 minutes to 2 hours or longer, depending on batch size, pan temperature and operator skill. Residual hydrogen cyanide is driven off as gas and moisture falls to a level that halts microbial spoilage. Final moisture content is reported at under 10 percent in current commercial practice.

Notice the phrase depending on batch size, pan temperature and operator skill. That is why frying time is not a safety guarantee. Two hours in a shallow pan over a good fire is a different process from two hours in a deep overloaded pan over a weak one.

Now the composition standards, which are firm and quotable. The Codex Standard for Gari sets moisture at 12.0 percent maximum, crude fibre at 2 percent maximum, ash at 2.75 percent maximum, and total titratable acidity between 0.6 and 1.0 percent as lactic acid.

That acidity range is worth dwelling on. Codex specifies a minimum acidity, not just a maximum. Under-fermented gari is out of specification on acidity, which means the standard itself carries a fingerprint of whether fermentation ran.

What the Codex gari standard does not do is state a numeric cyanide limit. It cross-refers to the general Codex contaminants standard, which this course did not retrieve. So do not present a Codex-sourced cyanide figure for gari as coming from the gari standard itself, because it does not.

What governs cyanide is your national standards body. Nigeria's Standard Organisation is cited at a maximum of 10 mg per kg cyanide in gari. That is an example of how such a rule is written in one country, not a universal number.

Now the hardest part of this lesson, and the reason it is safety-critical.

Traditional fermentation is cited at 4 to 6 days. Economically pressured processors are reported to sometimes shorten it to under 2 days. That is a direct link between commercial time pressure and reduced safety margin, and this course states it plainly rather than softening it. When roots are deteriorating, when the buyer wants delivery, when cash is short, the fermentation step is the one that gets cut, because nothing visible goes wrong when you cut it.

You will also see a figure of submerged fermentation at 48 to 60 hours at 30 to 35 degrees Celsius achieving good results in an optimisation study. Do not read that as permission. That is a controlled, monitored process at a managed temperature. It is not a validation that any two-day village fermentation is automatically safe. Those are two different claims and only one of them was tested.

What did measured gari actually contain? A systematic review of finished cassava products found gari averaging 5.7 mg per kg, with a range of 0 to 23.9. Below the commonly cited reference point on average, but with individual samples well above it. That is why testing, not assumption, is the only real safeguard.

Two products, and then the table that proves the whole argument of this module.

Fufu first. Fufu is made by peeling and soaking whole roots in water for several days, a step called retting, then pounding, sieving and cooking the resulting paste, or drying it into fufu flour to be reconstituted later.

The retting step relies on the same natural fermentation organisms and pH drop described in Lesson 2. Its safety logic is identical in mechanism to gari's fermentation step, with one addition: extended soaking in water allows both enzymatic detoxification and leaching of soluble cyanogenic breakdown products into the soak water.

Which gives the single most important instruction in fufu production. The soak water must be discarded, not reused, for exactly that reason. It is carrying out the toxin. Reusing it, or letting it drain back into the mash, or cooking with it, returns to the food the very thing the step existed to remove. A processor who reuses soak water to save time or water has undone the step while appearing to perform it.

This course could not retrieve in full a specific study of the biophysical and histological changes during retting, so the detailed physical description of what retting does to root tissue is not given here. What is taught is the mechanism, which is well founded, and the mechanism is enough to tell you what must not be done.

Now chips. Whole or sliced roots are sun-dried, with or without a preceding brief soak, to produce chips for animal feed, milling into flour, or sale.

Here the sourced numbers are alarming and you should not look away from them. A systematic review reported traditionally soaked chips averaging 46.6 mg per kg HCN, with a range of 10 to 200 mg per kg. That average is more than four times the commonly cited 10 mg per kg reference point, and the top of the range is twenty times it.

A separate survey of ready-to-eat commercial cassava-based chips, from an export market rather than African village production, is cited here for one reason only: it demonstrates that finished, ready-to-eat chip products can and do carry measurable residual cyanide. Following a process is not the same as producing a safe product. Finished-product testing is the real safeguard, and testing requirements are set by your national authority.

Now the table, which teaches more than any paragraph could. One systematic review compiled reported HCN reduction outcomes across processing methods.

• Boiling for about 30 minutes: 96.3 percent reduction, final level 5.1 mg per kg.
• Peeling plus soaking plus fermentation plus frying, combined: about 100 percent reduction, final level 1.5 mg per kg.
• Fermentation plus drying: 80 percent reduction, final level 10.2 mg per kg.
• Sun drying alone: 30 percent reduction, final level 30.5 mg per kg.
• Traditional soaking alone: 20 percent reduction, final level 50.7 mg per kg.

Read the bottom two lines again. Sun drying alone leaves 30.5. Soaking alone leaves 50.7. Both are far above the commonly cited 10 mg per kg reference point. A single step, done on its own, does not make bitter cassava safe.

And read the second line. Peel, soak, ferment and fry, all four together, reaches about 1.5 mg per kg. Combining mechanisms vastly outperforms any single step.

That is why traditional processing chains exist as chains and not as a menu you pick one item from. Every generation that developed gari or fufu was, without knowing the chemistry, assembling multiple mechanisms in sequence. When a modern processor drops a step to save time, they are not simplifying a tradition. They are dismantling a safety system.

Treat these figures as an illustrative comparison of method logic. They are compiled across multiple underlying studies of different varieties and conditions. No figure in that table guarantees that your batch, following those steps, reaches that endpoint.

Starch extraction and high-quality cassava flour, usually called HQCF, are the industrial end of this course. They are mechanically different from gari and fufu, and the difference has a safety consequence you must understand before you consider entering these markets.

Starch and HQCF rely less on fermentation and more on grating, sieving and washing, and controlled drying of unfermented root pulp. That is deliberate. Fermentation gives gari its sour taste and its character. An industrial buyer wanting a neutral flour to blend into bread, or a clean starch for paper or textiles, does not want that flavour or that acidity.

Now look at what that removes from the safety chain. Go back to Lesson 4's comparison. Fermentation plus drying gave 80 percent reduction, final level 10.2 mg per kg. Sun drying alone gave 30 percent, final level 30.5. The fermentation step was doing a great deal of work, and the unfermented routes do not have it.

What do they have instead? Peeling, which physically removes toxin. Grating, which triggers the reaction. Washing and sieving, which leach and carry away soluble compounds in the wash water, a real and substantial physical removal in starch extraction specifically, since starch is separated by repeated washing. And controlled drying, which drives off the gas.

So the unfermented routes are not without mechanisms. But they lean much harder on variety choice and on the drying being genuinely thorough. This is where Module 3's fourth axis of variety choice becomes a commercial decision with a safety edge: a variety destined for full gari-style fermentation and drying can safely be a higher-cyanogen variety than one destined for a route with no fermentation step. If you are producing HQCF, your variety choice and your wash and dry discipline are carrying weight that fermentation carries elsewhere, and residual-cyanide testing matters more, not less.

Now the commercial numbers, because this is a real business.

Conversion is reported at roughly 1 tonne of HQCF from 5.5 tonnes of fresh cassava roots, a ratio of about 18 percent by weight. That is consistent with root moisture content of roughly 60 to 65 percent being removed during processing. Plan your input requirement from that ratio and then check it against your own equipment, because process losses and deterioration losses both eat into it before a single unit is sold.

HQCF is reported to allow up to a 25 percent substitution of wheat flour in bakery products. That is the commercial argument you will make to a bakery. But the source does not specify the baked product type or the sensory and quality tradeoffs at that level, so treat 25 percent as a cited ceiling from one source, to be validated against your buyer's own product testing, not as a universal baking constant. A bakery will test it in their own product and their answer is the one that matters.

One programme's cost data: processing 288 tonnes of fresh roots into HQCF over one year cost US$17,238, about US$60 per tonne of fresh root processed. Treat that as unverified as a current figure and specific to one programme's cost structure and year. What you should copy is the method, which is tracking your own cost per tonne of fresh root input, not the number.

On market access: as of the cited source, HQCF and industrial starch technology was fully adopted in Nigeria and tested but not yet fully adopted in 13 other countries including Benin, Cameroon, DR Congo, Ghana, Liberia, Madagascar, Malawi, Rwanda, Sierra Leone, Tanzania, Togo, Uganda and Zambia. Useful context when you are assessing whether a buyer exists in your own country yet.

The Codex Standard for Edible Cassava Flour sets moisture at 13 percent maximum, crude fibre at 2.0 percent maximum, ash at 3.0 percent maximum, and particle size at a minimum of 90 percent passing a 0.60 mm sieve for fine flour and a 1.20 mm sieve for coarse flour. As with gari, the standard cross-refers to the contaminants standard rather than stating a cyanide number itself.

Finally, why starch has industrial value at all. Starch makes up 64 to 72 percent of root carbohydrate, composed of roughly 20 percent amylose and 70 percent amylopectin. That amylopectin-heavy composition gives cassava starch its particular clarity, thickening and gelling properties, which is why food, paper, textile and pharmaceutical industries want it. A fully sourced breakdown of each specific industrial end use was not available to this course, so ask your prospective buyer what property they are buying.

A learner who can only follow a recipe is helpless when something goes wrong. This lesson is about recognising failure, which is the skill that actually protects people.

Here are the failure signs, step by step.

Fermentation that did not run. The mash should sour. In one study pH fell from about 6.3 to about 4.0 and titratable acidity rose from about 0.08 percent to about 0.36 percent over 96 hours. You cannot measure pH without a meter, but you can smell and taste sourness, and Codex requires finished gari to have titratable acidity between 0.6 and 1.0 percent as lactic acid. A mash that still smells and tastes bland and starchy after the expected fermentation period has not fermented. Cold weather, a very clean sterilised vessel that carries no starter organisms, or too short a time can all cause it. The answer is to give it more time in warm conditions, never to proceed on schedule.

Remember why warmth matters: reported optimal linamarase activity is pH 5 to 8 and 30 to 40 degrees Celsius. A cold mash is a slow mash. The calendar days you counted do not mean the same thing in cold weather as in warm.

Pressing that did not finish. The cake should stop releasing liquid and hold its shape. If liquid still runs freely, dissolved cyanogenic breakdown products are still in the mash, and they will end up in the product.

Liquid that went back in. Press liquid, soak water and wash water carry toxin out. If any of it drains back into the mash, onto the drying floor, or into the pot, the removal step was cancelled. Check the physical layout of your work area for this, because it is usually a slope or a container placed badly rather than a decision anyone made.

Drying that did not finish. Sun drying is reported to reduce moisture to 8 to 12 percent, with 14 percent cited as a critical threshold for microbial control, and the Codex maxima are 12.0 percent for gari and 13 percent for cassava flour. A product that bends rather than snaps, that feels cool or soft, or that cakes together in the hand is not dry. Under-drying fails on two counts at once: the gas has not fully been driven off and the product will spoil.

Frying that was uneven. Frying time is reported at 30 minutes to 2 hours or longer depending on batch size, pan temperature and operator skill. An overloaded pan cooks the outside and leaves the middle. Smaller batches turned constantly are safer than large batches left to sit.

And the sign that should stop everything: a strong bitter taste, or any burning or tingling in the mouth, in a finished product. That is not a quality complaint. Stop the batch, do not sell it, and do not feed it to anyone. Investigate what step failed.

Now, testing. Residual cyanide testing is set by a national authority, not by this course. There is no universal threshold this document can give you, because the commonly repeated 10 ppm figure could not be traced to a primary Codex clause in this pass. What you do is contact your national standards body or national food safety authority and ask three questions: what residual cyanide limit applies to your product, what testing is required before it is sold, and which laboratories are accredited to do it. Keep the answers and the results as your own safety record, independent of what a customer or inspector later asks for.

Finally, cassava leaves, which are a separate safety topic and not a minor variant of the root products. Leaf cyanogenic content is reported far higher than root content in the same plant, as high as 2,000 mg per kg fresh weight in one source and 540 to 1,450 in another. One study found that even after processing and boiling for 30 minutes, finished leaf-dish cyanide levels ranged from 32 to 50 mg HCN per kg on a dry-matter basis, three to five times the commonly cited reference point, and that was the cooked result.

The study's authors judged that acceptable only because cassava leaf dishes are typically eaten in small quantities as a side dish, not as a staple. That is a judgement about portion size, not about safety of the concentration. The same source recommends against making cassava leaf a daily food, especially for children, and recommends extending cooking beyond the 30 minutes tested. Note also the tradeoff: the processing that reduces cyanide nearly eliminates vitamin C, while protein, iron, calcium and other minerals are retained.

Cook leaves thoroughly, eat modest portions, and do not make them a daily staple for children.