🍞 Flour and Ingredient Science

Flour is sold to you by name and used by you as a technical material. The names are not standardised across countries, so learn the numbers behind them and ask for the numbers, not the name.

The published protein bands are these. Bread flour sits at 12 to 14 percent protein and is the material for yeast-raised bread, rolls and buns. All-purpose flour sits at 8 to 11 percent and is general purpose, which means it is workable for bread and cake but optimal for neither. Pastry flour sits at 8 to 9 percent for pie crust, short pastry and some biscuits. Cake flour sits at 7 to 9 percent for delicate crumb. Cookie and biscuit flour is 8 to 10 percent, milled from soft red winter wheat. Vital wheat gluten, sold as gluten flour, is 40 to 45 percent protein and is used at low inclusion to fortify weak flour.

Codex sets a minimum of 7.0 percent protein on a dry-weight basis for wheat flour. Understand exactly what that means. It is a floor for the commodity so that what is sold as wheat flour really is wheat flour. It is not a bread-making specification. A 7 percent flour meets the Codex standard and will still not make good bread.

Now the point that separates a baker from a recipe follower. Protein quantity is not protein quality. A high-protein flour with weak gluten still bakes badly, and every experienced baker has met a 13 percent flour that behaved like an 10 percent one. Gluten quality is measured on an extensograph or alveograph, which measure resistance to stretching and extensibility, not on a protein number. Published extensograph readings show bread flour at resistance 400 and extensibility 190, giving a ratio of 2.11, while biscuit flour reads resistance 130 and extensibility 160, a ratio of 0.81. Read those two lines carefully. The biscuit flour is deliberately weak and extensible. It is not a worse flour. It is the correct flour for a product that must spread and stay flat rather than hold gas and rise.

This is why substitution is not free. Work an example. A baker in Nakuru, Kenya, cannot get bread flour and is offered all-purpose at 10 percent protein. Vital wheat gluten at 42 percent protein is available in 1 kg packs. She wants to lift 50 kg of the 10 percent flour toward 12 percent.

1. Protein now in 50 kg = 50 x 0.10 = 5.00 kg.
2. Target: a blend of total weight W at 12 percent, where W = 50 plus the gluten added, call it G.
3. Protein after = 5.00 + 0.42G, and this must equal 0.12 x (50 + G).
4. So 5.00 + 0.42G = 6.00 + 0.12G, giving 0.30G = 1.00, so G = 3.33 kg.
5. She adds about 3.3 kg of vital wheat gluten to 50 kg of flour, roughly 6.7 percent inclusion, to reach about 12 percent protein.

That arithmetic is correct and useful, but it comes with a warning she must respect. Raising the protein number does not guarantee she has raised the gluten quality, and adding vital wheat gluten also raises water absorption, so her dough will need more water. She must test bake before she commits a production run.

One piece of information you will not find in this course. Wet gluten percentage ranges by flour type were not available to us, and we will not invent them. Wet gluten is a genuinely useful figure, and the way to get it is to ask your miller for a certificate of analysis, or to have flour tested against ICC Standard 155 or AACC method 38-12.02 at a grain laboratory or an agricultural college. Ask for it by name and the laboratory will know what you mean.

If you learn only one purchase specification in this whole course, learn this one. The Falling Number test, also called the Hagberg test, is the single most useful thing a small baker in Africa can ask their miller for, and the reason is rain.

What it measures is alpha-amylase activity in the flour. Alpha-amylase is an enzyme that breaks starch into sugars. You need some of it, because it feeds the yeast and colours the crust. Too much of it destroys the crumb. When rain falls on standing wheat near harvest, the grain begins to sprout in the ear. Sprouting switches on alpha-amylase massively. That grain is still milled and still sold, and the flour looks completely normal in the bag.

The method is simple to picture. A flour and water slurry in a tube is heated in a boiling water bath and stirred. The time in seconds for a plunger to fall a set distance is the Falling Number. If the enzyme has thinned the starch paste, the plunger falls fast and the number is low. So remember one sentence: low seconds means high enzyme activity. The standard methods are AACC 56-81.03 and ICC 107/1, which are identical in approach. Published conditions use a 250 to 350 g sample of dry product, about 25 mL of water, a boiling bath, and agitation starting automatically after 5 seconds.

Now read the bands as a purchasing tool.

1. Below 220 seconds. High enzyme activity, typically sprouted wheat. Expect soft sticky dough that is hard to handle, large open holes, a sticky crumb that fouls slicer blades, an over-dark crust from sugar caramelisation, and deformed loaves.
2. Over 250 seconds. Described as suitable for most breadmaking processes.
3. 250 to 280 seconds. The stated optimal band for large-scale bread baking, and therefore the number to write into your purchase specification.
4. Above 300 seconds. Minimal enzyme activity, no sprouting. Expect long proof times, pale crust, low volume and dense crumb.
5. Above 350 seconds. Requires enzyme supplementation with malt or fungal alpha-amylase.

The real-world spread observed is wide: a minimum around 55 seconds, with some flours pushing into the high 400s and beyond. That range is why the number matters.

Here is how it plays out in practice. A baker in Blantyre, Malawi, had a good year and then, three weeks after the harvest rains, everything went wrong at once. The dough came out of the mixer soft and sticky. Loaves spread instead of standing up. The crust went almost black at the same oven setting she had used all year. Her slicer gummed up. She changed her yeast, cut her water, blamed her staff and lost most of a month.

Work what she should have done. Her dough was at 62 percent water. She dropped to 58 percent to fight the stickiness, so on a 50 kg flour batch she took out 50 x (0.62 minus 0.58) = 2 kg of water. The dough was still sticky, because the problem was not too much water, it was starch being digested by enzyme. Cutting water only made the loaves smaller and drier while the crumb stayed sticky. The correct action was to telephone the miller and ask one question: what is the Falling Number of the lot I received. When she finally did, the answer was 180 seconds. That is deep into the sprouted range and no adjustment of her mixing would have fixed it.

What to do with a low Falling Number flour, in order. First, ask the miller to blend it with a high Falling Number lot, which is what millers do routinely and is why they hold both. Second, if you must use it, blend it yourself at a small share with a sound flour and test bake. Third, reduce it to products where a soft sticky crumb matters less, though there are few. Fourth, refuse the delivery. That last option is only available to a baker who wrote the specification before the truck arrived.

So write it. On your purchase order, state: Falling Number 250 to 280 seconds, certificate of analysis to accompany each delivery. If your miller cannot supply the figure, ask which grain laboratory tests for them, or send a sample to an agricultural research station or grain laboratory yourself. The test is routine, quick and cheap compared with a month of ruined bread.

Two sacks of flour at the same price can hold very different amounts of water, and the one that holds more gives you more loaves from the same money. Water absorption is therefore a commercial property, not just a technical one.

The formal definition is the water needed to bring a dough to a standard consistency of 500 Brabender Units on a farinograph. The farinograph scale runs from 0 to 1,000 Brabender Units with 20 unit markings, and the machine records how a dough resists a mixing blade over time. Published bands read as follows: very strong flour above 63 percent absorption, strong above 58 percent, medium strength 54 to 60 percent, and weak below 55 percent. Those bands overlap as published, which is a useful reminder that the boundaries are conventions rather than laws.

What moves absorption up: higher protein, higher damaged starch, higher pentosans, added vital wheat gluten, and lower flour moisture. What moves it down: high enzyme or alpha-amylase activity from sprouting, which links this lesson straight back to Falling Number. A sprouted flour absorbs less water and then releases what it did absorb as the starch is digested, which is exactly why the dough goes slack in the trough.

Be careful about how much you claim to know here. The source that gives these bands explicitly declines to quantify the effect of any single factor. The quantitative effect of one extra percent of protein, or one extra percent of damaged starch, on absorption was not available for this course, and this course will not invent a coefficient. Do not let anyone sell you a rule of thumb dressed up as a measurement.

Expected starch damage sits at 6 to 9 percent for winter wheats and 7 to 10 percent for spring wheats. Damaged starch absorbs water where intact starch granules largely do not, which is why milling intensity changes absorption even when the wheat has not changed.

Why the money matters. Work it through. A baker in Bulawayo, Zimbabwe, is offered two flours at the same delivered price per 50 kg bag. Flour A absorbs 62 percent, flour B absorbs 56 percent. She uses 50 kg of flour and scales loaves at 900 g of dough.

1. Flour A dough weight, ignoring other ingredients for simplicity: 50 kg flour plus 31.0 kg water = 81.0 kg.
2. Flour B: 50 kg flour plus 28.0 kg water = 78.0 kg.
3. Loaves from A = 81,000 divided by 900 = 90 loaves.
4. Loaves from B = 78,000 divided by 900 = 86 loaves, rounding down.
5. Difference = 4 loaves per bag, from water, which costs almost nothing.

At a selling price of USD 1.00, that is USD 4.00 per bag of free revenue. Over 20 bags a month, USD 80. She must still check that flour A actually bakes well at 62 percent and does not go slack, but the arithmetic explains why absorption is worth asking about.

When you have no farinograph, and you will not have one, run these three checks instead.

1. The standing water test. Mix a fixed weight of flour with water added in small measured amounts until the dough reaches the consistency you know from experience. Record the water used as a percentage of flour weight. Do this identically for every new lot and you have your own comparative absorption figure, which is what you actually need.
2. The hand check. Squeeze a fistful of flour. Flour that packs into a lump that holds its shape is carrying more moisture than flour that falls apart. Higher flour moisture means lower absorption and less baking value per bag.
3. The slack test after 20 minutes. Mix a small test dough and leave it covered for 20 minutes. A dough that becomes noticeably slacker and stickier is telling you about enzyme activity, and you should ask for the Falling Number.

Record every one of these in the batch sheet from the records lesson, because a comparison you did not write down is a comparison you no longer have.

Three ingredients, each cheap, each capable of ruining a product when the level is wrong. Learn the numbers and the reasons together, because a level without a reason is a level you cannot adjust.

Salt first. Typical inclusion is 1.50 to 2.25 percent of flour weight, with an optimum of 1.75 to 2.25 percent. Below 1.50 percent bread tastes bland. Above 2.25 percent it tastes salty. Those are narrow limits, and they are the reason a baker weighs salt rather than scooping it.

Salt does four jobs. It gives flavour. It tightens gluten. It improves gas retention. And it regulates fermentation rate. The rheological effect is worth understanding because it sounds contradictory: salted dough is simultaneously more resistant to stretching and more able to extend without tearing once it is deformed. That combination is exactly what a bread dough needs, which is why unsalted bread is not merely flat-tasting but structurally worse.

Sugar next, and here the thresholds matter more than the level. A bread dough needs 3 to 3.5 percent fermentable solids for yeast activity. Below that the yeast has nothing to work on beyond what the flour's own amylases release. White pan bread typically carries about 8 percent sugar, within a range of 0 to 15 percent. Above 15 percent the dough is classified as a sweet dough. And there is a separate biological threshold: sugar at 5 to 10 percent baker's percent increases yeast activity, while above 10 percent yeast activity decreases. So a sweet bun dough at 18 percent sugar is fighting its own yeast and needs either more yeast or an osmotolerant strain.

Residual sugar, meaning the sugar left after fermentation, gives sweetness, drives crust browning by caramelisation, and extends shelf life because sugar is hygroscopic and holds moisture. Caramelisation begins at 160 degrees Celsius, while Maillard browning starts above 105 degrees and needs the crust surface above roughly 130 degrees. That is why a pale crust is very often a sugar problem rather than an oven problem.

Fat has six functions, and it is worth naming all six because bakers usually remember only the first. It tenderises by shortening gluten strands. It has creaming ability, incorporating air, and the better the creaming ability the lighter the cake. It has a plastic range, meaning it stays workable across a temperature span, which matters enormously in a hot bakery. It lubricates, letting the gluten network expand and easing mixing and handling. It moistens, retarding drying out, and pure shortening outperforms butter or margarine at this. And it provides nutrition as concentrated energy.

Inclusion levels by product: bread 2 to 5 percent, with a published pan bread formula at 3 percent; cookies 50 to 80 percent; sweet biscuits 25 to 32 percent; short pastry from 50 percent, which is the two parts flour to one part fat ratio, up to 100 percent for full puff.

Work a correction. A baker in Douala, Cameroon, complains that his bread tastes flat and collapses slightly in the pan. His formula uses 50 kg flour with 600 g salt and 300 g sugar.

1. Salt = 600 divided by 50,000, times 100 = 1.2 percent. That is below the 1.50 percent floor, so the bread will taste bland and gas retention will be poor. That is his collapse.
2. Correcting to 1.9 percent needs 50,000 x 0.019 = 950 g of salt, an increase of 350 g.
3. Sugar = 300 divided by 50,000, times 100 = 0.6 percent. That is far below the 3 to 3.5 percent fermentable solids the yeast needs, so fermentation is slow and the crust is pale.
4. Raising to 4 percent needs 50,000 x 0.04 = 2,000 g of sugar.
5. At an indicative USD 0.90 per kg for sugar and USD 0.35 per kg for salt, the extra 1.7 kg of sugar and 0.35 kg of salt cost about USD 1.65 on a 50 kg batch, spread over roughly 90 loaves. That is under two US cents a loaf to fix both faults.

That is the shape of most bakery problems. The fix is cheap; the diagnosis is the skill.

Enrichment means adding fat, sugar, eggs or dairy to a lean dough to make it richer, softer and longer keeping. It also makes the dough slower, more expensive and more hazardous, so each addition must earn its place.

Milk powder is the enrichment ingredient with the clearest published guidance. In fermented yeast doughs, use 2 to 8 percent based on flour. In cakes, up to 15 percent.

What it does to the dough: it increases water absorption, so take the dough softer from the mixer than you otherwise would; it slows fermentation; and it raises dough pH. What it does to the finished product: a darker crust from lactose caramelisation, improved volume, extended shelf life, a finer crumb and better sliceability. For a bakery selling sliced bread, that last pair is worth money.

There is one trap. Skim milk powder loses the fat-derived shelf-life benefit, and lactose does not ferment. So if you use skim powder and expect it to feed the yeast, you will be disappointed. You must add fermentable sugar separately or accept excessively slow fermentation. This is a common and expensive misunderstanding.

Work it. A baker in Accra, Ghana, wants to move her plain bread upmarket into a soft sliced sandwich loaf. She uses 50 kg flour per batch.

1. Milk powder at 4 percent = 50 x 0.04 = 2.0 kg per batch.
2. At an indicative USD 4.50 per kg for full-cream milk powder, that is 2.0 x 4.50 = USD 9.00 per batch. Verify with your own supplier.
3. Her batch makes about 90 loaves, so the added cost is 9.00 divided by 90 = USD 0.10 per loaf.
4. Because absorption rises, she adds water in small steps and takes the dough softer from the mixer, recording the new total water on her batch sheet.
5. Because fermentation slows, she must either extend the process or raise the dough temperature slightly, and because the crust will darken from lactose, she should expect to reduce her oven setting or her bake time and test bake before committing a production run.
6. She sells the enriched loaf at USD 0.25 more than her plain loaf. On 90 loaves that is USD 22.50 of extra revenue against USD 9.00 of extra cost, or USD 13.50 per batch, provided the customer accepts the higher price.

Now eggs, and here this course must be honest with you. Sourced egg inclusion ranges were thin. What was retrieved is limited: dry eggs at 5 percent in a published cookie formula, the classical cake ratio placing eggs at 100 percent of flour weight, and cookie and biscuit eggs at 10 to 30 percent. Egg functions, meaning aeration, coagulation and structure, emulsification through lecithin, colour and enrichment, and quantitative inclusion bands by product, were not available for this course. We will not invent them.

So do this instead. Obtain a food-science text chapter on eggs. The BCcampus open textbook Understanding Ingredients for the Canadian Baker, Part V on Eggs, is free online and appropriate for this level. Read it before you formulate egg-rich products, and build your own inclusion bands from test bakes recorded on your batch sheet.

Treat eggs as a food-safety topic as much as a formulation topic. Eggs are a declared allergen under Codex, and they are a raw animal product and therefore a microbiological hazard. Milk is also a declared allergen. That means an enriched dough turns your bakery into an allergen-dense environment, and once you add eggs and milk to some products but not others, you need separation in time and validated cleaning between them. In a one-room bakery, physical separation is usually impossible, so the workable route is scheduling: process products without the allergen before products with it, add allergenic ingredients as late in the process as possible, and run products with identical allergen profiles consecutively.

Finally, note that national codes may require a higher mandatory cook temperature for egg-containing products than the quality temperature a baker would choose. Ask your national food safety authority before you sell custards, quiches or egg-set fillings.

Buying flour well is a written activity. A baker who buys by handshake gets whatever the miller has surplus of that week, and every bad batch becomes an argument with no evidence on either side. A baker who buys on a specification gets consistency, and gets the right to refuse.

Your specification needs four lines and fits on half a page.

1. Protein percentage, with the band you require. For bread, 12 to 14 percent.
2. Falling Number in seconds. For bread, 250 to 280 seconds.
3. Moisture percentage, maximum. Codex sets a maximum of 15.5 percent by mass for wheat flour, and typical flour as milled is 14 to 15 percent. Storage problems begin above 16 percent.
4. Ash percentage, as agreed. Typical commercial flours sit below 0.6 percent, and patent flour goes as low as 0.35 percent. Note carefully that Codex sets no ash maximum at all and calls it buyer preference only, which means ash is a commercial specification you negotiate with your miller, never a legal one.

Ash matters because it rises with extraction rate, since minerals concentrate in the bran and germ. It is therefore a proxy for how much bran is in the flour, and by extension for colour and enzyme activity. A higher ash flour gives a darker crumb.

Avoid quoting European type numbers such as French T55, German 550 or Italian 00 to your supplier unless you can state the ash band you mean. Those are national ash-classification systems, and their exact bands were not available for this course. If you need them, obtain the French flour decree or German DIN 10355 rather than repeating trade folklore.

On delivery, check five things and write them on the batch sheet. Codex requires that at least 98 percent of wheat flour passes a 212 micrometre, number 70 sieve, so grossly coarse flour is out of specification. Squeeze a fistful for moisture feel. Smell for mustiness, which indicates damp storage. Look for insect activity and webbing. And record the lot number and delivery date, because that is your traceability link.

Contaminants are where you must go to your own authority rather than to this course. Codex sets a maximum of 5 micrograms per kilogram of ochratoxin A in raw wheat, barley and rye. Codex also requires flour to be free from heavy metals in hazardous amounts, to comply with pesticide and mycotoxin limits, and to be free from micro-organisms in hazardous amounts and free from parasites. But the maximum levels for deoxynivalenol, fumonisin and aflatoxin in cereals and flour were not available for this course. That is a real gap for an African bakery, because maize-containing and groundnut-containing bakery products carry genuine aflatoxin risk. Write to your national food safety authority and ask for the current national maximum levels for aflatoxin and deoxynivalenol in cereals and flour. Do not accept a figure from a foreign country as a substitute.

Flour treatment agents are the same story. Codex permits L-ascorbic acid to 300 mg/kg, benzoyl peroxide to 60 mg/kg, and azodicarbonamide to 45 mg/kg in flour for leavened bread. Several countries ban azodicarbonamide or benzoyl peroxide outright. Never treat a Codex number as a permission. Ask your national authority which treatment agents are permitted where you trade, and ask your miller in writing which ones they use.

Work the buying decision. A baker in Arusha, Tanzania, has two quotations for 50 kg bags. Supplier A is cheaper by an indicative USD 1.50 per bag but supplies no certificate of analysis. Supplier B is dearer but supplies protein, Falling Number, moisture and ash with every delivery.

1. Saving from A on 20 bags a month = 20 x 1.50 = USD 30.
2. One bad Falling Number lot costs him, from Module 1's arithmetic, roughly a week of production. At 200 loaves a day for six days at USD 0.50 full cost, that is 1,200 x 0.50 = USD 600 of cost against badly reduced revenue.
3. He would need twenty months of the saving to pay for one bad week.

That is the case for the certificate, in one calculation. Buy the paper, not just the flour.