How Food Is Analysed: Labels, Additives and Contaminants

Pick up any packet of food and you’ll find a table of numbers: energy, fat, saturates, carbohydrate, sugars, protein, salt. Behind every one of those numbers is an analytical method, often more than a century old but still in daily use. Food chemists also check that food is safe (free from dangerous levels of contaminants), that additives stay within legal limits, and that what’s in the packet matches what it says on the label.

Proximate analysis: the big five

The traditional starting point is proximate analysis, which splits food into five broad components:

1. water (moisture)
2. ash (minerals)
3. protein
4. fat
5. carbohydrate

The first four are measured; carbohydrate is often calculated by difference.

1. Moisture

A weighed sample is dried in an oven (commonly at about 100–105 °C) until its mass stops changing. The loss in mass is the water content. This is essentially gravimetric analysis.

For foods that decompose or lose other volatile compounds on heating, alternatives include vacuum ovens at lower temperatures and Karl Fischer titration, a specific chemical titration for water.

Moisture matters for shelf life (microbes need water) and for price: nobody wants to pay for added water.

2. Ash

The dried sample is burned in a furnace at around 500–600 °C until only a white or grey residue remains. This ash contains the inorganic minerals: calcium, potassium, iron, phosphorus and so on. Individual minerals can then be measured by atomic absorption spectroscopy or ICP-MS.

3. Protein: measuring nitrogen

Proteins are hard to measure directly, but they contain nitrogen (in their amino groups), and most other food components don’t. So protein is usually found by measuring total nitrogen and multiplying by a conversion factor.

The Kjeldahl method (1883)

1. Digestion: the sample is heated with concentrated sulfuric acid and a catalyst. Organic matter is destroyed, and nitrogen is converted to ammonium sulfate.
2. Distillation: excess sodium hydroxide is added, releasing ammonia gas: NH₄⁺ + OH⁻ → NH₃ + H₂O. The ammonia is distilled into a known amount of acid (often boric acid).
3. Titration: the trapped ammonia is titrated with standard acid. This is a classic acid–base titration; related logic appears in back titration.

The Dumas method

The sample is burned at very high temperature in oxygen, converting nitrogen to nitrogen gas, which is measured by a thermal conductivity detector. It’s faster and avoids hot concentrated acid, and it’s now widely used.

The conversion factor

Most proteins contain about 16% nitrogen, so protein ≈ nitrogen × 6.25 (since 100 ÷ 16 = 6.25). Different factors are used for some foods, such as about 6.38 for milk.

Worked example: a 2.000 g sample of breakfast cereal contains 0.0368 g of nitrogen. Protein = 0.0368 × 6.25 = 0.230 g, so protein = 0.230 ÷ 2.000 × 100 = 11.5%, or 11.5 g per 100 g.

A weakness exploited

Because these methods measure nitrogen, not protein, they can be fooled by adding nitrogen-rich non-protein compounds. In 2008, melamine (C₃H₆N₆, about 67% nitrogen by mass) was illegally added to diluted milk and infant formula in China to fake higher protein readings, causing serious illness in many infants. Labs now also test specifically for melamine using chromatography–mass spectrometry. See percent composition to check that nitrogen figure yourself.

4. Fat

Fat dissolves in nonpolar solvents while most other food components don’t. In the classic Soxhlet extraction, a dried sample sits in a thimble while hot solvent (such as petroleum ether) repeatedly condenses on it, dissolves the fat and siphons back into a flask. After several hours, the solvent is evaporated and the fat left behind is weighed. See polar vs nonpolar molecules.

For the types of fat on a label (saturates, and in some countries mono- and polyunsaturates and trans fats), the fats are converted into volatile methyl esters and separated by gas chromatography, which shows the individual fatty acids.

5. Carbohydrate and sugars

Total carbohydrate is often calculated by difference:

Carbohydrate = 100 − (moisture + ash + protein + fat) (and fibre, where it’s reported separately)

Individual sugars (glucose, fructose, sucrose, lactose) are measured by HPLC, which separates them so each can be quantified. Fibre is measured by digesting the sample with enzymes that mimic human digestion and weighing what survives.

Energy

Energy values on labels are usually calculated from the composition using standard factors, not measured by burning the food:

Component	kJ per gram	kcal per gram
Fat	37	9
Protein	17	4
Carbohydrate	17	4
Alcohol	29	7

(Fibre and some other components have their own factors.) A bomb calorimeter can measure the total energy released on combustion, but that includes energy the body can’t extract, so calculated values better reflect digestible energy. See exothermic vs endothermic reactions.

Salt

Labels usually show salt, calculated from sodium: salt = sodium × 2.5 (because NaCl is about 39% sodium by mass, and 58.44 ÷ 22.99 ≈ 2.54). Sodium is measured by flame photometry, atomic spectroscopy or ICP. Chloride can also be measured by titration with silver nitrate. See precipitation titrations.

Vitamins and additives

• Vitamins are measured mostly by HPLC. Vitamin C can also be measured by a redox titration with iodine or with the dye DCPIP.
• Preservatives (such as benzoates and sorbates), sweeteners and colours are separated and measured by HPLC to check they’re within legal limits.
• Sulfites are measured to protect people with sensitivities.

Contaminants

Safety testing looks for substances that shouldn’t be there, or only in tiny amounts:

• Pesticide residues: screened by GC-MS and LC-MS/MS, which can test for hundreds of pesticides in one run.
• Mycotoxins: toxins from moulds, such as aflatoxins in nuts and grains, measured by HPLC with fluorescence detection or LC-MS/MS.
• Heavy metals: lead, cadmium, mercury and arsenic by ICP-MS (for example, mercury in fish, arsenic in rice).
• Process contaminants: such as acrylamide, which forms when starchy foods are fried or baked at high temperatures.
• Allergens: traces of peanut, gluten or milk, usually detected with antibody-based tests or by mass spectrometry.

Food fraud: is it what it claims to be?

Expensive foods attract fraud. Analysts can check:

• Honey adulterated with cheap sugar syrups, using carbon isotope ratios (plants that make sugar in different ways have slightly different ¹³C : ¹²C ratios).
• Olive oil diluted with cheaper oils, from its fatty acid profile and minor components.
• Fish species substitution, by DNA testing.
• Geographic origin (for example of wine or coffee), from patterns of trace elements and isotopes.

Key takeaways

• Proximate analysis measures moisture, ash, protein and fat; carbohydrate is often found by difference.
• Protein is estimated from nitrogen (Kjeldahl or Dumas) × about 6.25, which fraudsters once exploited with melamine.
• Fat is extracted with nonpolar solvents; fatty acid types are separated by GC.
• Label energy is calculated with standard factors; salt is sodium × 2.5.
• Chromatography and mass spectrometry detect additives, contaminants and fraud at very low levels.