Food quality analysis

In short

Food quality analysis is the systematic measurement of a food's composition, physico-chemical properties, contaminants, additives and microbial load, against a declared or legal specification. It answers three commercial questions: does the product meet the FSSAI standard and the label claim, is it safe to eat, and is it the same today as the batch approved last month. In India the legal reference methods are the FSSAI *Manual of Methods of Analysis of Foods* (volume-wise, by commodity), supported by the relevant Indian Standards of BIS and the international AOAC and ISO methods. The topic carries heavy marks in B.Sc Food Science papers on food analysis and quality control, and almost every QC/QA interview — Nestlé, ITC, Britannia, Amul (GCMMF), Mother Dairy, Hindustan Unilever, Parle, Dabur, Cargill, SGS, Eurofins, TUV, Equinox and the state Food Testing Laboratories — opens with proximate analysis and a titration calculation.

Explained

What "quality" means and how it is measured

• Quality = fitness for purpose, i.e. conformity to a specification plus consumer acceptability. Safety is a non-negotiable subset: a safe food may still be of poor quality, but an unsafe food is never of acceptable quality.
• Quality attributes are grouped as: (i) physical/appearance — size, shape, colour, texture, defects; (ii) chemical/compositional — moisture, protein, fat, ash, fibre, carbohydrate, additives, contaminants; (iii) microbiological — indicator and pathogenic organisms; (iv) sensory; (v) hidden — nutritive value, toxins, adulterants, which the consumer cannot judge.
• Objective vs subjective methods: instruments give objective, reproducible numbers (refractometer Brix, Lovibond colour, texture analyser); human panels give subjective data. Quality control prefers objective methods correlated to the sensory attribute.
• Destructive vs non-destructive: Kjeldahl destroys the sample; NIR, X-ray, machine vision, ultrasonics and metal detection do not, so they can be used on 100% of production (in-line) instead of a sample.
• Method categories: *reference/official* methods (slow, legally defensible — FSSAI Manual, AOAC), *routine* methods (faster, factory floor), and *rapid/screening* methods (NIR, ELISA, strip tests, ATP swabs) which must be calibrated against a reference method.
• Method validation parameters: accuracy (closeness to true value), precision (repeatability and reproducibility), specificity, LOD (limit of detection) and LOQ (limit of quantitation), linearity/range, robustness. Testing labs are accredited to ISO/IEC 17025 by NABL; FSSAI notifies/recognises laboratories for legal samples.
• Sampling is the weakest link. Representative sampling (random, stratified or systematic), correct composite preparation, quartering/coning for solids, and sealed, labelled sample containers. Under the FSS Act a Food Safety Officer draws samples in prescribed quantity, divides them into parts and sends one to the Food Analyst; an appeal lies to the referral laboratory.

Proximate analysis — the five-component scheme

1. Moisture. *Hot-air oven*: dry a weighed sample at 100–105 °C to constant weight; loss in weight = moisture. *Vacuum oven* (60–70 °C, reduced pressure) for sugar-rich and heat-sensitive foods to avoid caramelisation. *Toluene/Dean–Stark distillation* for spices and onions (volatile oils would otherwise be counted as water). *Karl Fischer titration* for very low moisture (dried milk, oils, chocolate, honey) — iodine–sulphur dioxide reaction specific to water. *Rapid*: infrared moisture balance, NIR, microwave moisture meter.
2. Ash (total mineral matter). Incinerate 2–5 g in a silica/platinum crucible in a muffle furnace at 550 ± 25 °C till a white-grey ash remains; residue = total ash. Acid-insoluble ash (ash boiled with dilute HCl, filtered, re-ignited) measures sand/dirt — the classic test for adulteration of spices and tea; FSSAI prescribes maximum limits. *Sulphated ash* used for some additives.
3. Crude fat / ether extract. Soxhlet continuous extraction of dried sample with petroleum ether or hexane, 6–8 h; solvent distilled off, flask dried and weighed. Gerber method (sulphuric acid + amyl alcohol, butyrometer, centrifuge) for milk — fast and legal for payment. Röse–Gottlieb (ammonia + alcohol + ether/petroleum ether, gravimetric) is the reference method for milk and milk products. Weibull–Stoldt (acid hydrolysis then extraction) for bound fat in bakery, cereals, chocolate.
4. Crude protein. Kjeldahl: (a) *digestion* with concentrated H₂SO₄ + catalyst (CuSO₄, formerly with mercury or selenium) converts organic N to (NH₄)₂SO₄; (b) *distillation* after adding excess NaOH liberates NH₃, trapped in boric acid or standard acid; (c) *titration* gives N. Multiply by a nitrogen factor (6.25 general, 6.38 milk, 5.70 wheat, 5.95 rice, 5.46 peanut, 6.25 meat/egg) to get crude protein. Dumas combustion (N analyser) is the modern fast equivalent. Kjeldahl measures *total* nitrogen — this is why melamine adulteration of milk could fool it; true protein needs amino-acid or dye-binding/electrophoretic methods.
5. Crude fibre. Defatted sample digested with 1.25% H₂SO₄ then 1.25% NaOH; residue washed, dried, ashed; loss on ashing = crude fibre. Underestimates true dietary fibre; total dietary fibre uses the enzymatic-gravimetric AOAC 985.29/991.43 method (α-amylase, protease, amyloglucosidase) and is what must be declared on the label. Van Soest fractions (NDF, ADF, ADL) are used for feeds.
6. Carbohydrate "by difference" = 100 − (moisture + protein + fat + ash + crude fibre), all in g/100 g. Energy from Atwater factors.

Physico-chemical quality tests

• pH (glass electrode, calibrated with pH 4.0, 7.0, 9.2 buffers) measures hydrogen ion activity; titratable acidity (titration with 0.1 N NaOH to phenolphthalein or to pH 8.2) measures total acid. pH decides microbial risk and thermal process design (pH 4.6 is the low-acid/acid food boundary); acidity decides taste and is the legal specification in milk (% lactic acid), juices (% citric acid) and vinegar (% acetic acid).
• Total soluble solids (TSS) by hand/digital refractometer in °Brix (g sucrose equivalent per 100 g solution), temperature-corrected to 20 °C. Used for jam, squash, juice, sauce; FSSAI standards for fruit products specify minimum TSS and minimum fruit content.
• Water activity (a_w) = p/p₀, the ratio of the water vapour pressure of the food to that of pure water at the same temperature; measured by dew-point or resistive hygrometer. Controls microbial growth and shelf life far better than moisture content.
• Specific gravity / density: lactometer for milk (adulteration with water lowers it), hydrometer for syrups; milk SNF is computed from CLR and fat.
• Colour: Lovibond tintometer for oils and honey, Hunter/CIE L\*a\*b\* colorimeter, and ΔE as the total colour difference; in oils, colour is a refining index.
• Texture/rheology: texture analyser (TPA: hardness, cohesiveness, springiness, chewiness), penetrometer for fruit firmness, Brookfield viscometer for sauces, farinograph/extensograph/alveograph and falling number for wheat flour, amylograph for starch.
• Oil and fat quality: acid value/FFA (rancidity by hydrolysis), peroxide value (primary oxidation, meq O₂/kg), p-anisidine and TBA value (secondary oxidation), iodine value (degree of unsaturation), saponification value (mean chain length, inversely), unsaponifiable matter, smoke point, nitric acid test for argemone oil in mustard oil (the Baudouin/Villavecchia test detects sesame oil, not argemone), and slip melting point for vanaspati/bakery fats. FSSAI prescribes maximum FFA and peroxide limits for edible oils and fried products.
• Enzyme and process-indicator tests: phosphatase test (negative = correctly pasteurised milk), peroxidase test (negative = blanched vegetables / over-pasteurised milk), MBRT and resazurin for raw-milk hygiene, alcohol and clot-on-boiling (COB) tests for milk acceptance at the dock.

Instrumental and advanced analysis

• UV–Visible spectrophotometry — colorimetric assays following Beer–Lambert law: protein (Lowry, Biuret, Bradford), reducing sugars (DNS, Nelson–Somogyi), vitamin C, nitrite, phosphorus, lycopene.
• Chromatography — HPLC for vitamins, sugars, preservatives (benzoate, sorbate), sweeteners, synthetic colours, aflatoxins (with fluorescence detector), caffeine; GC/GC-MS for fatty acid profile (as FAME), pesticide residues, flavour volatiles, trans fat; LC-MS/MS for multi-residue pesticides, antibiotics and mycotoxins at ppb; TLC and paper chromatography for quick screening of permitted vs non-permitted colours.
• Atomic spectroscopy — AAS (flame/graphite furnace) and ICP-OES/ICP-MS for Ca, Fe, Zn and heavy metals (Pb, Cd, Hg, As, Sn); samples prepared by dry ashing or microwave-assisted acid digestion. FSSAI's Contaminants, Toxins and Residues Regulations, 2011 set the metal limits.
• Immunoassay — ELISA kits for allergens (gluten, peanut, milk protein), mycotoxins, antibiotic residues; lateral-flow strips for on-site screening.
• Molecular — PCR/real-time PCR for pathogen confirmation, species identification (meat speciation — pork/beef in a product), and GM detection.
• Spectroscopic rapid tools — NIR/FT-NIR for simultaneous moisture, protein, fat and ash on grain and flour (needs calibration against reference values and periodic re-validation); FTIR milk analysers (fat, protein, SNF, lactose, added water) at dairy chilling centres; e-nose and e-tongue for aroma/taste fingerprinting.
• Physical inspection — metal detector, X-ray inspection for stone/glass/bone, checkweigher for net content, vacuum/seal integrity and headspace gas analysis for packaging.

Microbiological quality

• Indicator organisms (hygiene, not disease): Standard Plate Count / Total Viable Count on plate count agar at 30–37 °C for 48 h; coliforms and E. coli (VRBA, MPN, Petrifilm) — indicate faecal/post-process contamination; yeast and mould (potato dextrose agar, acidified); Enterobacteriaceae; spore count for canning and milk powder.
• Pathogens (zero-tolerance in most ready-to-eat categories): *Salmonella*, *Listeria monocytogenes*, *Staphylococcus aureus* (and its toxin), *E. coli* O157:H7, *Vibrio cholerae/parahaemolyticus*, *Bacillus cereus*, *Clostridium perfringens* and *Cl. botulinum*. Classical scheme = pre-enrichment → selective enrichment → selective plating → biochemical/serological confirmation; rapid alternatives = PCR, immunoassay, chromogenic media.
• Expressing results: cfu/g or cfu/mL, reported to two significant figures, usually as powers of ten; commercial sterility for canned foods, not absolute sterility.
• Sampling plans — Codex/ICMSF two-class plan (n, c, m) for pathogens and three-class plan (n, c, m, M) for indicator counts. FSSAI's Food Products Standards and Food Additives Regulations, 2011 and the amended microbiological requirements specify n, c, m, M by category.
• Environmental and shelf-life work: swab/contact plates on surfaces, air sampling, ATP bioluminescence for cleaning verification, accelerated shelf-life studies at elevated temperature/RH, challenge studies.

Adulteration, authenticity and labelling checks

• Common Indian adulterants: water, starch, urea, detergent and maltodextrin in milk; argemone oil in mustard oil; metanil yellow in turmeric and dal; Sudan dyes/lead chromate in chilli and turmeric; brick powder in chilli powder; chicory in coffee; papaya seed in black pepper; exhausted tea leaves; chalk in wheat flour; vanaspati in ghee; melamine as a nitrogen booster.
• Screening kits: FSSAI's DART (Detect Adulteration with Rapid Test) manual gives household-level tests; Food Safety on Wheels mobile labs and the Food Safety Magic Box are used for field screening. A screening positive must be confirmed by a notified laboratory before any legal action.
• Authenticity tools: stable isotope ratio analysis (added cane sugar in honey, juice), DNA barcoding (fish and meat species), fatty-acid and sterol profiling (ghee vs vanaspati — the Bomer value and RM/Polenske values for butter fat), SMP vs whole milk powder.
• Label and declaration verification: net quantity (Legal Metrology (Packaged Commodities) Rules, 2011), nutrition panel per 100 g/100 mL plus per serve, RDA percentages, declaration of added sugar/trans fat as per the Labelling and Display Regulations, 2020, veg/non-veg symbol, FSSAI licence number and logo, allergen declaration, date marking, and claims (Advertising and Claims Regulations, 2018). Analytical tolerances apply to nutrient values — confirm the current tolerance table in the Labelling and Display Regulations before certifying a non-conformance.
• Statistical quality control closes the loop: control charts (X̄–R for measurements, p and c charts for defectives/defects), process capability Cp/Cpk, acceptance sampling by attributes using sampling plans (IS 2500 series / ISO 2859), and trend review in the internal audit.

Key points

• Quality – conformity to specification plus consumer acceptance; safety is a mandatory subset of it.
• FSSAI Manual of Methods – the legally referable analytical methods in India; AOAC, ISO and BIS methods support it.
• NABL / ISO 17025 – accreditation that makes a laboratory's result technically defensible; FSSAI notifies labs for legal samples.
• Proximate analysis – moisture, ash, crude fat, crude protein, crude fibre; carbohydrate by difference.
• Moisture – oven 100–105 °C; vacuum oven for sugary foods; distillation for spices; Karl Fischer for very low moisture.
• Ash at 550 °C; acid-insoluble ash is the sand-and-grit (adulteration) index for spices and tea.
• Kjeldahl – digestion, distillation, titration; total N × factor (6.38 milk, 5.70 wheat, 6.25 general) = crude protein.
• Soxhlet for fat generally; Gerber and Röse–Gottlieb for milk; Weibull–Stoldt for bound fat.
• Crude fibre ≠ dietary fibre – label declaration needs the enzymatic-gravimetric total dietary fibre method.
• pH vs titratable acidity – pH is intensity of acid (free H⁺); acidity is the total quantity; pH 4.6 divides acid from low-acid foods.
• °Brix – TSS by refractometer, corrected to 20 °C; the legal minimum for jams, squashes and sauces.
• Water activity – p/p₀; predicts microbial growth and shelf life better than moisture content.
• Oil deterioration – acid value/FFA = hydrolytic rancidity; peroxide value = primary oxidation; p-anisidine/TBA = secondary oxidation; iodine value = unsaturation.
• Phosphatase negative = properly pasteurised milk; peroxidase negative = adequately blanched vegetable.
• HPLC/GC/LC-MS-MS, AAS/ICP, ELISA, PCR, NIR – the instrumental set for additives, residues, metals, allergens, species and rapid composition.
• Indicator vs pathogen – TPC, coliform, yeast-mould indicate hygiene; *Salmonella* and *Listeria* are safety criteria with two-class (n, c, m) plans.
• Sampling error dominates – a perfect analysis of an unrepresentative sample is a wrong result.
• SQC – control charts, Cp/Cpk and acceptance sampling convert test data into process control.

Formulas and examples

Example 1 — Moisture and total solids of a biscuit. Empty dish 22.4500 g; dish + sample 27.4500 g; dish + dried sample 27.2250 g.

• Sample taken = 27.4500 − 22.4500 = 5.0000 g
• Water lost = 27.4500 − 27.2250 = 0.2250 g
• Moisture = (0.2250 / 5.0000) × 100 = 4.50 % → total solids = 95.50 %
• Dry basis = (0.2250 / 4.7750) × 100 = 4.71 % (dry basis is always the larger figure)

Example 2 — Kjeldahl protein in milk. 1.00 g milk digested; liberated ammonia titrated with 0.1 N HCl, 9.40 mL used, blank 0.20 mL.

• Net titre = 9.40 − 0.20 = 9.20 mL
• N % = (9.20 × 0.1 × 14.007 × 100) / (1.00 × 1000) = (9.20 × 0.1 × 14.007) / 10 = 1.2886 %
• Crude protein = 1.2886 × 6.38 = 8.22 % — far above the ~3.2 % expected for milk, so either the sample was a concentrate or the titre/sample weight is wrong; a QC analyst must reject such a result rather than report it.
• Re-worked with the realistic titre of 3.60 mL (blank 0.20): net 3.40 mL → N % = (3.40 × 0.1 × 14.007)/10 = 0.4762 % → protein = 0.4762 × 6.38 = 3.04 %, a normal buffalo/cow mixed milk value.

Example 3 — Acidity of milk, and FFA of an oil. (a) 10.0 g milk titrated with 0.1 N NaOH, 1.80 mL to a phenolphthalein end point; lactic acid E = 90.

• Acidity = (1.80 × 0.1 × 90 × 100) / (10.0 × 1000) = (16.2 × 100)/10000 = 0.162 % lactic acid — just above the usual 0.13–0.16 % range for fresh cow milk, so the milk is beginning to sour.

(b) 5.00 g of refined oil titrated with 0.1 N KOH, 0.45 mL.

• Acid value = (0.45 × 0.1 × 56.1) / 5.00 = 2.5245/5.00 = 0.505 mg KOH/g
• FFA as oleic = (0.45 × 0.1 × 28.2)/5.00 = 1.269/5.00 = 0.254 % (check: 0.505 × 0.503 ≈ 0.254 ✓). FSSAI's limit for refined edible oils is of the order of 0.5 % FFA — confirm the exact figure for the specific oil in the current Food Products Standards Regulations.

Example 4 — Proximate analysis completed, and energy. A wheat-flour biscuit, per 100 g: moisture 4.5, ash 1.2, crude fat 20.0, crude protein 7.5, crude fibre 1.8.

• Carbohydrate by difference = 100 − (4.5 + 1.2 + 20.0 + 7.5 + 1.8) = 100 − 35.0 = 65.0 g
• Energy = 4(7.5) + 4(65.0) + 9(20.0) = 30 + 260 + 180 = 470 kcal/100 g
• If the label declares 455 kcal, the deviation is (470 − 455)/455 = +3.3 %, normally inside the permitted analytical tolerance for energy — verify against the current tolerance table before raising a non-conformance.

Example 5 — Peroxide value and a dilution-plate count. (a) 5.00 g of stored ghee, 0.01 N sodium thiosulphate, titre 1.25 mL, blank 0.05 mL.

• PV = (1.25 − 0.05) × 0.01 × 1000 / 5.00 = (1.20 × 10)/5.00 = 2.4 meq O₂/kg — low, so oxidation has not progressed; typical regulatory ceilings for edible fats lie near 10 meq O₂/kg (check the current notified limit for the commodity).

(b) 1 mL of a 10⁻⁴ dilution of a milk powder suspension gives 68 colonies; duplicate plate 72.

• Mean = 70 colonies → count = 70 × 10⁴ = 7.0 × 10⁵ cfu/g. If the specification is m = 5.0 × 10⁴ and M = 5.0 × 10⁵ cfu/g, the sample exceeds M and is unacceptable, not merely marginal.

Where it is used

• Dairy — reception-dock testing at Amul/GCMMF, Mother Dairy, Hatsun and Nandini chilling centres: COB, alcohol, MBRT, lactometer CLR, Gerber fat, FTIR analysers, phosphatase on pasteurised milk, and added-water/urea/detergent screening.
• Edible oils and fats — FFA, peroxide value, iodine and saponification values, colour, nitric acid test for argemone, trans-fat by GC: Adani Wilmar, Marico, Ruchi, Patanjali, and the refineries' QC laboratories.
• Packaged foods and beverages — Brix, acidity, viscosity, preservative and synthetic-colour checks by HPLC, net content and seal integrity at Britannia, Parle, ITC, Nestlé, Coca-Cola and PepsiCo bottling plants; shelf-life and stability studies in their R&D centres.
• Export and third-party testing — pesticide residue, aflatoxin and heavy-metal testing for APEDA/Spices Board shipments of rice, spices, grapes and marine products at SGS, Eurofins, TUV SUD, Intertek and the Export Inspection Agency laboratories.
• Regulatory and public laboratories — state Food Testing Laboratories and FSSAI-notified/referral labs analysing Food Safety Officer samples; the Food Analyst's report is the legal evidence in prosecution under the FSS Act, 2006.
• Job roles — QC Analyst/Chemist, QA Executive, Food Analyst, Food Safety Officer (state PSC), R&D Executive (product development and shelf life), Laboratory Technician/Manager, auditor for FSSC 22000 or BRC schemes, and instrument/application specialist with analytical equipment suppliers.

Interview questions

• Q: What is proximate analysis and why is carbohydrate found "by difference"? A: It is the determination of moisture, ash, crude fat, crude protein and crude fibre; carbohydrate is obtained as 100 minus the sum of these because measuring every individual carbohydrate directly is slow and unnecessary for a label.
• Q: Why is a vacuum oven used instead of a hot-air oven for honey or jam? A: At 100–105 °C sugars decompose and caramelise, releasing water of decomposition and giving a falsely high moisture; a vacuum oven dries at 60–70 °C under reduced pressure.
• Q: Kjeldahl gives total nitrogen — what is the practical risk? A: Any non-protein nitrogen (urea, melamine, ammonium salts) is counted as protein, so a correct Kjeldahl result can still conceal adulteration; true protein needs amino-acid, electrophoretic or specific immunochemical methods.
• Q: Why 6.38 for milk and 5.70 for wheat instead of 6.25? A: The factor is 100 divided by the average nitrogen percentage of that food's proteins; casein is about 15.7 % N and wheat gluten about 17.5 % N, so the factors differ.
• Q: Difference between total ash and acid-insoluble ash? A: Total ash is all mineral residue after ignition at 550 °C; acid-insoluble ash is the fraction not dissolved by dilute HCl, essentially silica — sand and grit — making it the standard adulteration index for spices and tea.
• Q: A product has 20 % moisture yet does not spoil. Why? A: Water activity, not moisture content, governs microbial growth; bound water in sugar or salt solutions lowers a_w (as in intermediate-moisture foods) below the level moulds and bacteria need.
• Q: Acid value is normal but peroxide value is high. What has happened? A: Oxidative rancidity has begun without hydrolysis — peroxides are the primary oxidation products, while acid value reflects free fatty acids from hydrolytic breakdown; the oil will develop off-flavour even though FFA is in specification.
• Q: What does a positive phosphatase test on pasteurised milk mean? A: Pasteurisation was inadequate or raw milk contaminated the pasteurised product, because alkaline phosphatase is destroyed by a correct time–temperature treatment.
• Q: Difference between crude fibre and dietary fibre? A: Crude fibre is the residue after acid and alkali digestion and loses most hemicellulose and soluble fibre; total dietary fibre is measured enzymatically and is the figure permitted on the nutrition label — it is substantially higher.
• Q: What are m and M in a microbiological specification? A: In a three-class sampling plan m is the acceptable level, M the level beyond which the sample is unacceptable, n the number of sample units and c the number allowed to lie between m and M.
• Q: Which methods are legally valid for a food sample in India? A: Those prescribed in the FSSAI Manual of Methods of Analysis of Foods, performed in an FSSAI-notified or NABL-accredited laboratory; rapid kits such as DART are screening tools only.
• Q: Your NIR analyser and the Kjeldahl result differ by 0.8 % protein. What do you do? A: Treat Kjeldahl as the reference, check the NIR calibration set and whether the sample lies within the calibration range and particle-size/temperature conditions, then re-calibrate or bias-correct the NIR — never adjust the reference to the rapid method.

Common mistakes

• Reporting moisture on a wet basis when the question asks dry basis (or the reverse) — dry basis is always numerically higher.
• Forgetting the blank correction, or using the normality of the wrong reagent, in Kjeldahl, acidity, peroxide and iodine-value calculations.
• Using 6.25 for every food, and calling Kjeldahl nitrogen "true protein".
• Mixing units — g with mg, mL with L, cfu/mL with cfu/g — and dropping the dilution factor when calculating a plate count.
• Treating pH and titratable acidity as interchangeable, or quoting pH as a percentage.
• Assuming a screening kit result (DART strip, lateral flow) is a legal finding without confirmation in a notified laboratory.
• Analysing carefully but sampling carelessly — a non-representative or badly stored sample invalidates the best instrument.

Related topics

• FSSAI and food laws in India · HACCP and food safety management · Sensory evaluation · Food microbiology and spoilage · Food chemistry and biochemistry · Food packaging and shelf-life studies