Food Science & Technology · Food Process Engineering · also Unit Operations II, Mechanical Operations, Separation Processes in Food
The mechanical operations that clean, sort, reduce, mix, homogenise and separate food materials, including filtration, centrifugation and membrane processes, with the design equations used to size them.
In the exams: GATE XE-G Section 4 (mechanical operations); ICAR AIEEA-PG Code 18 Unit II (cleaning, sorting, size reduction, mixing, separation, filtration, membranes)
Before food can be heated, dried or packed, it must usually be cleaned, sorted, peeled, cut, ground, mixed, separated or conveyed. These mechanical unit operations change the size, shape, composition or position of food materials without relying mainly on heat. They decide yields and costs as much as any thermal process: a few per cent of peel loss, poor separation of cream or oil, a filter that blinds early, or a mixer that leaves ingredients unevenly distributed shows up directly in the profit and in the quality of the product.
This subject covers the operations named in GATE XE-G Section 4 (size reduction of solids, high-pressure homogenisation, filtration, centrifugation, settling, sieving, mixing and agitation of liquids) and in ICAR AIEEA-PG Code 18 Unit II (cleaning, de-husking, sorting and grading, peeling, size reduction, mixing and forming, separation, filtration and membrane processes, expression, extraction and leaching, crystallisation, distillation and material handling). Each chapter explains the principle, the governing equations used in exam problems and plant calculations, the main equipment types and their food applications.
Several neighbouring subjects meet this one. Agricultural Engineering's post-harvest-cereals-pulses-oilseeds covers cleaning, grading and milling machinery for grains, size-reduction energy laws as applied to grain, and conveyors and elevators for bulk grain; this subject applies the same principles across all foods and goes further into separations. Membrane processes introduced here are applied as concentration steps in evaporation-dehydration and in dairy processing; fluid flow and pumping are in food-transport-phenomena; cereal and oilseed product technology are in cereal-science-milling and fats-oils-technology. Most exam questions in this area are numerical — energy for size reduction, settling velocity, centrifugal force, filtration time, membrane flux and rejection, screen effectiveness and mixing power — with reasoning questions on choosing equipment.
Cleaning removes contaminants — soil, stones, sand, metal, glass, plant debris, insects, pesticide residues and microorganisms — from raw materials before processing. It protects consumers, protects equipment from damage (stones and metal in mills and pumps) and improves the efficiency of later operations. Dry cleaning methods include screening (separating by size), aspiration or winnowing (separating light chaff and dust in an air stream by differences in terminal velocity), magnetic separation (removing ferrous metal), abrasion (rubbing off adhering soil) and optical or X-ray detection of foreign bodies. Dry methods are cheap and leave the surface dry, which suits grains, nuts and pulses, but they create dust, with risks of recontamination and dust explosions.
Wet cleaning methods include soaking, spray washing, flotation washing and ultrasonic cleaning. They remove adherent soil and reduce microbial load more effectively and are gentler on fragile produce, but they use large volumes of water, create effluent and leave surfaces wet, which can favour spoilage unless produce is dried or processed quickly. Flotation separates sound produce from rotten or damaged pieces, stones and soil by differences in buoyancy; peas and beans, for example, are separated from heavier debris in froth flotation. Wash water may be chlorinated or treated with other sanitisers to prevent cross-contamination, within limits set by food law.
Sorting separates raw materials into categories by a single measurable physical property: size, shape, weight, colour or density. Size sorting uses screens (flat, vibrating, rotary drum) and roller or belt sorters with diverging gaps; weight sorting uses spring-loaded or electronic weighing heads, important for fruit and eggs sold by weight class; colour sorting uses photodetectors or cameras that compare reflected light at chosen wavelengths and remove off-colour items with air jets, widely used for rice, nuts, coffee beans and pulses. Machine vision systems combine size, shape, colour and surface defect detection, and near-infrared and hyperspectral systems can detect internal defects or composition.
Grading assesses overall quality by several attributes together and assigns produce to a grade, usually against official or commercial standards for each commodity — for example national grade standards for fruits, vegetables and grains, or the grades used in international trade. Grading may be manual, by trained inspectors at inspection belts, or increasingly by automated systems. The distinction matters in exam questions: sorting uses one property, grading an overall quality judgement.
Peeling removes skin, rind or husk with the least loss of edible flesh. The main methods are: flash steam peeling, in which produce is exposed to high-pressure steam for a short time and the pressure released suddenly, so that steam formed under the skin blows it off — widely used for potatoes and tomatoes, with low losses and little effluent; knife peeling, using fixed or rotating blades for fruits such as apples and citrus; abrasion peeling, in which rotating abrasive rollers or bowl surfaces rub off the skin, simple but with higher losses and large volumes of wet waste; caustic (lye) peeling, in which produce is dipped in hot dilute sodium hydroxide solution that softens and loosens the skin, which is then washed off — used for peaches, tomatoes and some root vegetables, but creating alkaline effluent; flame peeling, used for onions and peppers, in which the outer layers are burned off and the charred skin washed away; and enzymatic peeling of citrus segments.
Peeling efficiency is judged by the completeness of skin removal, the peeling loss (flesh removed with the peel), throughput, the quality of the peeled surface (no browning, softening or chemical residue) and the volume and strength of effluent. Losses vary widely with method and raw material, so methods are compared on trials with the actual crop.
Size reduction — cutting, crushing, grinding, milling, pulping and emulsifying — increases surface area, which speeds drying, heating, extraction and digestion; allows mixing and blending; produces particles or pieces of required size for a product; and separates components, as in flour milling. Most of the energy supplied to a mill is lost as heat and noise; only a small fraction creates new surface, so energy requirements are estimated by empirical laws relating energy to the change in particle size.
Rittinger's law (1867) states that the energy required is proportional to the new surface created: E = KR (1/x2 − 1/x1), where x1 and x2 are the initial and final particle sizes and KR is Rittinger's constant. It applies best to fine grinding, where new surface dominates. Kick's law (1885) states that the energy required to reduce size by a given ratio is the same whatever the initial size: E = KK ln(x1/x2). It applies best to coarse crushing, where energy goes mainly into elastic deformation before fracture. Bond's law (1952), intermediate between the two, states that energy is proportional to the change in the inverse square root of size: E = 10 Wi (1/√x2 − 1/√x1), with x in micrometres and E in kWh per tonne, where Wi is the work index — the energy, in kWh/t, to reduce a material from very large size to 80 per cent passing 100 µm. Bond's law is widely used for intermediate grinding and for scaling up mills.
Problems typically give the energy for one size reduction and ask for another. For example, if grinding sugar from 500 µm to 100 µm takes 3 kWh/t, Rittinger's law predicts that grinding the same material from 500 µm to 50 µm needs 3 × (1/50 − 1/500)/(1/100 − 1/500) = 3 × 0.018/0.008 = 6.75 kWh/t; Kick's law predicts 3 × ln(10)/ln(5) = 4.29 kWh/t. The difference shows why choosing the right law for the size range matters. With Bond's law and a work index of 12 kWh/t, reducing from 2,000 µm to 200 µm needs 10 × 12 × (1/√200 − 1/√2,000) = 120 × (0.0707 − 0.0224) ≈ 5.8 kWh/t. Particle sizes are usually characterised by a representative size such as the size through which 80 per cent of the material passes. In practice, none of the laws is exact for foods, which are often fibrous, moist or fatty rather than brittle; the constants are found by grinding trials on the actual material, and the laws are then used to predict the effect of changing feed or product size.
Equipment is chosen by the type of force it applies — compression, impact, shear (attrition) or cutting — and by the material's hardness, brittleness, fibre content, moisture and heat sensitivity. Crushers (jaw, gyratory, crushing rolls) use compression for coarse reduction of hard materials such as sugar cane or oilseed cake. Roller mills, with pairs of rolls turning at different speeds, combine compression and shear; fluted break rolls and smooth reduction rolls are the heart of wheat flour milling (cereal-science-milling). Hammer mills use high-speed swinging hammers to shatter brittle materials by impact, with a screen controlling product size; they are versatile for spices, sugar, grains and dried vegetables, but generate heat. Pin and disc mills use pins or discs rotating at high speed for fine grinding by impact and attrition. Ball mills grind by tumbling balls for very fine powders, though rarely in modern food plants.
Wet or fibrous foods need cutting and shearing equipment: slicers, dicers and shredders with sharp knives for fruits, vegetables and meat; bowl choppers (silent cutters) and mincers for meat; pulpers and finishers that force fruit through screens to separate pulp from skins and seeds; and colloid mills, in which product passes through a narrow gap between a fast rotor and a stator, for fine emulsions, nut butters and purées. Cryogenic grinding, in which spices or other heat-sensitive products are cooled with liquid nitrogen before milling, makes them brittle, prevents loss of volatile oils by heating and stops gumming of oily materials in the mill. Key selection factors are the required product size and distribution, throughput, heat generation, wear, hygiene and cleaning, dust and explosion risk, and energy use.
Sieving (screening) separates particles by size using a surface with openings of defined size. Particles smaller than the openings pass through as undersize (fines), and larger ones are retained as oversize. Industrial screens include vibrating flat screens, gyratory sifters (plansifters in flour mills), rotary drum (trommel) screens and centrifugal sifters; laboratory analysis uses a stack of standard test sieves of decreasing aperture shaken mechanically. Sieve apertures are specified in micrometres or by mesh number — the number of openings per linear inch — in national and international series (Tyler, US, ISO and others); because wire thickness differs between series, mesh numbers from different series do not give exactly the same aperture, and the aperture itself should be quoted.
Screening is never perfect: some undersize is carried over with the oversize, especially when screens are overloaded, particles are near the aperture size ('near-size' particles), the material is sticky or moist, or openings blind. Screen performance is measured by effectiveness. Using a mass balance with feed F, overflow (oversize) product D and underflow B, and the mass fraction of the desired coarse material xF, xD and xB in each stream, the effectiveness for the oversize is E_D = D xD/(F xF), for the undersize E_B = B(1 − xB)/(F(1 − xF)), and the overall effectiveness E = E_D × E_B.
A sieve analysis gives the mass fraction retained on each sieve. The results are plotted as a differential distribution (fraction in each size class) or a cumulative distribution (fraction smaller or larger than each size). Many ground food materials follow the Rosin–Rammler distribution or a log-normal distribution approximately, which allows the whole distribution to be summarised by two parameters.
Different mean sizes suit different purposes. The arithmetic mean size is simple but rarely the most useful. The surface–volume (Sauter) mean diameter, d32 = Σ nᵢ dᵢ³/Σ nᵢ dᵢ², is the diameter of a particle with the same surface-to-volume ratio as the whole population, and is used where surface area matters — in reaction, dissolution, drying, extraction and emulsions. The mass (volume) mean diameter d43 weights large particles more strongly. The size through which a stated percentage passes — d50 (median) or d80 — is used in milling, Bond calculations and specifications. In flour milling and cereal products, fineness and the proportion in each size range affect water absorption, baking performance and texture.
Modern particle sizing also uses laser diffraction, which measures the scattering pattern of light passing through a dispersion of particles and converts it to a volume-based distribution, and image analysis; these give much finer resolution than sieves, especially below about 50 µm, and are standard for emulsions, powders and fine flours. Results from different methods differ because they measure different properties, so the method must be stated.
The capacity and effectiveness of a screen depend on the aperture size and open area, screen motion (amplitude and frequency of vibration), the feed rate and bed depth, the particle shape, moisture and stickiness, and the proportion of near-size particles. Doubling the feed rate beyond the screen's capacity reduces effectiveness because the bed becomes too deep for fine particles to reach the mesh. Blinding — the plugging of openings by trapped particles — is reduced by ball or brush cleaners under the screen and ultrasonic vibration. In the food industry, sieving is used for grading flour and sugar, removing lumps and foreign matter as a safety check before packing (check sieves), grading of salt, spices, coffee and cocoa, and quality control of powders. Sieve analysis is also an official method in many product standards for granulation of semolina, sugar and similar products. Laboratory test sieves must themselves be checked periodically against calibration standards, because worn or damaged mesh changes the apparent particle-size distribution.
Mixing reduces non-uniformity in a material by distributing one component through another. It is used to blend ingredients in dry mixes, doughs and batters, to suspend solids, to dissolve sugar and salt, to disperse gases (aeration) and immiscible liquids (emulsification), and to improve heat and mass transfer in vessels. Mixing of low-viscosity liquids relies on convection and turbulent eddies; viscous liquids and pastes are mixed mainly by laminar shear and by repeated folding and cutting; dry solids are mixed by convection (bulk movement), diffusion (random movement of individual particles) and shear.
Solids mixing is complicated by segregation: particles that differ in size, density or shape tend to separate during mixing, discharge, conveying and storage, with small or dense particles percolating downward. A mixture of salt and herbs, for example, can be well mixed in the mixer and segregated in the hopper. Matching particle sizes, adding a little liquid to make particles cohere, minimising free fall and using mixers that avoid segregation help. The degree of mixing is measured by taking samples and comparing their composition variance with that of a completely segregated and a perfectly random mixture; the Lacey mixing index, M = (σ0² − s²)/(σ0² − σR²), runs from 0 (fully segregated) to 1 (perfectly random).
Low-viscosity liquids are mixed in tanks with impellers: propellers (axial flow, high speed, for blending and suspension), turbines (flat-blade Rushton turbines for radial flow and high shear; pitched-blade turbines for mixed flow) and paddles (large, slow, for gentle mixing). Baffles on the tank wall prevent the liquid from swirling as a whole and forming a vortex, converting rotation into top-to-bottom circulation. High-shear rotor–stator mixers disperse powders and form emulsions. Static mixers, fixed elements inside a pipe that repeatedly divide and recombine the flow, mix continuously with no moving parts.
Viscous liquids and pastes need close-clearance impellers: anchors and gates that sweep the wall (also improving heat transfer in jacketed vessels), helical ribbons and screws. Doughs and very viscous masses use kneaders (twin Z-blade or sigma mixers), planetary mixers, horizontal high-speed dough mixers, and continuous twin-screw mixers and extruders. Dry solids are blended in tumbling mixers (double-cone, V-blenders), which rely on gravity and are gentle but prone to segregation of dissimilar particles, and in convective mixers (ribbon blenders, paddle and plough-share mixers), which force material around and handle a wider range of materials, including small liquid additions.
The power drawn by an impeller depends on its type and diameter D, rotational speed N, the liquid density ρ and viscosity μ, and the tank geometry. Dimensional analysis gives two key groups: the impeller Reynolds number Re = ρ N D²/μ and the power number Np = P/(ρ N³ D⁵). For geometrically similar systems, Np is a function of Re only. In the laminar region (Re below about 10), Np is inversely proportional to Re, so P ∝ μ N² D³. In the fully turbulent region in a baffled tank (Re above about 10,000), Np is approximately constant for a given impeller type — commonly quoted as around 5 to 6 for a standard six-blade Rushton turbine and much lower, below 1, for propellers and hydrofoils — so P = Np ρ N³ D⁵. The steep dependence on speed and diameter means that doubling impeller speed in turbulent conditions raises power eightfold, and doubling diameter at the same speed raises it 32 times.
For example, a Rushton turbine with Np = 5.0 and D = 0.3 m turning at 3 rev/s in water (ρ = 1,000 kg/m³) draws P = 5.0 × 1,000 × 27 × 0.00243 ≈ 328 W. Note that N must be in revolutions per second. Without baffles, a vortex forms and the power number at high Re is much lower, and mixing poorer. Scale-up of mixing commonly keeps power per unit volume or impeller tip speed constant, depending on whether turbulence intensity or shear on fragile particles is critical.
Homogenisation reduces the size of dispersed droplets or particles in a liquid, most commonly fat globules in milk and cream, to give a stable, uniform product. In raw milk, fat globules have a mean diameter of a few micrometres (commonly quoted as about 3 to 4 µm) and rise to form a cream layer because they are less dense than the serum. Homogenisation reduces them to below about 1 µm, slowing creaming enormously, improving mouthfeel and colour (homogenised milk looks whiter because the many small droplets scatter more light) and creating new fat-globule surface covered by casein and whey proteins.
In a high-pressure valve homogeniser, a positive-displacement piston pump forces the liquid at high pressure through a narrow adjustable gap between a valve and its seat. As the liquid accelerates through the gap to very high velocity, the pressure energy is converted into kinetic energy and dissipated as intense turbulence, shear and cavitation, which break up the droplets. The droplets are then stabilised, if enough emulsifier is present, by rapid adsorption of proteins or emulsifiers on the new surface. Without stabilisation, small droplets may cluster or coalesce again, so many homogenisers have a second stage at lower pressure to break up clusters formed after the first stage.
For milk, typical total homogenising pressures are in the range of about 10 to 25 MPa, with the second stage at a fraction of the first; higher pressures give smaller droplets, approximately following an inverse power relation between mean droplet size and pressure. Milk is homogenised warm (commonly 55 to 80 °C) so that the fat is liquid; cold homogenisation of milk with partly solid fat is ineffective. Raw milk is usually homogenised before or after pasteurisation in the same line, because homogenisation of raw milk without heat treatment activates lipase on the new fat surface and produces rancid flavours rapidly.
Homogenisation is used for milk, cream, ice-cream mix, evaporated and condensed milk, infant formula, salad dressings, sauces, soups, fruit juices and nectars (to stabilise cloud and give body), peanut butter and other nut pastes, and plant-based beverages. Valve design, number of stages, temperature, pressure and the composition of the product (fat content, protein, emulsifiers) all affect the result. Ultra-high-pressure homogenisation, at pressures up to about 200 to 400 MPa, is being studied for producing very fine emulsions and for microbial inactivation, which arises from the intense shear and the heating that accompanies pressure release.
The reason homogenisation slows creaming is given by Stokes' law: the velocity at which a sphere rises or settles in a viscous liquid under gravity is v = d² (ρp − ρf) g/(18 μ), where d is diameter, ρp and ρf the densities of particle and fluid and μ the viscosity of the fluid. Velocity depends on the square of diameter, so reducing the mean fat-globule diameter from 4 µm to 1 µm cuts the creaming velocity by a factor of 16. Creaming can also be slowed by reducing the density difference or increasing the viscosity of the continuous phase, which is why thickeners stabilise dressings and beverages. Homogeniser efficiency is checked by measuring droplet size distribution (by laser diffraction) or by a creaming test, in which the fat content of the top layer after storage is compared with the rest. Homogenisers are among the most energy-intensive machines in a dairy, and their high-pressure pumps, valves and seals need careful maintenance, because worn valves give larger droplets and creaming complaints.
Gravity settling separates solid particles or immiscible droplets from a liquid (or gas) by the difference in density. For small particles in laminar conditions (particle Reynolds number below about 1), the terminal settling velocity follows Stokes' law, v = d²(ρp − ρf)g/(18μ). For example, a starch granule 20 µm in diameter with density 1,500 kg/m³ settling in water (ρ = 1,000 kg/m³, μ = 0.001 Pa s) has v = (20 × 10⁻⁶)² × 500 × 9.81/(18 × 0.001) ≈ 1.09 × 10⁻⁴ m/s, about 0.39 m per hour. At higher Reynolds numbers, drag coefficients must be used instead of Stokes' law.
Gravity settling is used in clarifying tanks for wash water and effluent, in separating starch from potato and maize slurries (though centrifuges now dominate), in settling yeast and trub in breweries, and in air classifiers and cyclones for dust. Settling tanks are designed on the overflow rate: a particle is removed if its settling velocity exceeds the liquid flow rate divided by the tank's surface area, so capacity depends on area rather than depth. In concentrated suspensions particles hinder each other, and settling is slower than predicted for single particles.
Centrifuges replace the acceleration due to gravity g with the centrifugal acceleration rω², where r is the radius and ω the angular velocity in rad/s (ω = 2πN/60 for N in rpm). The ratio rω²/g, the relative centrifugal force or g-number, is commonly several thousand in food separators. A bowl of radius 0.2 m spinning at 6,000 rpm (ω = 628 rad/s) gives rω² = 0.2 × 628² ≈ 79,000 m/s², about 8,000 times gravity. The settling velocity in a centrifuge is therefore v = d²(ρp − ρf) r ω²/(18μ), thousands of times greater than under gravity, allowing separation of fine particles and small droplets in seconds.
In liquid–liquid separation, as in a cream separator, the two phases form concentric layers, the denser outer and the lighter inner, divided at the neutral zone. The radius of the neutral zone, and so the positions of the outlets, is set by the densities of the phases and the outlet radii, and is adjusted by the outlet design (ring dams) to balance the purity of the two streams. Separators are also rated by the sigma (Σ) concept: a centrifuge's capacity equals the gravity settling velocity of the cut-size particle multiplied by Σ, a factor with units of area representing the equivalent settling area of the machine, which allows different centrifuges to be compared and scaled.
Tubular-bowl centrifuges are long narrow bowls spinning at very high speed, giving high g but small solids capacity; they are used for clarifying oils and separating fine solids at small scale. Disc-stack centrifuges contain a stack of conical discs that divides the liquid into thin layers, so particles need only travel a short distance to a disc surface before sliding outward; they are the standard for milk separation and clarification, bactofugation (removing bacterial spores from milk), juice and beer clarification, oil refining and yeast recovery, with solids discharged intermittently or continuously through nozzles or opening bowls. Decanter (scroll) centrifuges have a horizontal conical bowl with an internal screw conveyor turning slightly slower than the bowl, which continuously scrapes settled solids to one end; they handle high solids contents and are used for olive oil, fish meal and fish oil, starch, protein recovery, juice extraction and effluent sludge. Basket centrifuges with perforated walls act as centrifugal filters and are used to separate sugar crystals from molasses. Selection depends on solids content and particle size, the density difference, the required clarity and dryness of the solids, throughput and hygiene.
Filtration separates solids from a liquid by passing the suspension through a porous medium that retains the solids. In cake filtration, used when solids content is appreciable, a layer of retained solids — the cake — builds up on the filter medium (cloth, mesh, paper) and itself becomes the main filtering layer. In depth filtration, used for clarifying liquids with few solids (polishing beer, wine and juice), particles are trapped within a thick porous medium. Filtration may be driven by gravity, pressure, vacuum or centrifugal force.
The flow of filtrate through the cake and medium is laminar and follows Darcy's law: the rate is proportional to the pressure difference and inversely proportional to the liquid viscosity and the total resistance. For constant-pressure cake filtration this leads to Ruth's equation in the form t/V = (μ α c/(2 A² ΔP)) V + μ Rm/(A ΔP), where t is time, V the filtrate volume, A the filter area, ΔP the pressure difference, μ the filtrate viscosity, α the specific cake resistance, c the mass of dry cake per unit volume of filtrate and Rm the resistance of the filter medium. A plot of t/V against V from a laboratory test gives a straight line whose slope yields α and whose intercept yields Rm, and these are used to size full-scale filters. If the medium resistance is negligible, the time to collect a volume V at constant pressure is t = μ α c V²/(2 A² ΔP): doubling the filtrate volume takes four times as long, and doubling the area cuts the time to a quarter.
Many food cakes — of yeast, proteins, fruit pulp, fats or fine solids — are compressible: their specific resistance increases with pressure as the cake packs more densely, often approximated by α = α0 ΔP^s, where s is the compressibility (0 for incompressible, approaching 1 for highly compressible cakes). For highly compressible cakes, raising the pressure gives little or no increase in filtration rate, and slimy or gelatinous solids can blind the medium almost completely.
Filter aids overcome these problems. They are rigid, porous, inert powders, chiefly diatomaceous earth (kieselguhr), perlite and cellulose fibres, used in two ways: as a precoat, a thin layer deposited on the medium before filtration to protect it and give a clear filtrate from the start, and as body feed, mixed continuously with the suspension so that the growing cake stays porous and permeable. Filter aids are widely used in clarifying beer, wine, juices, sugar syrups and edible oils; disposal of spent diatomaceous earth and worker exposure to its dust are concerns that have encouraged the use of membrane and centrifugal alternatives.
Plate-and-frame filter presses consist of alternate plates and frames covered with filter cloths and clamped together; slurry is pumped in under pressure, the cake builds up in the frames, and the press is opened to remove the cake. They are simple and flexible, give dry cakes and handle high pressures, but they are batch machines needing labour for cake discharge. Leaf and candle filters, in which filter elements hang in a pressure vessel, are used with filter aids for beer and oils and can discharge cake automatically. Rotary vacuum drum filters rotate a cloth-covered drum partly submerged in the slurry; vacuum inside draws filtrate through and forms cake, which is washed, dewatered and scraped off as the drum turns — a continuous process used for starch, sugar juice and yeast. Belt filters and filter centrifuges are other continuous types. In all cases, washing the cake recovers soluble product (or removes impurities), and the choice of medium, pressure, cycle time and filter aid determines throughput and clarity.
Membrane processes separate components of a liquid by passing it under pressure over a semi-permeable membrane: part of the liquid, the permeate, passes through, and the retained part, the retentate (concentrate), is enriched in the larger or retained species. They work at low temperature, without phase change, and with lower energy use than evaporation. They are classified by the size of species retained. Microfiltration (MF), with pores of roughly 0.1 to 10 µm, retains bacteria, spores, fat globules and suspended solids at low pressures (commonly up to about 0.2 to 0.5 MPa); it is used to remove bacteria from milk and brine, clarify juices and beer, and fractionate milk proteins. Ultrafiltration (UF) retains macromolecules such as proteins, characterised by molecular-weight cut-off, typically from about 1 to several hundred kDa, at pressures commonly about 0.1 to 1 MPa; it passes water, lactose, salts and small molecules, and is used to concentrate milk and whey proteins, to standardise protein in cheese milk and to clarify juices. Nanofiltration (NF) retains sugars and divalent ions while passing water and monovalent ions; it is used to partly demineralise and concentrate whey. Reverse osmosis (RO) uses dense membranes that pass essentially only water, at high pressures (commonly several MPa), to concentrate milk, whey, juices and sap and to purify water.
The permeate flux J (volume per unit membrane area per unit time, often L/m² h) follows a resistance model: J = (ΔP − Δπ)/(μ Rt), where ΔP is the transmembrane pressure, Δπ the osmotic pressure difference across the membrane, μ the permeate viscosity and Rt the total resistance (membrane plus fouling and polarisation layers). In MF and UF the retained species are large and their osmotic pressure is small, so flux rises with pressure until a limiting flux is reached, set by the concentration polarisation and gel layer; in RO, osmotic pressure, which rises with solute concentration (π = i c R T), opposes the applied pressure, limiting the concentration achievable — commonly to about 25 to 30 per cent solids for fruit juices, as noted in evaporation-dehydration.
Separation performance is described by the rejection (retention) coefficient of each solute, R = 1 − Cp/Cf, where Cp and Cf are its concentrations in permeate and feed: R = 1 means complete retention, R = 0 free passage. The volume concentration factor VCF = Vfeed/Vretentate describes how far the feed has been concentrated: concentrating 1,000 L of skim milk to 250 L of retentate is a VCF of 4. For a completely retained solute, its concentration in the retentate rises by the VCF; partly retained solutes rise less. Diafiltration — adding water to the retentate and filtering again — washes out permeable components such as lactose and salts to raise the purity of retained protein, as in making whey protein isolates.
As permeate passes through, retained solutes accumulate at the membrane surface, forming a boundary layer more concentrated than the bulk: concentration polarisation. It raises the local osmotic pressure, adds resistance and can lead to a gel layer of proteins on the surface. Fouling — deposition and adsorption of proteins, fats, minerals (especially calcium phosphate in dairy) and microorganisms on or in the membrane — causes a further, often irreversible, decline in flux over a run. Both are reduced by cross-flow operation, in which the feed flows rapidly parallel to the membrane surface, sweeping away the boundary layer, in contrast to dead-end filtration; by controlling transmembrane pressure below the limiting flux, pretreatment of the feed, temperature and pH; and by regular cleaning-in-place with alkaline, acid and enzyme detergents compatible with the membrane material.
Membranes are made of polymers (polysulphone, polyethersulphone, polyamide thin-film composites, cellulose acetate) or ceramics (alumina, zirconia), which tolerate high temperatures, extreme pH and harsh cleaning and are favoured for MF of milk. Modules include spiral-wound (compact and economical, widely used for UF and RO of dairy liquids), hollow-fibre, tubular (tolerant of particles and viscous feeds) and plate-and-frame designs. Membrane plants are designed from flux data obtained on the actual product, because flux, fouling and cleaning behaviour vary widely between products.
Expression separates liquid from a solid by applying pressure, squeezing out juice or oil held within cells. It is used to extract fruit juices (apples, grapes, citrus, sugarcane), oils from oilseeds and fruits (mechanical pressing of rapeseed, groundnut, olive and coconut), and to dewater materials such as spent grains or pulp. The yield depends on the pressure, the time it is held, the rate at which pressure is applied, the disruption of cells beforehand, the temperature, the viscosity of the liquid, and the thickness and porosity of the pressed bed, through which liquid must escape.
Pre-treatments increase yield: size reduction (milling apples to a mash, flaking oilseeds), heating (cooking oilseeds lowers oil viscosity and denatures proteins that hold oil), and enzymatic treatment (pectinases break down pectin that holds juice in fruit mash and makes it slimy). Equipment includes batch presses such as rack-and-cloth and hydraulic presses, continuous belt presses that squeeze mash between belts over rollers, and screw presses (expellers), in which a rotating worm in a barrel with decreasing volume compresses the material; expellers are standard for oilseeds and leave an oil-rich cake that is either sold or solvent-extracted. Citrus juice extractors use reamers or cup-type extractors that separate juice from peel oil and peel.
Solid–liquid extraction, or leaching, dissolves a soluble component out of a solid into a solvent. Examples are extraction of sugar from sugar beet cossettes with hot water, of oil from oilseed flakes or press cake with hexane, of soluble coffee and tea solids with hot water, of caffeine from green coffee, and of pigments, flavours and bioactive compounds from plant materials. The process involves solvent penetrating the solid, dissolution of the solute, diffusion of the solution to the surface and transfer into the bulk solvent. The rate is increased by reducing particle size (shorter diffusion paths, though very fine particles impede drainage), raising temperature (higher solubility and diffusivity, lower viscosity), agitation, and maintaining a high concentration difference.
Extraction is usually carried out in stages. In a single stage, solid and solvent are mixed, allowed to reach (or approach) equilibrium, and separated; some solution always remains held in the solid (the underflow), limiting recovery. Multistage counter-current extraction, in which fresh solvent meets the most exhausted solid and the most concentrated solution meets fresh solid, gives much higher recovery with less solvent and a more concentrated extract. Its design uses equilibrium-stage calculations and diagrams that account for the solution retained by the solids. Industrial extractors include percolation extractors (solvent trickles through beds of flakes in moving baskets or on belts, as in oilseed extraction), immersion extractors and diffusers (continuous counter-current diffusers for sugar beet).
In oilseed processing, flakes or pre-pressed cake are extracted with food-grade hexane, which dissolves oil readily and is easily recovered by distillation because of its low boiling point. Solvent extraction leaves less than about 1 per cent oil in the meal, compared with several per cent after pressing alone. The miscella (oil–solvent solution) is distilled to recover oil, and the meal is desolventised and toasted, which also inactivates anti-nutritional factors such as trypsin inhibitors in soybean meal. Hexane is flammable, so plants need strict explosion protection, and residual solvent in oil and meal is limited by food and feed law. Alternatives such as ethanol, aqueous extraction and supercritical carbon dioxide (next chapter) are used or studied for particular products. Refining of the crude oil is covered in fats-oils-technology. For mechanically pressed 'cold-pressed' or virgin oils, solvent extraction is not used at all, and the lower yield is accepted in return for a product sold on its natural character.
Crystallisation forms solid crystals from a solution or melt. It is used to produce sugar, salt, lactose, citric acid and monosodium glutamate, to fractionate fats (separating solid stearin from liquid olein), to make products whose texture depends on crystals (fondant, fudge, chocolate, ice cream, butter) and, in freeze concentration, to remove water as ice. A solution can hold only a limited amount of solute at a given temperature, its solubility; for most sugars and salts solubility rises with temperature. A solution containing more solute than its solubility is supersaturated, and supersaturation is the driving force for crystallisation. It is created by cooling (for solutes whose solubility rises steeply with temperature, such as lactose), by evaporating solvent (as for sucrose and salt), or by adding a substance that reduces solubility.
Between the solubility curve and a second, higher curve lies the metastable zone, in which existing crystals grow but new nuclei form only slowly; above it, in the labile zone, spontaneous nucleation is rapid and uncontrolled. Controlled industrial crystallisation keeps the solution within the metastable zone, so that crystals grow on deliberately added seed rather than forming fine new nuclei.
Crystallisation has two stages. Nucleation is the formation of new crystals: primary nucleation occurs without existing crystals, homogeneously in a clear solution or heterogeneously on foreign particles and surfaces; secondary nucleation is induced by existing crystals, for example by fragments broken off by agitation. Crystal growth is the deposition of solute on existing crystals, controlled by diffusion of solute to the surface and its incorporation into the lattice. The balance between nucleation and growth determines the crystal size distribution: high supersaturation favours nucleation and many small crystals, low supersaturation favours growth of fewer, larger crystals.
The product requirement decides which is wanted. Granulated sugar needs large, uniform crystals that separate cleanly from the mother liquor; fondant, fudge and sweetened condensed milk need very many tiny crystals (below about 10 to 20 µm) so that they feel smooth rather than gritty; in ice cream, small ice and lactose crystals are wanted. Seeding with finely ground crystals, controlled cooling rates, agitation, and additives that inhibit or promote crystallisation (such as invert sugar or glucose syrup in confectionery, which interfere with sucrose crystallisation) are the tools for control.
Sucrose is crystallised in vacuum pans: concentrated syrup is boiled under vacuum at about 65 to 70 °C, seeded, and fed with more syrup while water is evaporated, keeping supersaturation in the metastable zone while crystals grow. The resulting massecuite — crystals in mother liquor — is centrifuged in basket centrifuges to separate the crystals, which are washed and dried; the mother liquor is crystallised again in further boilings, and the final exhausted liquor is molasses. Lactose is crystallised from concentrated whey by controlled cooling. Fats are fractionated by controlled cooling of melted fat and filtration of the solid fraction (dry fractionation), as in producing palm olein and stearin.
The theoretical yield of crystals from a cooling crystallisation follows from a mass balance using solubility data. For example, if 100 kg of water saturated with lactose at 60 °C holds about 59 kg of lactose (solubility 59 g per 100 g water) and is cooled to 20 °C where solubility is about 19 g per 100 g water, about 59 − 19 = 40 kg of anhydrous-basis lactose can crystallise, neglecting water of crystallisation; lactose actually crystallises as α-lactose monohydrate, which takes some water into the crystal and slightly changes the yield. Real yields are lower than theoretical because crystallisation does not reach equilibrium in the time available. Solubility values here are illustrative textbook figures; design uses measured data for the actual product.
Distillation separates the components of a liquid mixture by their different volatilities: when the mixture is partly vaporised, the vapour is richer in the more volatile component. For ideal mixtures, Raoult's law gives each component's partial pressure as its mole fraction in the liquid times its pure-component vapour pressure, and Dalton's law converts partial pressures to vapour composition. The relative volatility α of two components, the ratio of their vapour-to-liquid composition ratios, measures how easy the separation is: the larger α, the fewer stages needed; α = 1 means no separation by distillation. Many food mixtures are non-ideal. Ethanol and water form an azeotrope at about 95.6 per cent ethanol by mass (about 89 mole per cent) that boils at about 78.2 °C at atmospheric pressure, below the boiling point of either pure component; ordinary distillation cannot concentrate ethanol beyond this, and other methods (molecular sieves, entrainers) are used to make anhydrous ethanol.
In simple (batch or flash) distillation, the vapour from a single equilibrium stage is condensed; it gives only partial separation. Fractional distillation uses a column with trays or packing in which rising vapour and descending liquid (reflux) are repeatedly brought into contact, giving the effect of many equilibrium stages and producing nearly pure products at top and bottom. The number of theoretical stages needed is estimated graphically by the McCabe–Thiele method or by computer simulation, and the reflux ratio trades energy (more reflux, more boiling) against the number of stages.
Distillation is used to make spirits — whisky, brandy, rum, vodka and traditional spirits — from fermented mashes and wines, in pot stills (batch, retaining more congeners and flavour) or continuous column stills (higher strength, lighter spirit); to recover and concentrate aroma compounds (essence) from the first vapours of fruit-juice evaporators, as described in evaporation-dehydration; to recover solvent from oil miscella; to deodorise and strip volatile compounds from oils and other liquids; and to dealcoholise beer and wine under vacuum at low temperature. Steam distillation, in which live steam is passed through plant material, carries essential oils over at temperatures below their normal boiling points because the immiscible oil and water each exert their own vapour pressure; it is the main method for producing essential oils from spices, herbs, citrus peel and flowers (spices-flavours-essential-oils). Molecular (short-path) distillation under high vacuum is used for heat-sensitive compounds such as monoglycerides and vitamin concentrates.
Above its critical temperature and pressure, a substance becomes a supercritical fluid with properties between those of a gas and a liquid: liquid-like density and solvent power, combined with gas-like low viscosity and high diffusivity, which allow rapid penetration of solid materials. Carbon dioxide is by far the most used supercritical solvent in foods: its critical point is about 31.1 °C and 7.38 MPa, so it can be used at mild temperatures that protect heat-sensitive compounds; it is non-toxic, non-flammable, inexpensive and leaves no solvent residue because it returns to gas at atmospheric pressure. Its solvent power depends strongly on density, which is controlled by pressure and temperature, so selectivity can be tuned, and dissolved solutes are recovered simply by reducing the pressure in a separator.
Supercritical CO₂ is non-polar and dissolves mainly lipophilic compounds; small amounts of a co-solvent such as ethanol extend its range to more polar compounds. Its main food applications are decaffeination of coffee and tea, extraction of hop resins and oils for brewing, extraction of spice oleoresins and essential oils, removal of fat and cholesterol from some products, and extraction of carotenoids, omega-3 oils and other bioactives. The high-pressure equipment is expensive, so the method is used for high-value products or where solvent-free 'natural' extracts command a premium.
Material handling moves raw materials, intermediates, packaging and finished products through a plant. Good handling keeps products moving at the right rate between operations, minimises damage, contamination, spillage and labour, and is designed for cleaning and safety. Handling equipment is chosen by the nature of the material — bulk granular solids, powders, fragile pieces, packaged goods or liquids — the distance and direction (horizontal, inclined, vertical), the capacity, hygiene requirements and cost. Liquids are moved by pumps and pipelines (food-transport-phenomena); this chapter covers solids. The engineering of conveyors and elevators for bulk grain in silos and warehouses is also treated in Agricultural Engineering's storage-engineering-packaging. A useful first step in designing any line is to draw the flow of materials through the plant, identifying every transfer point, because each transfer is a potential source of damage, spillage, contamination and dust, and the best handling system is often the one with the fewest transfers.
Belt conveyors carry material on a continuous belt over idler rollers and are used for bulk solids, pieces and packages over long distances, horizontally or on moderate inclines; food-contact belts are made of approved plastics or stainless-steel mesh and designed for easy cleaning. Their capacity is the product of the cross-sectional area of the load on the belt, the belt speed and the bulk density. Screw (auger) conveyors move granular materials and powders along a trough or tube by a rotating helical screw; they are enclosed, compact and can mix and meter, but they can damage fragile materials and are harder to clean. Their volumetric capacity is approximately Q = (π/4)(D² − d²) p N φ, where D is the screw diameter, d the shaft diameter, p the pitch, N the speed and φ the trough loading fraction (commonly about 0.3 to 0.45 depending on material). For example, a screw of 0.2 m diameter, 0.05 m shaft, 0.2 m pitch running at 60 rpm (1 rev/s) at 0.4 loading moves (π/4)(0.04 − 0.0025) × 0.2 × 1 × 0.4 ≈ 0.00236 m³/s, about 8.5 m³/h.
Roller conveyors (gravity or powered) move packages and cartons. Vibratory conveyors move material in small hops along a vibrating trough and are gentle and easy to clean, suiting snacks, frozen products and fragile pieces. Pneumatic conveyors move powders and granules through pipes in an air stream: dilute-phase systems use high air velocity with material in suspension, simple but more abrasive and damaging; dense-phase systems move material as slugs at low velocity, gentler on products such as sugar, flour and coated pieces. Pneumatic systems are enclosed and hygienic but use more energy, and need dust filters and explosion protection. Hydraulic (water) flumes transport and wash fruits and vegetables simultaneously.
Bucket elevators raise bulk solids vertically in buckets fixed to a belt or chain. Centrifugal-discharge elevators run fast and throw material out of the buckets at the head pulley, suiting free-flowing grains; continuous-discharge elevators run slower with closely spaced buckets so that material slides off onto the back of the preceding bucket, suiting fragile or dusty materials. Inclined belt conveyors with cleats, and Z-type bucket conveyors with pivoting buckets, are used for gentle elevation of products such as snacks, vegetables and frozen foods. Fork-lift trucks, automated guided vehicles, overhead conveyors and robotic palletisers handle packaged goods and are increasingly automated.
All food-contact handling equipment must be hygienically designed: smooth, non-absorbent, corrosion-resistant surfaces, no crevices or ledges where product can lodge, easy dismantling or cleaning in place, and protection against contamination by lubricants and foreign bodies. Dust from conveying flour, sugar, starch and milk powder creates a dust-explosion hazard, controlled by enclosure, dust extraction, earthing to remove static, explosion venting or suppression and housekeeping. Handling equipment is also a common source of foreign-body contamination — belt fragments, metal from worn screws — so metal detectors and visual checks follow conveying steps in many lines.
A student uses this subject for the numerical questions on Rittinger, Kick and Bond energies, screen effectiveness, Stokes and centrifugal settling, filtration time, membrane flux, rejection and concentration factor, agitator power and conveyor capacity, and for equipment-choice questions. In the pilot plant, the same student runs a sieve analysis, grinds a material at different settings and compares the energy used, separates cream with a laboratory separator, filters a slurry at constant pressure to find cake resistance, and measures flux decline in a small ultrafiltration rig.
A process engineer or production manager five years in sets peeler conditions to balance skin removal against yield, maintains mills and screens to keep particle size on specification, tunes separators and clarifiers for fat and solids specifications, schedules membrane cleaning when flux falls, chooses filter aids and pressures for oil or juice filtration, adjusts homogeniser pressure for creaming tests, and designs or modifies conveying so that products are not damaged or contaminated, with metal detection after conveying steps. They also keep the yield records that show where product is being lost.
A senior engineer or plant director chooses separation technology for new lines — centrifuge versus filter versus membrane, expeller versus solvent extraction, pot versus column still — on the basis of yield, product quality, energy, effluent, safety (hexane, dust explosions) and capital cost. They lead yield-improvement programmes, since a fraction of a per cent in peeling loss, cream or oil recovery is worth large sums in high-volume plants, and they set hygienic-design and dust-safety standards for equipment purchases. They also manage effluent and by-product streams from peeling and separation.
A research leader works on separation science and equipment for future foods: membranes and chromatographic methods for fractionating plant and dairy proteins, low-solvent and green extraction (supercritical, pressurised liquid and enzyme-assisted extraction) of bioactives, wet and dry fractionation of pulses, energy-efficient size reduction, controlled crystallisation of fats and sugars, and process analytical technology for online particle and composition measurement. They advise industry on scale-up, supervise doctoral research and write examination questions that test the principles behind equipment. They also review evidence for new separation technologies before industry adopts them.
This manual does not give full equilibrium-stage design methods for multistage leaching or McCabe–Thiele construction, detailed centrifuge neutral-zone derivations, or drag-coefficient charts beyond Stokes' law. Grain cleaning and milling machinery in depth is in Agricultural Engineering's post-harvest subjects; flour milling in cereal-science-milling; oil refining in fats-oils-technology; essential oils in spices-flavours-essential-oils; hygienic design standards in food-plant-design-utilities. Numerical values given for pressures, pore sizes, loading fractions, solubilities and power numbers are typical textbook ranges for orientation; design must use supplier and product data.
Original practice questions written to the syllabus and pattern of the exams named in each question's examTags. They are not past papers and do not reproduce any board's questions. Every numerical answer was computed by script. Items tagged to licensing or recruitment exams whose syllabi Laisha has not yet verified are plant-level applications written to the expected scope.
These notes are free to read here, with no account.
What the Laisha app adds:
Free for ten days, then a subscription. We would rather write that here than have you find it later.