Login Register

Access the GIFSQ Portal

Select your user type to log in or register a new account.

Student Portal

Access your food safety courses and exams.

Instructor Portal

Manage courses, view student submissions, and grade quizzes.

Company Portal

Manage corporate setup, view employee logs, and access QA services.

GIFSQ ID: FTN-0059

← All subjects

Food Engineering

Unit 1: Fluid Flow

Note

Unit 1: Fluid Flow

Milk rushes through a pipe into a pasteuriser, juice is pumped up to a filling tank, and thick tomato paste creeps slowly through a different pipe. All three are fluids in motion, yet they behave in very different ways. Why does a thin liquid speed up when a pipe narrows, why does a pressure gauge read zero even though the air around us presses on everything, and why does ketchup refuse to pour until you shake the bottle? This unit answers those questions, and it gives you the tools to measure flow, to predict pressure, and to choose the right pipe and pump for a food plant.

Questions to think about

  • When a pipe narrows, does the liquid inside speed up or slow down, and what happens to its pressure?
  • Why do we need three different words, absolute, gauge, and vacuum, for one idea called pressure?
  • What makes tomato paste, yoghurt, and cream behave so differently from water when they flow?

1.1 Pressure: absolute, gauge, and vacuum

Pressure is force spread over an area. In symbols, p = F / A, and its unit is the pascal, Pa, which is one newton per square metre. Food plants talk about pressure all the time: steam pressure in a boiler, air pressure in a filler, the pressure a pump must build to push milk up to a tank on the second floor.

The air around us presses on everything with about 101,325 Pa at sea level. That value is called atmospheric pressure, and it is the natural zero point for most plant instruments. This gives three ways to state a pressure, and mixing them up is one of the most common mistakes in a plant.

  • Absolute pressure is measured from a perfect vacuum, from zero pressure. It is the true pressure of the fluid.
  • Gauge pressure is measured from atmospheric pressure. A gauge reads zero when the system is open to the air, even though the absolute pressure there is about 101 kPa.
  • Vacuum is a pressure below atmospheric pressure. It is usually stated as how far below atmospheric the pressure sits, so a vacuum of 40 kPa means the absolute pressure is about 61 kPa.

The link between them is simple: absolute pressure equals atmospheric pressure plus gauge pressure. A gauge reading of 200 kPa means an absolute pressure near 301 kPa. A vacuum of 40 kPa means an absolute pressure near 61 kPa. When a recipe or a machine manual gives a pressure, always ask which of the three it means before you act on it.

Quick check

A pressure gauge on a steam line reads 150 kPa. What is the absolute pressure, if atmospheric pressure is 101 kPa?

Check your answer

About 251 kPa. Absolute pressure equals gauge pressure plus atmospheric pressure: 150 + 101 = 251 kPa.

1.2 The fundamental equation of hydrostatics

A fluid at rest still pushes on its container, and the push grows with depth. Think of a tall milk storage tank: the wall near the bottom feels a much larger push than the wall near the top, because it carries the weight of all the milk above it.

For a liquid that does not move, the pressure at a depth h below the surface is:

p = p₀ + ρ g h

Here p₀ is the pressure at the surface, ρ is the density of the liquid, g is the acceleration due to gravity (9.81 m/s²), and h is the depth. This is the fundamental equation of hydrostatics. It says three plain things: pressure grows in direct proportion to depth, it depends on the density of the liquid, and at the same depth in one connected body of liquid the pressure is the same everywhere.

A worked example makes it concrete. Milk has a density near 1,030 kg/m³. At a depth of 2 m below the surface of an open tank, the extra pressure from the milk column is ρ g h = 1,030 x 9.81 x 2 = 20,209 Pa, about 20.2 kPa. Add the atmospheric pressure of 101 kPa pressing on the surface, and the absolute pressure at that depth is about 121 kPa. The gauge pressure there is 20.2 kPa, which is the number a gauge fitted at that point would show.

The same equation explains why a tall tank needs thicker walls at the bottom, why the level in a sight glass matches the level in the tank, and why a small leak near the base of a tank sprays much harder than a leak near the top.

How pressure grows with depth in a tank

Surface: air presses down, p₀ = 101 kPa
Half way down: p = p₀ + ρ g h, pressure is larger
Bottom: greatest depth, greatest pressure, thickest wall

Pressure at any depth depends only on depth and density, not on the shape of the tank.

Quick check

Two tanks hold the same liquid at the same depth: one is wide, one is narrow. Which has the higher pressure at the bottom?

Check your answer

Both have the same pressure. Pressure depends on depth and density only, not on the width or shape of the tank.

1.3 Measuring pressure difference: manometers

A manometer measures a pressure difference by balancing it against a column of liquid. It is one of the oldest instruments in the plant, and it is still used because it is simple, needs no power, and shows the answer directly.

U-tube manometer. A U-shaped glass tube holds a heavy liquid, often mercury or a coloured oil. One side connects to the point whose pressure is wanted, the other side is open to the air or to a second point. The pressure difference pushes the liquid down on one side and up on the other. The difference in height, Δh, gives the pressure difference as Δp = ρ g Δh, where ρ is the density of the manometer liquid. A large height difference means a large pressure difference.

Well type manometer. One side of the U is replaced by a wide well. The level in the well barely moves, so the reading is taken from the single narrow column. It is easier to read, and it is common on air ducts and filter housings.

Inclined tube manometer. The measuring tube leans at a shallow angle, so a small vertical rise stretches into a long movement along the tube. This magnifies small pressure differences, which makes it the right choice for the small pressure drops across air filters and through ventilation ducts.

Differential manometer. Both sides connect to two points in the same system, so it reads the difference between them directly. It is the instrument behind the venturimeter in section 1.6: it reads the pressure drop across a constriction, and from that drop the flow rate is worked out.

Practical note: when you read a manometer, read the levels with your eye at the level of the liquid, note which liquid is inside, and check that no air bubble is trapped in the tube, because a bubble changes the reading. Always record whether the value is a gauge value or a difference between two points.

Quick check

A U-tube manometer holds mercury. The level on one side is 0.15 m higher than on the other. Mercury has a density of 13,600 kg/m³. What is the pressure difference?

Check your answer

Δp = ρ g Δh = 13,600 x 9.81 x 0.15 = 20,012 Pa, about 20 kPa.

1.4 Steady flow: the continuity equation

When a liquid flows steadily through a pipe, the mass entering one end each second must equal the mass leaving the other end each second, because liquid cannot pile up inside the pipe. For a liquid whose density stays constant, this gives the continuity equation:

A₁ v₁ = A₂ v₂

Here A is the cross-sectional area of the pipe and v is the average velocity of the liquid. The product A v is the volume flow rate, Q, in m³/s. The message is plain: where the pipe narrows, the liquid must speed up, and where the pipe widens, the liquid must slow down.

A worked example: milk flows at 1 m/s in a pipe of 0.1 m diameter, then enters a pipe of 0.05 m diameter. The area falls to one quarter of its first value, since area depends on the square of the diameter, so the velocity rises to four times its first value, 4 m/s. This is why a nozzle makes a jet fast and thin, and why a sudden narrowing in a line raises the velocity and can raise wear on the pipe wall.

The same idea governs pumps and fillers: the volume flow rate fixes how fast a tank fills, how long a batch takes to move, and how much liquid a pump must deliver each second.

Quick check

Juice flows at 0.5 m/s in a pipe of diameter 0.08 m. It enters a pipe of diameter 0.04 m. What is the new velocity?

Check your answer

2 m/s. The diameter halves, so the area falls to one quarter, so the velocity rises four times: 0.5 x 4 = 2 m/s.

1.5 Bernoulli's equation and its uses

Bernoulli's equation is an energy balance for a flowing liquid. It says that along a streamline, the sum of three kinds of energy per unit weight stays constant: pressure energy, kinetic energy, and potential energy from height. In the form most used in food plants:

p / (ρ g) + v² / (2 g) + z = constant

Each term is a height, called a head: pressure head, velocity head, and elevation head. The equation holds for steady flow of a liquid with constant density, with no friction losses and no pump or turbine adding or removing energy between the two points. Real pipes have friction, so engineers add a loss term, but the ideal form already explains a great deal.

Two examples from the outline show its power.

Example 1: draining a tank (Torricelli's result). A tank holds liquid with its surface at a height H above a small outlet at the bottom. At the surface the velocity is nearly zero and the pressure is atmospheric. At the outlet the pressure is also atmospheric. Bernoulli's equation then gives the outlet velocity as v = √(2 g H). A tank with a 4 m head empties through its outlet at about √(2 x 9.81 x 4) = 8.86 m/s at first, slowing as the level falls. This is why a full tank drains fast and a nearly empty tank trickles.

Example 2: flow through a constriction. When a pipe narrows, continuity says the velocity rises. Bernoulli's equation says that when velocity rises, pressure must fall, because the total stays constant. A venturimeter uses exactly this pressure drop to measure flow, and section 1.6 explains how. The same effect lifts liquid in a spray nozzle and explains why a fast stream of air can suck liquid up a tube.

Bernoulli's trade: pressure, speed, and height

Wide pipe: slow liquid, high pressure

↓

Narrow throat: fast liquid, low pressure

↓

Wide pipe again: slow liquid, pressure partly recovers

Total head stays constant: pressure head + velocity head + elevation head.

Quick check

A pipe narrows and the liquid speeds up. What happens to the pressure in the narrow part?

Check your answer

It falls. Bernoulli's equation keeps the total head constant, so a rise in velocity head must be matched by a fall in pressure head.

1.6 Flow patterns: laminar flow, turbulent flow, and Reynolds number

Not all flow looks the same. In laminar flow, the liquid moves in smooth layers that slide past each other, with no mixing across the layers. A thin stream of honey pouring from a spoon is laminar. In turbulent flow, the liquid churns, eddies form, and layers mix. Water rushing from a fully open tap is turbulent. Laminar flow is quiet and orderly, turbulent flow mixes well and carries heat faster, but it also costs more pumping energy.

Osborne Reynolds found one number that predicts which pattern appears: the Reynolds number.

Re = (ρ v D) / μ = (v D) / ν

Here ρ is density, v is average velocity, D is the pipe diameter, μ is the dynamic viscosity, and ν is the kinematic viscosity, μ / ρ. The number has no units, so it works in any system of units. In a pipe, the pattern is roughly: Re below 2,100 means laminar flow, Re above 10,000 means turbulent flow, and values between are a transition zone where the pattern is unsteady and can flip either way. That wide transition band, 2,100 to 10,000, is the one used in the aseptic processing work in Irudayaraj, where a fully turbulent flow is only counted above 10,000.

A worked example: water at 20 °C has a density near 1,000 kg/m³ and a viscosity near 0.001 Pa·s. In a pipe of 0.05 m diameter at 1 m/s, Re = (1,000 x 1 x 0.05) / 0.001 = 50,000, well into turbulent flow. The same water at 0.02 m/s in the same pipe gives Re = 1,000, laminar. Thick liquids such as syrup and cream have high viscosity, so their Reynolds numbers stay low and their flow stays laminar even at fair speeds. That matters for heating: laminar flow heats unevenly, because the layers do not mix, so a heat exchanger for a thick product needs a design that stirs or scrapes the liquid.

Laminar flow has a second consequence for heating that matters in aseptic processing. In a straight tube in laminar flow, the liquid at the centre moves about twice as fast as the average velocity, so the fastest fluid spends the least time in the holding tube and gets the least heat treatment. A holding tube is therefore sized on that fastest particle, not on the average, so that even the quickest fluid is heated enough to be safe.

Quick check

Milk flows in a pipe with a Reynolds number of 800. Is the flow laminar or turbulent, and why does it matter for heating?

Check your answer

Laminar, since 800 is below 2,100. It matters because laminar layers do not mix, so heating is uneven unless the design stirs the milk.

1.7 Flow measuring equipment

Four meters cover most food plant needs, and each uses a different principle.

Venturimeter. A short section of pipe narrows to a throat and then widens again. Pressure taps at the inlet and at the throat connect to a differential manometer. From continuity, the throat velocity is higher; from Bernoulli, the throat pressure is lower. The pressure difference grows with the square of the flow rate, so measuring the difference gives the flow. It is accurate, causes only a small permanent pressure loss, and has no moving parts. Its drawback is size and cost, and it needs a straight run of pipe before it so the flow is settled.

Orifice meter. A flat plate with a sharp-edged hole sits across the pipe. The flow squeezes through the hole, the velocity rises, and the pressure falls across the plate. Two pressure taps, one before and one after the plate, read the difference. It is cheap and simple, but it causes a larger permanent pressure loss than a venturimeter, and the sharp edge wears over time, which changes the reading.

Pitot tube. A small tube points into the stream and brings the liquid at its mouth to a stop. The pressure at the stopped point, the stagnation pressure, is higher than the pressure in the moving stream by the velocity head, v² / (2 g). Measuring that difference gives the velocity at that one point. It is cheap and easy to install, but it reads one point only, not the average over the whole pipe, so it suits checks and large ducts more than precise metering.

Rotameter. A tapered vertical glass tube holds a float. Liquid flows upward, lifts the float, and the float settles where the drag of the upward stream balances its weight. The higher the flow, the higher the float sits. A scale on the tube reads the flow directly. It needs no power, is easy to read, and is common on small lines for water, air, and cleaning solutions. It must sit vertical, and it suits clean liquids better than liquids with solids that could jam the float.

Which meter fits the job

Need accuracy, low loss: venturimeter
Need cheap and simple: orifice meter
Need a quick spot check: pitot tube
Need a direct reading on a small line: rotameter

Quick check

Which meter has no moving parts and the smallest permanent pressure loss: venturimeter or orifice meter?

Check your answer

The venturimeter. The orifice plate causes a larger permanent pressure loss and its sharp edge wears.

1.8 Flow of fluid in tubes: viscosity

Viscosity is the internal friction of a fluid, the resistance it offers to flowing. Imagine two plates with liquid between them: pull the top plate and the liquid layers slide over each other. The force needed per unit area, the shear stress τ, is linked to how fast the velocity changes across the gap, the shear rate dv/dy. Newton's law of viscosity states:

τ = μ (dv/dy)

The constant μ is the dynamic viscosity, in Pa·s. Water at room temperature has μ near 0.001 Pa·s, milk near 0.002 Pa·s, cream higher still, and honey near 10 Pa·s. The higher the value, the thicker the fluid feels and the harder it is to pump.

Newtonian fluids follow this law with a constant μ. Their viscosity does not change when you stir faster. Water, milk, thin juice, and clear syrups are Newtonian, and most textbook equations assume this behaviour.

Non-Newtonian fluids do not follow it, and most real foods belong here.

  • Pseudoplastic (shear thinning): viscosity falls as shear rate rises. Ketchup, yoghurt, tomato paste, and fruit puree flow more easily when stirred or shaken, which is why shaking a ketchup bottle helps it pour.
  • Dilatant (shear thickening): viscosity rises as shear rate rises. A thick starch slurry can stiffen when stirred hard.
  • Bingham plastic: the fluid does not flow at all until the stress passes a yield value, then it flows. Tomato paste and some chocolate masses behave this way, which is why a paste can hold its shape on a spoon.
  • Thixotropic: viscosity falls with time under steady stirring and recovers when left still. Some sauces and creams show this.

This matters in plant design. A pump sized for water will struggle with tomato paste, not only because the paste is thick, but because its thickness changes with the speed of the pump. Heating, mixing, and filling all depend on which class the food falls into, and a food engineer measures the flow curve before choosing equipment.

Practical note. The practical work for this unit uses a venturimeter, an orifice meter, and a rotameter on the same line, so you can compare their readings at the same flow, and it measures Reynolds number to see the change from laminar to turbulent flow. Viscosity is measured with an Ostwald viscometer, which times how long a fixed volume of liquid takes to flow through a thin capillary: a longer time means a higher viscosity, and the result is compared against water run through the same instrument.

Quick check

Ketchup pours more easily after you shake the bottle. What class of fluid is it, and what does shaking do to its viscosity?

Check your answer

Pseudoplastic, or shear thinning. Shaking raises the shear rate, and the viscosity falls, so the ketchup flows more easily.

1.9 Derivation: Bernoulli's equation from work and energy

Section 1.5 states Bernoulli's equation. This section derives it, one small step at a time, so the formula is a result you can rebuild, not a line to memorise.

Step 1: follow a small parcel of liquid. Picture a small parcel of liquid of mass m moving along a streamline from point 1 to point 2. At point 1 it has pressure p₁, velocity v₁, and height z₁. At point 2 it has pressure p₂, velocity v₂, and height z₂. The parcel keeps its mass and its volume, because the liquid does not compress, and no friction steals energy on the way.

Step 2: count the work done on the parcel. Three things do work on the parcel as it moves. First, the pressure behind it pushes it forward, and the pressure ahead of it pushes back. The net pressure work is p₁V minus p₂V, where V is the volume of the parcel. Second, gravity does work as the parcel changes height: moving up costs energy, moving down gives it back. Third, nothing else does work, because there is no pump and no friction in this ideal case.

Step 3: count the change in the parcel's energy. The parcel's kinetic energy changes from (1/2) m v₁² to (1/2) m v₂². Its potential energy changes from m g z₁ to m g z₂. The work done on the parcel must equal the change in its energy: that is the work-energy principle.

Step 4: write the balance. Work by pressure plus work by gravity equals the change in kinetic energy plus the change in potential energy. In symbols: p₁V minus p₂V equals (1/2) m (v₂² minus v₁²) plus m g (z₂ minus z₁).

Step 5: divide by the weight of the parcel. The weight is m g, and the mass is density times volume, m = ρ V. Divide every term by m g. The pressure terms become p₁ / (ρ g) minus p₂ / (ρ g). The kinetic terms become v₂² / (2 g) minus v₁² / (2 g). The height terms stay as z₂ minus z₁.

Step 6: collect the terms at each point. Move all the terms for point 1 to one side and all the terms for point 2 to the other side. The result is p₁ / (ρ g) + v₁² / (2 g) + z₁ = p₂ / (ρ g) + v₂² / (2 g) + z₂. Each side is the same total, so along the streamline p / (ρ g) + v² / (2 g) + z = constant. That is Bernoulli's equation, and each step shows what it assumes: steady flow, constant density, no friction, and no pump or turbine between the two points. Break one of those assumptions and the equation needs a correction term, which is what section 1.10 does for friction.

Quick check

A liquid flows horizontally, so z₁ = z₂, and its velocity doubles from one point to the next. Does the pressure rise or fall, and why?

Check your answer

It falls. With height unchanged, Bernoulli's equation says a rise in v² / (2 g) must be matched by a fall in p / (ρ g), so the pressure drops.

1.10 Friction loss: Darcy-Weisbach, the Moody chart, and pump head

Real pipes have friction, and friction eats head. The pressure at the end of a long pipe is lower than Bernoulli's equation predicts, and the difference is the head loss. An engineer who ignores it sizes a pump too small, and the plant starves.

The Darcy-Weisbach equation. The head lost to friction in a straight pipe is h f = f (L / D) (v² / (2 g)), where f is the friction factor, L is the pipe length, D is the diameter, and v is the average velocity. The loss grows with the square of the velocity, which is why doubling the flow quadruples the friction loss, and it grows with length and falls with diameter. A long thin pipe is the worst case for friction.

Finding f from the Moody chart. The friction factor depends on the Reynolds number and on how rough the pipe wall is. In laminar flow the answer is simple: f = 64 / Re, and roughness does not matter, because the smooth layers never touch the wall bumps. In turbulent flow the engineer reads f from the Moody chart, a plot of f against Re with one curve for each relative roughness, ε / D, where ε is the height of the wall roughness. A smooth drawn tube has a low curve, a rough old steel pipe has a high one. The chart is read in three steps: work out Re, work out ε / D, and read f where the two meet. A typical value for water in a commercial steel pipe in turbulent flow is f near 0.02.

Minor losses. Bends, valves, entries, and exits each add a loss, written as h m = K (v² / (2 g)), where K is a loss coefficient for that fitting. A fully open gate valve has a small K, near 0.15, while a globe valve has a large K, near 10, which is why globe valves are avoided where pressure matters. The total head loss is the friction loss plus the sum of the minor losses.

Worked example. Milk flows at 1.5 m/s through 40 m of pipe of 0.05 m diameter. Take f = 0.02. Then h f = 0.02 x (40 / 0.05) x (1.5² / (2 x 9.81)) = 0.02 x 800 x 0.115 = 1.84 m. Add two bends and one valve with a combined K of 2.0: h m = 2.0 x 0.115 = 0.23 m. The total loss is about 2.1 m of head. A pump that must also lift the milk 5 m needs at least 7.1 m of head, before any safety margin, and that number is what the pump is chosen against.

Pump head and cavitation. A pump adds head to the liquid, and the pump curve shows the head it gives at each flow rate. The plant needs the pump's head to beat the lift plus the losses at the wanted flow. There is a second limit: at the pump inlet the pressure must stay above the vapour pressure of the liquid, or the liquid boils into bubbles that collapse inside the pump, a damaging effect called cavitation, which sounds like gravel in the pump and eats the impeller. The margin between the pressure at the inlet and the vapour pressure is the net positive suction head, NPSH. The pump maker states the NPSH the pump needs, and the plant must supply more than that, by keeping the suction line short, wide, and free of needless valves, and by keeping the suction tank level high enough.

Choosing a pump. A centrifugal pump suits thin liquids at high flow, such as milk and juice, and it is the common choice in food plants. A positive displacement pump, such as a lobe pump or a screw pump, suits thick liquids and pastes, such as tomato paste and cream, because it pushes a fixed volume with each turn whatever the thickness. A pump sized for water will fail on tomato paste, not only because the paste is thick, but because its viscosity changes with speed, as section 1.8 explains. The practical check is simple: measure the flow the plant gets at the real suction and discharge conditions, and compare it with the pump curve, rather than trusting the catalogue figure alone.

Quick check

A pipe run is doubled in length, with the same diameter and flow. What happens to the friction head loss?

Check your answer

It doubles. The Darcy-Weisbach loss is proportional to L / D, so twice the length gives twice the loss at the same velocity.

Try it yourself

Take a kitchen tap, a measuring jug, and a stopwatch. Run the tap gently into the jug for 10 seconds and record the volume. Then open the tap fully and repeat. Work out the volume flow rate, Q, in litres per second for each case. Now look at the stream: the gentle stream is smooth and glassy, the fast stream is rough and broken. Those are laminar and turbulent flow in your own kitchen. Estimate which case has the higher Reynolds number and explain why. Then shake a bottle of ketchup and a bottle of water, and note which one changes how easily it pours, and what that tells you about its viscosity class.

Key ideas to carry forward

  • Absolute pressure is measured from vacuum, gauge pressure from atmospheric pressure, and vacuum is a pressure below atmospheric pressure.
  • In a still liquid, pressure grows with depth as p = p₀ + ρ g h, and depends on depth and density only.
  • A manometer turns a pressure difference into a liquid column height, Δp = ρ g Δh.
  • Continuity, A₁ v₁ = A₂ v₂, says a narrower pipe means a faster liquid.
  • Bernoulli's equation trades pressure head, velocity head, and elevation head, and it explains tank draining and the venturimeter.
  • Reynolds number predicts laminar flow below about 2,100 and turbulent flow above about 10,000, with a transition zone between them.
  • Venturimeter, orifice meter, pitot tube, and rotameter each measure flow by a different principle, and each fits a different job.
  • Viscosity is internal friction, τ = μ (dv/dy). Newtonian fluids keep a constant viscosity, while most foods are pseudoplastic, dilatant, Bingham, or thixotropic.

Glossary

Absolute pressure

Pressure measured from a perfect vacuum, the true pressure of the fluid. Example: a boiler at 300 kPa absolute holds 300 kPa above zero pressure.

Gauge pressure

Pressure measured from atmospheric pressure, so a gauge reads zero when open to the air. Example: a steam gauge reading 150 kPa means 150 kPa above the air around us.

Vacuum

A pressure below atmospheric pressure, stated as the amount below it. Example: a vacuum packer pulls a vacuum to remove air from a pouch before sealing.

Hydrostatics

The study of fluids at rest, and how pressure changes with depth. Example: the pressure at the bottom of a milk tank is worked out from hydrostatics.

Manometer

An instrument that measures a pressure difference from the height of a liquid column. Example: a U-tube manometer on a filter shows the pressure drop across the filter.

Continuity equation

The rule that volume flow rate stays constant along a pipe, A₁ v₁ = A₂ v₂. Example: juice speeds up four times when a pipe diameter halves.

Volume flow rate

The volume of fluid passing a point each second, Q = A v, in m³/s. Example: a filler needs a flow rate of 2 litres per second to fill bottles on time.

Bernoulli's equation

An energy balance for steady flow: pressure head plus velocity head plus elevation head stays constant. Example: it predicts how fast a tank drains through its outlet.

Head

Energy per unit weight of fluid, stated as a height in metres. Example: a pump that adds 10 m of head can lift water 10 m up a pipe.

Laminar flow

Smooth flow in layers that slide past each other without mixing. Example: honey pouring in a smooth thread is laminar.

Turbulent flow

Churning flow with eddies that mix the liquid across the pipe. Example: water rushing from a fully open tap is turbulent.

Reynolds number

A number with no units, Re = (ρ v D) / μ, that predicts whether flow is laminar or turbulent. Example: a Reynolds number of 50,000 in a water pipe means turbulent flow.

Venturimeter

A flow meter that narrows a pipe to a throat and reads the pressure drop to find the flow rate. Example: a venturimeter on a milk line measures flow with no moving parts.

Orifice meter

A flow meter that uses a flat plate with a hole to create a pressure drop. Example: an orifice plate on a steam line gives a cheap flow reading.

Pitot tube

A tube that faces into a stream and measures velocity from the rise in pressure where the flow stops. Example: a pitot tube checks air speed in a drying duct.

Rotameter

A vertical tapered tube with a float that rises with flow rate, read straight off a scale. Example: a rotameter shows the flow of cleaning solution on a small line.

Viscosity

A fluid's internal friction, its resistance to flowing. Example: honey has a much higher viscosity than water, so it pours slowly.

Newtonian fluid

A fluid whose viscosity stays constant whatever the shear rate. Example: water and milk are Newtonian, so their flow follows the simple equations.

Non-Newtonian fluid

A fluid whose viscosity changes with shear rate or time. Example: ketchup is non-Newtonian, so it pours more easily after shaking.

Pseudoplastic

A fluid whose viscosity falls as it is stirred faster, also called shear thinning. Example: yoghurt thins when it is stirred, then thickens again when it rests.

Shear stress

Force per unit area acting along a surface, the push that makes fluid layers slide. Example: a pump impeller applies shear stress to the liquid it moves.

Shear rate

How fast velocity changes across a gap in a flowing fluid, in 1/s. Example: a fast stirrer gives a high shear rate in a sauce.

Friction factor

A number with no units, f, that sets the friction loss in a pipe in the Darcy-Weisbach equation. Example: a friction factor near 0.02 is typical for water in turbulent flow in a steel pipe.

Head loss

The head eaten by friction and fittings as liquid moves through a pipe, in metres. Example: a long thin pipe run can lose 2 m of head to friction alone.

Moody chart

A chart that gives the friction factor from the Reynolds number and the relative roughness of the pipe wall. Example: an engineer reads f off the Moody chart before sizing a pump.

Cavitation

Bubbles forming and collapsing inside a pump when the inlet pressure falls below the vapour pressure, damaging the impeller. Example: a rattling pump with a starved suction line is often cavitating.

Net positive suction head (NPSH)

The margin between the pressure at a pump inlet and the vapour pressure of the liquid, in metres. Example: the plant must supply more NPSH than the pump needs, or the pump cavitates.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Irudayaraj, Joseph M. (ed.). Food Processing Operations Modeling, Design and Analysis. Marcel Dekker, 2002.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.
  • Ravi Kumar, S. Dairy Engineering. AgriMoon course material.

Unit 2: Heat Transfer

Note

Unit 2: Heat Transfer

A pasteuriser heats milk to a safe temperature in seconds, a freezer pulls heat out of ice cream, and a hot pan browns the outside of a loaf while the centre is still cool. Heat moves by rules that can be measured and predicted, and every food process that heats or cools is designed with those rules. Why does a thin metal wall pass heat faster than a thick one, why does stirring a pot heat soup faster, and why does a counter flow heat exchanger beat a parallel flow one? This unit answers those questions.

Questions to think about

  • If you hold one end of a metal spoon in hot soup, why does the handle warm up, and what decides how fast?
  • Why does a stirred pot of soup heat faster than a still pot, even on the same burner?
  • A heat exchanger has hot milk on one side and cold milk on the other: which way should the two streams flow, and why?

2.1 Conduction and Fourier's law

Conduction is the movement of heat through a solid or a still fluid, passed from molecule to molecule without any bulk movement of the material. A metal spoon in hot soup warms by conduction. The wall of a steam kettle passes heat to the milk inside by conduction. The crust of a loaf is heated by conduction from the hot air touching it, then the heat travels inward by conduction through the crumb.

Fourier's law gives the rate of heat flow by conduction:

q = -k A (dT/dx)

Here q is the heat flow rate in watts, k is the thermal conductivity of the material in W/m·K, A is the area the heat crosses, and dT/dx is the temperature gradient, the change in temperature with distance. The minus sign says heat flows from hot to cold, down the temperature gradient. Thermal conductivity is a property of the material: metals conduct well, with k in the tens or hundreds, water conducts modestly, k near 0.6 W/m·K, and most foods conduct poorly, k near 0.5 W/m·K, close to water, because they are mostly water. Air is a poor conductor, k near 0.026 W/m·K, which is why trapped air in packaging insulates, and why a foam is a good insulator.

For a flat slab of thickness L with a steady temperature difference ΔT across it, Fourier's law becomes a plain working form:

q = k A ΔT / L

Double the thickness and the heat flow halves. Double the area and it doubles. A worked example: a steel plate 0.005 m thick, area 1 m², k = 45 W/m·K, with a temperature difference of 80 K across it, passes q = 45 x 1 x 80 / 0.005 = 720,000 W. A plastic wall of the same thickness with k = 0.2 W/m·K would pass only 3,200 W. That is why kettle walls are metal, not plastic, and why an insulated tank wall carries a thick layer of low conductivity material.

One more property decides how fast a temperature change spreads through a food: thermal diffusivity, α = k / (ρ Cp), where ρ is density and Cp is specific heat. It combines the three properties above into one number, in m²/s. A food with high thermal diffusivity warms through quickly, while a food with low thermal diffusivity warms slowly at its centre even when its surface is hot, which is why a thick piece of food can be scorched outside and cool inside.

Quick check

A slab passes 100 W. If its thickness is doubled and everything else stays the same, what is the new heat flow?

Check your answer

50 W. Heat flow by conduction is inversely proportional to thickness: q = k A ΔT / L.

2.2 Heat transfer through a slab and a cylinder

Real walls are rarely one material. A cold store wall is brick, insulation, and plaster. A pipe carries a hot liquid inside and loses heat through its wall to the air outside. Both cases are handled by the same idea: thermal resistance.

Slab, one layer. From q = k A ΔT / L, the resistance is R = L / (k A), and q = ΔT / R. Resistance grows with thickness and falls with conductivity and area, just like electrical resistance grows with the length of a wire.

Slab, layers in series. When heat must cross several layers one after another, the resistances add: R total = R₁ + R₂ + R₃, and q = ΔT / R total. The layer with the largest resistance controls the result. A 10 cm layer of foam insulation dominates the resistance of a cold store wall, even though the brick beside it is far thicker. That is the design lesson: add resistance where it counts, and do not pay for extra thickness in a layer that barely resists.

Cylinder. A pipe wall is a cylindrical shell, and its area grows with radius, so the resistance of a cylindrical layer is R = ln(r₂/r₁) / (2 π k L), where r₁ and r₂ are the inner and outer radii and L is the pipe length. For several layers, the resistances add as before. A worked example: a steam pipe of length 10 m, inner radius 0.05 m, outer radius 0.055 m, steel k = 45 W/m·K, gives R = ln(0.055/0.05) / (2 x 3.14 x 45 x 10) = 0.0953 / 2,826 = 0.0000337 K/W, a tiny resistance, so the steel wall barely slows the heat. Add 0.05 m of insulation with k = 0.04 W/m·K and the insulation resistance is ln(0.105/0.055) / (2 x 3.14 x 0.04 x 10) = 0.646 / 2.51 = 0.257 K/W, thousands of times larger. The insulation, not the steel, decides the heat loss.

Resistances in series: heat crosses one layer after another

Hot side: steam or hot milk, high temperature

↓

Layer 1: steel wall, small resistance

↓

Layer 2: insulation, large resistance, controls the heat flow

↓

Cold side: air, low temperature

Total resistance is the sum of the layers. Heat flow q = ΔT / R total.

Quick check

A wall has two layers in series with resistances 0.5 K/W and 2.0 K/W, and the temperature difference is 50 K. What is the heat flow?

Check your answer

20 W. Total resistance is 2.5 K/W, so q = 50 / 2.5 = 20 W.

2.3 Convection: natural and forced

Convection is heat carried by the bulk movement of a fluid. When a fluid moves past a surface, it picks up heat or gives it up. There are two kinds.

Natural convection is driven by density differences. Warm liquid near a heater becomes lighter and rises, cool liquid sinks to take its place, and a slow circulation starts with no pump and no fan. A pot of soup warming on a stove shows this: currents rise in the centre and fall at the sides. Natural convection is gentle, so it is slow.

Forced convection is driven by a pump, a fan, or a stirrer. Milk pumped through a pasteuriser, air blown over a cooling loaf, and a stirred kettle all use forced convection. It moves heat far faster than natural convection, because fresh fluid is always brought to the surface.

The rate of heat transfer by convection is written with Newton's law of cooling:

q = h A ΔT

Here h is the surface heat transfer coefficient, also called the film coefficient, in W/m²·K, A is the surface area, and ΔT is the temperature difference between the surface and the bulk fluid. The value of h depends on the fluid, its velocity, and the shape of the surface. Still air has h near 5 to 10 W/m²·K. Moving air reaches 10 to 100. Water in forced flow reaches 500 to 10,000, and condensing steam reaches 5,000 to 15,000. That spread explains why steam heating is so effective and why a product cools slowly in still air.

Quick check

Which gives faster heating of a soup surface: still air or stirred water, and why?

Check your answer

Stirred water. Its surface coefficient h is far larger, so q = h A ΔT is far larger for the same area and temperature difference.

2.4 The overall heat transfer coefficient

A real heat exchanger has hot fluid on one side, a wall in the middle, and cold fluid on the other side. Heat must cross three resistances in series: the hot fluid film, the wall, and the cold fluid film. The overall heat transfer coefficient, U, rolls all three into one number:

1 / U = 1 / h hot + L / k + 1 / h cold

per unit area, where L/k is the wall resistance. Then the heat flow is:

q = U A ΔT

The smallest film coefficient dominates U. If one side is steam condensing with h near 10,000 and the other side is a thick sauce with h near 200, then U is near 200 no matter how good the wall is. The design lesson is to improve the weakest side first: stir the sauce, scrape the wall, or raise its velocity, rather than polishing the side that is already fast.

Fouling adds a fourth resistance. Milk proteins and minerals bake onto hot surfaces as a thin deposit, and that deposit insulates. The fouling factor, R f, is added to the sum: 1 / U = 1 / h hot + L / k + 1 / h cold + R f. A fouled pasteuriser transfers less heat, so the milk leaves cooler than planned, and the plant must stop and clean. Section 2.7 returns to this.

Quick check

A heat exchanger has h hot = 5,000, h cold = 250, and a thin metal wall whose resistance is tiny. About what is U?

Check your answer

About 238 W/m²·K. The cold side dominates: 1 / U is close to 1 / 250, so U is close to 250.

2.5 Dimensionless numbers and their uses

Engineers predict the surface coefficient h without building the exchanger first, by using numbers that have no units. Three of them carry most of the work.

  • Reynolds number, Re = (ρ v D) / μ. It describes the flow pattern, laminar or turbulent, as in Unit 1. Turbulent flow gives a higher h, because eddies bring fresh fluid to the wall.
  • Prandtl number, Pr = (μ Cp) / k. It compares how fast momentum spreads with how fast heat spreads in a fluid. Water has Pr near 7 at room temperature, air near 0.7, and thick oils have Pr in the hundreds or thousands.
  • Nusselt number, Nu = (h D) / k. It is the dimensionless form of h itself: a large Nu means convection beats conduction across the fluid.

Correlations link them, in the general form Nu = C Re^m Pr^n, where C, m, and n are constants fitted to experiments for a given shape, such as flow inside a tube or across a flat plate. The routine is: work out Re and Pr from the fluid properties, velocity, and tube size, read Nu from the correlation, and then get h = Nu k / D. That value feeds the overall coefficient U and the sizing of the exchanger. For laminar flow inside a tube the correlation is different from turbulent flow, which is one more reason the Reynolds number matters in a food plant.

Quick check

Why does a turbulent flow give a higher surface coefficient than a laminar flow in the same tube?

Check your answer

Turbulence mixes the fluid and brings fresh fluid to the wall, so heat is carried away faster and h rises.

2.6 Heat exchangers: parallel flow, counter flow, LMTD, and NTU

A heat exchanger moves heat from one stream to another without letting them mix. Milk is heated by hot water, or cooled by cold water, through a metal wall. The two streams can run in two arrangements.

Parallel flow. Both streams enter at the same end and run side by side. The temperature difference is largest at the inlet and shrinks along the length, and the cold stream can never leave hotter than the hot stream leaves. It is simple but wasteful of temperature difference.

Counter flow. The streams enter at opposite ends and run against each other. The temperature difference stays more even along the length, and the cold stream can leave hotter than the hot stream leaves. For the same duty, a counter flow exchanger needs less area than a parallel flow one. That is why counter flow is the standard choice in food plants.

Because the temperature difference changes along the exchanger, a single ΔT will not do. The correct average is the log mean temperature difference, LMTD:

LMTD = (ΔT₁ - ΔT₂) / ln(ΔT₁ / ΔT₂)

where ΔT₁ and ΔT₂ are the temperature differences at the two ends. Then q = U A LMTD. A worked example: hot water enters at 90 °C and leaves at 60 °C, while milk enters at 10 °C and leaves at 50 °C in counter flow. At one end ΔT₁ = 90 - 50 = 40 K, at the other ΔT₂ = 60 - 10 = 50 K. LMTD = (40 - 50) / ln(40/50) = (-10) / (-0.223) = 44.8 K. If U = 800 W/m²·K and the duty is 100 kW, the area needed is A = q / (U x LMTD) = 100,000 / (800 x 44.8) = 2.79 m².

The NTU method is the second way to size and rate exchangers, and it suits the case where the outlet temperatures are unknown. Effectiveness ε is the actual heat transferred divided by the maximum possible, and NTU = (U A) / C min, where C min is the smaller of the two heat capacity rates, mass flow times specific heat. Charts and formulas link ε and NTU for each flow arrangement, so an engineer can find the outlet temperatures from the inlet temperatures and the exchanger size. LMTD is handiest when the outlet temperatures are known; NTU is handiest when they are not.

Quick check

In a counter flow exchanger, the cold stream leaves at 70 °C and the hot stream leaves at 60 °C. Is that possible, and what does it show?

Check your answer

Yes. In counter flow the cold stream can leave hotter than the hot stream leaves, which shows counter flow uses the temperature difference better than parallel flow.

2.7 Types of heat exchanger and fouling

Shell and tube heat exchanger. One fluid flows inside a bundle of tubes, the other flows around them inside a shell. A single shell pass with multi tube passes is common: the tube fluid goes back and forth through the bundle several times, which lengthens its path and raises its velocity and its surface coefficient. Baffles in the shell direct the shell fluid across the tubes. Shell and tube units are sturdy, handle high pressures, and are easy to clean on the tube side, which is why they are common in dairy and juice plants.

Plate heat exchanger. Thin corrugated plates are pressed together, and the two fluids flow in alternate narrow gaps. The area is large for the size, the gaps are narrow so the coefficients are high, and the unit opens for cleaning, which suits milk pasteurisation. Its limits are pressure and the gaskets between plates, which set the temperature ceiling.

Fouling. Deposits build on heat transfer surfaces: milk stone from proteins and calcium salts, scale from hard water, and burnt layers on overheated surfaces. Fouling adds resistance, lowers U, raises the pressure drop, and forces longer heating times or a shutdown for cleaning. It is controlled by keeping wall temperatures moderate, keeping velocities high enough to scour the surface, and cleaning on a set schedule, often clean in place. When sizing a new exchanger, the engineer includes a fouling factor so the unit still meets its duty when it is dirty, not only when it is new.

Practical note. The practical work for this unit runs material and energy balances on a heat exchanger: measure the inlet and outlet temperatures and flow rates of both streams, work out the heat given up by the hot stream and the heat gained by the cold stream, and compare the two. The difference is the heat lost to the surroundings, and a good experiment keeps it small. Fouling shows up in the same exercise as a fall in U over repeated runs.

Counter flow plate exchanger: two streams, opposite directions

Hot water in, 90 °C, top left
Thin plates pass heat across, streams never mix
Cold milk in, 10 °C, bottom right, flows upward against the hot stream

Counter flow keeps the temperature difference even along the plates.

Quick check

A plate pasteuriser is cleaned less often than the schedule says. What happens to U and to the milk outlet temperature?

Check your answer

U falls because fouling adds resistance, and the milk leaves cooler than planned because less heat crosses the plates.

2.8 Derivation: from Fourier's law to heat through a slab

Section 2.1 states that heat through a slab is q = k A ΔT / L. This section derives it from Fourier's law, one small step at a time.

Step 1: start with Fourier's law in differential form. At any point inside a solid, the heat flow is q = minus k A (dT / dx), where dT / dx is the rate at which temperature changes with distance x through the wall. The minus sign says heat flows from hot to cold, down the temperature slope, not up it.

Step 2: assume steady state and a flat slab. In steady state the heat flow q is the same at every depth in the slab, because heat is not piling up anywhere. The slab has thickness L, area A, and a constant conductivity k. The hot face sits at temperature T₁ and the cold face at T₂.

Step 3: separate the variables. Rearrange the law so that the temperature terms sit on one side and the distance terms on the other: dT = minus (q / (k A)) dx. This says that each thin slice of the slab, dx thick, drops the temperature by an amount proportional to q.

Step 4: integrate across the slab. Add up the slices from the hot face (x = 0, T = T₁) to the cold face (x = L, T = T₂). The left side integrates to T₂ minus T₁, and the right side integrates to minus (q / (k A)) L. So T₂ minus T₁ = minus q L / (k A).

Step 5: solve for q. Rearrange: q = k A (T₁ minus T₂) / L, which is k A ΔT / L, the formula section 2.1 uses. The derivation shows what the formula assumes: steady state, one dimensional flow, constant k, and no heat source inside the wall. It also shows why the resistance idea works: the slab's resistance is L / (k A), and the same integration through a cylinder gives a resistance with a logarithm in it, because the area grows with radius, which is why section 2.2 treats cylinders separately.

Quick check

A slab's thickness doubles, with the same area, conductivity, and temperature difference. What happens to the heat flow?

Check your answer

It halves. Heat flow is q = k A ΔT / L, so doubling L halves q.

2.9 Derivation: why the log mean, not the plain average

Section 2.6 states the log mean temperature difference, LMTD. This section shows why a plain average is wrong and where the logarithm comes from.

Step 1: the temperature difference changes along the exchanger. At the inlet end the two streams differ by ΔT₁, and at the outlet end they differ by ΔT₂. In between, the difference slides from one value to the other. A plain average, (ΔT₁ + ΔT₂) / 2, would be right only if the difference changed in a straight line. It does not.

Step 2: look at one short slice of the exchanger. In a short slice of area dA, the heat transferred is dq = U ΔT dA, where ΔT is the local difference between the streams at that slice. That heat warms the cold stream and cools the hot stream by a small amount each, so the local difference ΔT shrinks a little as you move along the exchanger.

Step 3: the difference shrinks in proportion to itself. The change in ΔT across the slice is proportional to ΔT itself: a large difference drives a large heat flow, which shrinks the difference fast, while a small difference drives a small heat flow, which shrinks it slowly. A quantity whose rate of change is proportional to itself changes exponentially, not linearly, and that is the key point.

Step 4: integrate along the whole exchanger. Adding up the slices, the total heat is q = U A times the correct average difference, and the integration of an exponential change gives a logarithm: LMTD = (ΔT₁ minus ΔT₂) / ln(ΔT₁ / ΔT₂). When the two end differences are close, the log mean is close to the plain average, which is why a rough estimate can use the average, but when they differ a lot, the plain average overstates the driving force and the exchanger comes out too small. The worked example in section 2.6 shows the size of the error: with end differences of 40 K and 50 K, the plain average is 45 K and the log mean is 44.8 K, close, but with end differences of 10 K and 50 K the plain average is 30 K while the log mean is only 24.9 K, and an exchanger sized on 30 K would be far too small.

Quick check

An exchanger has end temperature differences of 10 K and 50 K. Is the correct average closer to 30 K or to 25 K, and why?

Check your answer

Closer to 25 K, the log mean. The temperature difference shrinks exponentially along the exchanger, not linearly, so the plain average of 30 K overstates the driving force.

2.10 Predicting food properties from composition

A heat transfer calculation needs three properties of the food: density, specific heat, and thermal conductivity. Measuring all three for every new product is slow, so engineers predict them from the composition, using models built from published data. The models below come from the property prediction work collected by Heldman in Irudayaraj, and they are used in their original form, with the numbers as published.

Specific heat from composition. Water dominates the specific heat of a food, because water has a specific heat near 4.18 kJ/kg·K, far above fat near 1.68 to 2.09, protein near 1.55, carbohydrate near 1.42, and ash near 0.84. The simplest model, Siebel's equation, uses the moisture content alone: Cp = 0.837 + 0.034 times the moisture percentage, in kJ/kg·K, for high moisture foods. A fuller model adds the parts: Cp = 1.424 Mc + 1.549 Mp + 1.675 Mf + 0.837 Ma + 4.187 Mw, where Mc, Mp, Mf, Ma, and Mw are the mass fractions of carbohydrate, protein, fat, ash, and water. The coefficients are the specific heats of each component, and the sum is weighted by how much of each is present.

Worked example. A milk with 87 percent water, 4 percent fat, 3.5 percent protein, 4.8 percent carbohydrate, and 0.7 percent ash has Cp = 1.424 x 0.048 + 1.549 x 0.035 + 1.675 x 0.04 + 0.837 x 0.007 + 4.187 x 0.87 = 0.068 + 0.054 + 0.067 + 0.006 + 3.643 = 3.84 kJ/kg·K. Siebel's equation gives 0.837 + 0.034 x 87 = 3.80 kJ/kg·K, close to the fuller model, which is why the quick model is trusted for high moisture foods. The lesson is plain: a food that is mostly water has a specific heat close to water, and drying a food lowers its specific heat, so a dried product heats and cools faster per kg than the wet food it came from.

Thermal conductivity from composition. Thermal conductivity also follows the water content, because water conducts heat far better than fat or air. A component model of the same form gives k = 0.25 Mc + 0.155 Mp + 0.161 Mf + 0.135 Ma + 0.58 Mw, in W/m·K, where the coefficients are the conductivities of the components and water, at 0.58, dominates again. For the same milk, k = 0.25 x 0.048 + 0.155 x 0.035 + 0.161 x 0.04 + 0.135 x 0.007 + 0.58 x 0.87 = 0.012 + 0.005 + 0.006 + 0.001 + 0.505 = 0.53 W/m·K. A simpler rule for fruits and vegetables with more than 60 percent water is k = 0.148 + 0.439 Mw, which gives 0.53 W/m·K for the same milk, again in agreement.

What the models cannot do. These models predict the property at one composition and one temperature, and they assume the components act independently. They do not cover a phase change: when water freezes, the specific heat and the conductivity both change sharply, and ice conducts heat about four times better than liquid water, so a frozen food needs its own set of values. They also do not cover air: a porous food full of air, such as a foam or a dried powder, conducts heat far worse than its composition suggests, because still air is a good insulator. The professional habit is to use the model for a first estimate, to check it against a published value for a similar food, and to measure the property when the design depends on it closely, for example in a sterilisation calculation where the centre temperature decides safety.

Quick check

Two foods have the same composition except that one holds 80 percent water and the other holds 40 percent water. Which has the higher specific heat, and roughly by how much?

Check your answer

The wetter food. By Siebel's equation the difference is 0.034 x (80 minus 40) = 1.36 kJ/kg·K, so the wet food needs far more heat to warm by one degree.

Try it yourself

Fill two identical mugs with hot water at the same temperature. Wrap one mug in a thick towel and leave the other bare. Measure the temperature of each every 5 minutes for half an hour and plot the two cooling curves on paper. The towel is an added resistance in series, so its mug cools more slowly. Now stir one mug and leave the other still, and watch how stirring changes the cooling rate: stirring raises the surface coefficient, which is forced convection at work in your kitchen.

Key ideas to carry forward

  • Thermal diffusivity, α = k / (ρ Cp), decides how fast a temperature change spreads through a food.
  • Conduction follows Fourier's law, q = k A ΔT / L, and thermal conductivity decides how well a material passes heat.
  • Resistances in series add, and the largest resistance controls the heat flow through a wall.
  • Convection follows q = h A ΔT, with natural convection driven by density differences and forced convection driven by pumps, fans, or stirrers.
  • The overall coefficient U rolls the two films and the wall into one number, and the weakest side dominates it.
  • Reynolds, Prandtl, and Nusselt numbers predict h from correlations of the form Nu = C Re^m Pr^n.
  • Counter flow beats parallel flow, and the size of an exchanger follows q = U A LMTD, while NTU suits cases where outlet temperatures are unknown.
  • Fouling adds resistance, lowers U, and must be allowed for in design and cleaning schedules.

Glossary

Conduction

Heat moving through a material by molecule to molecule transfer, with no bulk movement. Example: heat travels through the metal wall of a kettle by conduction.

Fourier's law

The law that heat flow by conduction equals conductivity times area times temperature gradient. Example: a thicker wall passes less heat by Fourier's law.

Thermal diffusivity

How fast a temperature change spreads through a material, α = k / (ρ Cp), in m²/s. Example: a food with low thermal diffusivity stays cool at its centre while its surface is hot.

Thermal conductivity

A material property, k, that says how well it conducts heat, in W/m·K. Example: steel has a high thermal conductivity, foam has a low one.

Thermal resistance

The opposition a layer offers to heat flow, R = L / (k A). Example: insulation has a high thermal resistance, so it slows heat loss.

Convection

Heat carried by the bulk movement of a fluid past a surface. Example: hot water flowing past a cold pipe wall heats it by convection.

Natural convection

Convection driven by density differences, warm fluid rising and cool fluid sinking. Example: currents in a pot of soup warming on a stove.

Forced convection

Convection driven by a pump, fan, or stirrer. Example: milk pumped through a pasteuriser is heated by forced convection.

Surface heat transfer coefficient

The film coefficient h, in W/m²·K, that links heat flow to area and temperature difference in convection. Example: condensing steam has a very high surface coefficient.

Overall heat transfer coefficient

The single coefficient U that rolls the two films and the wall into one number. Example: a pasteuriser is sized from its U value.

Reynolds number

A number with no units that describes the flow pattern, laminar or turbulent. Example: turbulent flow in a tube gives a higher heat transfer coefficient.

Prandtl number

A number with no units comparing momentum spread with heat spread in a fluid. Example: thick oil has a high Prandtl number.

Nusselt number

A number with no units, Nu = (h D) / k, the dimensionless form of the surface coefficient. Example: a correlation gives Nu, and h follows from it.

Heat exchanger

Equipment that moves heat from one stream to another without mixing them. Example: a plate heat exchanger heats milk with hot water.

Parallel flow

An exchanger arrangement where both streams enter at the same end. Example: parallel flow gives a large temperature difference at the inlet and a small one at the outlet.

Counter flow

An exchanger arrangement where the streams enter at opposite ends. Example: counter flow pasteurisers need less area for the same duty.

LMTD

Log mean temperature difference, the correct average temperature difference along an exchanger. Example: exchanger area is found from q = U A LMTD.

NTU

Number of transfer units, (U A) / C min, used to rate exchangers when outlet temperatures are unknown. Example: NTU charts predict the outlet temperatures of a new exchanger.

Shell and tube heat exchanger

An exchanger with one fluid inside a bundle of tubes and the other around them in a shell. Example: juice plants use shell and tube units for sturdy high pressure duty.

Plate heat exchanger

An exchanger of thin corrugated plates with fluids flowing in alternate gaps. Example: milk pasteurisation commonly uses plate heat exchangers.

Fouling

Deposits that build on heat transfer surfaces and add resistance. Example: milk stone fouling on a pasteuriser lowers its U value.

Fouling factor

An added thermal resistance, R f, allowed for fouling in design. Example: a designer includes a fouling factor so the exchanger still works when dirty.

Prediction model

A published equation that estimates a food property from its composition and temperature, instead of measuring it each time. Example: Siebel's equation predicts specific heat from moisture content alone.

Siebel's equation

Cp = 0.837 + 0.034 times the moisture percentage, in kJ/kg·K, a quick estimate of specific heat for high moisture foods. Example: a food with 87 percent moisture has Cp near 3.8 kJ/kg·K by Siebel's equation.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Irudayaraj, Joseph M. (ed.). Food Processing Operations Modeling, Design and Analysis. Marcel Dekker, 2002.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.
  • Ravi Kumar, S. Dairy Engineering. AgriMoon course material.

Unit 3: Evaporation

Note

Unit 3: Evaporation

Tomato juice becomes tomato paste, milk becomes condensed milk, and fruit juice becomes a thick concentrate, all by boiling away water. Evaporation is one of the oldest and most used operations in food processing: it removes water to save transport, to make a product keep longer, and to build flavour and body. Why is it done under vacuum, why does the boiling point rise as the liquid thickens, and why do plants chain several evaporators together? This unit answers those questions.

Questions to think about

  • Why would a plant boil juice at 60 °C instead of 100 °C, when water boils at 100 °C?
  • As juice thickens into concentrate, why does it become harder and harder to boil?
  • If one evaporator can do the job, why do plants often use three?

3.1 The mechanism of evaporation

Evaporation is the removal of a solvent, usually water, from a solution by boiling it off as vapour. It is different from drying, which removes water from a solid, and from distillation, which separates liquids by their different boiling points. In evaporation the aim is to concentrate the liquid that stays behind, and the vapour is usually condensed and discarded or reused.

Heat must be supplied to boil the water, and the amount is set by the latent heat of vaporisation, about 2,257 kJ per kg of water at 100 °C. The heat arrives through a heating surface, usually steam condensing on the other side of a metal wall, and the vapour leaves the top of the vessel. The concentrated liquid leaves from the bottom. Three things set the rate: the area of the heating surface, the temperature difference between the heating steam and the boiling liquid, and the overall heat transfer coefficient, which falls as the liquid thickens and fouls the surface.

Food liquids are sensitive to heat. Vitamins are lost, colours darken, and flavours change when juice or milk sits hot for a long time. That is why food evaporators boil under vacuum: lowering the pressure lowers the boiling temperature, so the product can boil at 50 to 70 °C instead of 100 °C, and the quality stays much closer to the fresh food.

Quick check

Why is evaporation usually done under vacuum in food plants?

Check your answer

Vacuum lowers the boiling temperature, so the food boils at 50 to 70 °C and suffers far less heat damage to colour, flavour, and vitamins.

3.2 Mass and heat balance: the single effect evaporator

Every evaporator is designed from two balances that must both close: a mass balance and a heat balance.

Mass balance. What enters must leave, either as concentrate or as vapour. For a feed F with solids fraction x F, a product P with solids fraction x P, and vapour V:

F = P + V, and F x F = P x P

The solids do not evaporate, so all the solids in the feed appear in the product. A worked example: 1,000 kg/h of juice at 10 percent solids is concentrated to 50 percent solids. The solids flow is 1,000 x 0.10 = 100 kg/h, so the product flow is P = 100 / 0.50 = 200 kg/h, and the vapour flow is V = 1,000 - 200 = 800 kg/h. The economy of the evaporator is the kg of water evaporated per kg of steam used, and a single effect evaporator has an economy near 0.8 to 0.95.

Heat balance. The heat supplied by condensing steam must cover the heat needed to warm the feed to its boiling point plus the latent heat of the vapour formed, plus small losses. In symbols, for steam flow S with latent heat λ s:

S λ s = F Cp (T boil - T feed) + V λ v

where Cp is the specific heat of the feed and λ v is the latent heat of the vapour. The term F Cp (T boil - T feed) is the sensible heat, and V λ v is the latent heat, which is far larger. Feeding the liquid already hot, or preheating it with the vapour, saves steam, which is why plants preheat feed with condensate and with outgoing product.

Quick check

1,000 kg/h of milk at 12 percent solids is evaporated to 40 percent solids. What is the product flow rate?

Check your answer

300 kg/h. Solids flow is 120 kg/h, so P = 120 / 0.40 = 300 kg/h, and 700 kg/h of water is evaporated.

3.3 Boiling point elevation

Pure water at a given pressure boils at one temperature. A solution boils at a higher temperature, because the dissolved solids hold the water back. This rise is the boiling point elevation, BPE. A sugar solution at 50 percent solids boils about 4 K above pure water at the same pressure, and the elevation grows as the liquid concentrates, reaching 5 to 10 K as the solids climb further. Salt solutions, sugar syrups, and milk concentrates all show it.

Boiling point elevation matters twice. First, it shrinks the temperature difference available for heat transfer, because the boiling liquid is hotter than the vapour temperature, so q = U A ΔT falls unless the steam temperature is raised. Second, it is a practical way to follow the concentration: measuring the boiling temperature in the evaporator tells the operator how concentrated the product has become, without taking a sample. The practical work for this unit measures boiling point elevation for a sugar solution at several concentrations, so you can see the rise grow with concentration.

Single effect evaporator: feed in, vapour out, product out

Steam in: condenses on the heating surface, gives up latent heat

↓

Feed in: boils under vacuum, water leaves as vapour

↓

Vapour out the top to the condenser; thick product out the bottom

Mass balance: F = P + V. Heat balance: steam heat = sensible heat + latent heat.

Quick check

A juice concentrate boils at 68 °C while pure water at the same pressure boils at 60 °C. What is the boiling point elevation, and what does it do to heat transfer?

Check your answer

8 K. It shrinks the temperature difference across the heating surface, so the heat flow falls unless the steam temperature is raised.

3.4 Single effect and multiple effect evaporation

Single effect. One vessel, one steam supply, and the vapour goes to a condenser. It is simple and suits small plants and heat sensitive products, but its economy is below 1, so it uses nearly a kilogram of steam per kilogram of water evaporated. It is the right choice when the volume is small or when the product cannot stand long heating.

Multiple effect. Several vessels are chained so that the vapour from the first effect heats the second, the vapour from the second heats the third, and so on. Each later effect runs at a lower pressure and a lower boiling temperature, so the temperature steps down the chain. A triple effect evaporator can reach an economy near 2.5, which means one kilogram of steam evaporates about 2.5 kg of water. The steam is used three times over. The cost is complexity, more vessels, more pumps, and a longer path for the product, which can mean more heat exposure. The balance is struck by the price of steam against the price of equipment and the sensitivity of the product.

Forward feed sends the product in the same direction as the vapour, from the hottest effect to the coolest, which suits heat sensitive liquids because the hottest treatment comes when the liquid is thinnest. Backward feed sends the product against the vapour flow, which suits viscous liquids, because the thickest liquid meets the hottest effect where heat transfer is hardest. Mixed feed is used in larger plants to get the best of both.

Quick check

A triple effect evaporator has an economy of 2.4. How much water does 100 kg of steam evaporate?

Check your answer

About 240 kg. Economy is kg of water evaporated per kg of steam, so 100 x 2.4 = 240 kg.

3.5 Working principles of common evaporators

Central circulating evaporator. A vertical vessel holds a bundle of short tubes, the calandria, with a large central downtake tube. Steam heats the outside of the tubes, the liquid inside boils, and the bubbles lift the liquid up the tubes. The liquid then falls back down the central tube, and the circulation repeats. It handles liquids that crystallise or scale, because the circulation keeps the surfaces swept, and it suits sugar solutions and dairy concentrates.

Climbing film evaporator. The feed enters at the bottom of long vertical tubes and is carried upward by the vapour it makes, as a thin film on the tube wall. The contact time is short, seconds rather than minutes, and the heat transfer is high because the film is thin. It suits heat sensitive liquids such as fruit juice and milk, and it is the standard choice for large juice concentration plants.

Falling film evaporator. The feed is spread at the top of long vertical tubes and flows down as a thin film under gravity, while steam heats the outside. The contact time is again short, and the liquid sees a small temperature difference, so the quality stays high. It handles viscous liquids well and is common for dairy products, tomato paste, and fruit concentrates. Many modern plants use falling film as the main design, sometimes with a climbing film first effect.

Each design is a trade between contact time, heat transfer, and the thickness of the liquid. Thin films and short times protect quality; circulation and agitation protect against fouling and crystals.

Multiple effect evaporation: steam used three times over

Effect 1: fresh steam in, hottest, highest pressure, vapour made

↓

Effect 2: vapour from Effect 1 heats it, cooler, lower pressure

↓

Effect 3: vapour from Effect 2 heats it, coolest, lowest pressure, vapour to condenser

Economy rises with the number of effects; temperature and pressure step down the chain.

Practical note. The practical work for this unit runs mass and energy balances on an evaporator: record the feed rate and solids, the product rate and solids, the steam used, and the temperatures, then check the solids balance and work out the economy. The boiling point elevation practical measures the boiling temperature of a solution at several concentrations and plots the rise, which is the same rise that shrinks the driving temperature difference in a real evaporator.

Quick check

Which evaporator suits a heat sensitive fruit juice best: a central circulating evaporator or a falling film evaporator, and why?

Check your answer

The falling film evaporator. Its thin film and short contact time limit heat damage, while the central circulating type holds the liquid hot for longer.

3.6 Vapour recompression and the factors that set the evaporation rate

The economy of an evaporator can be raised further by recompressing the vapour and using it again as heating steam. In thermal vapour recompression, a steam jet pulls in part of the vapour from the evaporator, mixes it with fresh high pressure steam, and sends the mixture back to the heating surface. In mechanical vapour recompression, a compressor squeezes the vapour to a higher pressure and temperature, and it returns as the heating medium. Both routes reuse the latent heat of the vapour instead of sending it to a condenser, and the economy can rise well above the number of effects, reaching 10 or more in a mechanical recompression plant. The cost is the steam for the jet or the power for the compressor, and the choice turns on the local price of steam against the price of power.

Several factors set the evaporation rate in any evaporator. A larger heating area gives a larger rate, which is why film evaporators use long tubes. A larger temperature difference between the steam and the boiling liquid gives a larger rate, but the boiling point elevation and the vacuum level limit it. A higher overall coefficient U gives a larger rate, and U falls as the liquid thickens, as the viscosity rises, and as fouling builds on the surface. A faster circulation, as in a forced circulation evaporator, raises U by scouring the surface and by keeping the liquid moving. The feed rate also matters: a feed that is too fast floods the tubes and carries liquid into the vapour line, while a feed that is too slow lets the liquid dry on the tubes and burn.

A forced circulation evaporator adds a pump that drives the liquid through the heating tubes at high speed, several metres per second. It handles viscous liquids, liquids that crystallise, and liquids that scale, because the fast flow keeps the surfaces swept and stops deposits settling. It uses more pumping power than a natural circulation unit, but it is the standard choice for tomato paste, for salt crystallisation, and for any liquid that would foul a film evaporator.

Quick check

Why does a mechanical vapour recompression evaporator use far less fresh steam than a single effect evaporator?

Check your answer

It compresses the vapour it makes and uses it again as heating steam, so the latent heat is reused instead of being thrown away in a condenser.

3.7 Worked example: the full heat balance of a single effect evaporator

Section 3.2 sets up the mass balance. This section adds the heat balance, with numbers, so you can see where every kJ goes.

The job. A juice feed of 1,000 kg/h at 10 percent solids and 20 °C is concentrated to 40 percent solids. Steam at 120 °C condenses on the heating surface. The juice boils at 70 °C under vacuum, and its boiling point elevation is 5 K, so the liquid boils at 75 °C while pure water at that pressure would boil at 70 °C. The latent heat of vaporisation is about 2,333 kJ/kg at 70 °C, and the latent heat of the steam is about 2,202 kJ/kg at 120 °C. The specific heat of the juice is about 3.9 kJ/kg·K.

Step 1: mass balance. Solids in are 1,000 x 0.10 = 100 kg/h, and solids stay in the product, so the product flow is 100 / 0.40 = 250 kg/h. Water evaporated is 1,000 minus 250 = 750 kg/h.

Step 2: sensible heat to warm the feed. The feed must be warmed from 20 °C to its boiling point of 75 °C: q sensible = 1,000 x 3.9 x (75 minus 20) = 214,500 kJ/h.

Step 3: latent heat to boil the water. Evaporating 750 kg/h needs q latent = 750 x 2,333 = 1,749,750 kJ/h. This dwarfs the sensible heat, which is the usual pattern: boiling the water costs about eight times as much heat as warming the feed.

Step 4: steam needed. Total heat is 214,500 + 1,749,750 = 1,964,250 kJ/h. Each kg of steam gives up 2,202 kJ, so the steam flow is 1,964,250 / 2,202 = 892 kg/h. The economy is 750 / 892 = 0.84 kg of water per kg of steam, below 1 because part of the steam heat warms the cold feed and part is lost. Preheating the feed with waste heat, or with vapour from the evaporator, lifts the economy toward 1, which is why plants preheat feed whenever they can.

Step 5: check the temperature difference. The steam condenses at 120 °C and the juice boils at 75 °C, so the driving difference is 45 K. Without the boiling point elevation it would be 50 K, so the elevation has cost about a tenth of the driving force. As the juice thickens toward 40 percent solids, the elevation grows and the heat transfer coefficient falls, so the same surface passes less heat near the end of the run, and the operator sees the evaporation rate sag even though the steam pressure has not changed.

Quick check

In the example above, which part of the heat load is larger, warming the feed or boiling the water, and by roughly how much?

Check your answer

Boiling the water, by about eight times: about 1,750,000 kJ/h against about 215,000 kJ/h for warming the feed.

3.8 Dühring's rule, and what goes wrong in practice

Dühring's rule. Boiling point elevation is awkward to measure for every concentration, so engineers use Dühring's rule: at a given concentration, the boiling point of a solution is a straight line function of the boiling point of pure water at the same pressure. In practice this means a Dühring chart, a set of straight lines, one per concentration, on which the boiling point of the solution is read off from the boiling point of water. Two measured points fix one line, and the chart then gives the boiling point at any pressure, which is how the operator knows the boiling temperature when the vacuum changes during a run. The rule holds well for sugar solutions and salt solutions, and it breaks down only when the solute changes its nature with temperature, which most food solutes do not.

Foaming. Some liquids, especially those with proteins, foam as they boil, and the foam climbs into the vapour pipe and carries product into the condenser, which loses solids and fouls the condenser. The fixes, in order of cost, are to lower the boiling rate so the foam has time to break, to add a foam breaker, a spinning device or a mesh that bursts bubbles at the top of the vessel, and to change the feed point so the liquid enters below the boiling surface rather than splashing onto it. A small dose of a food grade antifoam is a last resort, because it must be declared and it can affect later steps.

Entrainment. Even without foam, fast vapour can tear droplets off the boiling surface and carry them away, a loss called entrainment. It rises with vapour velocity, so it is worst at high evaporation rates and in vessels with a small vapour space. The fix is a separator, a wider space above the liquid where the vapour slows and the droplets fall back, or a set of baffles that force the vapour to turn, throwing the droplets out by their own weight. A plant that finds solids in its condensate has an entrainment problem, and the condensate is no longer clean water for reuse.

Scaling and fouling. Dissolved salts and proteins deposit on the heating surface as the liquid concentrates, and the deposit insulates the surface, so the heat transfer coefficient falls over the run. Milk stone and calcium salts are the usual culprits. The signs are a falling evaporation rate at a fixed steam pressure and a rising temperature difference across the wall. The fix is a cleaning cycle on a set schedule, and in design, a higher liquid velocity across the surface, which scours the deposit, and a heating surface that can be opened and cleaned. An evaporator that is never cleaned slowly starves the plant of capacity, and the cost shows up as steam use per kg of water, which is why the economy is tracked every shift.

Quick check

A plant finds solids in the condensate from its evaporator. What is the fault called, and what part of the vessel should be checked first?

Check your answer

Entrainment. The vapour space and separator above the boiling liquid should be checked, because fast vapour is carrying droplets into the condenser.

Try it yourself

Put a measured volume of salt water, say 500 ml at 5 percent salt, in a wide shallow pan on a stove, and weigh the pan before heating. Boil it gently for 10 minutes, then weigh it again. The weight lost is the water evaporated. Work out the new salt percentage from the mass balance. Now taste a drop, carefully, and note how the boiling behaviour changes as it thickens. You have just run a single effect evaporator and measured its mass balance with kitchen tools.

Key ideas to carry forward

  • Evaporation concentrates a liquid by boiling off water, usually under vacuum so the food boils at 50 to 70 °C and keeps its quality.
  • The mass balance is F = P + V with F x F = P x P, and the heat balance sets steam use against sensible heat plus latent heat.
  • Boiling point elevation raises the boiling temperature as solids rise, and it shrinks the driving temperature difference for heat transfer.
  • Multiple effect evaporation reuses vapour from one effect to heat the next, raising economy to near 2.5 for three effects.
  • Central circulating, climbing film, and falling film evaporators differ in contact time, film thickness, and the liquids they suit best.

Glossary

Evaporation

Removing water from a liquid by boiling it off as vapour, to concentrate what stays behind. Example: juice is evaporated to make concentrate for transport.

Latent heat of vaporisation

The heat needed to turn a liquid into vapour at its boiling point, with no change in temperature. Example: about 2,257 kJ turns 1 kg of water at 100 °C into steam.

Mass balance

The rule that what enters a unit must leave it, as product, vapour, or waste. Example: the solids in the feed must equal the solids in the product from an evaporator.

Heat balance

The rule that heat supplied must equal heat used plus heat lost. Example: steam heat in an evaporator covers warming the feed plus boiling off water.

Economy

The kg of water evaporated per kg of steam used. Example: a triple effect evaporator with an economy of 2.4 evaporates 240 kg of water with 100 kg of steam.

Boiling point elevation

The rise in boiling temperature of a solution above pure water at the same pressure. Example: a 50 percent sugar solution boils several degrees above water.

Single effect evaporator

One evaporator vessel heated by fresh steam, with vapour going to a condenser. Example: a small plant concentrates juice in a single effect unit.

Multiple effect evaporator

Several vessels chained so vapour from one heats the next. Example: a triple effect plant concentrates milk with much less steam per kg of water.

Forward feed

Feeding product in the same direction as vapour flow, hottest effect first. Example: forward feed suits heat sensitive juice because the thin liquid meets the hottest effect.

Backward feed

Feeding product against the vapour flow, so the thickest liquid meets the hottest effect. Example: backward feed suits viscous liquids that need the most heat at the end.

Calandria

The bundle of heating tubes inside an evaporator, with steam on the outside. Example: the calandria of a central circulating evaporator holds short vertical tubes.

Climbing film evaporator

An evaporator where liquid is carried up long tubes as a thin film by its own vapour. Example: fruit juice is concentrated in a climbing film evaporator for short contact time.

Falling film evaporator

An evaporator where liquid flows down long tubes as a thin film under gravity. Example: milk concentrate is made in falling film evaporators.

Central circulating evaporator

An evaporator with a tube bundle and a central downtake, where the liquid circulates by natural convection. Example: sugar solutions are evaporated in central circulating units.

Sensible heat

Heat that changes the temperature of a material without changing its state. Example: warming feed from 20 °C to its boiling point uses sensible heat.

Vapour

The gas formed when a liquid boils, carrying the latent heat away. Example: vapour from the first effect heats the second effect in a multiple effect plant.

Dühring's rule

The rule that the boiling point of a solution at a given concentration is a straight line function of the boiling point of pure water at the same pressure. Example: a Dühring chart gives the boiling point of a sugar solution at any vacuum from the boiling point of water.

Entrainment

Droplets of liquid carried away by fast vapour into the condenser, losing product and fouling the condensate. Example: solids in the condensate are a sign of entrainment in an evaporator.

Scaling

Deposits of salts or proteins building on a heating surface, insulating it and cutting the heat transfer rate. Example: milk stone scaling on an evaporator tube lowers its economy over a run.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.
  • Ravi Kumar, S. Dairy Engineering. AgriMoon course material.

Unit 4: Drying and Dehydration

Note

Unit 4: Drying and Dehydration

A grape becomes a raisin, milk becomes powder, and a fresh herb becomes a dried spice that keeps for a year. Drying removes water from a solid so microbes cannot grow and the food becomes light to store and ship. Why does drying slow down halfway through, why does the surface of a drying slice sometimes harden and trap water inside, and why does a dryer need hot dry air rather than just hot air? This unit answers those questions.

Questions to think about

  • Why does a wet cloth dry fast at first and then slower and slower, even though the air stays the same?
  • What is the difference between water that is free to leave a food and water that is bound tight to it?
  • Why can a food be dry on the outside and still wet in the middle?

4.1 Mass balance and moisture content

Drying is a mass transfer operation: water moves from inside the food to its surface, then leaves the surface into the air. The mass balance is simple. A feed of wet solid F with moisture fraction w F gives a dried product P with moisture fraction w P and evaporated water W:

F = P + W, and F w F = P w P + W

Moisture content is stated two ways, and mixing them up is a common error. Wet basis moisture is the mass of water divided by the total mass of the wet material, times 100. Dry basis moisture is the mass of water divided by the mass of dry solids, times 100. Dry basis is the engineer's choice for drying calculations, because the dry solids stay constant while the water leaves. A food with 80 percent moisture wet basis holds 4 kg of water per kg of dry solids, which is 400 percent dry basis.

A worked example: 100 kg of sliced apple at 85 percent moisture wet basis is dried to 15 percent moisture wet basis. The dry solids are 15 kg. In the product, those 15 kg are 85 percent of the mass, so the product mass is 15 / 0.85 = 17.65 kg, and the water removed is 100 - 17.65 = 82.35 kg. That is the mass balance of a dryer in one line, and it sets the size of the air heater, because every kilogram of water removed needs its latent heat supplied by the air.

Quick check

A food holds 3 kg of water per kg of dry solids. What is its moisture content on a dry basis?

Check your answer

300 percent dry basis. Dry basis moisture is water mass divided by dry solids mass, times 100.

4.2 Equilibrium moisture content

A food left in air of a given humidity will dry, or take up water, until it reaches a moisture content where it is in balance with that air. That value is the equilibrium moisture content, EMC. It depends on the temperature and the relative humidity of the air, and on the food itself. At high relative humidity the EMC is high, at low humidity it is low, and the curve linking them is the sorption isotherm.

The EMC is the floor for drying: no dryer can dry a food below its EMC in that air. To reach a lower moisture, the air must be drier or hotter. It also governs storage. A dried food stored in damp air will pick up water until it reaches the EMC for that air, and then it may cake, go soft, or grow mould. That is why dried foods are packed in moisture proof bags and stored in dry rooms.

Hysteresis is a real feature of the isotherm: a food drying down to a given humidity reaches a different EMC from the same food wetting up to the same humidity. The drying curve sits above the wetting curve, so the history of the food matters, not only the air around it now.

Quick check

Air at 60 percent relative humidity gives a food an EMC of 12 percent dry basis. Can a dryer using that air dry the food to 8 percent dry basis?

Check your answer

No. The food cannot go below its equilibrium moisture content in that air. The air must be made drier or hotter.

4.3 Bound water and free water

Not all water in a food is the same. Free water sits in pores and capillaries and behaves much like pure water: it flows easily, freezes near 0 °C, dissolves salts, and lets microbes grow. It leaves first in drying. Bound water is held tight to proteins, starches, and sugars by hydrogen bonds and other forces. It does not freeze at 0 °C, does not dissolve salts in the usual way, and microbes cannot use it. It leaves last, and removing it takes much more energy per kilogram.

The line between them is not sharp, and it moves with the food. Fresh fruit is mostly free water. A dried fruit at low moisture holds mostly bound water. The drying rate curve in section 4.5 is really the story of free water leaving fast, then bound water leaving slowly.

Quick check

Which leaves first in drying, free water or bound water, and why?

Check your answer

Free water. It is held loosely in pores, while bound water is held tight to proteins and starches and needs more energy to remove.

4.4 Principles of drying

Drying needs two steps working together. First, heat must reach the water, to supply the latent heat of vaporisation. Second, the vapour must be carried away, so the air at the surface stays dry enough for more water to leave. If either step is slow, drying is slow.

Hot air does both jobs at once in most dryers: it carries heat in and carries vapour out. The air temperature, its humidity, and its velocity all matter. Hotter air carries more heat and can hold more vapour. Drier air pulls water off the surface faster. Faster air sweeps vapour away and thins the still layer of air clinging to the surface. Raising any one of the three speeds drying, but raising temperature also risks damage: colour darkening, flavour loss, and case hardening, where the surface dries into a hard skin that traps water inside.

The heat can arrive in other ways too. In a drum dryer, heat comes by conduction through a hot metal drum. In a freeze dryer, heat arrives gently while water leaves by sublimation under vacuum. In a spray dryer, hot air meets tiny droplets, and the huge surface area dries them in seconds. The principle is the same in every case: supply latent heat, and keep the vapour moving away.

What drying needs: heat in, vapour out

Hot dry air in: brings heat, carries little vapour

↓

Wet food: heat turns water to vapour at the surface

↓

Moist air out: carries the vapour away, keeps the surface dry

If the air is already damp, or too slow, drying stalls.

Quick check

A dryer is running with hot but very humid air. Will the food dry fast? Why?

Check your answer

No. Damp air cannot take up much more vapour, so water leaves the surface slowly even though heat is plentiful.

4.5 The drying curve and the drying rate curve

If you weigh a food sample at intervals while it dries in steady air, and plot moisture content against time, you get the drying curve: moisture falls fast at first, then more slowly, then levels off near the EMC. If you plot the drying rate, the water removed per unit area per unit time, against moisture content, you get the drying rate curve, and it has three clear parts.

Warming up period. The food warms to the wet bulb temperature of the air, and the rate rises briefly. This is short and often ignored in design.

Constant rate period. The surface is wet, water reaches the surface as fast as it evaporates, and the rate stays constant. The surface sits at the wet bulb temperature, and the rate is set by the air, not by the food: hotter, drier, faster air gives a faster constant rate. The food dries like a free water surface in this period. The moisture content at the end of this period is the critical moisture content, X c.

Falling rate period. Below the critical moisture content, the surface starts to dry out, water must travel from inside the food to the surface, and that internal movement is slow. The rate falls steadily as the moisture falls, and the food warms above the wet bulb temperature. This period is set by the food, not the air: its thickness, its pore structure, and how fast water diffuses inside it. That internal movement is very slow: in grain, the moisture diffusivity during drying is about 1.3 x 10⁻⁷ m²/hr, and during wetting it is about ten times slower still, near 1.2 x 10⁻⁸ m²/hr. Most of the drying time is spent here, and most of the quality damage happens here, because the food is hot and dry.

The drying rate in the constant rate period can be worked out from a heat balance: the heat arriving from the air, h A (T air - T surface), supplies the latent heat of the water leaving, so the rate is N c = h (T air - T surface) / λ, where λ is the latent heat. In the falling rate period, the rate is often taken as falling in direct proportion to the free moisture still present, which lets the drying time be worked out by integration.

4.6 Calculation of drying rate and drying time

Drying rate. In the constant rate period, rate per unit area N c = h (T air - T wet bulb) / λ, in kg of water per m² per second. Multiply by the exposed area to get the total rate in kg/s.

Drying time. The total time is the sum of the time in the constant rate period and the time in the falling rate period. For a batch of dry solids S with area A, the constant rate time is:

t c = S (X 0 - X c) / (A N c)

where X 0 is the starting moisture on a dry basis and X c is the critical moisture on a dry basis. For the falling rate period, if the rate falls in direct proportion to the free moisture, the time is:

t f = S (X c - X e) / (A N c) x ln[(X c - X e) / (X - X e)]

where X e is the equilibrium moisture and X is the final moisture, all on a dry basis. A worked example: S = 10 kg, A = 2 m², N c = 0.0005 kg/m²·s, X 0 = 4.0, X c = 1.0, X e = 0.1, final X = 0.2. The constant rate time is t c = 10 x (4.0 - 1.0) / (2 x 0.0005) = 30 / 0.001 = 30,000 s, about 8.3 hours. The falling rate time is t f = 10 x (1.0 - 0.1) / (2 x 0.0005) x ln[(1.0 - 0.1) / (0.2 - 0.1)] = 9,000 x ln(9) = 9,000 x 2.197 = 19,774 s, about 5.5 hours. Total drying time is about 13.8 hours, and the falling rate period takes a large share even though it removes far less water.

The design lessons follow from the equations. Thinner slices dry faster because water has less far to travel. More area dries faster. Hotter, drier air helps the constant rate period, but the falling rate period is set by the food, so the only levers there are thickness, temperature, and patience.

Practical note. The practical work for this unit measures a drying curve and a drying rate curve directly: weigh slices of a food at set times in a tray dryer, plot moisture against time, then plot rate against moisture, and read off the critical moisture content and the constant rate. The same data give the drying time calculation above, so the experiment tests the theory in one run.

The drying rate curve: fast, then slowing

Constant rate: surface wet, rate set by the air, food at wet bulb temperature

↓

Critical moisture content: surface starts to dry out

↓

Falling rate: water must travel from inside, rate falls, food warms up

Most drying time is spent in the falling rate period.

Quick check

Why does doubling the air temperature not halve the total drying time?

Check your answer

Hotter air speeds only the constant rate period. The falling rate period is set by water moving inside the food, which temperature helps less, so the total time falls by less than half.

4.7 Dryer types in food plants

The principles of section 4.4 are built into several dryer designs, and the choice turns on the form of the food and the scale.

  • Tray dryer. Trays of food sit in a cabinet with hot air blown over them. It is a batch dryer, simple and flexible, suited to small batches of fruit, herbs, and spices, but it needs labour to load and unload, and the trays nearest the heater dry faster than the rest.
  • Tunnel dryer. Trays or a belt move through a long tunnel with hot air, so the food enters wet at one end and leaves dry at the other. It runs without stopping and suits larger volumes of fruit and vegetables.
  • Spray dryer. A liquid is sprayed into fine drops in a tall chamber of hot air, and the drops dry in seconds to a powder that falls to the bottom. It is the standard dryer for milk powder, instant coffee, and egg powder, because the contact time is so short that heat damage is small even though the air is hot.
  • Drum dryer. A thin film of liquid is spread on a hot rotating drum, dries in one turn, and is scraped off as flakes. It suits viscous liquids and pastes, such as potato flakes and baby food, and it gives a cooked flavour that suits those products.
  • Freeze dryer. The food is frozen, then put under vacuum so the ice leaves by sublimation, straight from solid to vapour. It keeps colour, flavour, and shape better than any other dryer, because the food never gets hot and the structure does not collapse, but it is slow and costly, so it suits high value foods such as instant coffee, berries, and meals for special uses.

Quick check

Why does a spray dryer suit milk powder even though its air is very hot?

Check your answer

The drops are tiny and dry in seconds, so the contact time is very short and the milk suffers little heat damage despite the hot air.

4.8 Derivation: drying time in the constant rate and falling rate periods

Section 4.6 states how drying rate and drying time are calculated. This section derives the drying time formulas, so you can see where they come from and what they assume.

Constant rate period: start from the heat balance. In the constant rate period the surface stays wet, so all the heat arriving at the surface is used to evaporate water. The heat arrives by convection from the air: q = h A (T air minus T surface), where h is the surface coefficient and T surface is the wet bulb temperature. The water evaporated per second is q divided by the latent heat, λ. The drying rate, in kg of water per kg of dry solids per second, is therefore R c = h A (T air minus T wet bulb) / (λ W s), where W s is the mass of dry solids. The time to dry from moisture X₁ to moisture X₂ in this period is the moisture removed divided by the rate: t = W s (X₁ minus X₂) / (h A (T air minus T wet bulb) / λ). The formula shows three plain things: the time grows with the mass of solids and the moisture to be removed, it falls when the air is hotter or drier, and it falls when the surface area is larger, which is why thin layers and small pieces dry fast.

Worked example. A tray holds 10 kg of dry solids of sliced fruit at 3.0 kg water per kg solids, to be dried to the critical moisture content of 1.0 kg/kg in the constant rate period. The tray area is 2 m², h is 25 W/m²·K, the air is at 70 °C, the wet bulb temperature is 30 °C, and λ is 2,400 kJ/kg. The heat rate is q = 25 x 2 x (70 minus 30) = 2,000 W, which evaporates 2,000 / 2,400,000 = 0.000833 kg/s of water. The water to remove is 10 x (3.0 minus 1.0) = 20 kg, so the time is 20 / 0.000833 = 24,000 s, about 6.7 hours. That long time is the reason fruit drying is done in thin layers with fast moving air, and why a thicker layer or still air would take far longer.

Falling rate period: the rate follows the moisture left. Below the critical moisture content the surface dries out, and the rate is set by water moving from inside the food to the surface. In the simplest useful model the rate is proportional to the free moisture still present, R = k (X minus X e), where X e is the equilibrium moisture content and k is a constant for the food and the air. This is the same mathematics as a cooling curve: the moisture falls fast at first, then ever more slowly as it nears equilibrium, and it never quite reaches equilibrium in finite time.

Deriving the falling rate time. The rate of moisture loss is minus W s (dX / dt) = A R, with R as above. Separating the variables gives dX / (X minus X e) = minus (k A / W s) dt. Integrating from the critical moisture X c at time zero to moisture X at time t gives ln((X c minus X e) / (X minus X e)) = (k A / W s) t, so t = (W s / (k A)) ln((X c minus X e) / (X minus X e)). The logarithm is the warning: each equal step down in moisture takes longer than the one before, so drying the last few percent of water costs more time than drying the first half, and chasing a very low moisture content is expensive. In practice the constant k is found from a drying test on the real food, not from a book, because it depends on the structure of the food and how it was cut.

Quick check

A food is dried from the critical moisture content toward its equilibrium moisture content. Does the last 10 percent of the removable water take the same time as the first 10 percent, and why?

Check your answer

No, it takes far longer. In the falling rate period the rate is proportional to the moisture still above equilibrium, so the rate falls as the food dries, and the time follows a logarithm.

4.9 Inside the food: diffusion, Fick's law, and the sorption isotherm models

Fick's law. Water moves inside a food by diffusion, from wet regions to dry regions, and Fick's law states the rate: the flux of water, in kg per m² per second, equals minus D times the gradient of moisture concentration, where D is the moisture diffusivity, in m²/s. The minus sign says water flows down the concentration slope, from wet to dry. The diffusivity is not a constant: it rises with moisture content and with temperature, and it falls steeply as the food dries, which is the deep reason the falling rate period slows down. Modelling work on grains shows the same pattern, with diffusivity falling as the grain dries and rising when the grain is rewetted, and with different values for the parts of a grain, such as the bran and the inner kernel, because their structures differ.

Why diffusion sets the limit. In the constant rate period the surface is wet, so the air sets the rate. In the falling rate period the surface is dry, so the inside sets the rate, and no amount of hotter air can push water out faster than diffusion can bring it to the surface. This is why blowing hotter air at a nearly dry food wastes energy: the heat arrives, but the water is not there to carry it away, so the food heats up instead, and heat sensitive foods scorch or case harden. The professional response is to dry fast while the food is wet, then to dry gently, with milder air, once the falling rate period starts.

Sorption isotherm models. Section 4.2 describes the sorption isotherm as a curve. Three named models turn that curve into an equation, and each has a use. The BET model describes water bound in layers on a surface, and it fits the low humidity end of the curve, up to about 50 percent relative humidity. The GAB model extends BET with a third layer, and it fits most foods over the full range, from dry to damp, which is why it is the standard model in food work. The Oswin model is a simple two constant equation that fits many foods well enough for storage calculations. The engineer uses the model to answer practical questions: at what humidity will a powder cake, how dry must the air be to hold a dried food below its safe moisture, and how much water will a package pick up through its film over its shelf life. The model is fitted to measured data for the real food, because two foods with the same moisture content can hold their water very differently.

Reading the psychrometric chart. The practical work for this unit uses the psychrometric chart, and it is worth knowing how to read it in four moves. First, find the dry bulb temperature on the bottom scale and go straight up. Second, follow the wet bulb line, which slopes up to the left, to the saturation curve, to read the wet bulb temperature. Third, read the humidity ratio, the kg of water per kg of dry air, off the right hand scale, and the relative humidity off the curved lines. Fourth, read the enthalpy off the diagonal scale. With those four readings you can work out how much water a kg of air can pick up, how much heat the air carries, and whether the air leaving a dryer is near saturation, which tells you whether the dryer is using its air well or throwing hot damp air away.

Quick check

A dryer is running with very hot air, but the food is nearly dry and its temperature is climbing. Is more heat the fix, and why?

Check your answer

No. In the falling rate period diffusion inside the food sets the rate, not the air, so extra heat only overheats the food. Gentler air and more time are the fix.

Try it yourself

Slice an apple into thin slices and thick slices. Weigh each batch, lay them on a tray in a warm oven with the door slightly open, and weigh them every 15 minutes. Plot moisture against time for both batches. The thin slices will reach low moisture much sooner, and their drying rate curve will show the constant rate period ending earlier. Note also the colour change: the slices that dried slowly in the falling rate period brown more, because they stayed warm and moist for longer.

Key ideas to carry forward

  • Moisture is stated wet basis or dry basis, and dry basis is the right choice for drying calculations because dry solids stay constant.
  • Equilibrium moisture content is the floor for drying in a given air, and it sets the safe moisture for storage.
  • Free water leaves first and easily, bound water leaves last and costs more energy.
  • Drying needs heat in and vapour out, and hot, dry, fast air speeds it, while case hardening traps water inside.
  • The drying rate curve has a constant rate period set by the air and a falling rate period set by the food, split at the critical moisture content.
  • Drying time is the sum of the two periods, and thinner slices and larger area cut it most directly.

Glossary

Drying

Removing water from a solid so microbes cannot grow and the food keeps. Example: grapes are dried to raisins.

Dehydration

Drying by removing water, often with heat, to a low moisture content. Example: milk is dehydrated to milk powder.

Moisture content, wet basis

Mass of water divided by total mass of the wet material, times 100. Example: fresh apple at 85 percent moisture wet basis is mostly water.

Moisture content, dry basis

Mass of water divided by mass of dry solids, times 100. Example: 4 kg of water per kg of dry solids is 400 percent dry basis.

Equilibrium moisture content

The moisture a food reaches when it is in balance with the air around it. Example: a dried food in damp air picks up water until it reaches its EMC.

Sorption isotherm

The curve linking equilibrium moisture content to the relative humidity of the air. Example: the isotherm shows how damp the air must be before a powder cakes.

Free water

Water held loosely in pores, behaving like pure water and available to microbes. Example: the juice in fresh fruit is mostly free water.

Bound water

Water held tight to proteins, starches, and sugars, not available to microbes. Example: the last water in a dried fruit is bound water.

Wet bulb temperature

The temperature a wet surface reaches when evaporation cools it in moving air. Example: the surface of food in the constant rate period sits at the wet bulb temperature.

Constant rate period

The first part of drying, where the surface stays wet and the rate stays steady. Example: the constant rate period is set by the air, not the food.

Critical moisture content

The moisture at which the constant rate period ends and the falling rate period begins. Example: below the critical moisture content the surface starts to dry out.

Falling rate period

The later part of drying, where the rate falls because water must travel from inside the food. Example: most drying time is spent in the falling rate period.

Drying curve

A plot of moisture content against time during drying. Example: the drying curve falls fast at first, then levels off near the EMC.

Drying rate curve

A plot of drying rate against moisture content, showing the constant and falling rate periods. Example: the drying rate curve shows where the critical moisture content sits.

Case hardening

A hard dry skin forming on the surface during drying, trapping water inside. Example: drying fruit too hot causes case hardening.

Latent heat

Heat that changes the state of water without changing its temperature. Example: each kg of water dried off needs about 2,257 kJ of latent heat.

Fick's law

The law that water moves inside a food from wet regions to dry regions, at a rate set by the moisture diffusivity and the concentration gradient. Example: the falling rate period is slow because diffusion inside the food is slow.

Moisture diffusivity

A property, D, in m²/s, that says how fast water diffuses through a food. Example: diffusivity falls steeply as a food dries, which is why the last water is the slowest to remove.

GAB model

A three layer sorption isotherm model that fits most foods over the full humidity range, used to predict safe storage moisture. Example: the GAB model predicts the humidity at which a powder starts to cake.

BET model

A sorption isotherm model for water bound in layers on a surface, fitting the low humidity end of the curve. Example: the BET model is used for the dry end of an isotherm, below about 50 percent relative humidity.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Irudayaraj, Joseph M. (ed.). Food Processing Operations Modeling, Design and Analysis. Marcel Dekker, 2002.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.

Unit 5: Distillation

Note

Unit 5: Distillation

A fermented liquid holds alcohol and water mixed together, and a flavour maker wants the aroma compounds out of a spice without the rest. Distillation separates liquids by boiling them: the more volatile part leaves as vapour first, and the vapour is condensed into a richer product. Why does the vapour hold more alcohol than the liquid it came from, why does a tall column with many plates do better than a single pot, and why can some mixtures never be fully separated by boiling? This unit answers those questions.

Questions to think about

  • If you boil a mix of alcohol and water, why is the vapour richer in alcohol than the liquid?
  • Why does a tall column separate better than a single boiling pot?
  • What happens when a mixture reaches a point where the vapour and the liquid have the same composition?

5.1 Introduction and types

Distillation is the separation of a liquid mixture into its parts by boiling and condensing. The part that boils more easily, the more volatile component, is enriched in the vapour, and when the vapour is condensed, the condensate is richer in that component than the liquid it came from. Repeating the step enriches it further, and a column does the repetition in one vessel.

Food uses of distillation include recovering alcohol from fermented liquids, making spirits and flavour extracts, recovering solvents used in extraction, and removing off flavours from oils and fats by steam stripping. The two liquids most often met in food distillation are ethanol, C₂H₅OH, and water, H₂O, and this unit uses that pair as its running example, while the methods apply to any pair.

The main types are:

  • Simple distillation. One boiling pot and one condenser. It separates liquids whose boiling points are far apart, and it gives only a rough separation.
  • Fractional distillation, or rectification. A column with plates or packing, where vapour and liquid meet many times. It separates liquids with closer boiling points and gives a much purer product.
  • Steam distillation. Steam is blown through the liquid, and volatile compounds leave with the steam at a temperature below their normal boiling point. It suits heat sensitive flavours and essential oils that would be damaged by direct boiling.
  • Vacuum distillation. The column runs at reduced pressure, so the liquids boil cooler. It suits heat sensitive products and high boiling compounds.

Quick check

Which type of distillation suits a delicate flower aroma that would be damaged by boiling at 100 °C?

Check your answer

Steam distillation, or vacuum distillation. Both let the volatile compounds leave at a lower temperature than direct boiling.

5.2 Vapour-liquid relationship: Raoult's law and boiling point diagrams

The key to distillation is the link between the composition of a liquid and the composition of the vapour above it. For an ideal mixture, Raoult's law states that the partial pressure of each component in the vapour equals its mole fraction in the liquid times its vapour pressure as a pure liquid at that temperature:

p A = x A P A*, p B = x B P B*

Here x is the liquid mole fraction, P* is the pure component vapour pressure, and p is the partial pressure in the vapour. The total pressure is the sum of the partial pressures, and the vapour mole fraction of A is y A = p A / P total. Because the more volatile component has the higher pure vapour pressure, its vapour is richer than its liquid: y A is greater than x A when A is more volatile.

Two diagrams carry the whole story for a binary mixture at constant pressure.

The t-x-y diagram. Temperature is plotted against composition. Two curves appear: the lower curve is the boiling point of the liquid, the bubble point curve, and the upper curve is the dew point of the vapour. Between the curves, liquid and vapour exist together. Pick a liquid composition, go up to the bubble point curve to read its boiling temperature, then go across to the dew point curve to read the composition of the vapour in balance with it. For ethanol and water, the vapour is richer in ethanol than the liquid at every point below the azeotrope, which is why distillation works; above the azeotrope the vapour is leaner in ethanol than the liquid, so the separation runs the other way.

The x-y diagram. Vapour composition y is plotted against liquid composition x. The diagonal line y = x marks no separation. The equilibrium curve sits above the diagonal, and the gap between the curve and the diagonal measures how easy the separation is. A wide gap means easy separation and few plates; a narrow gap means hard separation and many plates. When the curve meets the diagonal, the mixture is an azeotrope, and the vapour has the same composition as the liquid, so no further separation by ordinary distillation is possible. Ethanol and water form an azeotrope near 95.6 percent ethanol by weight, which is why ordinary distillation cannot make pure ethanol from a fermented liquid.

Quick check

In an x-y diagram, what does it mean when the equilibrium curve touches the diagonal line?

Check your answer

That mixture is an azeotrope: vapour and liquid have the same composition, so ordinary distillation cannot separate it further.

5.3 Rectification: the column and its accessories

A rectification column carries out many simple distillations in one vessel. Liquid feed enters near the middle. Vapour rises through the column, liquid flows down, and on each plate the two meet, exchange heat and mass, and move toward balance. The vapour grows richer in the volatile component as it rises, and the liquid grows richer in the less volatile component as it falls.

The main parts are:

  • Plates or trays. Perforated plates, bubble cap plates, or sieve plates hold a shallow layer of liquid that the rising vapour bubbles through. Each plate is one contact stage.
  • Packing. An alternative to plates: rings or structured packing give a large surface for vapour and liquid to meet, with a low pressure drop. Packing suits small columns and vacuum duty.
  • Reboiler. A heater at the bottom that boils the liquid leaving the column and sends vapour back up. It supplies the heat for the whole separation.
  • Condenser. A cooler at the top that condenses the vapour leaving the column. Part of the condensate returns to the column as reflux, and the rest leaves as the top product, the distillate.
  • Reflux. The liquid returned to the top of the column. It flows down and washes the rising vapour, which is what makes the separation sharp. More reflux means purer products but more heat use, so the reflux ratio is a key operating choice.

The section above the feed is the rectifying section, where the vapour is purified, and the section below the feed is the stripping section, where the liquid is stripped of its volatile part. The bottom product is rich in the less volatile component, and the top product is rich in the volatile one.

A rectification column: vapour rises, liquid falls

Top: condenser, distillate out, part returns as reflux

↓

Rectifying section: plates wash the rising vapour, it grows richer

↓

Feed enters near the middle

↓

Stripping section: falling liquid is stripped of volatile part

↓

Bottom: reboiler boils liquid back up, bottom product out

Quick check

What is reflux, and what does raising it do to the separation?

Check your answer

Reflux is condensed top vapour returned to the column. Raising it improves the separation and the purity, but it uses more heat in the reboiler.

5.4 Operating lines and the effect of feed temperature

The McCabe-Thiele method works on the x-y diagram with two straight lines, the operating lines, one for each section of the column.

Rectifying section operating line. A balance over the top of the column gives y = [R / (R + 1)] x + x D / (R + 1), where R is the reflux ratio, the reflux flow divided by the distillate flow, and x D is the distillate composition. Its slope is R / (R + 1), always below 1, and it passes through the point (x D, x D) on the diagonal. A larger reflux ratio makes the slope closer to 1, which pushes the line closer to the diagonal and gives a sharper separation.

Stripping section operating line. A balance over the bottom gives a line with a slope above 1, passing through the point (x B, x B), where x B is the bottom product composition. It is found from the feed line and the rectifying line.

The feed line, or q-line. The feed condition sets where the two operating lines meet. The value q is the heat needed to bring one mole of feed to its boiling point, divided by the molar latent heat. A cold liquid feed has q above 1, and its q-line slopes steeply upward. A saturated liquid feed at its boiling point has q = 1, and its q-line is vertical. A partly vaporised feed has q between 0 and 1, and a saturated vapour feed has q = 0, with a horizontal q-line. A cold feed is the common case in food plants, and it needs extra heat in the column, which the reflux supplies, so a cold feed increases the load on the reboiler.

Quick check

A feed enters the column as a cold liquid, below its boiling point. Is q above or below 1, and what does that mean for the reboiler?

Check your answer

q is above 1. Extra heat is needed to bring the feed to its boiling point, so the reboiler load rises.

5.5 Number of ideal plates: McCabe-Thiele, Lewis, and minimum reflux

McCabe-Thiele method. On the x-y diagram, draw the equilibrium curve, the diagonal, the two operating lines, and the q-line. Then step off stages between the equilibrium curve and the operating lines, starting at the distillate composition and stepping down to the bottom composition. Each step is one ideal plate, a plate where the leaving vapour and liquid are in balance. Count the steps to get the number of ideal plates. The feed enters at the plate where the stepping crosses the q-line. It is a graphical method, easy to see and to check, and it is the standard teaching method.

Lewis method. The Lewis method does the same stepping by calculation rather than by drawing, working plate by plate from the top with material balances and the equilibrium relation. It gives the same count as McCabe-Thiele when the same assumptions are used, and it suits a spreadsheet where drawing is awkward.

Minimum reflux. If the reflux ratio is lowered, the operating lines move away from the diagonal, more plates are needed, and at one low value the lines touch the equilibrium curve at a pinch point. That is the minimum reflux ratio: below it, the separation is impossible with any number of plates, and at it, the column would need an infinite number of plates. Real columns run at a reflux ratio 1.2 to 1.5 times the minimum, a balance between the cost of plates and the cost of heat.

Total reflux. At the other extreme, with all the condensate returned and no product taken, the operating lines lie on the diagonal and the number of plates is at its minimum. A column is often tested at total reflux to measure its plate efficiency.

Quick check

Why does lowering the reflux ratio raise the number of plates needed?

Check your answer

A lower reflux ratio moves the operating lines away from the diagonal, so each plate does less separation, and more plates are needed to reach the same products.

5.6 Column diameter

The diameter of a column is set by the vapour flow: the vapour must rise fast enough to bubble through the liquid on each plate, but not so fast that it carries liquid up with it, which is called flooding. The design uses the maximum allowable vapour velocity, found from the flooding velocity for the plate type and the vapour and liquid densities, with a safety factor, often 70 to 85 percent of flooding.

The vapour volume flow is worked out from the molar flow and the vapour density, and the area needed is the volume flow divided by the chosen velocity. A wider column costs more but runs safely below flooding, while a narrow column is cheaper but risks flooding and poor separation. The height is set by the number of plates times the plate spacing, plus space for the feed, the disengagement zones at top and bottom, and the reboiler and condenser connections.

Practical note. The practical work for this unit runs a mass and energy balance on a distillation: record the feed rate and composition, the distillate and bottom rates and compositions, and the heat supplied, then check the overall and component balances. It also finds the number of ideal plates by the Lewis method and by the McCabe-Thiele diagram for the same separation, so you can see the two methods agree, and it studies a laboratory rectification column to match the parts named in section 5.3 to real equipment.

5.7 Derivation: the rectifying operating line from a material balance

Section 5.4 states the rectifying operating line, y = [R / (R + 1)] x + x D / (R + 1). This section derives it from a balance, one small step at a time.

Step 1: draw a box around the top of the column. The box takes in the vapour rising from the plate below, V, and the reflux flowing down from the condenser, L. It gives out the distillate product, D, and the vapour going up to the condenser is the same V that leaves the box at the top. In steady state, what enters the box must leave it, so V = L + D. This is the total balance over the top of the column.

Step 2: balance the volatile component. Let y be the mole fraction of the volatile part in the rising vapour, x be its mole fraction in the falling liquid, and x D be its mole fraction in the distillate. The volatile part entering the box is V y, and the volatile part leaving is L x plus D x D. In steady state these are equal: V y = L x + D x D.

Step 3: solve for y. Divide by V: y = (L / V) x + (D / V) x D. This is already a straight line on the x-y diagram, with slope L / V and intercept (D / V) x D.

Step 4: bring in the reflux ratio. The reflux ratio is R = L / D, the reflux returned for each unit of distillate taken. From the total balance, V = L + D = D (R + 1), so L / V = R / (R + 1) and D / V = 1 / (R + 1). Putting these in gives y = [R / (R + 1)] x + x D / (R + 1), the operating line of section 5.4. The derivation shows why the slope is always below 1, why the line passes through the point (x D, x D) on the diagonal, and why raising the reflux ratio pushes the slope toward 1 and the line toward the diagonal, giving a sharper separation at the cost of more heat.

Quick check

A column runs at a reflux ratio of 3. What is the slope of its rectifying operating line?

Check your answer

3 / (3 + 1) = 0.75. The slope is R / (R + 1), always below 1.

5.8 Real plates: Murphree efficiency and what makes a column misbehave

An ideal plate is a promise: the vapour and liquid leaving it are in balance. A real plate falls short, because the vapour and liquid meet for only a short time, and the contact is never perfect. The gap is measured by the Murphree plate efficiency, E = (y n minus y n-1) / (y* n minus y n-1), where y n is the vapour composition actually leaving plate n, y n-1 is the vapour entering it from below, and y* n is the vapour composition that would leave if the plate were ideal, in balance with the liquid on the plate. An efficiency of 0.7 means the plate does 70 percent of the separation an ideal plate would do. Typical values for food and spirit columns are 0.5 to 0.8, and the number of real plates is the number of ideal plates divided by the efficiency: 10 ideal plates at an efficiency of 0.7 need about 14 real plates. The efficiency is measured by running the column at total reflux, counting the separation it actually gives, and comparing it with the ideal count, which is the test section 5.5 mentions.

What lowers efficiency. Efficiency falls when the vapour and liquid do not mix well on the plate, when the liquid is foamy and carries over, when the vapour rate is too low for the liquid to be held up, and when the plate spacing is too tight for the two phases to part. Viscous liquids and foaming feeds are the usual causes in food columns, and the fix is a plate design with better contact, such as a sieve plate with smaller holes or a packed section, or a lower feed rate so the column runs inside its good range.

Flooding, weeping, and dumping. Three faults bound the working range of a column. Flooding is the top end: the vapour rises so fast it carries liquid upward, the pressure drop climbs, and the column stops separating. Weeping is the bottom end: the vapour rate is so low that liquid drips through the plate holes instead of staying on the plate, and the contact is lost. Dumping is worse weeping, where the liquid pours through the plate in a stream. A column must run between the weeping point and the flooding point, and the diameter is chosen so the normal vapour rate sits well inside that range, usually at 70 to 80 percent of the flooding velocity, leaving room for a surge in feed.

Priming and entrainment. Priming is liquid carried up the column as a spray, usually because the reboiler is boiling too hard, and it shows up as a high liquid level and a poor separation at the top. The fix is to ease the reboiler duty and to check that the feed is not flashing violently on entry. A column that is priming and a column that is flooding look alike from the outside, and the operator tells them apart by the pressure drop: flooding raises it sharply and steadily, while priming comes and goes with the boiling in the reboiler.

The t-x-y diagram: boiling and condensing curves for a binary mix

Top curve: dew point, the temperature at which vapour first condenses
Middle band: liquid and vapour live together, one boiling, one condensing
Bottom curve: bubble point, the temperature at which liquid first boils

The gap between the curves is the separation a single boil gives: wider gap, easier split.

Quick check

A column needs 10 ideal plates, and its measured Murphree efficiency is 0.5. About how many real plates are needed?

Check your answer

About 20. Real plates equal ideal plates divided by efficiency: 10 / 0.5 = 20.

Try it yourself

Mix 100 ml of water with 100 ml of a clear spirit of known strength, and note the smell. Heat the mix gently in a pot with a lid held at a slant, and catch the first few drops of condensate on a cold plate held above the pot. Compare the smell and the strength of the drops with the liquid left in the pot: the drops will be richer in alcohol, which is the single stage of distillation in your kitchen. Now think about what a column adds: the vapour would meet falling liquid many times, and each meeting would enrich it a little more.

Key ideas to carry forward

  • Distillation separates liquids by boiling, because the vapour is richer in the more volatile component than the liquid.
  • Raoult's law links vapour composition to liquid composition, and the t-x-y and x-y diagrams carry the whole separation story.
  • An azeotrope is a mixture where vapour and liquid have the same composition, and ordinary distillation cannot pass it.
  • A rectification column uses plates, a reboiler, a condenser, and reflux to repeat the separation many times in one vessel.
  • Operating lines and the q-line set the McCabe-Thiele stepping, and the count of steps is the number of ideal plates.
  • Minimum reflux needs infinite plates, total reflux needs the fewest plates, and real columns run at 1.2 to 1.5 times minimum reflux.
  • Column diameter is set by the vapour flow and the flooding velocity.

Glossary

Distillation

Separating a liquid mixture by boiling and condensing, so the vapour is richer in the volatile part. Example: alcohol is recovered from a fermented liquid by distillation.

Volatile

Tending to evaporate easily, with a high vapour pressure and a low boiling point. Example: ethanol is more volatile than water.

Raoult's law

The rule that a component's partial pressure in the vapour equals its liquid mole fraction times its pure vapour pressure. Example: Raoult's law predicts the vapour over an ethanol and water mix.

Bubble point

The temperature at which a liquid first starts to boil at a given pressure. Example: the bubble point curve is the lower curve on a t-x-y diagram.

Dew point

The temperature at which a vapour first starts to condense at a given pressure. Example: the dew point curve is the upper curve on a t-x-y diagram.

Azeotrope

A mixture whose vapour has the same composition as its liquid, so distillation cannot separate it further. Example: ethanol and water form an azeotrope near 95.6 percent ethanol.

Rectification

Fractional distillation in a column with plates or packing, where vapour and liquid meet many times. Example: a spirit is purified by rectification.

Reflux

Condensed top vapour returned to the column to wash the rising vapour. Example: raising the reflux improves purity but uses more heat.

Reflux ratio

Reflux flow divided by distillate flow. Example: a column running at a reflux ratio of 3 returns 3 kg of reflux for each kg of distillate.

Reboiler

A heater at the bottom of a column that boils liquid back up as vapour. Example: the reboiler supplies the heat for the whole separation.

Ideal plate

A plate where the vapour and liquid leaving it are in balance. Example: the McCabe-Thiele method counts ideal plates by stepping on the x-y diagram.

McCabe-Thiele method

A graphical method for finding the number of ideal plates by stepping between the equilibrium curve and the operating lines. Example: stepping from the distillate to the bottom composition counts the plates.

Lewis method

A plate by plate calculation method for finding the number of ideal plates. Example: the Lewis method gives the same plate count as McCabe-Thiele for the same separation.

Minimum reflux ratio

The lowest reflux ratio that can do a separation, needing infinite plates. Example: real columns run at 1.2 to 1.5 times the minimum reflux ratio.

Operating line

A straight line on the x-y diagram from the balance over one section of a column. Example: the rectifying operating line has slope R / (R + 1).

q-line

The feed line on the x-y diagram, set by the thermal condition of the feed. Example: a saturated liquid feed has a vertical q-line.

Flooding

A column fault where vapour rises so fast it carries liquid upward and the column stops separating. Example: a column diameter is chosen to stay well below the flooding velocity.

Steam distillation

Distillation where steam carries volatile compounds out at a temperature below their boiling point. Example: essential oils are won by steam distillation.

Murphree efficiency

The fraction of the ideal separation a real plate achieves, usually 0.5 to 0.8 in food columns. Example: 10 ideal plates at an efficiency of 0.7 need about 14 real plates.

Weeping

A column fault where the vapour rate is so low that liquid drips through the plate holes instead of staying on the plate. Example: a column run far below its design rate starts weeping and loses separation.

Priming

Liquid carried up a column as a spray because the reboiler is boiling too hard. Example: easing the reboiler duty stops priming and restores the separation.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.

Unit 6: Crystallization

Note

Unit 6: Crystallization

Sugar crystals grow in a pan, lactose crystals set the texture of ice cream, and salt crystals form as seawater evaporates. Crystallization turns a dissolved solid into a clean, ordered solid, and the size and shape of the crystals decide how the product feels, pours, and keeps. Why do some crystals grow large and even while others stay tiny and gritty, and why does a single seed crystal change everything? This unit answers those questions.

Questions to think about

  • Why does sugar sometimes form large clear crystals and sometimes a mass of tiny gritty ones?
  • What decides whether a new crystal starts on its own or grows on a crystal that is already there?
  • Why does ice cream turn gritty when it is stored badly?

6.1 Basic principles: solubility and supersaturation

Crystallization starts with a solution. The solubility of a solid is the most that will dissolve in a given amount of solvent at a given temperature, and it is usually shown on a solubility curve. Most food solids, sugar and salt among them, dissolve more as the temperature rises, so a hot solution can hold more solid than a cold one.

A solution that holds exactly its solubility limit is saturated. A solution that holds more than the limit is supersaturated, and it is unstable: the extra solid wants to come out, and it will, as crystals, once it gets a start. Supersaturation is the driving force of crystallization, and it is made in three ways: cooling a hot saturated solution, evaporating solvent from it, or adding a substance that lowers the solubility.

The ratio of the actual concentration to the solubility at that temperature is the supersaturation ratio, S. When S is just above 1, crystals grow slowly and evenly. When S is far above 1, new crystals burst into being all at once, and the product is a mass of tiny crystals. The art of crystallization is holding S in the right band: high enough to grow crystals at a useful rate, low enough to avoid a shower of new tiny ones.

Quick check

A sugar solution at 20 °C can hold 200 g of sugar per 100 g of water, and it holds 220 g. What is the supersaturation ratio, and what will happen?

Check your answer

S = 220 / 200 = 1.1. The solution is supersaturated, so the extra sugar will come out as crystals once nucleation starts.

6.2 The super-solubility curve and the metastable zone

Between the solubility curve and a second curve above it, the super-solubility curve, lies the metastable zone. Inside this zone, a solution is supersaturated, but new crystals do not start on their own: existing crystals will grow, but no new ones appear. Above the super-solubility curve lies the labile zone, where new crystals start spontaneously and in large numbers.

This gives the operator a clear rule. To grow large, even crystals, work inside the metastable zone: add seed crystals, and let them grow without letting the supersaturation climb into the labile zone. To make many tiny crystals, push into the labile zone, or stir hard, or cool fast. Sugar boilers watch the supersaturation closely for this reason: a moment of carelessness that pushes the batch into the labile zone fills the pan with fine grain that cannot be grown large again.

Cooling, evaporation, and seeding all move the solution through the diagram. Cooling a saturated solution moves it up through the metastable zone. Evaporating water moves it up too, because the concentration rises. Seeding drops the solution back into the metastable zone by giving the extra solid somewhere to go.

Solubility, metastable zone, and labile zone

Labile zone: new crystals start on their own, many tiny crystals
Metastable zone: existing crystals grow, no new ones start, the working zone
Below the solubility curve: unsaturated, crystals dissolve

Hold the batch in the metastable zone to grow large even crystals.

Quick check

Why is the metastable zone the right place to grow large sugar crystals?

Check your answer

Inside it, existing crystals grow but no new ones start, so the solid goes onto the crystals you already have instead of making new tiny ones.

6.3 Nucleation and growth rate

Nucleation is the birth of a new crystal. In primary nucleation, a crystal starts on its own from the solution, which needs a high supersaturation and a clean start, and it happens in the labile zone. In secondary nucleation, new crystals start on the surface of crystals that are already there, helped by stirring, by crystals knocking together, or by a seed crystal added on purpose. Secondary nucleation needs far less supersaturation, and it is the usual source of new crystals in a plant crystallizer. Seeding is secondary nucleation put to work: the operator adds a weighed amount of fine crystals of known size, and the batch grows from those, which gives control of the final crystal size.

Growth rate is how fast a crystal adds new layers, usually stated as the increase in a crystal dimension per unit time. It rises with supersaturation and with temperature, and it depends on how fast solute moves to the crystal surface and how fast it locks into the crystal lattice. Two steps work in series: diffusion of solute through the liquid to the surface, and the surface reaction that fits the molecule into the crystal. Whichever step is slower sets the growth rate. Stirring speeds the diffusion step, so gentle stirring grows crystals faster, while very hard stirring breaks crystals and makes secondary nuclei.

The factors affecting growth rate are supersaturation, temperature, agitation, impurities, and viscosity. Impurities can block growth on some crystal faces and change the crystal shape, which is why a trace of an impurity can turn a free flowing sugar into a sticky mass. High viscosity slows diffusion, so a thick syrup grows crystals slowly even at high supersaturation.

Quick check

Why does a sugar boiler add seed crystals instead of waiting for crystals to start on their own?

Check your answer

Seeding gives secondary nucleation at low supersaturation, so the batch grows from known crystals in the metastable zone instead of bursting into many tiny crystals in the labile zone.

6.4 Mass balance and energy balance

Mass balance. The solid in the feed must appear in the crystals, in the liquid left behind, the mother liquor, or both. For a feed F with solute fraction x F, crystals C, and mother liquor M with solute fraction x M:

F = C + M, and F x F = C x C + M x M

If the crystals are pure, x C = 1. A worked example: 1,000 kg of sugar solution at 70 percent sugar is cooled and evaporated until the mother liquor holds 65 percent sugar and 500 kg of pure crystals have formed. Then M = 1,000 - 500 = 500 kg, and the check is 1,000 x 0.70 = 700 kg of sugar in, against 500 x 1.0 + 500 x 0.65 = 825 kg, which does not close, so the numbers are inconsistent, and the operator knows a measurement is wrong. That is the value of the balance: it catches errors before they become a bad batch. With consistent numbers, the balance gives the crystal yield from the feed and the mother liquor compositions, which set the recovery of the plant.

Energy balance. Heat must be removed to cool a solution, or supplied to evaporate water from it, and crystallization itself releases heat, the heat of crystallization, as the solid forms its ordered lattice. The balance is: heat in with the feed plus heat of crystallization equals heat out with the products plus heat removed by cooling or taken up by evaporation, plus losses. In an evaporative crystallizer the largest term is the latent heat of the water evaporated, as in Unit 3. In a cooling crystallizer the largest term is the sensible heat removed to drop the temperature, plus the smaller heat of crystallization.

Quick check

Why does a mass balance that does not close tell the operator something is wrong?

Check your answer

Mass is conserved, so the solid in must equal the solid out in crystals plus mother liquor. A gap means a flow or a composition was measured wrong.

6.5 Crystallizers

A crystallizer makes supersaturation, holds the crystals while they grow, and lets the product be taken out at the right size. The main types are:

  • Batch cooling crystallizer. A stirred tank with a cooling jacket. The hot saturated solution is cooled slowly along a set curve, seed is added at the right point, and the crystals grow over hours. It gives good control and suits small batches and high value products.
  • Evaporative crystallizer. A vessel like an evaporator, where boiling off water makes the supersaturation. It suits solids whose solubility changes little with temperature, such as common salt, where cooling alone would give almost no crystals.
  • Vacuum crystallizer. Evaporation under vacuum, so the boiling is gentle and the temperature stays low. It suits heat sensitive products and gives even crystals.
  • Draft tube baffle crystallizer. A large vessel with an internal draft tube and a baffle that classifies crystals by size: small crystals are carried back for more growth, large ones are taken out as product. It gives a narrow size range and runs without stopping, which suits sugar and salt plants.
  • Scraped surface crystallizer. A cooled cylinder with a rotating scraper that keeps the wall clear. It suits viscous products that would foul a plain wall, and it is used for fats, margarine, and some confectionery masses.

The choice turns on the solubility curve, the heat sensitivity of the product, the size wanted, and whether the plant runs batch or continuous. In every type, the control variables are the same: the supersaturation, the seed, the temperature path, and the time the crystals spend growing.

Practical note. The practical work for this unit studies crystallization in the laboratory: make a supersaturated solution by cooling or by evaporation, seed it, watch crystals appear and grow, and measure the yield against the mass balance. Note the difference between a batch left still, which gives a few large crystals, and a batch stirred hard, which gives many small ones, and link that to nucleation and the metastable zone.

How a batch crystallizer is run

Hot saturated solution in, clear, no crystals

↓

Cool slowly into the metastable zone, add seed crystals

↓

Hold and stir gently: crystals grow, supersaturation stays in the zone

↓

Separate crystals from mother liquor, wash, dry

Quick check

Which crystallizer suits common salt, whose solubility barely changes with temperature: a cooling crystallizer or an evaporative crystallizer?

Check your answer

An evaporative crystallizer. Cooling alone makes almost no supersaturation for salt, so water must be boiled off to force crystals out.

6.6 Crystal size, shape, and product quality

The size and shape of crystals decide how a product looks, feels, and behaves. Large even sugar crystals pour freely and dissolve at a steady rate, while fine sugar dissolves fast and packs tight. Small lactose crystals in ice cream are smooth on the tongue, while large ones feel gritty and sandy, which is the classic fault of badly stored ice cream: small ice and lactose crystals grow larger during temperature swings, a process called Ostwald ripening, where small crystals dissolve and large ones grow at their expense. Salt crystals are judged by their shape too: cubic crystals pour well, while flaky crystals dissolve fast on the tongue.

The magma density, the mass of crystals per unit volume of the mixture in the crystallizer, is a key operating number. A high magma density means many crystals sharing the supersaturation, so each grows slowly and the size stays even. A low magma density means few crystals growing fast, which can give large crystals but also risks a burst of new nuclei if the supersaturation climbs. The operator controls magma density through the seed charge and the time the batch spends growing.

Crystal shape is set by which faces of the crystal grow fastest. Impurities can slow one face and speed another, which changes the shape from a cube to a needle or a plate. That matters for filtering and washing: needle crystals pack badly and trap mother liquor, while even cubes filter and wash cleanly. The same impurity that changes the shape can also slow the whole growth rate, which is why the purity of the feed is checked before a crystallizer is charged.

Storage matters as much as making. A crystalline product that takes up water from damp air will cake, because tiny bridges of dissolved solid form between crystals and then harden. A product that dries too far can lose its water of crystallisation and crumble to dust. Both faults are controlled by packing in a moisture proof container and storing at a steady temperature and humidity.

Quick check

Why does ice cream turn gritty when its storage temperature swings up and down?

Check your answer

Small crystals dissolve and large ones grow at their expense, a process called Ostwald ripening, so the crystals get larger and feel gritty on the tongue.

6.7 Crystal growth: the growth rate law and controlling size

Nucleation starts crystals, but growth decides their size, and growth follows its own law. The linear growth rate, G, the rate at which a crystal face moves outward, in m/s, is linked to the supersaturation by G = k g (S minus 1) to the power g, where S is the supersaturation ratio of section 6.1, k g is a growth rate constant for the substance and the conditions, and the exponent g is usually between 1 and 2. The form matters more than the numbers: growth rises steeply with supersaturation, so a small rise in supersaturation grows the crystals much faster, but the same rise also pushes the batch toward the labile zone, where new crystals start and steal the solid. The operator's job is to hold the supersaturation in the narrow band where growth is fast and nucleation is quiet, and that band is the metastable zone of section 6.2.

Seeding practice. Seed crystals are added at a size and a number that set the final size: the solid that comes out of solution is shared between the seeds, so more seeds mean smaller final crystals and fewer seeds mean larger ones. A sugar boiler seeds the pan with a weighed amount of fine crystals, then holds the supersaturation steady by feeding syrup at the rate the crystals consume it, so the crystals grow evenly without new ones starting. The seed size matters too: seeds that are too fine dissolve before they can grow, because small crystals have a higher solubility than large ones, which is the same Ostwald ripening effect of section 6.6 working against the operator.

Temperature and stirring. Growth speeds up with temperature, because the molecules move faster and attach to the crystal face more readily, but cooling is what creates the supersaturation in the first place, so the operator balances the two: cool enough to stay supersaturated, warm enough for the faces to grow. Stirring helps by bringing fresh solution to the crystal faces and sweeping away the depleted layer around each crystal, but stirring too hard breaks crystals and the fragments become new seeds, a secondary nucleation that fills the batch with fine grain. Gentle, even stirring is the rule, and a broken impeller or a sudden speed change shows up in the crystal size within the batch.

Quick check

A batch is growing well in the metastable zone, and the operator raises the supersaturation sharply to speed it up. What two things happen?

Check your answer

Growth speeds up, because the growth rate rises steeply with supersaturation, and new crystals start, because the batch is pushed into the labile zone, so the product turns fine grained.

6.8 What goes wrong: false grain, caking, and gritty ice cream

False grain. A sudden burst of tiny crystals in a sugar pan, called false grain, is the classic crystallizer fault. It happens when the supersaturation jumps into the labile zone, often because the feed was added too fast, the cooling was too sharp, or a cold spot on the wall dropped the temperature locally. The fine crystals cannot be grown large again, so the batch must be redissolved or accepted as a lower grade. The prevention is a slow, even feed, a steady cooling rate, and a wall temperature close to the bulk temperature, so no cold spot forms.

Caking. Dry crystals cake into lumps when the surface of the crystals picks up moisture, dissolves a little, and then dries again, cementing the crystals together. It is worst for small crystals, which have the most surface, and for storage in damp air or with temperature swings that condense water on the surface. The fixes are to dry the crystals to a low, even moisture, to cool them before packing so no moisture condenses inside the bag, and to store them in dry air. A free flowing agent, a fine powder that keeps the crystals apart, is used for some salts and sugars where the rules allow it.

Gritty ice cream and sandy texture. In ice cream the fault is lactose, which crystallises slowly during storage, especially when the temperature swings, as section 6.6 explains. The prevention is to keep lactose below its solubility limit in the mix, to freeze fast so the crystals stay small, and to hold the storage temperature steady and low, so Ostwald ripening has no warm spells to work in. A product that was smooth leaving the factory and turns gritty in a shop freezer has almost always been through temperature swings in storage or transport.

Sticking and wall crusts. Crystals grow on the cold wall of a crystallizer faster than in the bulk, forming a crust that insulates the wall and slows cooling. Scraped surface crystallizers exist for this reason: the scraper keeps the wall clear, so the heat keeps flowing and the crust never builds. In a batch pan the operator watches the wall and stops to clean it before the crust thickens, because a thick crust means a long batch and an uneven product.

Quick check

A bag of dry sugar cakes into a hard lump after storage in a damp room. What has happened at the crystal surfaces?

Check your answer

The surfaces picked up moisture, dissolved a little, and dried again, cementing the crystals together. Dry, cool, steady storage prevents it.

Try it yourself

Dissolve as much sugar as you can in a cup of hot water, then let it cool slowly with a string hanging in it. Add a few grains of sugar to the string as seed. Leave the cup still for a day, then look at the string: crystals will have grown on it. Now repeat with a second cup, but stir it hard and cool it fast in a cold water bath. Compare the crystals: the still cup gives fewer, larger crystals, and the stirred cup gives many tiny ones. You have seen nucleation, the metastable zone, and growth rate with kitchen tools.

Key ideas to carry forward

  • Supersaturation is the driving force of crystallization, made by cooling, evaporation, or adding a substance that lowers solubility.
  • The metastable zone is the working zone: existing crystals grow there and no new ones start.
  • Primary nucleation starts new crystals on its own in the labile zone, while secondary nucleation starts them on existing crystals at lower supersaturation.
  • Growth rate rises with supersaturation and temperature, and is set by diffusion to the surface and fitting into the lattice.
  • Mass balance gives the crystal yield, and energy balance covers cooling, evaporation, and the heat of crystallization.
  • Crystallizer choice turns on the solubility curve, heat sensitivity, wanted crystal size, and batch or continuous running.

Glossary

Crystallization

Forming a solid crystal from a solution, by cooling, evaporation, or changing the solvent. Example: sugar crystals grow from a supersaturated syrup.

Solubility

The most solid that will dissolve in a given amount of solvent at a given temperature. Example: sugar dissolves more in hot water than in cold water.

Supersaturation

A solution holding more dissolved solid than its solubility limit, unstable and ready to crystallise. Example: a cooled sugar solution is supersaturated.

Supersaturation ratio

Actual concentration divided by solubility at that temperature. Example: a ratio of 1.1 means the solution holds 10 percent more than its limit.

Metastable zone

The band between the solubility curve and the super-solubility curve, where crystals grow but no new ones start. Example: sugar boilers work inside the metastable zone.

Labile zone

The region above the super-solubility curve, where new crystals start spontaneously. Example: fast cooling pushes a solution into the labile zone and makes many tiny crystals.

Nucleation

The birth of a new crystal in a solution. Example: nucleation starts the crystallization of lactose in ice cream.

Primary nucleation

New crystals starting on their own from the solution, needing high supersaturation. Example: primary nucleation fills a pan with fine grain when cooling is too fast.

Secondary nucleation

New crystals starting on crystals already present, needing less supersaturation. Example: seed crystals cause secondary nucleation in a sugar pan.

Seed crystal

A small crystal added on purpose to start growth in the metastable zone. Example: a sugar boiler seeds the pan to control the final crystal size.

Growth rate

How fast a crystal adds new layers, usually as growth in size per unit time. Example: growth rate rises with supersaturation.

Mother liquor

The liquid left behind after crystals are removed. Example: mother liquor from a sugar crystallizer is reprocessed to recover more sugar.

Heat of crystallization

Heat released when a dissolved solid forms its ordered crystal lattice. Example: the heat of crystallization warms a batch slightly as crystals form.

Draft tube baffle crystallizer

A continuous crystallizer that classifies crystals by size and returns small ones for more growth. Example: sugar plants use draft tube baffle crystallizers for even grain.

Scraped surface crystallizer

A cooled cylinder with a rotating scraper that keeps the wall clear of solid. Example: margarine is crystallised in a scraped surface unit.

Crystal yield

The mass of crystals won from a feed, found from the mass balance. Example: the yield tells the plant how much sugar was recovered.

Growth rate

How fast a crystal face moves outward, in m/s, rising steeply with supersaturation. Example: holding supersaturation steady in the metastable zone keeps the growth rate even across a batch.

False grain

A sudden burst of tiny crystals when supersaturation jumps into the labile zone, which cannot be grown large again. Example: feeding syrup too fast into a sugar pan makes false grain.

Caking

Dry crystals cementing into lumps when their surfaces pick up moisture and dry again. Example: sugar stored in damp air cakes, while dry steady storage keeps it free flowing.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.

Unit 7: Separation Techniques

Note

Unit 7: Separation Techniques

Milk is split into cream and skim, juice is cleared of pulp, and a cloudy broth is filtered bright. Separation is the family of operations that divides a mixture into its parts without changing them chemically, using size, density, or a spinning force. Why does a filter slow down as it collects solids, why do fine particles settle so slowly, and why does a spinning bowl separate cream in seconds when gravity needs hours? This unit answers those questions.

Questions to think about

  • Why does a filter run fast when it is clean and slower and slower as it loads up?
  • Why does a fine dust settle almost never, while a coarse grain drops at once?
  • How can a spinning machine separate cream from milk when the density difference between them is so small?

7.1 Filtration: theory

Filtration separates solids from a liquid by passing the mixture through a porous medium, the filter cloth or membrane, that lets the liquid through and holds the solids back. The liquid that passes is the filtrate, and the solids that stay are the cake. The driving force is a pressure difference across the medium, made by gravity, by a pump, or by vacuum.

The flow of filtrate follows the same idea as flow through a porous bed: the rate is proportional to the pressure difference and the area, and inversely proportional to the viscosity of the liquid and the resistance. As the cake builds, its resistance grows, so the rate falls. The total resistance is the sum of the medium resistance and the cake resistance, and the cake resistance grows with the mass of cake and with its specific resistance, which is a property of the particles and how tightly they pack. A cake of fine, soft particles has a high specific resistance and filters slowly; a cake of coarse, hard particles has a low specific resistance and filters fast.

Two operating modes are used. In constant pressure filtration, the pressure is held steady and the flow rate falls as the cake builds. In constant rate filtration, the flow rate is held steady by raising the pressure as the cake builds, until the pressure limit is reached. The practical work for this unit studies both modes on a plate and frame filter press, so you can see the falling rate in one mode and the rising pressure in the other.

The classic filtration equation for constant pressure, in its integrated form, is:

t / V = (μ α c / (2 A² Δp)) V + (μ R m / (A Δp))

where t is time, V is filtrate volume, μ is viscosity, α is the specific cake resistance, c is the solids concentration, A is the area, Δp is the pressure difference, and R m is the medium resistance. A plot of t/V against V is a straight line, and its slope gives α while its intercept gives R m. That plot is the standard way to measure the filterability of a slurry in the laboratory.

Quick check

A filter cake doubles in thickness during a run at constant pressure. What happens to the filtration rate, and why?

Check your answer

It falls. The cake resistance grows with the mass of cake, so the total resistance rises and the rate falls at the same pressure difference.

7.2 Filtration equipment: plate and frame press and rotary vacuum filter

Plate and frame filter press. A row of plates and frames is pressed together in a frame, with filter cloth between them. Slurry is pumped into the frames, the liquid passes through the cloth and out of the plates, and the solids stay in the frames as a cake. When the frames are full, the press is opened, the cakes are dropped out, and the cycle starts again. It is a batch operation, it handles high solids well, it gives a dry cake, and it is common for juice clarification, syrup filtration, and edible oil refining. Its design turns on the area needed, found from the filtration equation for the time allowed per cycle.

Rotary vacuum filter. A large drum covered with filter cloth turns slowly in a trough of slurry, with vacuum inside the drum. As each part of the drum dips into the slurry, liquid is sucked through the cloth and the cake builds on the outside. As the drum turns out of the trough, the cake is washed, dried by the air drawn through it, and then scraped off by a knife. It runs without stopping, it suits slurries that filter easily, and it is used for starch, yeast, and some juice pulps. Its drawback is the thin cake and the cost of the vacuum system.

Plate and frame filter press: slurry in, clear filtrate out

Slurry pumped into the frames between plates

↓

Liquid passes through the cloth, solids stay as a cake in the frame

↓

Frames fill, press opens, cakes drop out, cycle starts again

A batch operation that gives a dry cake and a bright filtrate.

Quick check

Which filter suits a slurry that must run without stopping, and which suits a batch that must give a dry cake?

Check your answer

The rotary vacuum filter runs without stopping, while the plate and frame press is a batch unit that gives a dry cake.

7.3 Sedimentation: terminal velocity and drag

Sedimentation separates solids from a liquid by gravity: a particle heavier than the liquid sinks, and a particle lighter than the liquid rises. A particle falling through a liquid speeds up until the drag force on it balances its weight minus the buoyancy, and then it falls at a steady speed called the terminal velocity.

For a small spherical particle falling slowly, Stokes' law gives the drag, and the terminal velocity is:

v t = g d² (ρ p - ρ f) / (18 μ)

where d is the particle diameter, ρ p is the particle density, ρ f is the fluid density, and μ is the fluid viscosity. Three lessons sit in that equation. Terminal velocity grows with the square of the diameter, so a particle twice as wide falls four times as fast, which is why fine dust settles almost never. It grows with the density difference, so a particle barely heavier than the liquid falls slowly. And it falls as viscosity rises, so a thick syrup holds particles up far longer than water does.

The drag coefficient, C D, is the general way to state the drag on a particle at any speed: drag force = C D A (ρ f v² / 2), where A is the projected area of the particle. At low Reynolds numbers, C D = 24 / Re, which is Stokes' law in another form. At higher Reynolds numbers the drag coefficient levels off and must be read from a chart, and the terminal velocity is found by trial, guessing a velocity, working out Re and C D, and checking the balance until it closes.

Free settling is one particle falling alone in a large body of liquid. Hindered settling is many particles falling together in a thick slurry: each particle is slowed by the liquid pushed up by its neighbours, and the settling rate falls as the solids concentration rises. A thickener works in the hindered regime, and its design uses the settling rate measured at the working concentration, not the free settling rate of one particle.

Quick check

Two particles have the same density. One is twice the diameter of the other. How much faster does the larger one settle?

Check your answer

Four times as fast. Terminal velocity grows with the square of the diameter in Stokes' law.

7.4 Thickeners: working principle and design

A thickener is a large, shallow, circular tank that separates a slurry into a clear liquid overflow and a thickened underflow. Slurry enters at the centre and spreads outward slowly. Solids settle to the floor, a slowly turning rake scrapes them toward a central outlet, and clear liquid rises and leaves over a weir at the rim. The tank is sized by the settling rate: the area needed is the feed flow of solids divided by the settling flux, and the depth must hold the slurry for long enough for the solids to reach the floor. Thickeners are used for fruit pulp, starch slurries, and wastewater solids, where a large volume must be clarified gently and cheaply, with no filter cloth and no high pressure.

The design steps are: measure the settling curve of the slurry in a tall cylinder, read the settling velocity at the working concentration, work out the area from the solids throughput, and add depth for the thickening zone where the settled solids are compressed by their own weight. A thickener that is too small overflows cloudy liquid; one that is too large wastes floor space and capital.

Quick check

A thickener is fed more solids than its area can settle. What will the overflow look like, and why?

Check your answer

Cloudy. The solids loading is above the settling flux the area can handle, so solids are carried over the weir with the overflow.

7.5 Centrifugal separation

A centrifuge replaces gravity with a spinning force. A bowl spinning at high speed makes a force many hundred or thousand times gravity, so a particle that would take hours to settle under gravity separates in seconds. The separation factor is the centrifugal acceleration divided by gravity, ω² r / g, where ω is the angular velocity and r is the radius. A bowl of 0.2 m radius spinning at 6,000 revolutions per minute gives a factor near 8,000, so cream separates from milk almost at once.

The working principle is the same as sedimentation, with the spinning force in place of gravity. In a cream separator, milk enters a stack of conical discs inside the spinning bowl. The discs split the milk into thin layers, so a fat globule has only a short distance to travel to reach a disc surface, and it then slides up and out as cream, while the heavier skim milk moves outward and leaves by a separate outlet. The thin layers and the high force together give a fast, sharp separation. The same machine, with different outlets, clarifies milk by throwing out dirt and cells as sludge.

Food uses of centrifugal separation include cream separation, milk clarification, oil and juice clarification, starch washing, and yeast recovery. The separation efficiency is measured by comparing the fat left in the skim milk with the fat in the feed: a good separator leaves less than 0.05 percent fat in the skim.

Practical note. The practical work for this unit covers constant rate and constant pressure filtration on a plate and frame filter press, the determination of terminal velocity and drag coefficient during unhindered settling, and the study of a centrifugal cream separator with its separation efficiency. Run the filter press at one pressure and record volume against time, then plot t/V against V to find the specific cake resistance from the slope. Drop a single sphere through a tall column of liquid and time its steady fall to find its terminal velocity, then compare it with Stokes' law.

Cream separator: thin layers, strong spinning force

Whole milk in at the centre, split into thin layers by conical discs

↓

Light fat globules move inward and up, leaving as cream

↓

Heavy skim milk moves outward and leaves by its own outlet

The spinning force is thousands of times gravity, so separation takes seconds.

Quick check

Why does a cream separator use a stack of thin conical discs instead of one open bowl?

Check your answer

The discs split the milk into thin layers, so a fat globule travels only a short distance to a surface, and separation is fast and sharp.

7.6 Worked example: sizing a plate and frame filter press

A worked example shows how the filtration equation is used. A juice slurry is filtered at a constant pressure of 200 kPa through a press with 10 m² of area. In a laboratory test on the same slurry, a plot of t/V against V gave a straight line with a slope of 2.0 x 10⁶ s/m⁶ and an intercept of 500 s/m³. The volume filtered after time t follows from the line: t/V = 2.0 x 10⁶ V + 500. To filter 5 m³, t/V = 2.0 x 10⁶ x 5 + 500 = 10,000,500 s/m³, so t = 10,000,500 x 5 = 50,002,500 s, which is far too long, and that result tells the designer the laboratory numbers were taken on a small area and must be scaled by the area ratio before use. The correct scaling uses the filtration equation in full: the slope scales with 1/A² and the intercept with 1/A, so on 10 m² the time to filter 5 m³ is a small fraction of the laboratory time, and the press is sized by working the equation at the real area, the real pressure, and the cycle time allowed. The lesson is to scale the equation, not the raw times, and to check the answer against the plant's cycle: a press that needs 8 hours to filter a batch that must be done in 2 hours is the wrong size, whatever the laboratory test seemed to say.

The same care applies to sedimentation design. A settling test in a tall cylinder gives the settling velocity at the working concentration, and the thickener area is the solids flow divided by that velocity. A test on a dilute sample gives a faster velocity than the thickener will see, because hindered settling slows the real slurry, so the test must be run at the working concentration or the thickener will be undersized and its overflow will run cloudy.

Quick check

Why must a thickener settling test be run at the working solids concentration, not on a dilute sample?

Check your answer

Hindered settling slows a thick slurry, so a dilute test gives a faster settling velocity than the real thickener will see, and the thickener would be undersized.

7.7 Derivation: terminal velocity from a force balance (Stokes' law)

Section 7.3 states Stokes' law for terminal velocity. This section derives it from a force balance, one small step at a time.

Step 1: list the forces on a falling particle. Three forces act on a small sphere sinking in a liquid. Gravity pulls it down with the weight of the sphere, (π/6) d³ ρ p g, where d is the diameter, ρ p is the density of the particle, and g is gravity. Buoyancy pushes it up with the weight of the liquid it displaces, (π/6) d³ ρ f g, where ρ f is the density of the fluid. Drag pushes it up, resisting the motion, and for a small sphere moving slowly Stokes found the drag is 3 π μ d v, where μ is the viscosity and v is the velocity.

Step 2: the particle stops speeding up. When the particle is first dropped it speeds up, because gravity beats the other two forces. As it speeds up, the drag grows, because drag is proportional to velocity. Soon the three forces balance, the net force is zero, and the particle stops speeding up and falls at a steady speed. That steady speed is the terminal velocity.

Step 3: write the balance. At terminal velocity, weight equals buoyancy plus drag: (π/6) d³ ρ p g = (π/6) d³ ρ f g + 3 π μ d v t, where v t is the terminal velocity.

Step 4: solve for v t. Move the buoyancy term to the left: (π/6) d³ (ρ p minus ρ f) g = 3 π μ d v t. Divide both sides by 3 π μ d: v t = d² (ρ p minus ρ f) g / (18 μ). That is Stokes' law. The derivation shows what it assumes: a small rigid sphere, slow flow so the drag stays proportional to velocity, a still liquid, and a particle far from walls and other particles. It also shows why the diameter matters so much: the velocity grows with d², so a particle twice as wide falls four times as fast, which is the quick check in section 7.3, and why a particle slightly denser than the liquid falls slowly, while a much denser one falls fast.

Step 5: check the assumption. Stokes' law holds only while the particle Reynolds number, Re p = (ρ f v t d) / μ, stays below about 1. After working out v t, the engineer checks Re p: if it is above 1, the drag is no longer proportional to velocity, the drag coefficient of section 7.3 must be used instead, and the velocity is found by trial, as section 7.3 describes.

Quick check

Two spheres of the same density fall in the same liquid. One is 0.2 mm across, the other is 0.1 mm across. How much faster does the larger one fall?

Check your answer

Four times as fast. Terminal velocity grows with the square of the diameter, and (0.2 / 0.1)² = 4.

7.8 Centrifuges: g-force, the sigma factor, and a worked example

Gravity settling is slow, so plants spin the mixture. A centrifuge replaces g with a centrifugal acceleration, ω² r, where ω is the angular velocity in radians per second and r is the radius from the centre. The relative centrifugal force, RCF, is the ratio of that acceleration to gravity: RCF = ω² r / g, a number with no units that says how many times stronger than gravity the spin is.

Worked example. A disc stack cream separator spins at 6,000 revolutions per minute. The angular velocity is ω = 2 π x 6,000 / 60 = 628 rad/s. At a radius of 0.10 m the acceleration is ω² r = 628² x 0.10 = 39,438 m/s². Dividing by g = 9.81 m/s² gives an RCF of about 4,020. The fat globules feel a force four thousand times gravity, which is why a separation that takes hours in a settling tank takes seconds in the separator. The same calculation shows why speed matters so much: the force grows with the square of the speed, so doubling the speed quadruples the force, and a small rise in speed buys a large rise in separation.

The sigma factor. One centrifuge is not compared with another by speed alone, because the size and the flow matter too. The sigma factor, Σ, rolls the geometry of the machine into a single area, in m², that can be compared with the area of a settling tank doing the same job. A centrifuge with a large sigma can pass more flow at the same separation, and the design rule is simple: the flow rate divided by the sigma factor, Q / Σ, equals the settling velocity the machine can capture, so a particle with a smaller settling velocity needs a larger sigma or a lower flow. The maker publishes Σ for each machine, and the engineer uses it to size the unit and to compare two machines fairly.

Choosing gravity or spinning. Gravity settling costs almost nothing to run and suits large volumes of coarse solids, such as a clarifier holding juice pulp. Centrifugal separation costs power and a dearer machine, and it suits fine particles, emulsions, and cases where time matters, such as cream separation and yeast recovery. The rule is to settle by gravity where the particles are large and time is plentiful, and to spin where they are fine, the density difference is small, or the product cannot wait.

Quick check

A centrifuge spins twice as fast as before, at the same radius. What happens to the relative centrifugal force?

Check your answer

It quadruples. The force grows with the square of the angular velocity, so twice the speed gives four times the force.

Try it yourself

Stir a spoonful of fine flour into a tall glass of water and watch it settle. Time how long the cloud takes to clear halfway. Now stir the same amount into a glass of thick syrup thinned with a little water, and compare: the thicker liquid holds the particles up far longer, which is viscosity at work in Stokes' law. Then shake a bottle of unhomogenised milk and leave it in the fridge overnight: the cream will rise to the top by gravity, slowly. A cream separator does the same job in seconds by spinning.

Key ideas to carry forward

  • Filtration rate falls as the cake builds, because cake resistance grows with the mass of cake, and a plot of t/V against V gives the specific cake resistance.
  • The plate and frame press is a batch filter giving a dry cake, while the rotary vacuum filter runs without stopping.
  • Terminal velocity follows Stokes' law for small particles, growing with the square of the diameter and the density difference, and falling with viscosity.
  • Hindered settling is slower than free settling, and thickeners are sized from the settling rate at the working concentration.
  • A centrifuge replaces gravity with a force thousands of times stronger, and a disc stack cream separator uses thin layers for a fast, sharp split.

Glossary

Filtration

Separating solids from a liquid by passing the mix through a porous medium that holds the solids back. Example: juice is filtered bright through a filter cloth.

Filtrate

The liquid that passes through a filter. Example: clear juice leaving a filter press is the filtrate.

Filter cake

The layer of solids held back on a filter medium. Example: the cake in a plate and frame press is dropped out when the frames are full.

Specific cake resistance

A property of the cake, α, that says how much it resists flow per unit mass. Example: a fine soft cake has a high specific cake resistance and filters slowly.

Constant pressure filtration

Filtration with the pressure held steady, so the flow rate falls as the cake builds. Example: a gravity filter runs at constant pressure.

Constant rate filtration

Filtration with the flow rate held steady, so the pressure rises as the cake builds. Example: a pump driven filter can be run at constant rate until its pressure limit.

Plate and frame filter press

A batch filter of plates and frames pressed together with cloth between them. Example: syrup is filtered in a plate and frame press.

Rotary vacuum filter

A turning drum filter that works without stopping, with vacuum inside the drum. Example: starch slurry is filtered on a rotary vacuum filter.

Sedimentation

Separating solids from a liquid by letting them sink, or rise, under gravity. Example: pulp settles to the bottom of a juice tank.

Terminal velocity

The steady speed a particle reaches when drag balances its weight minus buoyancy. Example: a grain of sand reaches its terminal velocity soon after it starts falling in water.

Stokes' law

The law giving drag and terminal velocity for a small slow particle. Example: Stokes' law predicts that a particle twice as wide falls four times as fast.

Drag coefficient

A number, C D, that states the drag on a particle at any speed. Example: at low Reynolds numbers C D equals 24 / Re.

Free settling

One particle falling alone in a large body of liquid. Example: a single sphere timed in a tall column is in free settling.

Hindered settling

Many particles falling together in a thick slurry, each slowed by its neighbours. Example: a thickener works in the hindered settling regime.

Thickener

A large shallow tank that settles solids to the floor and overflows clear liquid. Example: fruit pulp slurry is thickened in a circular thickener.

Centrifugal separation

Separating by spinning, where the spinning force replaces gravity. Example: cream is separated from milk in a centrifuge.

Separation factor

Centrifugal acceleration divided by gravity, ω² r / g. Example: a cream separator runs at a separation factor near 8,000.

Separation efficiency

How completely a separator removes the wanted part, often as fat left in skim milk. Example: a good separator leaves less than 0.05 percent fat in the skim.

Relative centrifugal force (RCF)

The centrifugal acceleration in a centrifuge divided by gravity, a number with no units saying how many times stronger than gravity the spin is. Example: a cream separator at 6,000 revolutions per minute reaches an RCF near 4,000.

Sigma factor

A single area, in m², that sums up the geometry of a centrifuge, used to size it and to compare machines. Example: the flow divided by the sigma factor equals the settling velocity the machine can capture.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.
  • Ravi Kumar, S. Dairy Engineering. AgriMoon course material.

Unit 8: Size Reduction

Note

Unit 8: Size Reduction

Wheat becomes flour, spices become powder, and a coarse sugar becomes a fine dusting sugar. Size reduction breaks large pieces into smaller ones, and the size of the pieces decides how a food mixes, dissolves, bakes, and feels in the mouth. Why does grinding take so much energy, why do different mills suit different foods, and how do you describe a pile of powder that holds a thousand different sizes? This unit answers those questions.

Questions to think about

  • Why does grinding a tonne of grain use far more energy than simply cutting it in half?
  • How can you describe the size of a powder when no two particles are the same size?
  • Why does a hammer mill suit grain while a roller mill suits the same grain for a different product?

8.1 Why size is reduced, and the energy laws

Size reduction, or comminution, is done to make a food easier to mix, to make it dissolve or cook faster, to release oil or juice from cells, and to give a wanted texture. Flour must be fine to bake well, spice must be fine to spread its flavour, and sugar must be fine to dissolve fast. The cost is energy: grinding is one of the hungriest operations in a food plant, and much of the energy ends as heat, which can damage the product.

Three laws link the energy used to the change in size, and each fits a different range.

Rittinger's law. The energy used is proportional to the new surface area created. For fine grinding, where new surface dominates, E = K R (1/d 2 - 1/d 1), where d 1 and d 2 are the starting and final particle sizes and K R is a constant for the material and the mill. Halving the size of a fine powder takes far more energy than halving the size of a coarse grain, because the new surface grows fast as the pieces get small.

Kick's law. The energy used is proportional to the volume reduced, or to the logarithm of the size ratio: E = K K ln(d 1 / d 2). It fits coarse crushing, where the same energy halves the size whatever the starting size.

Bond's law. The energy used is proportional to the change in the square root of size: E = K B (1/√d 2 - 1/√d 1). Bond's work index, W i, is the energy needed to reduce a material from a very large size to 100 microns, and it is the standard way to compare how hard materials are to grind. Bond's law sits between Rittinger and Kick, and it is the most used of the three for sizing mills.

No law is exact for every food. They are guides for comparing mills, for scaling up from a laboratory test, and for estimating the power a mill will draw. The practical lesson is simple: fine grinding is costly, so grind only as fine as the product needs.

Quick check

Which law says the energy is proportional to the new surface created, and which range does it fit?

Check your answer

Rittinger's law. It fits fine grinding, where the new surface created dominates the energy use.

8.2 Equipment for size reduction

The choice of mill turns on the food: hard or soft, brittle or tough, oily or dry, and the size wanted.

  • Hammer mill. A rotor spins hammers at high speed inside a chamber, and the hammers strike the feed and throw it against a screen. Particles leave through the screen when they are small enough. It is fast, it handles grain, spices, and dried foods, and the screen sets the top size. Its drawbacks are heat from the impacts, dust, and wear on the hammers.
  • Disk mill, or pin mill. Two discs, one fixed and one spinning, carry pins or grooves that shear the feed between them. It suits fine grinding of spices and dry foods, and it gives a narrow size range.
  • Roller mill. Pairs of rolls turn against each other with a set gap, and the feed is crushed and sheared in the gap. It is the classic flour mill: successive breaks with sieving between them, each roll pair set finer than the last. It gives a controlled size, uses less energy than impact mills, and runs cool, which protects the flour.
  • Ball mill. A turning drum holds steel balls that tumble and crush the feed. It gives very fine powders and suits chocolate and some spice grinding, but it is slow and uses much energy.
  • Colloid mill. A rotor spins close to a stator with a tiny gap, and a liquid or paste is sheared in the gap. It does not grind dry solids; it makes fine dispersions, such as peanut butter, sauces, and emulsions, by breaking droplets and particles in a liquid.

A mill is chosen by the feed, the product size, the heat sensitivity, and the capacity. A hammer mill for coarse fast work, a roller mill for flour, a disk mill for fine spice, a ball mill for very fine work, and a colloid mill for pastes and liquids.

Which mill fits the job

Coarse and fast, grain and dried foods: hammer mill
Controlled flour, cool running: roller mill
Fine spice powder: disk mill
Very fine powder, slow: ball mill
Pastes and liquids, fine dispersion: colloid mill

Quick check

Why does a flour mill use roller mills rather than hammer mills for its main breaks?

Check your answer

Roller mills give a controlled size, use less energy, and run cool, which protects the flour, while hammer mills are hotter and less controlled.

8.3 Sieve analysis and particle size distribution

A ground product is never one size. It is a spread of sizes, and sieve analysis measures that spread. A stack of sieves with smaller and smaller openings is shaken, the sample is split between the sieves, and each fraction is weighed. The result is a table of mass against size range, and from it come the cumulative curves: the fraction finer than a size, and the fraction coarser than a size.

Standard sieves are made in a series where each opening is a fixed ratio of the one above it, often the square root of 2, so the sizes step evenly on a log scale. The opening is stated in microns or in mesh number, the number of openings per inch.

From the distribution come several single numbers that describe the powder:

  • Mean particle size. The mass average size, worked out from the sieve fractions.
  • Cut-off diameter. The size at which a separator or a sieve divides the stream: particles larger than it go one way, smaller ones go the other.
  • Fineness modulus. A single number that summarises the coarseness of a powder, found by adding the cumulative percentages held on a set of standard sieves and dividing by 100. A higher fineness modulus means a coarser powder.
  • Uniformity index. A number that says how narrow the size spread is. A powder where most particles sit in a narrow band has a high uniformity index, and it flows and packs more evenly than a wide spread powder.

A wide spread packs tighter, because small particles fill the gaps between large ones, while a narrow spread flows more freely. A baker wants a known fineness and a narrow spread, so every batch of flour bakes the same way.

A worked example shows how the numbers are found. A 100 g flour sample is shaken through a sieve stack, and the masses held are: 5 g on the 500 micron sieve, 25 g on the 250 micron sieve, 40 g on the 125 micron sieve, 20 g on the 63 micron sieve, and 10 g in the pan. The cumulative percentage finer than 250 microns is the mass that passed that sieve, 40 + 20 + 10 = 70 g, or 70 percent. The cumulative percentage finer than 125 microns is 30 percent. The median size, where 50 percent is finer, sits between 125 and 250 microns, near 180 microns by interpolation. The fineness modulus is found by adding the cumulative percentages held on the standard sieves and dividing by 100: the cumulative held on 500 microns is 5 percent, on 250 microns is 30 percent, on 125 microns is 70 percent, and on 63 microns is 90 percent, giving a sum of 195, so the fineness modulus is 1.95. A second flour with a fineness modulus of 2.60 is coarser, and the miller knows at once which batch is which without reading the whole table.

Quick check

A 100 g sample holds 70 g finer than 250 microns. What percentage is finer than 250 microns, and what does the median size mean?

Check your answer

70 percent. The median size is the size where 50 percent of the mass is finer and 50 percent is coarser.

Quick check

A powder has most of its mass on two adjacent sieves. A second powder spreads its mass over eight sieves. Which has the higher uniformity index?

Check your answer

The first powder. Its size spread is narrow, so its uniformity index is higher.

8.4 Worked example: sizing a mill from Bond's law

Bond's law in its working form for a mill is:

E = 10 W i (1/√P - 1/√F)

where E is the energy in kWh per tonne, W i is the work index, P is the product size at which 80 percent passes, in microns, and F is the feed size at which 80 percent passes, in microns. A worked example: a spice with a work index of 12 kWh per tonne is ground from a feed where 80 percent passes 2,000 microns to a product where 80 percent passes 200 microns. Then E = 10 x 12 x (1/√200 - 1/√2,000) = 120 x (0.0707 - 0.0224) = 120 x 0.0483 = 5.8 kWh per tonne. For a mill taking 5 tonnes per hour, the power is about 29 kW, before allowing for the efficiency of the motor and drive. That number sizes the motor and checks the running cost, and it shows why grinding finer than needed is money lost as heat.

Practical note. The practical work for this unit studies particle size distribution from a pin mill, a disk mill, a hammer mill, and a roller mill: take a sample from each, run it through a sieve stack, weigh each fraction, and plot the cumulative distribution. Compare the fineness modulus and the uniformity index of the four products, and match the mill to the product: the roller mill flour will be the most even, the hammer mill product the widest spread, and the disk mill the finest.

Quick check

A mill must grind a feed twice as fine as before, from a product size of 200 microns to 100 microns, with the same feed size of 2,000 microns and the same work index of 12 kWh per tonne. Does the energy per tonne double, stay the same, or rise by less than double?

Check your answer

It rises by less than double. The new energy is E = 10 x 12 x (1/√100 - 1/√2,000) = 120 x (0.1000 - 0.0224) = 9.3 kWh per tonne, against 5.8 kWh per tonne before, a rise of about 1.6 times, not a doubling.

8.5 Energy, heat, and dust: running a mill safely

Grinding uses far more energy than the new surface it makes would need in theory, because most of the energy ends as heat in the mill and the product. A hammer mill can raise the product temperature by 10 to 20 K in one pass, and that heat can damage flavour, drive off aroma, and, in the case of fat rich seeds, smear oil across the mill and choke it. Mills are therefore run with air flowing through them to carry heat away, and heat sensitive products are ground in cooled mills or in short passes with cooling between them. Cryogenic grinding, where the feed is cooled with liquid nitrogen, N₂, before milling, keeps spices and oily seeds brittle and protects their aroma, at the cost of the nitrogen.

Dust is the second running problem. Fine food dust in air can catch fire and explode, and flour, starch, and sugar dust are all known hazards. A mill must therefore be vented, its dust collected in a cyclone or a bag filter, and its electrical equipment rated for a dusty place. The dust collector also recovers product: the fine fraction that would otherwise be lost is worth money, and it is also the fraction that sets the fineness of the product, so losing it changes the size distribution.

Wear is the third running cost. Hammers, screens, rolls, and balls wear as they work, and worn parts change the product: a worn screen lets coarse pieces through, a worn roll gives an uneven grind, and a worn hammer draws more power for less work. The mill is checked on a schedule, and the sieve analysis from section 8.3 is the test that shows wear before it becomes a quality complaint.

Quick check

Why is air drawn through a hammer mill during grinding, and what else does the air do?

Check your answer

It carries away the heat of grinding, which protects flavour and stops oily feeds smearing. It also carries the fine dust to a collector, which recovers product and controls the explosion risk.

8.6 Derivation: one law behind Rittinger, Kick, and Bond

Section 8.1 states three energy laws for size reduction. This section shows they are three cases of one law, and derives each from it.

Step 1: start with the differential form. Suppose the energy needed to shrink a particle by a small amount dx is dE = minus K dx / x to the power n, where x is the particle size, K is a constant for the material and the machine, and n is a number that says how the energy depends on size. The minus sign says the size falls as energy is spent. The three laws are three choices of n.

Step 2: Rittinger's law, n = 2. Put n = 2 and integrate from the feed size x₁ to the product size x₂: E = K R (1/x₂ minus 1/x₁). The energy is proportional to the new surface created, because surface area per unit mass grows as 1/x. Rittinger's law fits fine grinding, where most of the energy goes into making new surface.

Step 3: Kick's law, n = 1. Put n = 1 and integrate: E = K K ln(x₁ / x₂). The energy is proportional to the logarithm of the size ratio, which means the same energy halves the size whatever the starting size. Kick's law fits coarse crushing, where the energy goes into breaking the particle through its volume rather than into new surface.

Step 4: Bond's law, n = 1.5. Put n = 1.5 and integrate: E = 2 K B (1/√x₂ minus 1/√x₁), which is Bond's law in the form section 8.1 uses, with the constant written through the work index W i. Bond's law sits between the other two, and it fits the middle range of grinding, which is why section 8.4 uses it to size a mill.

Step 5: choose the law by the size range. The derivation gives a practical rule: use Kick's law for coarse crushing, above about 1 mm, Bond's law for intermediate grinding, from about 1 mm down to about 50 microns, and Rittinger's law for fine grinding, below about 50 microns. Using Rittinger's law for a coarse crusher overstates the energy, and using Kick's law for a fine mill understates it, which is how motors get sized wrong. The work index of section 8.1 is the anchor for Bond's law: it is measured in a standard laboratory mill, and it lets the engineer compare materials and scale the laboratory result to the plant mill.

A sieve stack: the sample passes down, coarse to fine

Top sieve, 500 microns: holds the coarsest 5 g

↓

250 micron sieve: holds 25 g, the rest passes on

↓

125 micron sieve: holds 40 g, the largest fraction

↓

Pan: catches the finest 10 g that pass every sieve

Each fraction is weighed, and the cumulative curve is built from the weights.

Quick check

A mill grinds a feed from 2 mm to 0.2 mm, then a second mill grinds the same feed from 0.2 mm to 0.02 mm. Which step needs more energy per tonne, and which law fits each step?

Check your answer

The second step needs far more, because energy rises steeply as size falls. Kick's law fits the coarse first step, and Rittinger's law fits the fine second step.

Try it yourself

Take a handful of dry rice and crush half of it in a mortar, then shake both the whole rice and the crushed rice through a kitchen sieve or colander. Weigh what passes and what stays. The crushed batch has a wider spread of sizes, and the fine dust passes at once while the coarse pieces stay. Now stir each batch into a glass of water: the fine powder clouds the water and settles slowly, while the whole grains sink at once. You have seen size distribution, settling, and the link between size and behaviour in one test.

Key ideas to carry forward

  • Size reduction is done to improve mixing, dissolving, cooking, and texture, and fine grinding is costly in energy.
  • Rittinger's law fits fine grinding, Kick's law fits coarse crushing, and Bond's law, with its work index, is the most used for sizing mills.
  • Hammer, disk, roller, ball, and colloid mills each suit a different feed and a different product size.
  • Sieve analysis measures the particle size distribution, and the fineness modulus and uniformity index summarise it in single numbers.
  • Cut-off diameter is the size at which a separator divides a stream into coarse and fine parts.

Glossary

Size reduction

Breaking large pieces into smaller ones, also called comminution. Example: wheat is size reduced to flour.

Comminution

Another name for size reduction. Example: spice comminution turns whole peppercorns into powder.

Rittinger's law

The law that grinding energy is proportional to the new surface area created. Example: Rittinger's law fits fine grinding.

Kick's law

The law that grinding energy is proportional to the logarithm of the size ratio. Example: Kick's law fits coarse crushing.

Bond's law

The law that grinding energy is proportional to the change in the square root of size. Example: Bond's law sizes mills from the work index.

Work index

Bond's measure of how hard a material is to grind, in kWh per tonne. Example: a high work index means a hard, costly material to grind.

Hammer mill

A mill where spinning hammers strike the feed and throw it against a screen. Example: grain is ground fast in a hammer mill.

Disk mill

A mill where the feed is sheared between a fixed and a spinning disk. Example: spices are finely ground in a disk mill.

Roller mill

A mill where the feed is crushed between pairs of rolls with a set gap. Example: flour milling uses successive roller mills.

Ball mill

A turning drum holding steel balls that tumble and crush the feed. Example: chocolate is ground very fine in a ball mill.

Colloid mill

A mill that shears a liquid or paste in a tiny gap between a rotor and a stator. Example: peanut butter is made smooth in a colloid mill.

Sieve analysis

Measuring particle size distribution by shaking a sample through a stack of sieves and weighing each fraction. Example: sieve analysis gives the size spread of a flour.

Particle size distribution

The spread of sizes in a powder, shown as mass against size range. Example: a flour's distribution decides how it bakes.

Cut-off diameter

The size at which a separator divides a stream into coarse and fine parts. Example: a sieve's cut-off diameter is its opening size.

Fineness modulus

A single number summarising the coarseness of a powder from sieve data. Example: a higher fineness modulus means a coarser powder.

Uniformity index

A number saying how narrow the size spread of a powder is. Example: a high uniformity index means most particles sit in a narrow band.

Mesh number

The number of sieve openings per inch, a way to state sieve size. Example: a higher mesh number means a finer sieve.

Differential grinding law

The single equation, dE = minus K dx / x to the power n, from which Rittinger's, Kick's, and Bond's laws all follow by choosing n. Example: n = 2 gives Rittinger's law, n = 1 gives Kick's law, n = 1.5 gives Bond's law.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.

Unit 9: Mixing

Note

Unit 9: Mixing

A cake batter must hold its air, a spice blend must taste the same in every spoonful, and a sauce must stay smooth and not split. Mixing brings materials together until they are even, and it is one of the most common operations in a food plant, yet it is also one of the hardest to judge: when is a mix mixed enough, and when is it mixed too much? This unit answers those questions.

Questions to think about

  • How do you know when a spice blend is truly even, and not just even on the surface?
  • Why does over mixing a cake batter make it tough, when mixing is meant to help?
  • Why does the same mixer work well for a thin soup and badly for a thick dough?

9.1 Homogeneity and the mechanism of mixing

A mixture is homogeneous when every small sample taken from it has the same composition as the whole, within an agreed tolerance. A spice blend is homogeneous when a spoonful from the top, the middle, and the bottom all hold the same share of each spice. A batter is homogeneous when the flour, sugar, and liquid are spread evenly through it. Homogeneity is checked by taking samples from several points and measuring a marker, such as the salt content or the colour, and looking at the spread of the results. As mixing goes on, the spread shrinks, and the mix is judged done when the spread falls inside the limit the product needs.

Mixing works by three mechanisms, and a real mixer uses all three at once.

  • Convection. Bulk movement carries groups of particles from one place to another. A paddle turning in a bowl moves whole regions of the mix, and this is the main mechanism in the early stages, when the parts are still far apart.
  • Diffusion. Single particles wander across boundaries between regions, by random movement and by small local flows. Diffusion is slow, but it finishes the job at the smallest scale, where convection cannot reach.
  • Shear. Layers of the mix slide past each other, stretching and folding the material so that the distance between unlike parts shrinks. Shear is the main mechanism in pastes and doughs, where the material is too thick for convection.

A mixer that only stirs in circles will leave dead zones where the mix never moves. Good mixers move material in three directions, fold it over, and scrape the wall, so no region is left still.

Quick check

Why does a mixer that only turns in circles leave parts of the batch unmixed?

Check your answer

Circular stirring moves material around but does not fold it or scrape the wall, so dead zones stay still and never join the mix.

9.2 Rate of mixing and energy consumption

Mixing is not instant, and the rate at which a mix becomes even follows a pattern: fast at first, when the parts are far apart and convection does the work, then slower and slower as the remaining unevenness shrinks and diffusion must finish it. The degree of mixing is often tracked by the variance of the composition in samples, which falls toward a limit set by the random arrangement of particles. A perfect mix is never reached in practice: the aim is a spread inside the tolerance the product needs.

The power a mixer draws depends on the impeller, its speed, the size of the vessel, and the properties of the mix. For a stirred tank, the power number, N P = P / (ρ N³ D⁵), is a number with no units that links power P to density ρ, impeller speed N, and impeller diameter D. In laminar flow the power number falls as the Reynolds number rises, and in turbulent flow it levels off to a constant for a given impeller shape. The practical result is that power rises with the cube of the speed in turbulent flow, so doubling the speed needs about eight times the power, and with the fifth power of the diameter, so a larger impeller is far hungrier than a small one.

The energy used per batch is power times time, and the design balance is between a short, hard mix that uses much power and risks damage, and a long, gentle mix that uses less power at each moment but runs longer. Over mixing costs money and can damage the product: a batter loses its air, a dough gets tough, and an emulsion can break.

Quick check

A mixer runs in turbulent flow. If the impeller speed doubles, about how much does the power rise?

Check your answer

About eight times. Power rises with the cube of the speed in turbulent flow.

9.3 Liquid-liquid, solid-liquid, and solid-solid mixing

Liquid-liquid mixing. Two liquids that mix easily, such as water and a syrup, are blended by an impeller that moves them together and breaks one into streaks inside the other. Two liquids that do not mix, such as oil and water, need more: the impeller must break one liquid into small drops inside the other, and an emulsifier must be present to stop the drops joining again. A dressing and a mayonnaise are both liquid-liquid mixes, and the drop size sets their texture and their stability.

Solid-liquid mixing. A solid is stirred into a liquid to dissolve it, to suspend it, or to disperse it. Sugar dissolving in water needs the solid carried to fresh liquid so the layer of saturated solution around each crystal is swept away. A powder that floats or clumps, such as cocoa in milk, needs a mixer that pulls the powder down under the surface and wets it, or it will sit on top in dry lumps. Keeping a solid suspended, such as pulp in juice, needs enough agitation to stop settling, but not so much that air is beaten in.

Solid-solid mixing. Dry powders are mixed in tumbling drums, ribbon blenders, or paddle mixers. The particles must move past each other, and the mix is judged by the spread of samples. Two risks dominate: segregation, where particles of different size or density separate again after mixing, and over mixing, where a blend that was even starts to separate because the mixer keeps sorting the particles. A spice blend with a fine powder and a coarse grain will segregate if it is vibrated or poured from a height, so the handling after mixing matters as much as the mixing itself.

Three kinds of mixing, three different jobs

Liquid-liquid: blend streams, break one into drops in the other
Solid-liquid: dissolve, suspend, or wet a powder into a liquid
Solid-solid: tumble powders together, then handle gently so they do not separate again

The same word, mixing, covers three different jobs.

Quick check

A spice blend is even when it leaves the mixer, but after it is poured from a height into a bin it is uneven again. What happened?

Check your answer

Segregation. Particles of different size or density separated as they fell, so the blend must be handled gently after mixing.

9.4 Homogenization

Homogenization is a special kind of mixing: a liquid is forced through a narrow valve under high pressure so that fat globules and droplets are broken into very small, even drops. Milk is the classic case. Raw milk holds fat globules up to 10 microns wide, and they rise and form cream. Homogenised milk holds globules near 1 micron wide, and they stay spread through the milk because their rise is so slow and they are kept apart by the protein layer that forms on their new surface.

A homogenizer has a high pressure pump and a homogenizing valve. The pump drives the liquid to 10 to 25 MPa, and the valve forces it through a gap a fraction of a millimetre wide. In that gap the liquid speeds up, the pressure falls sharply, and the fat globules are torn apart by shear, by turbulence, and by cavitation, the collapse of tiny vapour bubbles. A second stage at a lower pressure breaks up any clusters that formed in the first stage.

The result is a product that does not cream, that feels smoother, that looks whiter because the smaller globules scatter light, and that digests more easily. The same machine makes sauces, dressings, and juice drinks smooth and stable. The power used is high, and the valve wears, but for milk and many liquid foods there is no simpler way to get a stable, even product.

Quick check

Why does homogenised milk not form a cream layer, while raw milk does?

Check your answer

Homogenization breaks the fat globules to near 1 micron, so they rise so slowly and stay so well spread that no cream layer forms.

9.5 Mixer types in food plants

  • Propeller and turbine agitators. A spinning impeller in a tank, suited to thin liquids. Baffles on the tank wall stop the liquid turning as a solid body and force it to mix.
  • Paddle and anchor agitators. Slow, wide paddles that scrape the wall, suited to thick liquids and pastes where a fast impeller would only cut a hole in the middle.
  • Ribbon blender. A horizontal trough with a helical ribbon that moves powder along the trough and back, suited to dry powders and spice blends.
  • Planetary mixer. A beater that turns on its own axis while circling the bowl, suited to batters, doughs, and creams in batch work.
  • Static mixer. A tube with fixed elements that split and fold the stream as it flows through, with no moving parts. It suits liquids that are already flowing in a line and need a final blend.
  • High shear mixer. A rotor spinning close to a stator with a tiny gap, suited to making emulsions and dispersions with very small drops.

The choice turns on the viscosity, the batch size, whether the mix is a liquid, a paste, or a powder, and whether the product can stand shear. A thin soup takes a turbine, a thick sauce takes an anchor, a spice blend takes a ribbon, and a mayonnaise takes a high shear mixer. The wrong choice shows up fast: a turbine in a thick paste cuts a hole in the middle and leaves the rest still, while an anchor in a thin liquid turns too slowly to blend it in a reasonable time, and a high shear mixer in a delicate cream breaks the structure it was meant to build. The mixer is therefore chosen from the product, not from the catalogue, and the choice is tested on a small batch before the plant commits to it.

Quick check

What happens when a fast turbine impeller is used in a very thick paste?

Check your answer

It cuts a hole in the middle and leaves the rest still, because the paste is too thick for the impeller to move it. A slow anchor or paddle suits a thick paste.

9.6 Mixing time, scale-up, and checking the mix

Mixing time is the time a mixer needs to reach the wanted homogeneity, and it is found by test, not by guess. Samples are taken at set times from several points in the vessel, a marker such as salt content is measured, and the spread of the results is plotted against time. The spread falls steeply at first and then levels off, and the mixing time is the point where the spread first enters the tolerance band and stays there. Running longer than that wastes power and can damage the product, and running shorter leaves the batch uneven.

Scale-up is the step from a laboratory mixer to a plant mixer, and it is where most mixing mistakes are made. The rules that work at one size do not carry over simply, because power, speed, and time scale differently. Two common scale-up rules are constant power per unit volume, which keeps the intensity of mixing the same, and constant impeller tip speed, which keeps the shear at the blade the same. A plant mixer scaled by power per volume may shear a delicate product too hard, while one scaled by tip speed may mix too slowly. The safe route is to scale by the rule that fits the product, and then to test the full size batch by sampling, as in the laboratory test, before the batch is released.

Checking the mix in a plant uses the same sampling idea. A thief sampler pulls small samples from several depths and positions, the marker is measured, and the spread is compared with the limit. A single sample from the top proves nothing: a mix can be even on the surface and uneven at the bottom, especially when the parts differ in density. The number of samples matters too: too few samples can hide a bad mix, and the standard is to sample enough points to cover the whole vessel.

Quick check

Why is a single sample from the top of a mixer not enough to judge a batch?

Check your answer

A mix can be even on the surface and uneven at the bottom, especially when the parts differ in density, so samples must come from several depths and positions.

9.7 Worked example: sizing a mixer from the power number

Section 9.2 states the power number, N P = P / (ρ N³ D⁵). This section uses it to size a mixer, with numbers, and to work out the mixing time.

The job. A stirred tank holds 500 litres of a sauce with a density of 1,050 kg/m³ and a viscosity low enough that the flow is turbulent. The impeller is a Rushton turbine with a diameter of 0.30 m, and its power number in turbulent flow is about 5, a published value for that impeller shape. The impeller turns at 2 revolutions per second.

Step 1: power. P = N P ρ N³ D⁵ = 5 x 1,050 x 2³ x 0.30⁵. Now 2³ = 8, and 0.30⁵ = 0.00243, so P = 5 x 1,050 x 8 x 0.00243 = 102 W. That is the power drawn by the impeller alone, before the motor and gearbox losses, so a motor of about 150 to 200 W covers it with a margin. The number is small because the impeller is small and slow, and it shows how fast the power grows: doubling the speed to 4 revolutions per second would raise the power eight times, to about 816 W, while doubling the impeller diameter would raise it thirty-two times.

Step 2: mixing time. In turbulent flow the mixing time, t m, follows a simple rule: the product N t m is roughly constant for a given impeller and tank shape, often near 40 to 60 for a turbine in a baffled tank. With N = 2 revolutions per second and N t m = 50, the mixing time is t m = 50 / 2 = 25 s. That is the time to reach a chosen level of uniformity, usually 95 percent mixed, and it is the number the operator waits for before sampling. The rule also explains scale-up: if the impeller speed is halved on scale-up, the mixing time doubles, so a plant batch mixes more slowly than the laboratory batch that proved the recipe, and the schedule must allow for it.

Step 3: check the regime. The power number of 5 holds only in turbulent flow. The mixer Reynolds number is Re = (ρ N D²) / μ. With a viscosity of 0.05 Pa·s, Re = (1,050 x 2 x 0.09) / 0.05 = 3,780, which is turbulent enough for the constant power number to hold. With a viscosity of 1 Pa·s, Re would be 189, laminar, and the power number would be far higher and falling with Re, so the power calculation would need the laminar curve for that impeller instead. The check takes a minute and stops the commonest mixer sizing mistake: using a turbulent power number for a laminar mix.

Step 4: tip speed and shear. The tip speed of the impeller is π D N = 3.14 x 0.30 x 2 = 1.88 m/s. The tip speed sets the shear at the blade, and for a shear sensitive product, such as a cream that must not be overworked or a suspension of fragile pieces, the scale-up rule of section 9.6 is constant tip speed, not constant power per volume. The two rules disagree, and the engineer picks the one that protects the product, then checks the other one still gives an acceptable mixing time.

Impeller types, and the liquid each suits

Propeller: thin liquids, fast, low shear, blending and heat transfer
Rushton turbine: thin to medium liquids, strong mixing, gas dispersion
Paddle: medium liquids, gentle, keeps solids off the bottom
Anchor: thick pastes, slow, scrapes the wall so nothing sticks

Pick the impeller by the thickness of the liquid, not by what is in stock.

Quick check

A mixer draws 100 W at 2 revolutions per second in turbulent flow. About what power will it draw at 4 revolutions per second?

Check your answer

About 800 W. Power rises with the cube of the speed in turbulent flow, and 2³ = 8, so 100 x 8 = 800 W.

Try it yourself

Mix a teaspoon of cocoa powder into a glass of cold milk with a spoon, and watch how it clumps and floats: the powder resists wetting. Now mix the same cocoa into a little hot water first to make a smooth paste, then stir the paste into the milk: it blends at once. You have seen the solid-liquid mixing problem and its fix. Then whisk a little oil into water with a fork, and watch the drops join again in seconds: without an emulsifier, a liquid-liquid mix does not stay mixed.

Key ideas to carry forward

  • Homogeneity means every small sample matches the whole within tolerance, and it is checked by sampling from several points.
  • Mixing works by convection, diffusion, and shear, and a good mixer moves material in three directions and scrapes the wall.
  • Mixing is fast at first and slow at the end, and power rises with the cube of the impeller speed in turbulent flow.
  • Liquid-liquid mixing blends or emulsifies, solid-liquid mixing dissolves, suspends, or wets, and solid-solid mixing tumbles powders that can segregate again.
  • Homogenization forces liquid through a narrow valve at high pressure, breaking fat globules to near 1 micron so milk does not cream.

Glossary

Homogeneity

The state where every small sample of a mixture has the same composition as the whole. Example: a spice blend is homogeneous when spoonfuls from top and bottom match.

Convection

Bulk movement carrying groups of particles from one place to another during mixing. Example: a paddle moving whole regions of batter is convection.

Diffusion

Single particles wandering across boundaries between regions, finishing mixing at the smallest scale. Example: diffusion evens out the last small differences in a blend.

Shear

Layers of material sliding past each other, stretching and folding the mix. Example: shear in a dough mixer stretches the dough and spreads the yeast.

Power number

A number with no units, N P = P / (ρ N³ D⁵), linking mixer power to speed and impeller size. Example: the power number levels off in turbulent flow.

Segregation

A mixed powder separating again because particles differ in size or density. Example: a spice blend segregates when poured from a height.

Dead zone

A region in a mixer where the material barely moves and never joins the mix. Example: the corners of a square tank can be dead zones.

Emulsion

A mix of two liquids that do not blend, with one broken into drops in the other. Example: mayonnaise is an emulsion of oil drops in water.

Emulsifier

A substance that sits at the boundary between drops and stops them joining again. Example: egg yolk is the emulsifier in mayonnaise.

Homogenization

Forcing a liquid through a narrow valve at high pressure to break fat globules into tiny even drops. Example: milk is homogenised so it does not form cream.

Homogenizer valve

The narrow gap in a homogenizer where the liquid speeds up and the globules are torn apart. Example: the valve gap is a fraction of a millimetre wide.

Cavitation

Tiny vapour bubbles forming and collapsing in a fast liquid, tearing droplets apart. Example: cavitation in the homogenizer valve helps break fat globules.

Ribbon blender

A horizontal trough mixer with a helical ribbon, suited to dry powders. Example: spice blends are mixed in a ribbon blender.

Planetary mixer

A mixer whose beater turns on its own axis while circling the bowl. Example: cake batter is mixed in a planetary mixer.

Static mixer

A tube with fixed elements that split and fold a flowing stream, with no moving parts. Example: a static mixer blends a flavour into a juice line.

High shear mixer

A mixer with a rotor spinning close to a stator, making very small drops. Example: a dressing is made smooth in a high shear mixer.

Mixing time

The time a mixer needs to reach a chosen level of uniformity, usually 95 percent mixed, in seconds. Example: with N t m near 50 and an impeller speed of 2 revolutions per second, the mixing time is about 25 s.

Tip speed

The speed of the outer edge of an impeller, π D N, in m/s, which sets the shear at the blade. Example: a tip speed of 1.88 m/s suits a gentle mix, while a higher tip speed shears harder.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.
  • Ravi Kumar, S. Dairy Engineering. AgriMoon course material.

Unit 10: Conveying

Note

Unit 10: Conveying

Grain flows into a mill, sugar rises to a packing machine, and flour moves from a silo to a mixer without a single person lifting a sack. Conveying moves bulk solids and liquids from one place to another inside a plant, and the choice of conveyor decides whether the product arrives whole, clean, and on time, or broken, dusty, and late. Why does one plant use a belt and another use air, and why does a screw conveyor suit flour but not whole fruit? This unit answers those questions.

Questions to think about

  • Why would a plant move grain with air instead of a belt, when air needs so much power?
  • What happens to a fragile food when it is dropped from a height, and how do conveyors avoid that?
  • Why does a screw conveyor work well for flour but badly for whole apples?

10.1 Introduction to conveying

Conveying is the movement of materials within a plant, between receiving, storage, processing, and packing. The material may be a liquid, a powder, a grain, or a piece of food, and each form needs a different conveyor. The design questions are always the same: how much material per hour, how far, how high, how fragile is it, how dusty is it, and how clean must the conveyor be?

A good conveyor moves the rated capacity without spilling, without breaking the product, without letting dust escape, and with parts that can be cleaned. A poor one wastes power, damages the product, and spreads dust that can carry microbes and, in the case of fine flour dust, can even catch fire. The four conveyors in this unit, the belt conveyor, the bucket elevator, the screw conveyor, and the pneumatic conveyor, cover most food plant needs between them.

Quick check

A plant must move whole biscuits from an oven to a packing machine. Which matters more: speed or gentle handling?

Check your answer

Gentle handling. Whole biscuits break easily, so the conveyor must move them without drops, impacts, or vibration, even if that means a slower line.

10.2 Belt conveyor

A belt conveyor carries material on a continuous loop of belt, running over rollers and driven by a head pulley. The belt may be flat or troughed, where the edges are raised by angled rollers to hold the load. Material is loaded at the tail end and carried to the head end, where it leaves over the pulley.

Belt conveyors suit long horizontal runs and gentle slopes, up to about 20 degrees for most bulk foods, and they carry large capacities, thousands of tonnes per hour in grain handling, with low power per tonne. They are open, so they suit material that is not dusty and not easily contaminated, and they are easy to inspect and clean. Their limits are the slope, the need for a straight path, and the risk of spillage at the loading point and at the discharge. Food belts are made from food grade materials, and the belt must be cleaned and dried on a schedule, because a damp belt grows microbes.

The capacity is set by the belt width, the belt speed, and the load cross section, and the power is set by the friction of the belt and rollers plus the lift of the material. A belt that runs too fast throws material off at the head pulley; a belt that runs too slow needs a wider belt for the same capacity.

Quick check

Why is a belt conveyor a poor choice for a steep lift of 30 metres straight up?

Check your answer

Belts suit gentle slopes up to about 20 degrees. A steep lift needs a bucket elevator or a pneumatic conveyor instead.

10.3 Bucket elevator

A bucket elevator lifts material straight up. Buckets fixed to a belt or a chain run up inside a casing, scoop material at the bottom, carry it to the top, and throw it out as they tip over the head pulley. It is the standard way to lift grain, sugar, flour, and other bulk solids from a pit to a silo or to the top of a mill.

The buckets must be filled without choking the boot, the bottom housing, and they must empty cleanly at the top without material falling back down the down leg, which is called back legging. The speed matters: too slow and the buckets do not fill or throw properly, too fast and the material is flung against the casing and damaged. Elevators are enclosed, so they control dust, but they need explosion relief and cleaning access, because a casing full of flour dust is a hazard if it is never cleaned.

Bucket elevators suit vertical lifts of tens of metres, they use little floor space, and they lift gently compared with a pneumatic line. Their limits are the single vertical path and the wear on the belt and buckets.

Quick check

What is back legging in a bucket elevator, and why is it a problem?

Check your answer

Material falling back down the down leg instead of leaving at the top. It wastes power, lowers capacity, and damages the product by repeated handling.

10.4 Screw conveyor

A screw conveyor, or auger, moves material with a helical screw turning inside a trough or a tube. The screw pushes the material forward a little with each turn, like a bolt turning through a nut. It suits powders and small grains, flour, sugar, salt, and ground spice, over short distances, horizontal or on a slope up to about 30 degrees.

Screw conveyors are compact, enclosed, cheap, and easy to fit into a tight plant. The screw can also mix as it moves, and a ribbon screw can blend a powder while it conveys it. Their limits are the short distance, the wear on the screw and trough, and the damage they do to fragile material: the screw grinds and crushes as it pushes, so whole grains, nuts, and pieces of food are broken. They also need a steady feed, because a flooded inlet chokes the screw and stalls the motor.

The capacity is set by the screw diameter, the pitch, the speed, and the fill of the trough, and the power rises with the length, the lift, and the friction of the material. A screw that is too long needs a larger motor and a centre bearing that can itself choke the flow.

Which conveyor fits the job

Long flat run, large capacity: belt conveyor
Straight up, tens of metres: bucket elevator
Short enclosed run, powder: screw conveyor
Long route with bends, dusty powder: pneumatic conveyor

Quick check

Why is a screw conveyor a poor choice for moving whole almonds?

Check your answer

The screw grinds and crushes as it pushes, so fragile whole nuts would be broken. A belt or a gentle bucket elevator suits them better.

10.5 Pneumatic conveyor

A pneumatic conveyor moves material in a stream of air through a pipe. The air may push the material, in a pressure system, or pull it, in a vacuum system, and the material travels as a cloud of particles in the air. It suits powders and small grains, flour, sugar, starch, and ground spice, over long routes with bends, up and down, where a belt cannot go.

Pneumatic conveying has three flow regimes. In dilute phase, the air is fast and the particles are spread thin, which suits light powders and long runs but can damage fragile material by impact on bends. In dense phase, the air is slower and the material moves as slugs or plugs, which is gentler and uses less air, but needs a higher pressure and a careful design to avoid blockages. In between lies a transition where the flow is unsteady.

The strengths are clear: the pipe can run anywhere, the system is enclosed so dust is controlled, and one line can feed several points. The costs are also clear: the fan or blower uses much more power per tonne than a belt, the particles wear the bends, and fragile material breaks. A vacuum system is safer for dust control, because a leak pulls air in rather than blowing dust out, while a pressure system suits longer runs and higher capacities.

The design sets the air velocity above the minimum that keeps particles moving, the saltation velocity, with a margin, and sizes the blower for the pressure drop of the pipe, the bends, and the separator at the end, usually a cyclone that spins the air and drops the powder out.

The separator at the end of the line deserves attention, because it decides how much product is recovered and how much dust escapes. A cyclone spins the air stream so the particles are thrown to the wall and fall to an outlet, while the clean air leaves from the top. It recovers most of the powder, but the finest dust passes through, so a bag filter is often fitted after the cyclone to catch what is left. An air lock, a rotating valve, lets the powder out of the cyclone without letting air leak in or out, which keeps the pressure in the line steady. A line without a proper air lock will either blow powder out through the outlet or suck air in and lose its carrying velocity, and both faults show up at once as lost product and a blocked pipe.

Quick check

Why is a bag filter often fitted after the cyclone at the end of a pneumatic line?

Check your answer

The cyclone recovers most of the powder but lets the finest dust through, so the bag filter catches what is left and stops it escaping as dust.

Practical note. The practical work for conveying is observational: study a belt conveyor, a bucket elevator, a screw conveyor, and a pneumatic line in a plant or a teaching rig, note the feed, the path, the discharge, and the product damage at each stage, and match each conveyor to the material it carries. Count the drops and the bends on a pneumatic line, and link them to the breakage you see in the product.

Quick check

Why does a pneumatic conveyor use more power per tonne than a belt conveyor?

Check your answer

It must move a large volume of air fast enough to carry the particles, and the blower works against the pressure drop of the whole pipe and separator, while a belt only overcomes friction and lift.

10.6 Choosing a conveyor and keeping it clean

Choosing between the four conveyors is a comparison, not a guess. A worked comparison makes the trade clear. A plant must move 20 tonnes per hour of sugar over 40 metres horizontally and then 15 metres up to a packing machine. A belt conveyor covers the horizontal run cheaply and gently, but it cannot do the 15 metre lift, so a bucket elevator is added for the vertical part. A screw conveyor could do the horizontal run in a smaller space, but it would grind the sugar crystals and make dust, so it is the wrong choice for a product sold on crystal size. A pneumatic line could do the whole route in one pipe, but it would use far more power and break more crystals on its bends, so it is the costly choice. The belt plus elevator pair wins on power, on product quality, and on cleaning, because both are open enough to inspect and wash.

Hygiene decides the details. A food conveyor must be built from food grade materials, with no cracks or pockets where product can lodge and grow microbes, and with parts that can be reached for cleaning. Belts are washed and dried on a schedule, elevator boots are emptied and cleaned so stale material does not sit and ferment, screw troughs are opened and brushed out, and pneumatic lines are purged with air and checked for build-up at bends. A conveyor that cannot be cleaned will contaminate the next product that runs through it, whatever its mechanical merits.

Dust control is the other running duty. Transfer points, where material falls from one conveyor to another, are the main source of dust, and they are enclosed and fitted with extraction. A drop of more than a metre is avoided where the product is fragile, and a curved chute is used to guide the material gently instead of letting it fall straight. The design rule is simple: move the product as little as far as needed, with as few drops as possible, and keep the dust inside the system.

Quick check

A plant must move sugar crystals that are sold on their size. Why is a screw conveyor the wrong choice for the horizontal run?

Check your answer

The screw grinds and crushes as it pushes, so it would break the crystals, make dust, and spoil the size the customer pays for.

10.7 Worked example: sizing a belt conveyor

Section 10.2 states that capacity is set by belt width, belt speed, and load cross section. This section sizes a belt, with numbers.

The job. A plant must move 200 tonnes per hour of grain with a bulk density of 750 kg/m³ up a 10 m lift on a belt running at 1.5 m/s.

Step 1: volume flow. 200 tonnes per hour is 200,000 kg/h, or 55.6 kg/s. Dividing by the bulk density gives the volume flow: 55.6 / 750 = 0.074 m³/s.

Step 2: load cross section. The load cross section is the volume flow divided by the belt speed: A = 0.074 / 1.5 = 0.049 m². A troughed belt of about 0.6 m width carries a load of roughly this cross section at a normal surcharge angle, so a 0.6 m belt is the right size, and a 0.5 m belt would need a higher speed or a deeper trough to carry the same flow. The check is to compare the needed cross section with the belt maker's table for that width and trough angle, and to keep a margin, because a belt loaded to its very edge spills at the slightest surge.

Step 3: power to lift. The power to lift the grain is mass flow times g times height: 55.6 x 9.81 x 10 = 5,454 W. With a drive efficiency of 0.8, the motor must give 5,454 / 0.8 = 6,818 W for the lift alone. Friction in the belt and rollers adds more, often as much again on a long run, so the motor is sized at about 12 to 15 kW for this duty. The lesson is that the lift dominates the power on a short steep run, while friction dominates on a long flat run, and a conveyor that looks cheap on a flat drawing can be dear once the lift is counted.

Step 4: check the discharge. At the head pulley the grain leaves the belt with the belt speed, 1.5 m/s, and follows a trajectory into the discharge chute. If the speed were much higher, the grain would be thrown hard against the chute and break; if it were much lower, the grain would fall straight down and choke the chute. The belt speed is therefore chosen with the product in mind, not only with the capacity, and fragile foods run slower than grain.

A belt conveyor, from loading to discharge

Tail end: material loaded onto the belt, centred, no spillage

↓

Carrying run: troughed belt on rollers, steady speed, even load

↓

Head pulley: material leaves over the pulley into the discharge chute

The return belt runs back underneath, empty, to the tail end.

Quick check

A belt must carry twice the flow at the same belt speed. What must happen to the load cross section?

Check your answer

It must double. Volume flow equals cross section times belt speed, so twice the flow at the same speed needs twice the cross section, a wider belt or a deeper trough.

10.8 Pneumatic conveying: air speed, saltation, and choosing the phase

A pneumatic conveyor moves solids in a stream of air, and the air speed decides whether the system works or blocks. The key speed is the saltation velocity: the lowest air speed at which particles stay suspended and keep moving. Below it, particles fall out of the stream, collect on the bottom of the pipe, and build a bed that grows until the pipe chokes. Above it, the particles fly along, carried by the air. The design rule is to run well above the saltation velocity, usually with a margin of about 50 percent, because a dip in air flow, a worn fan, or a surge in feed can drop the speed below the limit and block the line in seconds.

Dilute phase and dense phase. In dilute phase conveying the air is fast, 15 to 30 m/s, and the particles are spread thin, with a solids to air mass ratio below about 15. It suits light powders and long runs, and it is simple to run, but the fast particles wear bends and break fragile material. In dense phase conveying the air is slow, 2 to 10 m/s, and the material moves as slugs or plugs that fill the pipe, with a solids to air ratio far higher. It is gentler and uses less air, but it needs a higher pressure, a stronger pipe, and a careful design, because a plug that stops is a blockage that must be cleared by hand. The choice is a trade: dilute phase for robust material over a long run, dense phase for fragile or abrasive material over a shorter run where the gentler ride is worth the dearer system.

Air needed. The air flow follows from the solids flow and the chosen solids to air ratio. Moving 1,000 kg/h of flour in dilute phase at a ratio of 10 needs 100 kg/h of air, about 83 m³/h at normal density, and the fan must supply that flow at the pressure the line needs, which is the pressure drop of the pipe, the bends, the cyclone, and the filter added together. A filter at the end is not optional: it catches the fines that would otherwise leave with the air, and a blinded filter raises the pressure drop until the fan cannot hold the air speed and the line salts out. The operator watches the filter pressure drop every shift for this reason.

Why air, despite the power. Section 10.1 asks why a plant would move grain with air when air needs so much power. The answers are that the line can bend around the plant, climb vertically, and split to several destinations, that the system is enclosed so dust is contained, and that the same air can dry or cool the product on the way. The power is the price of that flexibility, and the engineer pays it only where a belt or a screw cannot do the job.

Quick check

The air speed in a dilute phase line falls below the saltation velocity. What happens in the pipe?

Check your answer

Particles fall out of the stream, collect on the bottom, and build a bed that chokes the pipe, so the line blocks.

Try it yourself

Pour a cup of dry rice slowly from a height of one metre into a bowl, then pour another cup gently down a tilted board into the bowl. Compare the broken grains in each bowl: the drop breaks more grains, which is why conveyors avoid drops. Now blow a straw across a spoonful of flour on a table and watch the fine particles lift and travel with the air: that is dilute phase pneumatic conveying on a small scale, and it shows why dust control matters.

Key ideas to carry forward

  • Conveying moves material inside a plant, and the choice turns on capacity, distance, height, fragility, dust, and cleaning.
  • Belt conveyors suit long flat runs with large capacity and low power per tonne, but only gentle slopes.
  • Bucket elevators lift bulk solids straight up, tens of metres, in an enclosed casing, but need care to avoid back legging.
  • Screw conveyors suit short enclosed runs of powder, are cheap and compact, but crush fragile material.
  • Pneumatic conveyors move powder in air through pipes over long routes with bends, in dilute or dense phase, but use more power and can break fragile material.

Glossary

Conveying

Moving material from one place to another inside a plant. Example: grain is conveyed from receiving to the mill.

Belt conveyor

A conveyor carrying material on a continuous loop of belt over rollers. Example: grain moves along a belt conveyor to storage.

Bucket elevator

A vertical conveyor of buckets on a belt or chain that lift material straight up. Example: a bucket elevator lifts sugar to the top of a silo.

Back legging

Material falling back down the down leg of a bucket elevator instead of leaving at the top. Example: back legging wastes power and damages grain.

Screw conveyor

A conveyor that pushes material forward with a helical screw turning in a trough. Example: flour moves from a bin to a mixer in a screw conveyor.

Auger

Another name for a screw conveyor. Example: an auger feeds ground spice into a packing machine.

Pneumatic conveyor

A conveyor that moves material in a stream of air through a pipe. Example: flour is blown through a pneumatic line to the bakery.

Dilute phase

Pneumatic conveying with fast air and particles spread thin. Example: dilute phase suits light powders on long runs.

Dense phase

Pneumatic conveying with slow air and material moving as slugs. Example: dense phase is gentler on fragile powder.

Saltation velocity

The lowest air velocity that keeps particles moving in a pneumatic line. Example: the air must run above the saltation velocity or the line blocks.

Cyclone separator

A device that spins an air stream so particles drop out of it. Example: a cyclone at the end of a pneumatic line drops the flour out of the air.

Troughed belt

A belt whose edges are raised by angled rollers to hold the load. Example: a troughed belt carries grain without spilling.

Head pulley

The driven pulley at the discharge end of a belt conveyor. Example: material leaves the belt over the head pulley.

Tail pulley

The pulley at the loading end of a belt conveyor. Example: material is loaded onto the belt near the tail pulley.

Boot

The bottom housing of a bucket elevator where buckets scoop material. Example: a choked boot stalls the elevator.

Dust explosion

A fire or explosion from fine dust in air, a hazard with flour and starch. Example: cleaning and explosion relief control the dust explosion risk.

Saltation velocity

The lowest air speed at which particles stay suspended in a pneumatic conveyor. Example: running below the saltation velocity lets particles settle and choke the pipe.

Bulk density

The mass of a bulk solid per unit of volume it occupies, including the air between the particles, in kg/m³. Example: grain with a bulk density of 750 kg/m³ needs 0.074 m³/s of belt capacity for 200 tonnes per hour.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.

Unit 11: Leaching and Extraction

Note

Unit 11: Leaching and Extraction

Oil is pressed and washed out of seeds, caffeine is taken out of coffee beans, sugar is washed out of beet slices, and flavour is drawn out of vanilla pods. Leaching and extraction move a wanted substance out of a solid or a liquid into a solvent, and the choice of solvent and equipment decides how much is won, how pure it is, and how much is left behind. Why does hot water pull more out of tea than cold water, why do plants wash the same solid several times with fresh solvent, and how can carbon dioxide act as a liquid? This unit answers those questions.

Questions to think about

  • Why does a second wash with fresh water pull out more sugar than one long wash with the same water?
  • What makes one solvent a good choice for pulling oil out of a seed, and another a bad one?
  • How can a gas like carbon dioxide be used as a liquid solvent?

11.1 Introduction: leaching and extraction

Leaching is the removal of a soluble substance from a solid by a liquid solvent. Sugar is leached from beet slices with hot water, caffeine is leached from coffee beans with water, and oil is leached from oilseed flakes with a solvent. The solvent soaks into the solid, dissolves the wanted substance, and carries it out.

Liquid extraction, also called solvent extraction, is the removal of a substance from one liquid by a second liquid that does not mix with the first. A flavour compound is pulled from a watery liquid into an oil, or an acid is pulled from an oil into water. The substance moves because it prefers the second liquid, and that preference is measured by the distribution coefficient.

Both operations are driven by a difference in concentration: the substance moves from where it is strong to where it is weak. Both are helped by small particle size, which shortens the distance the substance must travel, by heat, which speeds diffusion and raises solubility, and by fresh solvent, which keeps the concentration difference large. Both end with a separation step: the solid is drained, the two liquids are settled apart, or the solvent is evaporated to leave the product behind.

Quick check

Sugar is leached from beet slices with hot water. Is this leaching or liquid extraction, and why?

Check your answer

Leaching. The sugar is removed from a solid, the beet slice, by a liquid solvent, hot water.

11.2 Principles of leaching

Leaching has four steps, and the slowest step sets the rate.

  1. The solvent reaches the solid. It wets the surface and soaks into the pores. A solvent that does not wet the solid will leach slowly, whatever else is done.
  2. The solute dissolves. The wanted substance passes into the solvent at the solid surface. A hotter solvent dissolves more and faster.
  3. The solute diffuses out. It travels through the pores of the solid to the surface, then across the still layer of liquid clinging to the surface into the bulk liquid. This internal diffusion is usually the slow step, which is why small particles leach faster than large ones: the distance is shorter.
  4. The extract is separated. The solid is drained, pressed, or filtered, and the liquid carrying the solute, the extract, is taken away.

The equilibrium of leaching is simple: if the solid and the liquid stay in contact long enough, the solute spreads evenly, and the concentration in the liquid leaving the solid matches the concentration held in the wet solid. A single contact leaves much solute behind, because the liquid and the solid reach the same concentration. Fresh solvent in a second contact pulls out more, and several contacts in a chain, called cross current or counter current leaching, win far more of the solute for the same total solvent.

Counter current leaching is the standard plant arrangement: the solid moves one way through a chain of stages, the solvent moves the other way, so the freshest solvent meets the most exhausted solid, and the strongest extract meets the freshest solid. The concentration difference stays large at every stage, and the solvent use is far lower than washing the whole batch in one go.

Quick check

Why does washing a solid twice with fresh solvent win more solute than one wash with the same total solvent?

Check your answer

Each fresh wash restores a large concentration difference, so more solute moves out. One wash reaches balance and leaves much solute behind in the solid.

11.3 Leaching equipment

  • Batch tank with agitator. The solid and the solvent are stirred together in a tank, then settled and drained. It is simple and suits small batches and high value solids.
  • Percolation battery, or diffusion battery. A row of vessels holds the solid, and the solvent flows from one vessel to the next, so each vessel sees fresh solvent in turn and the extract grows stronger as it passes along the row. It is the classic sugar beet system, run as a batch of vessels in sequence.
  • Continuous counter current extractor. A moving belt, a rotating drum, or a screw carries the solid one way while the solvent flows the other way. It runs without stopping and suits large volumes of oilseed and sugar beet.
  • Rotocel extractor. A rotating carousel of baskets carries the solid under a series of solvent sprays, with the extract collected and pumped forward from stage to stage. It is a common oilseed extractor.
  • Soxhlet extractor. A laboratory unit where the solvent is boiled, condensed, and dripped through the solid again and again, so a small sample is leached with a small volume of solvent. It is the standard laboratory method for measuring the oil content of a seed.

The choice turns on the volume, the value of the product, and how much solvent can be used. Large oilseed plants use continuous counter current extractors, sugar plants use diffusion batteries, and laboratories use Soxhlet units.

Counter current leaching: solid one way, solvent the other

Fresh solid in, meets strong extract, gives up solute easily

↓

Middle stages: solid weakens, solvent strengthens, difference stays large

↓

Spent solid out, meets fresh solvent, last solute is won

Fresh solvent in at the spent end, strong extract out at the fresh end.

Quick check

In a counter current leaching chain, which end gets the fresh solvent: the fresh solid end or the spent solid end?

Check your answer

The spent solid end. The freshest solvent meets the most exhausted solid, so the last solute is won.

11.4 Liquid extraction: principles and equipment

In liquid extraction, two liquids that do not mix are brought together, the solute moves from one to the other, and the liquids are then settled apart. The distribution coefficient, K, is the ratio of the solute concentration in the extract phase to its concentration in the raffinate phase at balance: K = y / x. A large K means the solute strongly prefers the extract phase, and a few stages will win most of it. A K near 1 means many stages are needed.

The equipment mirrors the leaching equipment, with the solid replaced by a second liquid:

  • Mixer-settler. A stirred tank mixes the two liquids, then a settler lets them separate by gravity. Several mixer-settlers in a chain, run counter current, give a sharp separation. It is the most common liquid extraction unit in food plants.
  • Spray column. One liquid is sprayed as drops through the other, which flows the other way. The drops give a large surface for transfer, and the column is simple, but the drops can coalesce and the capacity is limited.
  • Packed column. Packing breaks the liquids into thin films and drops and gives a large contact area in a small volume. It suits systems where the liquids separate easily after contact.
  • Centrifugal extractor. A spinning bowl mixes the liquids and separates them at once by the spinning force. It suits liquids that separate slowly by gravity, such as emulsions, and it is compact.

A solvent for liquid extraction must be selective for the wanted substance, must not mix with the feed liquid, must separate easily, must be safe with food, and must be easy to recover by distillation so it can be used again. Those demands together narrow the choice sharply, which is why so few solvents are used in food work.

Quick check

A solute has a distribution coefficient of 10 between the extract and the raffinate. What does that mean in plain words?

Check your answer

At balance, the solute is 10 times as concentrated in the extract phase as in the raffinate phase, so it strongly prefers the extract.

11.5 Special extraction techniques: supercritical fluid extraction

A supercritical fluid is a substance held above its critical temperature and pressure, where it is neither a liquid nor a gas. It flows like a gas and dissolves like a liquid, and its solvent power can be tuned by changing its pressure. Carbon dioxide, CO₂, is the usual choice for food: its critical point is mild, 31 °C and 7.4 MPa, it is not toxic, it leaves no residue because it turns back into a gas when the pressure is released, and it does not damage heat sensitive flavours.

Supercritical CO₂ extraction is used to take caffeine out of coffee beans, to win flavours and essential oils from spices, and to pull oil from seeds without a liquid solvent. The process is: CO₂ is compressed above its critical point, passed through the solid, where it dissolves the wanted substance, then the pressure is dropped in a separator, the CO₂ turns to gas and leaves, and the extract is collected pure. The CO₂ is then compressed again and reused.

Its strengths are a clean product with no solvent residue, a low temperature that protects flavour, and a solvent power that can be tuned. Its costs are the high pressure equipment and the energy of compression, so it suits high value products, flavours, aromas, and decaffeinated coffee, rather than bulk oil.

Other special techniques include extraction with hot water under pressure, called subcritical water extraction, and extraction helped by ultrasound or microwaves, which break cell walls and speed the movement of solute out of the solid. Each of them trades equipment cost against a cleaner product or a faster run. Subcritical water is water held liquid above its normal boiling point by pressure, at 100 to 200 °C, where it behaves more like an organic solvent and pulls out compounds that cold water leaves behind, without any solvent residue at all. Ultrasound sends sound waves through the solvent that make tiny bubbles collapse against the solid surface, cracking cell walls and letting the solvent in faster, which cuts the extraction time for flavours and colours. Microwave heating warms the water inside the cells directly, so the cells burst and release their contents into the solvent. All three are used for high value flavours, colours, and plant extracts, where a shorter time or a cleaner label is worth the extra equipment.

Quick check

How does ultrasound help extraction, and what does it do to the solid?

Check your answer

Sound waves make tiny bubbles that collapse against the solid surface, cracking cell walls so the solvent gets in faster and the extraction time falls.

Quick check

Why does supercritical CO₂ leave no solvent residue in the extract?

Check your answer

When the pressure is released, the CO₂ turns back into a gas and leaves, so nothing of the solvent stays in the product.

11.6 Worked example: a leaching mass balance

A worked example shows how a leaching balance is done. A plant leaches 1,000 kg of beet slices holding 16 percent sugar and 75 percent water, with the rest insoluble fibre, using hot water. After leaching and pressing, the spent pulp leaving the system weighs 600 kg and holds 2 percent sugar. The sugar in the feed is 1,000 x 0.16 = 160 kg. The sugar left in the spent pulp is 600 x 0.02 = 12 kg. So the sugar won in the extract is 160 - 12 = 148 kg, and the recovery is 148 / 160 = 92.5 percent. The extract mass is the feed plus the solvent water added, minus the spent pulp, and its sugar content follows from the same balance. If the plant wants a higher recovery, it can wash the pulp again with fresh water, which is a second cross current stage, or it can run the pulp and the water counter current, which wins more sugar for the same water, as section 11.2 explains.

The same balance governs liquid extraction. For a feed of 1,000 kg holding 5 percent of a flavour compound, extracted with a solvent where the distribution coefficient is 10, the first contact with an equal mass of solvent leaves most of the compound in the solvent, because the compound prefers the extract phase ten to one. A second contact with fresh solvent wins most of what is left. The number of stages needed is worked out from the coefficient and the solvent ratio, and a counter current chain of mixer-settlers uses the least solvent for a given recovery, just as in leaching.

Quick check

A leaching run wins 148 kg of sugar from a feed holding 160 kg. What is the recovery, and what would a second wash with fresh water do?

Check your answer

Recovery is 92.5 percent. A second wash with fresh water would win part of the 12 kg left in the pulp, raising the recovery further.

11.7 Choosing a solvent: what a good solvent must do

A leaching or extraction solvent is chosen against five tests, and a solvent that fails one of them is the wrong choice whatever its other merits.

Selectivity. The solvent must take the wanted solute and leave the rest behind. A solvent that pulls out everything gives a dirty extract that needs a second separation, which costs more than the extraction saved. Selectivity is measured by comparing the distribution coefficients: a good solvent has a high coefficient for the wanted solute and a low one for the rest. Water is selective for sugars and acids and poor for oils, hexane is selective for oils and poor for sugars, and ethanol sits in between, which is why ethanol is the usual choice for plant extracts that hold both water loving and oil loving compounds.

Low viscosity and good wetting. The solvent must flow into the pores of the solid and out again, carrying the solute with it. A thick solvent wets slowly, drains slowly, and leaves solute behind in the pores. Warming the solvent lowers its viscosity and speeds the whole step, which is why leaching is often run warm, but warming also raises the solubility of unwanted parts, so the temperature is a balance between speed and selectivity.

Easy recovery. The solvent must be won back from the extract and from the spent solid, cheaply and cleanly. A solvent that boils at a low temperature is easy to distil off, but it is also easy to lose to the air, and its vapour must be caught. A solvent that boils at a high temperature is hard to remove, and the heat needed to boil it off can damage the extract. The recovery step is often the most expensive part of the plant, and a solvent that is cheap to buy but dear to recover is a dear solvent.

Safety and residue. The solvent must not leave a harmful residue in the food, and it must be safe to handle in the plant. Food grade solvents are few: water, ethanol, and carbon dioxide are the safe core, and any other solvent must be removed to a level the rules allow, with the removal proved by testing, not assumed. Flammable solvents need a plant built for them, with no ignition source and with vapour extraction, which is a capital cost that must be counted in the choice.

Cost and supply. The solvent must be affordable at the scale used, available steadily, and stable in storage. A solvent that degrades in light or air, or that needs a special store, adds a running cost that the extraction must carry. Water wins on all five tests for many jobs, which is why it is the first solvent tried, and a plant moves to ethanol or carbon dioxide only when water cannot do the job.

Quick check

A solvent pulls the wanted flavour out of a spice, but it also pulls out a bitter compound that water leaves behind. Which of the five tests has it failed?

Check your answer

Selectivity. A good solvent takes the wanted solute and leaves the rest behind, and this one does not.

11.8 Tuning supercritical CO₂: pressure, density, and staged separators

Section 11.5 states that the solvent power of supercritical CO₂ can be tuned by changing its pressure. This section shows how the tuning works and how a plant uses it.

Density is the dial. Above its critical point, 31 °C and 7.4 MPa, carbon dioxide has a density that rises steeply with pressure at a fixed temperature. Near the critical point a small rise in pressure gives a large rise in density, and the solvent power follows the density: a dense fluid dissolves like a liquid, a thin fluid dissolves almost nothing. The operator therefore sets the pressure to set the density, and the density to set what dissolves. At a low density the fluid takes only the lightest, most soluble compounds, such as the volatile oils of a spice. At a high density it takes the heavier compounds too, such as the waxes and the heavier flavour parts. The temperature matters as well: raising the temperature at a fixed pressure lowers the density, so it weakens the solvent, even though it raises the vapour pressure of the solute, which strengthens extraction, and the two effects pull in opposite directions, which is why the pressure and temperature are tuned together, not one at a time.

Staged separators. The extract is won back by dropping the pressure in steps, not all at once. In the first separator the pressure is dropped a little, so the density falls a little, and the heaviest compounds, the waxes, fall out first. In the second separator the pressure is dropped further, and the flavour compounds fall out. In the last separator the pressure is dropped to near atmospheric, the CO₂ turns to gas, and the lightest compounds are caught. Each separator gives a different cut of the extract, so the plant can sell the flavour cut on its own and keep the waxes out of it, which a single separator could not do. The CO₂ is then compressed again and returned to the extractor, so the same gas is used again and again, and the running cost is the power for the compressor, not the gas.

Co-solvents. Pure CO₂ is a poor solvent for water loving compounds, because it is non polar. Adding a small amount of ethanol, usually 5 to 10 percent, as a co-solvent changes that: the ethanol carries the water loving compounds into the fluid, so a spice extract that pure CO₂ would leave pale comes out full and strong. The price is that the ethanol must be separated from the extract afterwards, usually by a gentle evaporation, and the extract is no longer solvent free, which matters where the label must say so.

Quick check

An extractor is run at a higher pressure, with the temperature held fixed. Does the CO₂ dissolve more or less, and why?

Check your answer

More. A higher pressure at fixed temperature raises the density of the fluid, and the solvent power follows the density.

Try it yourself

Put a tea bag in a cup of cold water and another in a cup of hot water, and watch the colour spread. The hot cup colours far faster, because heat raises solubility and speeds diffusion, which is leaching in your kitchen. Now make a second cup with the same tea bag and fresh hot water: the second cup still colours, which shows that one contact leaves solute behind and fresh solvent wins more. Then shake a little cooking oil with water and a drop of food colour: the colour stays in the water, which shows that a solute moves to the phase it prefers, the idea behind the distribution coefficient.

Key ideas to carry forward

  • Leaching removes a soluble substance from a solid with a liquid solvent, and liquid extraction moves a substance between two liquids that do not mix.
  • Leaching has four steps, and internal diffusion is usually the slow one, so small particles and heat speed it.
  • Counter current contact keeps the concentration difference large at every stage and wins more solute with less solvent.
  • The distribution coefficient K sets how many stages a liquid extraction needs.
  • Supercritical CO₂ extraction uses a fluid above its critical point to win flavours and caffeine with no solvent residue and low heat damage.

Glossary

Leaching

Removing a soluble substance from a solid with a liquid solvent. Example: sugar is leached from beet slices with hot water.

Extraction

Moving a wanted substance from one phase into another, often with a solvent. Example: oil is extracted from seed flakes with a solvent.

Solvent

The liquid that dissolves and carries the wanted substance. Example: hot water is the solvent that leaches sugar from beet.

Solute

The substance that is dissolved and moved. Example: sugar is the solute leached from beet slices.

Extract

The liquid carrying the solute away from the solid or the feed. Example: the strong sugar solution leaving a diffusion battery is the extract.

Raffinate

The liquid left behind after the solute has been extracted from it. Example: the weak liquid leaving a mixer-settler is the raffinate.

Distribution coefficient

The ratio of solute concentration in the extract phase to that in the raffinate phase at balance. Example: a coefficient of 10 means the solute strongly prefers the extract.

Counter current

An arrangement where the two streams move in opposite directions. Example: counter current leaching wins more solute with less solvent.

Cross current

An arrangement where fresh solvent is used at each stage. Example: cross current washing is simple but uses more solvent than counter current.

Diffusion battery

A row of vessels where solvent flows from one to the next, leaching the solid in turn. Example: sugar beet is leached in a diffusion battery.

Rotocel extractor

A rotating carousel extractor that carries solid under solvent sprays. Example: oilseed is extracted in a Rotocel unit.

Soxhlet extractor

A laboratory unit that leaches a sample again and again with condensed solvent. Example: the oil content of a seed is measured with a Soxhlet extractor.

Mixer-settler

A unit that mixes two liquids then lets them separate by gravity. Example: flavour compounds are pulled from a liquid in a chain of mixer-settlers.

Supercritical fluid

A substance above its critical temperature and pressure, neither liquid nor gas. Example: supercritical CO₂ dissolves caffeine from coffee beans.

Critical point

The temperature and pressure above which a substance becomes a supercritical fluid. Example: CO₂ reaches its critical point at 31 °C and 7.4 MPa.

Decaffeination

Removing caffeine from coffee or tea. Example: supercritical CO₂ is used to decaffeinate coffee beans.

Selectivity

How well a solvent takes the wanted solute and leaves the rest behind. Example: water is selective for sugars, hexane for oils, and ethanol sits in between.

Co-solvent

A small amount of a second solvent, usually ethanol, added to supercritical CO₂ to help it dissolve water loving compounds. Example: 5 to 10 percent ethanol lets CO₂ win a full spice extract that pure CO₂ would leave pale.

References

  • Singh, R. Paul, and Dennis R. Heldman. Introduction to Food Engineering. 5th edition.
  • Earle, R. L. Unit Operations in Food Processing.
  • Fellows, P. J. Food Processing Technology: Principles and Practice. 4th edition.
  • Toledo, Romeo T. Fundamentals of Food Process Engineering. 3rd edition.
  • Geankoplis, Christie John. Transport Processes and Separation Process Principles. 4th edition.