In short
A conventional jacket is cheap and easy to clean but gives low coolant velocity and a low pressure rating. A limpet (half-pipe) coil forces the fluid along a spiral path at 1–3 m/s, giving better heat transfer and higher pressure capability, and suits thermal oil and steam. An internal coil adds area and has the best coefficient, but it is harder to clean and risks leaks into the batch.
Every batch reactor needs a way to add or remove heat. The three common choices are a conventional jacket, a limpet (half-pipe) coil welded to the outside of the shell, and a coil inside the vessel. Each has a different heat transfer coefficient, pressure rating, cost and cleaning burden.
In short: a jacket is cheapest and simplest but has poor fluid velocity; a limpet coil gives controlled velocity, better heat transfer and a higher pressure rating, and is the usual choice for thermal oil and steam on steel reactors; an internal coil adds the most area and the best coefficient, but it is harder to clean and puts a utility fluid inside your batch.
This guide explains how each works, how to estimate whether the area is enough, and how to choose for a given chemical reactor.
What are the options?
Conventional jacket. An outer shell welded around the vessel, leaving an annular gap of typically 50–100 mm. Fluid enters at one nozzle and leaves at another. Without help, velocity in the annulus is very low, so the jacket-side coefficient is poor. Agitation nozzles (tangential inlets) or a spiral baffle in the annulus raise the velocity.
Dimple jacket. A thin sheet pressed with dimples and spot-welded to the shell. It is light and cheap and gives turbulence, but its pressure rating depends on the dimple pattern and welding. Common in food and pharma stainless vessels.
Limpet (half-pipe) coil. A pipe split lengthwise, commonly 50–100 NB, welded in a helix around the shell and bottom dish. Fluid follows the full spiral, so velocity can be set at 1–3 m/s. The coil also stiffens the shell. Large vessels often have the coil split into two or three zones with separate inlets.
Internal coil. A helical coil, or vertical harp coils, inside the vessel. Both sides of the tube see good velocity: utility fluid inside, agitated batch outside. The tube wall is thin compared with the vessel shell, so wall resistance is low.
External loop. Not a jacket at all: the batch is pumped through an external heat exchanger and back. Used when the heat load is very large, for example in large exothermic reactions. See our shell and tube heat exchanger guide for that option.
How do they compare?
| Factor | Conventional jacket | Dimple jacket | Limpet coil | Internal coil |
|---|---|---|---|---|
| Utility-side velocity | Low unless baffled | Moderate | High, controlled (1–3 m/s) | High |
| Typical overall U, water cooling, SS vessel, W/m²·K | 250–450 | 350–550 | 450–800 | 600–1,000 |
| Pressure capability on utility side | Low to moderate | Moderate | High | High |
| Thermal oil service | Poor (low velocity, hot spots) | Limited | Very good | Good |
| Steam service | Good at low pressure | Good | Good at higher pressure | Good |
| Glass-lined reactor | Standard | Not used | Not common | Not used |
| Cleaning (process side) | Smooth wall, easy | Easy | Easy | Difficult; deposits on coil |
| Risk of utility leak into batch | Very low | Very low | Very low | Higher |
| Interference with mixing | None | None | None | Acts as baffle; can create dead zones |
| Cost | Lowest | Low | Moderate | Moderate, plus fabrication inside vessel |
The U values are indicative for clean, well-agitated, water-like batches. Viscous batches have much lower process-side coefficients, and then the choice of utility-side arrangement matters less than the agitator. See our agitator engineering guide and agitator design tool.
Why does heat transfer area become a problem on scale-up?
For geometrically similar vessels, volume grows with D3 but jacket area grows with D2:
AV∝1D
Approximate figures for vessels with a straight side equal to the diameter and a 2:1 ellipsoidal bottom, jacketed on the straight side and bottom dish:
| Nominal volume | Diameter | Jacket area | Area per m³ |
|---|---|---|---|
| 1 KL | 1.05 m | about 4.6 m² | about 4.5 m²/m³ |
| 5 KL | 1.8 m | about 13 m² | about 2.6 m²/m³ |
| 20 KL | 2.8 m | about 33 m² | about 1.65 m²/m³ |
So a 20 KL reactor has roughly 40% less area per cubic metre than a 5 KL one, and about a third of a 1 KL vessel's. A reaction that was easy to control at small scale can run away at plant scale if this is ignored. This is the most common reason to move from a jacket to a limpet coil, or to add an internal coil or external loop. Our article on interlocks and safety systems for exothermic reactors covers the protection side.
How do you check whether the area is enough?
Use the steady-state balance during the peak heat release:
Q=U A ΔTlm
with
ΔTlm=(Tb−tin)−(Tb−tout)ln[(Tb−tin)/(Tb−tout)]
where Tb is the batch temperature and tin, tout the utility inlet and outlet temperatures. For batch heating or cooling over time, use the time-dependent form; our batch heating time calculator does this.
Worked example (hypothetical)
A 5 KL stainless reactor runs an exothermic step at 60 °C. Peak heat release is 150 kW. Cooling water is available at 32 °C and may return at 37 °C.
- ΔTlm: (60 − 32) = 28 °C, (60 − 37) = 23 °C, so ΔTlm≈25.4 °C
- Required UA=150,000/25.4≈5,900 W/K
- Cooling water flow: 150/(4.18×5)≈7.2 kg/s, about 26 m³/h
Now compare the options:
| Option | Assumed U, W/m²·K | Area needed | Available |
|---|---|---|---|
| Plain jacket | 400 | 14.8 m² | about 13 m² at full level, less at part fill: not enough |
| Limpet coil (same coverage) | 650 | 9.1 m² | about 11–12 m² (coil pitch leaves gaps): enough, small margin |
| Jacket + internal coil | 400 on jacket, 800 on 4 m² coil | jacket must supply 2,700 W/K → 6.8 m² | comfortable |
Two practical points come out of this. First, the area actually wetted at the start of a semi-batch addition may be far less than the full jacket area. Check the heat balance at the lowest working level. Second, if cooling water arrives warmer in summer, say 34 °C, the margin shrinks quickly. Chilled water or brine for the peak, or a slower addition rate, may be needed.
How do you size a limpet coil?
- Choose the half-pipe size, usually 50, 65, 80 or 100 NB. Larger pipe means lower pressure drop but a lower velocity for the same flow.
- Set the velocity. Aim for about 1.5–3 m/s for water and 1.5–2.5 m/s for thermal oil. Below about 1 m/s the coefficient falls and fouling increases.
- Calculate the flow area of the half-pipe (roughly half the pipe's internal area, less the effect of the weld) and the resulting velocity.
- Check pressure drop. The developed length of a spiral on a large vessel can exceed 100 m. If pressure drop is too high, split the coil into two or three zones fed in parallel.
- Set pitch. The gap between turns is commonly 25–50 mm. A tighter pitch gives more coverage but more welding.
- Check mechanical design. The coil adds stiffness, but the shell, coil and welds must be checked for the utility pressure, the vessel's design pressure and vacuum, and thermal stresses. See our pressure vessel guide.
For steam heating, feed steam at the top and drain condensate from the bottom of each zone through its own trap. For cooling water, feed at the bottom and leave at the top so the coil stays full and vents air.
When is an internal coil the right choice?
Add an internal coil when the external area is not enough and an external loop is not justified. It suits:
- Clean, low-viscosity batches that do not leave deposits
- Duties where rapid cooling is needed, such as quench or crash cooling for crystallisation
- Large vessels where the jacket alone cannot meet the peak load
Avoid it when:
- The batch is a slurry, sticky, polymerising or viscous. Material builds up on the coil and is hard to clean.
- The chemistry is water-reactive (for example chlorosulphonic acid, oleum, acid chlorides or Grignard steps) and the coil would carry water. A pinhole leak becomes a hazard. Use a non-reactive thermal fluid in the coil, or keep the cooling outside the vessel.
- Product changeovers need fast, validated cleaning, as in multi-product pharma plants.
Internal coils also act as baffles and change the flow pattern. Allow for this when selecting the impeller and power.
Steam, thermal oil, water and brine: which arrangement for which utility?
| Utility | Typical range | Preferred arrangement | Notes |
|---|---|---|---|
| Low-pressure steam | up to about 3 bar(g) | Jacket or limpet | Good condensate drainage is essential |
| Medium-pressure steam | 3–10 bar(g) | Limpet coil | Jacket needs a thick outer shell at higher pressure |
| Thermal oil | 150–300 °C | Limpet coil | High velocity avoids oil cracking and hot spots |
| Cooling water | 30–37 °C | Limpet, or jacket with agitation nozzles | Watch fouling and scaling on the water side |
| Chilled water | 5–10 °C | Limpet or baffled jacket | Insulate to stop condensation |
| Brine or glycol | −20 to 0 °C | Limpet or jacket | Higher viscosity, so check velocity and pressure drop |
Many reactors use several utilities in one batch: steam to heat, water to cool, brine for the final cooling. Provide a sequence to drain and blow out each utility before switching, and avoid sudden temperature swings on glass-lined vessels, which have strict thermal-shock limits. Our comparison of glass-lined vs Hastelloy vs SS316L reactors covers material limits. For hot water heating loops, see the hot water generation system.
Common problems and fixes
| Problem | Likely cause | Fix |
|---|---|---|
| Slow heating with steam | Condensate logging; failed trap; air in jacket | Trap per zone, vent at top, check trap sizing |
| Slow cooling | Low coolant velocity; scaled water side; warm cooling water | Agitation nozzles or limpet; descale; check cooling tower |
| Hot spots and oil cracking | Low thermal oil velocity in plain jacket | Use limpet coil at 1.5–2.5 m/s |
| Temperature overshoot | Large thermal mass of jacket; slow control | Cascade control on jacket outlet temperature; split-range valves |
| Crust on internal coil | Unsuitable product for internal coil | Remove coil, use external area or loop |
| Cracks at limpet welds | Thermal cycling; poor weld profile | Full-penetration or properly profiled welds, stress check for cycling |
Conclusion
A conventional jacket is enough for many low-pressure and glass-lined duties, a limpet coil is the workhorse for steel reactors on thermal oil, steam and tight cooling duties, and an internal coil adds area when external surface runs out, provided the product is clean and the chemistry tolerates it. Send your batch size, heat release and utility data to SEMCO and our engineers will check the heat balance and recommend the right arrangement.
Frequently asked questions
What is a limpet coil on a reactor?
A limpet coil, or half-pipe coil, is a pipe cut lengthwise and welded in a spiral around the outside of a reactor shell. The heating or cooling fluid flows through it at a controlled velocity, which gives better heat transfer than a plain jacket and lets the coil take higher pressures.
Which is better, a jacket or a limpet coil?
For thermal oil, higher-pressure steam or duties where heat transfer is tight, a limpet coil is usually better. For glass-lined reactors, low-pressure duties and products where cost matters most, a conventional jacket is often enough, especially with agitation nozzles or a spiral baffle to raise velocity.
When should I add an internal coil to a reactor?
When the jacket or limpet area is not enough for the heat load, which often happens on scale-up because area per unit volume falls as the reactor grows. Avoid internal coils for slurries, sticky or viscous products, and water-reactive chemistry if the coil carries water.
Why does a reactor that worked at pilot scale overheat at plant scale?
Jacket area grows with the square of the diameter while volume grows with the cube, so a larger reactor has less heat transfer area per cubic metre. A 20 KL reactor has roughly 40% less jacket area per cubic metre than a 5 KL reactor of similar shape.
Free tools for this topic
Related Engineering Articles
- LABSA & SLES Production Plant Engineering: SO3 Falling Film Reactor Design & Process Optimization
- Alkyd & Polyester Resin Plant Reactor Design: Polycondensation Kinetics, Thermal Oil Heating & Thinning Systems
- Glass-Lined vs. Hastelloy C-276 vs. SS316L Chemical Reactors: Technical Buyer's Guide
- Turnkey Reaction Engineering for Fatty Alcohol Ethoxylates: Kinetics, EO Safety, and Reactor Scale-Up