# How a chemical process works: reactions, separations, catalysts and flowsheets: Unit operations, their energy bills, and batch versus continuous

Source page: https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/p4/

[Understanding Chemicals](https://www.materialinsights.org/learn/chemicals/) / [Module 2: How a chemical process works: reactions, separations, catalysts and flowsheets](https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/) Part 4 of 5

## Unit operations, their energy bills, and batch versus continuous

Around 1910 engineers at MIT noticed that the physical steps in chemical plants repeat across industries: heat transfer, fluid flow, distillation, filtration and so on. They called them unit operations. Most of a plant's energy goes into these steps rather than into the reaction. How they are run also depends on scale: tonnes a year in batches, or hundreds of thousands of tonnes continuously.

### 4.1 Every separation exploits a difference in a physical property

A separation works by finding something that differs between the components and using it. Distillation uses differences in boiling point. Absorption dissolves one gas selectively into a liquid, as when carbon dioxide is scrubbed from ammonia synthesis gas. Extraction dissolves one component of a liquid into a second, immiscible liquid; Shreve describes lubricating oil treated with furfural at 50 to 140 C, giving a solvent-rich extract holding the impurities and a cleaner raffinate. Adsorption holds molecules on a solid surface such as activated carbon or molecular sieves. Crystallisation uses differences in freezing point or solubility. Filtration and centrifuging separate solids from liquids, and drying removes the last liquid from a solid.

The xylenes, feedstock for Paraxylene and then Purified terephthalic acid, show how the property chosen decides the cost. Wittcoff gives boiling points for the four C8 aromatics all within 9 C of each other. Ortho-xylene, boiling 5 C higher, can be split off by distillation, but only on a huge column of 150 to 200 plates with heavy reflux. Para- and meta-xylene are almost inseparable this way. Their freezing points, though, differ widely, so the older route crystallises para-xylene by cooling the dried mixture progressively to about minus 68 C, then melts, washes and recrystallises to 99.5% purity. Wittcoff notes that this needs a lot of refrigeration energy and the solid fouls the cooling walls. The newer route uses molecular sieve adsorption in a simulated moving bed.

Kent's green engineering heuristics add two rules. Do easy separations first, so that the hardest step handles the smallest stream. And do not reach for distillation by habit when a membrane or other method suits small volatility differences better.

Data The main unit operations and where their energy goes

| Operation | Property exploited | Example from the sources | Main energy demand |
|---|---|---|---|
| Distillation | boiling point | C2 and C3 splitters in an ethylene plant | steam to reboilers; cooling or refrigeration to condensers |
| Absorption | solubility of a gas in a liquid | CO2 removal from ammonia syngas | steam to regenerate the solvent |
| Extraction | solubility between two liquids | furfural treatment of lube oil | distilling the solvent back out |
| Adsorption | affinity for a solid surface | molecular sieves for para-xylene; PSA for hydrogen | pressure or temperature swing to release |
| Crystallisation | freezing point or solubility | para-xylene at low temperature | refrigeration or evaporation |
| Evaporation | volatile solvent versus dissolved solid | sugar juice, caustic soda, salt brine | steam, reduced by multiple effects |
| Filtration and drying | particle size; volatility of residual liquid | wax removal; fine chemical isolation | pumping power; heat to evaporate liquid |
| Heat exchange | temperature difference | gas-gas exchangers in acid plants | pumping; losses grow as surfaces foul |
| Compression |  | ammonia synthesis gas | electricity or steam turbines |

Compiled from the sources cited; the energy column summarises their descriptions rather than giving numbers.
 Source: Shreve (1984), Ch 37; Wittcoff (2004), Ch 9; Kent (2012), Ch 10 and Ch 26, Shreve pdf pp. 751-752; Wittcoff pdf pp. 346-347; Kent pdf pp. 327, 1022

- Shreve (1984), Ch 1 and Ch 37 Separation processes · pdf pp. 19, 751-752
- Wittcoff (2004), Ch 9 Chemicals from xylenes · pdf pp. 346-348
- Kent (2012), Ch 5 Green Engineering · pdf p. 176

### 4.2 Distillation and evaporation are paid for in steam

Distillation separates liquids by boiling the mixture and letting the vapour rise through a column fitted with trays or packing. Liquid flows down, vapour flows up, and at each tray they exchange molecules: the lighter component concentrates at the top, the heavier at the bottom. A reboiler at the base supplies heat, usually from steam. A condenser at the top removes heat, usually into cooling water. Part of the condensed liquid is returned to the column as reflux. The closer the boiling points, the more trays and the more reflux are needed, and reflux means boiling the same liquid again and again. That is why distillation is energy hungry.

Kent's flowsheet of a naphtha cracker shows how extreme this gets. After cracking at about 800 C the gas is compressed and chilled to minus 100 C at 3.5 MPa to strip out hydrogen and methane. Then comes a train of columns: a deethaniser, an ethane-ethylene splitter of 115 trays, a depropaniser and a propane-propylene splitter of 200 trays split across two columns. Much of the condensing has to be done by refrigeration, which costs compressor power. This is where much of the energy of Ethylene and Propylene production goes.

When components form an azeotrope, a mixture that boils without changing composition, plain distillation stops working. Shreve describes adding a solvent (extractive distillation, such as furfural to separate butenes from butanes) or an entrainer (azeotropic distillation, such as benzene to dry 95% ethanol).

Evaporation boils water off a solution of a solid. The trick for saving energy is multiple effects: vapour from one vessel heats the next, which runs at lower pressure so that it boils at a lower temperature. Kent says a four-effect evaporator evaporates about 4 lb of water per lb of steam. Vapour recompression, used in salt plants, does a similar job with a compressor.

Process flowchart Naphtha cracker: cracking followed by a long separation train Inputs **Naphtha**about 2.7 t per t ethylene (indicative) **Dilution steam**mixed with feed **Fuel**fires the cracking furnace Process steps 1 **Cracking furnace**

about 800 C; large molecules break into ethylene, propylene and others
 2 **Quench and primary fractionator**

rapid cooling to stop reactions; heavy ends and pyrolysis gasoline removed
 3 **Multi-stage compression and caustic wash**

gas compressed; acid gases removed
 4 **Chilling train and demethaniser (30-50 trays)**

down to about -100 C at 3.5 MPa; hydrogen and methane taken off
 5 **Deethaniser (40-50 trays) and acetylene hydrogenation**

C2 cut separated; traces of acetylene converted
 6 **Ethane-ethylene splitter (115 trays)**

polymer-grade ethylene at about -33 C, 1.9 MPa; ethane recycled to the furnace
 7 **Depropaniser (55-60 trays) and propane-propylene splitter (200 trays in two columns)**

propylene product; propane recycled
 8 **Debutaniser**

C4 cut and C5+ cut separated
 Outputs **Ethylene**about 31.5% of naphtha by weight **Propylene**about 15.3% **Methane-rich fuel gas**about 14.6% **Butadiene, C4s, aromatics, fuel oil**balance

Yields are typical naphtha cracking yields from Kent Table 10.9; naphtha per tonne of ethylene from Brennan's indicative table.
 Source: Kent (2012), Ch 10 Synthetic Organic Chemicals, Fig. 10.12 and Table 10.9; Brennan (2020), Ch 4 Table 4.2, Kent pdf p. 327; Brennan pdf p. 106  Worked example

#### Steam for concentrating sugar juice

Kent describes clarified cane juice at about 85% water being concentrated to a syrup of about 65% solids. Take 100 tonnes of juice: it holds 15 tonnes of solids. A syrup at 65% solids weighs 15 divided by 0.65, or 23.1 tonnes, so 76.9 tonnes of water must be boiled off. In a quadruple-effect evaporator at 4 tonnes of water per tonne of steam, that takes about 19 tonnes of steam. By the same rule, a single effect would need roughly four times as much. Kent adds that retrofitting older evaporators with plate packs can cut a cane mill's total steam demand by almost 30%.

- Kent (2012), Ch 10 Synthetic Organic Chemicals · pdf p. 327
- Shreve (1984), Ch 37 Separation processes · pdf p. 751
- Kent (2012), Ch 35 Sugar, Evaporation · pdf p. 1409
- Kent (2012), Ch 26 Salt · pdf p. 1022

### 4.3 Heat exchange and compression decide the energy bill

Heat exchangers pass heat from a hot stream to a cold one through a metal wall without mixing them. In a well-designed plant, hot reactor outlet gas preheats cold feed, raises steam or reheats a stream leaving an absorber. Shreve notes that exchanger performance falls day by day as surfaces foul, so refineries check it daily against records. Kent's green engineering checklist asks of every design whether heat integration networks are in place.

The sulphuric acid plant shows what heat recovery is worth. About 97% of its energy input comes from burning sulphur and 3% from electricity to drive the blower. Kent says most plants recover 55 to 60% of the energy as high-pressure steam at 40 to 60 bar and 400 to 480 C, but about 40% is lost as warm water from the acid coolers. Better economisers and hotter absorbers have raised recovery to about 70% in some plants. Shreve calls a modern acid plant an electrical generating system.

Compression is the other big energy user. Gas must be squeezed to synthesis pressure, and the work goes in as electricity or through steam turbines. Shreve describes the 1970s switch in ammonia plants from reciprocating to centrifugal compressors, which allowed single-train plants of up to 1,650 tonnes a day and cost cuts approaching 50%. Energy per tonne fell from 55.9 GJ in 1950 to 40 to 45 GJ for plants operating in the early 1980s, with designs available at 27 GJ.

Nitric acid plants face a pressure dilemma. Kent explains that ammonia oxidation prefers low pressure while absorbing nitrogen dioxide into water prefers high pressure. Dual-pressure plants add a compressor between the two stages, spending capital and power to gain efficiency at both ends.

Data Raw material and utility use per tonne of nitric acid, three plant designs

| Item (per t of acid) | High-High (11 bar) | Medium-Medium (9 bar) | Medium-High (4.5/12 bar) |
|---|---|---|---|
| Typical gauze run time | 70 days | 120 days | 210 days |
| Ammonia | 0.291 t | 0.285 t | 0.279 t |
| Platinum (net of recovery) | 50 mg | 35 mg | 30 mg |
| Power | 1 kWh | 2 kWh | 2 kWh |
| High-pressure steam exported | 0.51 t | 0.60 t | 0.54 t |
| Low-pressure steam imported | 0.27 t | 0.21 t | 0.06 t |
| Cooling water | 130 m3 | 120 m3 | 110 m3 |

Ammonia includes NOx abatement to 200 ppmv. Add 25 kW for HVAC, lighting and instruments. Export steam at 42 kg/cm2 g and 400 C; cooling water at 11 C temperature rise. Power is low because tail-gas expanders recover energy.
 Source: Kent (2012), Ch 22 Synthetic Nitrogen Products, Table 22.18 (Monsanto Enviro-Chem), pdf p. 915  In the data

#### Pricing the nitric acid utilities

Apply Brennan's 2018 utility prices to the dual-pressure column of Kent's table. Exported high-pressure steam, 0.54 t at $10 to $16, earns a credit of about $5 to $9 per tonne of acid. Imported low-pressure steam costs under $1. Power, 2 kWh at $70 to $140 per MWh, costs under 30 cents. Cooling water, 110 m3 at $0.10 to $0.20, costs $11 to $22, the largest utility line. All of these are small next to the 0.279 tonnes of ammonia, which is why ammonia efficiency drives the choice of design.

- Kent (2012), Ch 25 Sulfur and Sulfuric Acid · pdf p. 1017
- Shreve (1984), Ch 18 and Ch 37 · pdf pp. 325-328, 751
- Kent (2012), Ch 22 Nitric acid · pdf pp. 913-915
- Brennan (2020), Ch 4 Table 4.4 · pdf p. 110
- Kent (2012), Ch 5 Green Engineering · pdf p. 177

### 4.4 Scale and product type decide between batch and continuous

In a batch process a vessel is filled, the reaction is run, the vessel is emptied and cleaned, and the cycle repeats. In a continuous process feed flows in and product flows out without stopping, often for years. Shreve sets out the trade-off. Batches can be measured very precisely and the same equipment can make many products, but temperature control can be troublesome. Continuous plants need smaller, cheaper equipment per tonne, hold far less material in process, and give more uniform product, but they need good instruments and control. Wittcoff describes how cheap olefin feedstocks after World War II moved organic chemistry from batch processes to continuous 'hot tube' plants making hundreds of thousands of tonnes a year.

Fine chemicals and pharmaceuticals stay batch because volumes are small and products change. Pollak describes multipurpose plants with stainless steel and glass-lined reactors, usually 4 to 6 cubic metres, working from minus 20 to 200 C and up to 3 bar. As a rule of thumb, a cubic metre of reactor makes 15 to 30 tonnes a year of a one-step product. His example plant has six 4 cubic metre reactors, two filters and two dryers, and cost $21 million, or $0.9 million per cubic metre. By that rule its 24 cubic metres make 360 to 720 tonnes a year of a single step.

Compare Shreve's 2,800 tonne-per-day sulphuric acid plant of 1982: at 330 days a year that is about 924,000 tonnes, more than a thousand times larger. Pollak also lists limits on batch size: campaigns shorter than about 10 working days waste time on changeovers, big batches take too long to heat and cool, and a single batch of an expensive product can be worth more than $1 million. Kent adds that very large vessels are harder to control and more prone to runaway reactions.

Data Batch versus continuous at a glance

| Feature | Multipurpose batch plant | Continuous single-product plant |
|---|---|---|
| Typical output | 360-720 t/yr per step (24 m3 reactor volume) | about 924,000 t/yr (2,800 t/day acid plant) |
| Products | many, made in campaigns | one, or a fixed slate |
| Equipment | 4-6 m3 stirred vessels, filters, dryers | dedicated reactors, columns, exchangers |
| Capital intensity | about $29,000-58,000 per annual tonne (2010) | about $2,500 per annual tonne for a small 40,000 t/yr acetic acid plant (2018) |
| Strengths | flexibility, precise dosing | low cost per tonne, uniform product, small inventory per tonne |
| Weaknesses | changeover and cleaning time, heat transfer limits | needs large stable market and strong instrumentation |

Capital per tonne is calculated here from Pollak's $21 million plant and Brennan's $100 million example; the years and scopes differ, so compare orders of magnitude only.
 Source: Pollak (2011), Ch 5 and Ch 7; Shreve (1984), Ch 19; Brennan (2020), Ch 4, Pollak pdf pp. 74-75, 97-98; Shreve pdf p. 347; Brennan pdf p. 126

- Shreve (1984), Ch 2 Batch versus continuous processing · pdf pp. 25-26
- Wittcoff (2004), Ch 16 introduction · pdf p. 553
- Pollak (2011), Ch 5 Facilities and plants; Ch 7 Cost calculation · pdf pp. 74-75, 97-99
- Kent (2012), Ch 2 Equipment sizing · pdf p. 72

### See it in the table
[H₂O Water](https://www.materialinsights.org/commodities/water/)[St Steel](https://www.materialinsights.org/commodities/steel/)[NaCl Salt](https://www.materialinsights.org/commodities/salt/)[H2SO4 Sulphuric acid](https://www.materialinsights.org/chemicals/sulphuric_acid/)[HNO3 Nitric acid](https://www.materialinsights.org/chemicals/nitric_acid/)[AC Activated carbon](https://www.materialinsights.org/chemicals/activated_carbon/)[H2 Hydrogen](https://www.materialinsights.org/chemicals/hydrogen/)[CO2 Carbon dioxide](https://www.materialinsights.org/chemicals/carbon_dioxide/)[NH3 Ammonia](https://www.materialinsights.org/chemicals/ammonia/)[C2H4 Ethylene](https://www.materialinsights.org/chemicals/ethylene/)[C3H6 Propylene](https://www.materialinsights.org/chemicals/propylene/)[Bz Benzene](https://www.materialinsights.org/chemicals/benzene/)[PX Paraxylene](https://www.materialinsights.org/chemicals/paraxylene/)[PTA Purified terephthalic acid](https://www.materialinsights.org/chemicals/pta/)

### Check your understanding

Why is para-xylene not separated from meta-xylene by ordinary distillation?
 Wittcoff gives boiling points of 138.3 and 139.1 C. Crystallisation by freezing point, or adsorption on molecular sieves, works far better.

In Kent's nitric acid utility table, which utility costs the most per tonne at Brennan's prices?
 110 to 130 cubic metres at $0.10 to $0.20 is $11 to $26 per tonne, far above power (1 to 2 kWh) or imported steam.

Which factor does Pollak give as a limit on making batch reactors bigger?
 Heat transfer area grows more slowly than volume, so big batches spend too long heating and cooling. Short campaigns, inventory and the value at risk in one batch also cap size.

[← Part 3: Catalysts: what they change and what they cannot](https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/p3/)[Mark read · Part 5: Reading a flowsheet, and safety as a design input →](https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/p5/)

## Other sections

- [Module overview](https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/index.md)
- [The anatomy of a chemical plant](https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/p1/index.md)
- [Reaction basics that decide the economics](https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/p2/index.md)
- [Catalysts: what they change and what they cannot](https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/p3/index.md)
- [Reading a flowsheet, and safety as a design input](https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/p5/index.md)
- [Key terms, quiz and sources](https://www.materialinsights.org/learn/chemicals/how-chemical-processes-work/review/index.md)