Separation Processes

Separation processes are the part of chemical engineering that takes a mixture and splits it into its components.

11 frameworks, from 1900 to today · Chemical Engineering · Engineering

still a working view ended · each leads to its entry below

Overview

Separation processes are the part of chemical engineering that takes a mixture and splits it into its components1. Almost every industrial chemical plant is, in bulk, a separation plant: crude oil must be fractionated into fuels and feedstocks, fermentation broths must be stripped of water and purified, air must be split into oxygen and nitrogen, and ores must be leached and refined into metals. What is at stake is enormous. Distillation alone is often estimated to account for a large share of industrial energy use worldwide, and the economics of most chemical products depend more on how cheaply the product can be purified than on how cleverly it was made. The history of the subfield is the story of how engineers learned to design these splits deliberately—first by treating each separation as a craft, then by finding general principles that could be taught, calculated, and eventually simulated.

Distillation Before Theory: The Craft Tradition

Distillation is ancient. Hellenistic Alexandria produced distilled water and essential oils, and by the medieval period alchemists and physicians across the Islamic world and Europe were refining the apparatus—the alembic, the retort, the water-cooled condenser—that would remain recognizable for centuries. The practical knowledge was empirical: how to pack a still, how to control the fire, how to judge a cut by taste or smell. Distillation of alcohol, of turpentine, of mercury, and later of coal tar and petroleum grew into large industries without a general theory of what was happening inside the column.

The decisive change came with the petroleum industry in the nineteenth and early twentieth centuries. Continuous fractionating columns—tall vessels with trays on which vapor and liquid contact repeatedly—replaced batch stills, and their scale made guesswork expensive. Engineers needed to predict how many trays a column would need and how pure the products would be.

Unit Operations: Finding the General in the Particular

The intellectual foundation of the whole field was laid in the early twentieth century by Arthur D. Little and, above all, by Warren K. Lewis and his colleagues at MIT. Their idea, articulated in the 1910s and 1920s, was the unit operation: the claim that despite the bewildering variety of industries, the individual steps of processing fall into a small number of recurring types—distillation, absorption, extraction, filtration, drying, crystallization, evaporation—and that each type obeys the same physical principles regardless of what is being separated. This was a profound simplification. It meant a student could learn distillation once and apply it to whiskey, gasoline, or acetone.

The unit-operations framework, codified in textbooks such as Walker, Lewis, and McAdams's Principles of Chemical Engineering (1923), turned separation from a collection of trades into a discipline with a shared vocabulary2. It also set the agenda: if separations are general, they can be analyzed with the tools of physics and mathematics.

Equilibrium Stages and the Design of Columns

The first great analytical tool was the equilibrium stage. The idea is to model a tray or a section of packing as a place where vapor and liquid meet, mix, and leave in equilibrium with each other—so that the compositions of the two streams are related by thermodynamic data rather than by the details of how they contacted3. A column then becomes a stack of such stages, and the designer's job is to count how many are needed to reach a target purity.

The graphical method for doing this—the McCabe–Thiele diagram, published by Warren McCabe and Ernest Thiele in 1925—let an engineer plot the operating line (the material balance relating liquid and vapor flows) against the equilibrium curve and step off stages with a ruler. It was simple, visual, and remarkably durable; it is still taught. For absorption and stripping, the Kremser equations (from the work of Arthur Kremser in the 1930s) gave closed-form stage counts. These methods made column design a calculable engineering activity, and they dominated practice for decades.

When Simple Distillation Fails: Azeotropes and Extractive Distillation

Equilibrium-stage design exposed a hard problem. Many mixtures form an azeotrope—a composition at which vapor and liquid have the same composition, so ordinary distillation cannot pass through it4. Ethanol and water are the classic case: at about 95% ethanol by weight, the mixture boils without changing composition, which is why pure alcohol requires something other than simple distillation.

Engineers answered with altered separations. Azeotropic distillation adds a third component (an entrainer) that shifts the equilibrium and carries one component overhead, as in the benzene entrainer historically used to dry ethanol5. Extractive distillation adds a heavy solvent that changes the relative volatility of the two key components without itself vaporizing, as in the use of solvents to separate close-boiling hydrocarbons or to recover aromatics6. These methods required better thermodynamic data and more careful solvent selection, and they pushed designers toward systematic reasoning about how a solvent changes the phase behavior of a mixture—a theme that would grow into computer-aided molecular design much later.

Solvent Extraction and the Hydrometallurgical Tradition

Not every separation suits distillation. When components are heat-sensitive, close-boiling, or present in dilute aqueous solution, contacting the mixture with a second liquid that preferentially dissolves one component can be far cheaper. Solvent extraction (liquid–liquid extraction) developed alongside the unit-operations framework, with mixer-settlers and packed or pulsed columns as the workhorses.

Its most consequential application was in metallurgy. Hydrometallurgy—extracting metals from ores by leaching them into solution and then recovering them—grew from nineteenth-century cyanide gold leaching into a major industry. The recovery step often used solvent extraction or cementation, and later ion exchange. The field matured through the twentieth century with the processing of copper, uranium, and rare earths, and with the development of selective extractants designed for particular metal ions. Extraction also became central to nuclear fuel reprocessing, where the PUREX process (using tributyl phosphate) separates uranium and plutonium from fission products—a separation whose stakes are political as well as technical.

Thermodynamics as the Foundation

All of these methods rest on knowing how mixtures behave. The equilibrium-stage approach is only as good as the vapor–liquid equilibrium (VLE) data and models behind it7. Early design relied on measured data and simple correlations, but the middle of the twentieth century brought a systematic framework: fugacity and activity coefficients as the language for non-ideal mixtures, and equations of state for vapor and supercritical phases.

Key developments included the Margules and van Laar activity-coefficient models (early twentieth century), the concept of the regular solution (Joel Hildebrand, 1920s), and later the local-composition models—NRTL (Renon and Prausnitz, 1968) and UNIQUAC (Abrams and Prausnitz, 1975)—which could represent strongly non-ideal mixtures like alcohol–water. The UNIFAC group-contribution method (Fredenslund and colleagues, 1975) allowed prediction of activity coefficients from molecular structure when data were missing. These models, combined with equations of state such as the Peng–Robinson equation (1976), gave designers the ability to compute phase equilibria for mixtures they had never measured. Thermodynamic modeling became inseparable from separation design: a column could not be designed without a reliable equilibrium model, and a bad model could make a good column fail.

Rate-Based Design: Beyond the Equilibrium Stage

The equilibrium-stage model has an obvious weakness: real trays and packings do not reach equilibrium. The stage is an idealization, corrected after the fact by an efficiency factor—a fudge that hid the actual physics of how fast mass crosses from vapor to liquid.

From the 1980s onward, a rate-based (or non-equilibrium) approach gained ground. Instead of assuming equilibrium, it models the actual rates of mass and heat transfer across the interface between phases, using mass-transfer coefficients and interfacial area. This required understanding of mass transfer—how a component diffuses through a film or a turbulent boundary layer—which had been developed earlier by figures such as Thomas Chilton and Allan Colburn (the Chilton–Colburn analogy, 1934) and codified in the transport-phenomena tradition. Rate-based models, championed in packed-column design and supported by computational power, allowed more accurate prediction of column performance, especially for structured packings and reactive distillation where equilibrium assumptions break down badly.

Adsorption and Pressure-Swing Separation

Some separations are better done by sticking one component to a solid surface. Adsorption—the clinging of molecules to a solid—had long been used for decolorizing and drying, but the modern era began with synthetic zeolites (molecular sieves) in the mid-twentieth century. Zeolites have uniform pores that admit some molecules and exclude others, enabling shape-selective separation8.

The great industrial breakthrough was pressure-swing adsorption (PSA), developed commercially in the 1960s (with Charles Skarstrom's process patents). The principle is elegant: adsorb a component at high pressure, then release it by lowering the pressure, cycling between beds so the process runs continuously. PSA transformed industrial gas production—hydrogen purification and, above all, the separation of air into oxygen and nitrogen—and it made on-site gas generation practical. Temperature-swing adsorption and simulated moving-bed chromatography (used for separating xylene isomers and sugars) followed related logic. Adsorption brought separation into the realm of cyclic, transient processes, which required new design methods beyond the steady-state equilibrium stage.

Membrane Separation

A membrane is a barrier that lets some molecules through faster than others. Membranes had been laboratory curiosities since the nineteenth century, but industrial membrane separation arrived in the 1960s and 1970s. The key developments were the asymmetric and thin-film composite membranes (associated with Sidney Loeb and Srinivasa Sourirajan's work on reverse osmosis in the early 1960s), which made desalination and gas separation economically viable.

Membranes spread rapidly: reverse osmosis for water desalination, ultrafiltration and microfiltration for concentrating and purifying, gas-separation membranes for hydrogen recovery and nitrogen generation, and later pervaporation and electrodialysis. The appeal was that membranes avoid the energy cost of phase change—no boiling, no condensing—and can be modular and compact. Their design, however, required understanding of diffusion through polymers and of concentration polarization (the buildup of solute at the membrane surface), which tied membrane work closely to mass-transfer theory.

Cryogenic Separation

When gases must be separated at very high purity and very large scale, cooling them until they liquefy and then distilling them remains the method of choice9. Cryogenic separation—distillation at temperatures far below ambient—grew with the industrial gas industry in the early twentieth century. Carl von Linde's air-liquefaction process (patented in the 1890s) and Georges Claude's later improvements made it possible to liquefy air and separate oxygen and nitrogen by fractional distillation. Cryogenic plants became the backbone of industrial gas supply and later of natural-gas processing, where they recover helium and separate light hydrocarbons. The engineering challenges—insulation, refrigeration cycles, materials that stay tough at low temperature—made cryogenics a distinct specialty within separations.

Simulation and Flowsheet Modeling

As the number of possible process configurations grew, hand calculation became untenable. The computer changed design practice. From the late 1950s and 1960s, chemical engineers built process simulators—programs that model a whole flowsheet as a network of unit operations, each with its own model, linked by material and energy balances. Early systems such as the Monsanto-developed FLOWTRAN and later commercial packages (Aspen Plus, HYSYS, and others) let engineers compute the behavior of a plant before building it.

For separations, simulation meant that rigorous stage-by-stage calculations—solving the equilibrium and mass-balance equations for every tray simultaneously—became routine. It also meant that thermodynamic models, rate-based models, and unit-operation models had to be integrated into a single computational framework. Simulation shifted the designer's work from hand-stepping McCabe–Thiele diagrams to building and validating models, and it made the optimization of complex separation sequences (which order to separate components, which method to use for each) a tractable problem.

Process Intensification: Doing More in Less

The most recent strand of the field's history is process intensification: the drive to achieve a given separation in smaller, more efficient equipment, often by combining functions that were traditionally separate. The term gained currency from the 1980s onward, associated with figures such as Colin Ramshaw, who championed compact devices like the HiGee rotating packed bed (using centrifugal force to intensify mass transfer) and the static mixer.

Intensification takes many forms: reactive distillation, in which reaction and separation happen in the same vessel so that equilibrium-limited reactions are driven forward by removing product; membrane reactors; dividing-wall columns that combine two distillations in one shell; and microchannel devices. The motivation is economic and environmental—less capital, less energy, less solvent. Intensification also forced a rethinking of the unit-operation framework itself: if a single device performs reaction and separation simultaneously, the clean division into separate operations becomes a design choice rather than a natural boundary.

The Present: A Computational, Energy-Conscious Field

Today's separation practice descends directly from these layers. Equilibrium-stage and rate-based models coexist in commercial simulators; thermodynamic models like NRTL, UNIQUAC, and cubic equations of state are standard tools; membranes, adsorption, extraction, and cryogenics each have mature design methods. The dominant concerns have shifted toward energy and sustainability. Distillation remains the largest energy consumer among separation methods, and much current work targets its replacement or improvement—through heat integration, dividing-wall columns, hybrid membrane–distillation systems, and intensified equipment.

Open questions remain substantial. Designing solvents and materials (membranes, adsorbents, extractants) from first principles rather than by trial is an active goal, drawing on molecular simulation and machine learning. Separations of ever more complex mixtures—biological products, rare earths, lithium for batteries, carbon dioxide from flue gas—pose problems the classical methods handle imperfectly. And the integration of separations with renewable and variable energy supplies is reshaping how plants are designed and operated. The field that began with alchemists' stills now runs on thermodynamic models and flowsheet simulators, but its central task is unchanged: taking a mixture apart, cleanly and at a cost the world can afford.

References

  1. Separation process, Wikipedia
  2. Unit operation (History), Wikipedia
  3. Theoretical plate, Wikipedia
  4. Azeotrope, Wikipedia
  5. Azeotropic distillation, Wikipedia
  6. Extractive distillation, Wikipedia
  7. Vapor–liquid equilibrium, Wikipedia
  8. Molecular sieve, Wikipedia
  9. Air separation, Wikipedia

The frameworks of Separation Processes

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1900-Present

    Cryogenic Separation

    How can gases be split by cooling them to liquid?

    Fly to its tower in Nousfarer

  2. 1915-Present

    Unit operation

    What basic steps recur across all chemical processes?

    Fly to its tower in Nousfarer

  3. 1925-Present

    Theoretical plate

    How many stages does a column need?

    Fly to its tower in Nousfarer

  4. 1925-Present

    Thermodynamic Phase-Equilibrium Modeling

    How do mixtures distribute between vapor and liquid?

    Fly to its tower in Nousfarer

  5. 1930-Present

    Azeotropic and Extractive Distillation

    How can mixtures that resist distillation be split?

    Fly to its tower in Nousfarer

  6. 1930-Present

    Solvent Extraction and Hydrometallurgy

    How can a dissolved substance be pulled into another liquid?

    Fly to its tower in Nousfarer

  7. 1960-Present

    Adsorption and Cyclic Swing Separation

    How do porous solids trap and release selected molecules?

    Fly to its tower in Nousfarer

  8. 1960-Present

    Membrane technology

    How can a barrier selectively pass one component?

    Fly to its tower in Nousfarer

  9. 1960-Present

    Process simulation

    Can a whole plant be modeled on a computer?

    Fly to its tower in Nousfarer

  10. 1980-Present

    Rate-Based (Non-Equilibrium) Design

    How fast does mass actually cross between phases?

    Fly to its tower in Nousfarer

  11. 1985-Present

    Process Intensification

    Can separation equipment be made smaller and more efficient?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Cryogenic SeparationUnit operationTheoretical plateThermodynamic Phase-Equilibrium Model…Azeotropic and Extractive DistillationSolvent Extraction and HydrometallurgyAdsorption and Cyclic Swing SeparationMembrane technologyProcess simulationRate-Based (Non-Equilibrium) DesignProcess Intensification

Beyond the field

Where Separation Processes's ideas came from in other fields, and what they led to there: the strongest 12 of 35 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Separation Processes is a planet in Nousfarer

Its city has a tower for each of these 11 frameworks. Land, walk in, and climb one concept at a time.

Land on Separation Processes

Related fields

Fields its ideas are linked to

More in Chemical Engineering

See where Separation Processes sits on the Map of Human Knowledge