The New Distillation Playbook: When to Go Conventional, Add a Solvent, or Swing the Pressure

Conventional distillation remains the workhorse for mixtures with sufficiently different boiling points, but the economics flip fast once relative volatility (α, a measure of ease of separation) drops toward 1.05–1.10. When α is small or an azeotrope (a constant‑boiling mixture) forms, tray counts and reflux shoot up — and energy use explodes (www.scielo.br) (www.scielo.br). Ethanol–water is the classic: it forms an 87.2% ethanol azeotrope at 1 atm, so conventional distillation alone cannot yield anhydrous ethanol (www.scielo.br).

The stakes are big. Conventional columns already dominate plant costs: roughly one‑third of a petrochemical or chemical plant’s capital and over 50% of its energy are spent on distillation (www.scielo.br). A recent analysis cites 33% of CAPEX and >50% of energy in typical plants (www.scielo.br). The good news: rigorous column design and heat integration can cut energy use by 20–50% (www.scielo.br).

But when mixtures are close‑boiling (α barely above ~1.1), traditional towers become extremely tall or numerous. That is when advanced separation techniques — extractive distillation (adding a selective solvent/entrainer) or pressure‑swing distillation (changing pressure to move an azeotrope) — come into play.

Conventional distillation boundaries

In simple or fractionating columns, absent an azeotrope and with α≥1.05–1.10, one column can achieve high recovery. If an azeotrope is present or α is small, the column demands escalate quickly and energy cost “explodes” as noted in the literature (www.scielo.br) (www.scielo.br). Energy integration can still help: case studies show 20–50% energy cuts with rigorous design and heat integration (www.scielo.br).

Extractive distillation fundamentals

Extractive distillation (ED) introduces a high‑boiling entrainer (a selective solvent) to alter relative volatilities so a target component is more easily removed. It typically uses two columns: a main column where the entrainer is added and a solvent recovery column. ED is widely applied to azeotropic or close‑boiling systems like ethanol–water dehydration (using glycols, DMSO, or ionic liquids), isopropanol–water, and acetone–methanol (with water or other polar entrainers). With a suitable solvent, ED can reach very high purities (often >99%) (pmc.ncbi.nlm.nih.gov).

Solvent selection is critical: the entrainer should have high selectivity, low volatility (to exit in bottoms), chemical stability and non‑corrosivity, and reasonable cost (pmc.ncbi.nlm.nih.gov). In practice, solvents like DMSO, glycerol, ethylene glycol, or “green” ionic liquids are used. One study found the ionic liquid 1‑butyl‑3‑methylimidazolium bromide ([BMIM][Br]) produces a strong “salting‑out” effect on the ethanol–water azeotrope, enabling nearly azeotropic ethanol with modest reflux (jurnal.ugm.ac.id). That simulation reported an optimal design of 28 theoretical stages and a reflux ratio of ~1.5 to reach ~99.9% ethanol at minimum energy (jurnal.ugm.ac.id).

Because ED literally doses an entrainer, implementation hinges on stable addition; plants specify accurate chemical dosing equipment such as dosing pumps to control solvent flow into the main column.

ED performance and trade‑offs

ED often requires more capital and energy than a single column (two columns and solvent recovery), but it enables separations otherwise infeasible. A comparative acetone–methanol study (using water as an entrainer) achieved the same 99.5% product purity at about 15% lower total annual cost (TAC) than a comparable pressure‑swing scheme (www.researchgate.net). However, ED introduced trace water into both products and made even higher purity harder (ternary vapor–liquid equilibrium, VLE, constraints), limiting how pure one can get without infinite stages (www.researchgate.net).

In practice, ED is chosen when an appropriate solvent exists and entrainer residues are acceptable or easily removed. For water‑bearing mixtures (e.g., many alcohols), entrainers like glycol or glycerol are cheap and non‑toxic. ED remains the method of choice for large‑scale ethanol dehydration (bioethanol fuel) despite high energy demand (www.scielo.br), because its overall economy and maturity beat alternatives like membranes, azeotropic distillation, or pressure‑swing in that application (www.scielo.br).

ED energy and environmental notes

ED’s drawback is high reboiler duty (often >100 kW·h per 1000 L starting mixture) and the need to recover the solvent. Process intensification — heat pumps, mechanical vapor recompression, dividing‑wall columns — can mitigate. Literature shows energy integration in advanced distillation can slash heat duty by 30–45% (pubs.acs.org). One case study on a fully thermally integrated pressure‑swing sequence achieved ~32–45% energy savings (pubs.acs.org) — integration strategies apply similarly to ED if columns are coupled.

Because ED uses a solvent, there is contamination risk or added waste streams. “Green” entrainers (ionic liquids, deep eutectic solvents) are being actively researched to reduce environmental impact (jurnal.ugm.ac.id) (pmc.ncbi.nlm.nih.gov). Where solvent handling requires fluid polishing and sanitary containment, engineers specify 316L stainless housings such as stainless cartridge housings in solvent loops.

Pressure‑swing distillation fundamentals

Pressure‑swing distillation (PSD) exploits azeotropes that shift composition with pressure. One or more columns run at different pressures: the mixture is distilled at one pressure to give an overhead azeotrope, then sent to a second column at a higher (or lower) pressure where the azeotrope composition differs, enabling further fractionation. PSD is very effective when the azeotrope shifts significantly with pressure (often large for polar or hydrogen‑bonded systems). C&CE (2008) reports the acetone–methanol minimum‑boiling azeotrope (77.6% acetone at 1 atm) shifts to 37.5% acetone at 10 atm — about a 40‑point jump — making PSD feasible for that case (www.researchgate.net). Inflating overhead pressure to 10–15 atm or more is common in PSD setups.

Well‑known industrial examples include ethanol–water (azeotrope @94.6% at 101 kPa shifts enough at ~10–15 atm) and some alcohol–water or solvent–solvent systems. For high‑pressure service around ancillary filtration, plants use industrial housings like steel filter housings.

PSD advantages and trade‑offs

The primary advantage is no solvent: nothing foreign to contaminate products and no solvent recovery step, often described as “eco‑friendly” (pmc.ncbi.nlm.nih.gov). PSD frequently uses two simpler columns instead of the three usually needed in ED. In the acetone–methanol study, PSD reached the same 99.5% purity in both products without any third component, whereas ED left trace water in each (www.researchgate.net).

The downside is energy: PSD often requires substantial compression of overhead vapor to achieve the pressure swing, implying high electrical energy for compressors or condensers. In most comparative studies, PSD’s energy usage is higher than ED (though it saves on solvent cost and containment). One THF–water study noted that while PSD gave easier control, its energy consumption was significantly higher than ED (pubs.acs.org). In practice, PSD is chosen when solvent bleed‑through is unacceptable or when a mixture is strongly pressure‑sensitive.

PSD performance data and integration

Published comparisons (steady‑state design) show PSD can meet tight purities with modest equipment: often two columns of 20–30 trays each. When energy is accounted, ED is often cheaper: the acetone–methanol case found ED’s TAC 15% lower than PSD’s (www.researchgate.net). Another THF–water study also found ED lower cost (pmc.ncbi.nlm.nih.gov).

PSD has one virtue for heat integration: the high‑pressure (HP) column’s condenser is much hotter than the low‑pressure (LP) column’s reboiler (e.g., ΔT ≈ 60 K in one study), so heat from one can drive the other (www.researchgate.net). Fully heat‑integrated PSD flowsheets can cut energy use ~30–40% (pubs.acs.org).

Selection guide: data‑first checkpoints

• Azeotrope presence and volatility behavior: Identify if the binary/ternary mixture forms an azeotrope at operating pressure. If not, conventional distillation (with enough trays and reflux) is likely sufficient. If an azeotrope is formed, check if its composition shifts with pressure by calculating VLE (vapor–liquid equilibrium) at two pressures (e.g., 1 atm and 5–10 atm). If the azeotrope composition moves substantially — as in 77.6→37.5% acetone for acetone–methanol from 1 atm to 10 atm — PSD is technically feasible (www.researchgate.net). If the shift is small or the required pressure is impractically high, PSD won’t work and ED (or another method) is needed.
• Purity requirements and contamination: If the target is extremely high purity (≥99.9%) and even trace entrainer is forbidden (e.g., pharmaceuticals or ultra‑pure solvents), PSD has an edge — in acetone–methanol, ED could not easily reach beyond 99.5% due to ternary equilibrium, whereas PSD more readily achieves targets (www.researchgate.net). Conversely, if tiny residual solvent is tolerable or easily removed, ED remains viable. Where purity policing extends to filtration hardware, pharma and food plants specify sanitary 316L components such as stainless cartridge housings.
• Entrainer availability and environmental constraints: Confirm a suitable entrainer exists (compatibility, safety/regulatory fit, economic handling). Examples include water, glycols, DMSO, glycerol, or ionic liquids; some have issues (DMSO is flammable, glycols can cause foaming, etc.). Local regulations (including Indonesian B3 chemical rules) may restrict certain solvents. “Green” options (bio‑based glycols, ionic liquids) are increasingly attractive to meet stricter standards (jurnal.ugm.ac.id) (pmc.ncbi.nlm.nih.gov).
• Energy and economics: Quantify both capital and operating costs. Comparative studies suggest TAC for ED can be 10–20% lower than for PSD in many cases (www.researchgate.net) (pmc.ncbi.nlm.nih.gov), because the extra columns’ capital are often offset by lower steam per unit. PSD has far less expensive reflux solution (no entrainer) and removes solvent make‑up/purification. As a rule, if energy cost dominates and a solvent is cheap, ED may win; if capital is tight and energy cost acceptable, PSD may win. In THF–water, ED was cheaper (pmc.ncbi.nlm.nih.gov), and for acetone–methanol, ED was ~15% cheaper (www.researchgate.net).
• Process complexity and control: Conventional distillation is simplest, ED adds a solvent recovery column, PSD adds pressure control and possibly compressors. Dynamic‑control studies show ED and PSD have practically similar controllability (www.researchgate.net). PSD can simplify heat integration: HP condenser temperatures can be ~60 K above the LP reboiler, easing coupling (www.researchgate.net).
• Industry precedent and scale: For large‑scale ethanol plants, extractive dehydration (or hybrid ED/azeotropic with ethylene glycol) is the norm (www.scielo.br). Many petrochemical firms also implement PSD for specialty separations (especially where solvent use is discouraged). Proven solutions that comply at the intended capacity are lower‑risk choices. Ancillary utilities for implementation can include supporting equipment for handling solvent‑related side streams.

Decision summary and energy integration

Use conventional distillation whenever possible (simple, lowest CAPEX). If an azeotrope or close‑boiling pair blocks separation, first check PSD feasibility (significant azeotrope shift with pressure) because it avoids solvents. If PSD cannot meet purity/throughput or imposes excessive compression costs, plan ED with a suitable entrainer. For ethanol–water dehydration, many engineers prefer ED (e.g., with glycol or an ionic liquid) for its lower total cost (www.scielo.br), despite PSD’s solvent‑free appeal. Always back up the choice with simulation and TAC minimization (e.g., rigorous design studies).

In all cases, implement energy‑saving measures: heat integration and vapor recompression can cut distillation heat duty by roughly one‑third (pubs.acs.org). Given that distillation typically dominates plant energy use, designing with integration upfront can pay back quickly. For dosing, containment, and pressure service around columns, engineers match ancillaries to duty — from dosing pumps for entrainer addition to steel filter housings for industrial filtration around high‑pressure services.

References: Authoritative reviews and case studies were used above (www.scielo.br) (www.scielo.br) (www.scielo.br) (pmc.ncbi.nlm.nih.gov) (jurnal.ugm.ac.id) (jurnal.ugm.ac.id) (www.researchgate.net) (www.researchgate.net) (pmc.ncbi.nlm.nih.gov) (pubs.acs.org), as cited in‑line.