What's the typical enrichment factor for foam fractionation?

For rhamnolipids: 5-15x over broth concentration. For other strong surfactants (surfactin from B. subtilis, some lipopeptides): 10-40x. For sophorolipids or MEL: under 3x, which means foam fractionation is not the right DSP choice for those products. The enrichment factor depends on the product's surface activity, the foam stability, the gas flow rate, and the column height. The platform defaults to 8x for rhamnolipid and 15x for surfactin, which are typical literature values.

Does foam fractionation replace the rest of the DSP train or just the first step?

Replaces the first concentration step (primary recovery), not the entire chain. You still need polishing: centrifugation or filtration of the residual broth to recover non-foamed product, UF/DF to concentrate the collapsed foam to commercial spec, optionally solvent extraction for final purification. Overall DSP train is 3-5 unit operations instead of 5-7, which simplifies the operation and saves 15-30% on total DSP cost.

What's the capital cost for foam fractionation at industrial scale?

An inline foam fractionation column for a 20,000 L reactor typically runs $150-400k installed, including the column, demister, collapse chamber, level control, and automation. That's 20-40% less than an equivalent-capacity centrifuge plus solvent extraction train it replaces. Ongoing maintenance is lower because there are fewer moving parts than a centrifuge and no solvent loop to manage.

Can the platform model different foam fractionation configurations?

The foam fractionation unit operation exposes tunable parameters: enrichment factor, primary yield, column residence time, gas flow rate, and a cost model for consumables and labor. Users can run sensitivity analysis across any of these. We don't model column-internal fluid dynamics (that would need CFD), but the unit-operation-level output (yield, concentration, cost) captures what you need for TEA.

What products should NOT use foam fractionation?

Anything that doesn't foam strongly: sophorolipids (aggregate into dense phase instead), non-surfactant products (enzymes, organic acids, ethanol), intracellular products (need cell lysis before recovery anyway). Also avoid foam fractionation if your product is heat-sensitive and you can't break foam with heat or mechanical means without degradation. For most mAb and recombinant protein processes, foam fractionation is not applicable.

Foam fractionation as primary DSP for biosurfactants: economics and design

In most biosurfactant fermentations, foam is a liability: it caps the working volume, crashes kLa, costs money in antifoam, and introduces contamination risk through overflow paths. Foam fractionation inverts this. The foam becomes the product stream, concentrated 5-15x relative to broth before any other DSP step. For rhamnolipids and other strong surfactants, this is the highest-leverage DSP decision available today. This post covers the physics, the design parameters, and when the approach pays off.

The physics

Surfactants concentrate at gas-liquid interfaces because their amphiphilic structure lowers surface tension. When you sparge a gas through a surfactant-containing liquid, the rising bubbles carry a disproportionately high amount of surfactant in the interfacial film. At the top of the column, bubbles form foam and the lamellae between bubbles are enriched in surfactant. Collapse the foam (by mechanical shear, reduced pressure, or mild heat) and you get a concentrated liquid.

The enrichment factor scales with:

Surface activity. Higher = more enrichment. Rhamnolipids, surfactin, and lipopeptides foam with strong enrichment. Sophorolipids in low-water broth don't foam this way.
Foam stability. Longer foam residence time in the column allows more liquid to drain back out of the foam, which concentrates the remaining material further.
Column height. Taller column = more drainage = higher concentration. 2-3 m columns are typical industrially.
Gas flow rate. Low gas rates maximize enrichment (good drainage). High gas rates maximize throughput (but wash out drainage). The optimization target depends on whether you're chasing maximum concentration or maximum product-per-hour.

Design parameters that matter

For a first-pass design at 10,000-50,000 L scale:

Column diameter: 0.5-1.2 m
Column height: 2-3 m
Superficial gas velocity: 1-3 cm/s
Residence time in column: 30-120 s
Demister at top: to prevent liquid carryover into offgas
Collapse method: mechanical (rotating disk), thermal (60-70°C heated plate), or vacuum
Level control: continuous drain of collapsed liquid to holding tank

The single most sensitive parameter is residence time. Too short (high gas rate or small column): foam is wet, enrichment factor drops. Too long: column level rises, spillover risk. Most industrial installations aim for a stable 60-90 s residence time at steady state.

Yield vs enrichment tradeoff

Foam fractionation is not free of tradeoffs. Higher enrichment (driven by low gas rate, tall column, long residence time) comes with lower overall primary yield because more product stays in the broth rather than leaving as foam.

Typical operating points for rhamnolipid:

High enrichment (12-15x), moderate yield (70-80%):Low gas rate, high column, long residence. Best for when broth volume is a constraint and you want a concentrated primary stream to simplify downstream.
Balanced (8-10x enrichment, 82-88% yield):Most common industrial operating point. Gas rate tuned to the organism's oxygen demand, column sized for good drainage.
High yield (5-7x enrichment, 90-93% yield):Higher gas rate, shorter residence. Best for when you have plenty of downstream concentration capacity and want to recover as much product as possible from the foam.

The 82-88% yield balanced point is where most industrial processes operate. Below 80% primary yield the residual broth becomes a major disposal/recovery headache. Above 90% yield the enrichment drops enough that you lose much of the DSP-simplification benefit.

DSP chain with foam fractionation

A typical foam-fractionation-primary DSP for rhamnolipid:

Foam fractionation (inline to reactor): 85% yield, 8x enrichment. Broth at 40 g/L goes in, collapsed foam at ~320 g/L comes out.
Centrifugation of residual broth: recovers 85% of the remaining 15%, so overall primary + centrifuge yield is ~98% on foamed product.
UF/DF: concentrates from 320 g/L to 500-600 g/L, removes residual salts and low-MW impurities. 92% yield.
Spray dry or crystallize to final powder or solid. 95% yield.

Overall DSP yield: 70-75%. Solvent consumption: minimal (compared to 2-3 tonnes per batch for solvent extraction). Labor: roughly half of a conventional centrifuge+extract+crystallize chain.

COGS impact

For a 20,000 L, 30 batch/year, 40 g/L rhamnolipid process on glycerol feedstock:

Conventional DSP (antifoam + centrifuge + extract + crystal): $32-48/kg COGS
Foam fractionation primary DSP: $18-28/kg COGS

A 35-45% reduction from the DSP choice alone. Most of the savings are from solvent cost, DSP yield improvement, and simplified operation. Capital cost for the foam fractionation column is recovered in typically under 18 months at commercial-scale operation.

What Augur models

Foam fractionation is a registered unit operation in the platform (foam_fractionation in the DSP registry). Parameters exposed for override:

Primary yield (default 85%, range 70-95%)
Enrichment factor (default 8x for rhamnolipid, adjustable)
Gas flow rate and residence time (feeds into yield-vs-enrichment model)
Capital cost model (scales with column diameter and height)
Labor hours per batch (typically lower than centrifuge+extract)

Users can swap between foam fractionation and solvent extraction as the primary recovery step in a scenario comparison to see the COGS delta directly. The platform also flags products where foam fractionation is not appropriate (sophorolipid, non-surfactant products) and defaults to alternative DSP chains for those organisms.

When foam fractionation pays off

Three conditions need to be true:

Product is strongly surface-active in the broth (rhamnolipids, surfactin, some lipopeptides, some saponins).
Production is at scale above ~500 L. Below that, the capital cost of the column doesn't amortize.
Process runs continuously-fed or fed-batch long enough that continuous foam harvest is operationally sensible. For short batch processes (under 48 h) the setup-and-teardown overhead is higher than the savings.

For any biosurfactant producer meeting those criteria, foam fractionation is the single highest-leverage DSP decision available. The platform can model it, override it per-scenario, and compare the economics against conventional alternatives.

If you're designing the DSP train for a new biosurfactant process and want to see the foam fractionation economics for your specific strain and scale, request access.