Technological breakthroughs

Compiled by Bao Hien

A research team assessed the life cycle of a laundry detergent containing a fermentation-derived glycolipid. At baseline, its greenhouse gas emissions are higher than those of a conventional detergent, but they can fall sharply if the wash is run colder with a smaller dose.

The results were published online on December 17, 2025 in RSC Sustainability (2026 issue), using a cradle-to-grave life cycle model for the German market.

A surfactant made by fermentation

The compound studied is mannosylerythritol lipid (MEL), a glycolipid produced by microorganisms during fermentation. The feedstocks are rapeseed oil and glucose. MEL is currently made at laboratory to pilot scale, so the team used experimental data scaled up according to a projection scenario from earlier work. In the model, MEL replaces conventional surfactants one-to-one by mass.

The functional unit is one wash of 4.5 kg of fabric in a 6 kg machine, using 75 g of detergent and 50 liters of water. Electricity per cycle is 0.3 kWh at 30 °C, 0.5 kWh at 40 °C and 0.75 kWh at 60 °C.

Baseline results

At 40 °C with 75 g of detergent, emissions are 0.44 kg CO₂-equivalent per wash. The conventional detergent used as a reference is 0.36 kg, so the MEL detergent is about 22% higher (calculated from the paper’s two figures).

Emissions concentrate in two stages: machine electricity accounts for about 49% and production of the MEL-containing detergent about 33%. Wastewater treatment contributes about 6%, packaging 2% and domestic transport under 1%.

Wash temperature decides the outcome

The team tested several dose and temperature scenarios against the baseline:

• 33% lower dose (50 g, 40 °C): about 15% lower emissions.
• 60 °C: about 25% higher emissions. A 33% dose cut at 60 °C does not offset the extra electricity.
• 37.5 g at 30 °C: about 42% lower.
• 37.5 g at 30 °C with green electricity: about 67% lower.

The team also computed break-even points against the conventional detergent. At the same dose, the wash temperature would need to fall to about 21 °C. At the same temperature, the dose would need to fall by 50%. Combining a 33% dose cut with a temperature about 7 °C lower (around 33 °C) gives comparable emissions. These break-evens hold for climate change only, not for acidification or eutrophication.

In the 30 °C scenario, freshwater eutrophication falls by about 20% and the other seven impact categories by 40–50%.

What has to be true for the advantage to appear

The advantage of MEL does not come from the molecule itself, but from the assumption that it cleans better and so permits colder, smaller washes. According to the authors, no formulations tailored to MEL exist yet, and washing performance has not been fully measured at full scale.

Other limitations the authors state: the one-to-one replacement has no experimental basis for multi-surfactant systems; biodegradation is assumed complete but is uncertain; ecotoxicity is out of scope; scale effects such as process integration and energy recovery are not fully captured. The comparison with conventional surfactants is a preliminary approximation, and part of the background data comes from a database restricted from publication.

Inconsistencies

The abstract cites a reduction of up to 65% under optimistic assumptions, but the body does not tie that figure to a specific scenario; the closest are the two 30 °C scenarios at 42% and 67%. The paper also does not give the MEL share of the formulation (it is in the supplementary information). Separately, one reference scenario lists 9.5 liters of water in a table but 39.5 liters in the introduction. The 0.36 kg figure for the reference detergent is also lower than the 0.52 kg the paper cites from an industry reference guideline, owing to differences in scope and assumptions.

Source:

Gavez, L., Bippus, L., Briem, A., & Albrecht, S. (2026). “Biosurfactant-containing products from an environmental perspective – life cycle assessment of a liquid laundry detergent and a personal care product.” RSC Sustainability. https://doi.org/10.1039/D5SU00168D