This article is for informational purposes only and is not medical advice. Full disclaimer →
In one sentence: A 2023 biopharmaceutical study shows that virus-killing detergents can inactivate enveloped viruses even below their critical micelle concentration, challenging the long-held assumption that micelle formation is required.
Farcet JB, Karbiener M, et al. · International journal of molecular sciences (2023)
PubMed 37175626 · DOI · PMC full text
Why this paper matters
The paper's finding that detergents can inactivate enveloped viruses below their critical micelle concentration has practical consequences for biopharmaceutical manufacturing — the CMC-based thresholds used to validate viral safety steps may not reflect the mechanism actually driving inactivation. For a monolaurin-focused site, the direct connection is that monolaurin is an amphiphilic molecule from the same class studied here, and its envelope-disruption activity (established by Thormar in 1987) likely operates through similar sub-CMC mechanisms. The study tested only five nonionic detergents against a single model virus (BVDV), so generalization to ionic detergents or fatty-acid monoglycerides specifically remains unvalidated.
Background
Since the 1980s HIV and Hepatitis C contamination disasters, biopharmaceutical manufacturers have treated plasma-derived and cell-derived medicines with solvent/detergent (S/D) mixtures to inactivate enveloped viruses. The prevailing scientific model held that detergents must reach their critical micelle concentration (CMC) — the threshold at which molecules self-assemble into spherical clusters called micelles — to effectively disrupt viral lipid envelopes. CMC values are traditionally measured in pure water at 25°C, conditions far removed from real manufacturing environments containing salts, proteins, and different temperatures.
The EU placed Triton X-100 on the REACH Annex XIV authorization list in 2017, with a sunset date of January 4, 2021, prohibiting its use without authorization because its environmental metabolites act as endocrine disruptors. This has triggered an urgent industry-wide search for replacement detergents, with 2025 publications from Lonza (Yadav et al.) and other groups reflecting the accelerating transition. Understanding whether micelle formation is truly required for virus killing — and how CMC shifts in real manufacturing matrices — is now critical for designing the next generation of viral safety steps.
The CMC Dogma
For decades, scientists assumed detergents needed micelles to kill viruses. This study says: not necessarily.
The critical micelle concentration has long been treated as a key reference parameter for virus inactivation, yet it is routinely measured only in deionized water at approximately 25°C — conditions bearing little resemblance to actual manufacturing environments.[1, 2] CMC is traditionally determined in deionized water at approximately 25°C, while biopharmaceutical virus inactivation occurs in complex matrices containing salts, proteins, and buffers at controlled temperatures such as 14–17°C. The primary paper explicitly highlights this disconnect as a critical gap in the field, noting that S/D treatment has been in use since the mid-1980s response to HIV and HCV transmission through plasma-derived products.
A study by Gooran et al. using a supported lipid bilayer platform concluded that TX-100 and its potential replacement Simulsol SL-11W were only membrane-active at or above their CMC, but that study measured physical membrane disruption rather than biological virus infectivity.[1, 5] The Gooran et al. study, published in MDPI and employing Quartz Crystal Microbalance-Dissipation (QCM-D) measurements on supported lipid bilayers, found that both TX-100 and Simulsol SL-11W exhibited concentration-dependent membrane interactions only at or above their CMC values. The primary paper directly challenges the applicability of this conclusion to actual virus killing, citing evidence that sub-CMC detergent concentrations still yield measurable virus reduction factors in infectivity assays. The SLB platform measures physical membrane disruption, not whether a virus retains infectivity — these are meaningfully different endpoints.
CMC Shifts in Real-World Matrices
The same detergent can have wildly different micelle-forming behavior depending on what's dissolved around it.
Nereid's CMC varied by 4.2-fold between pure water at 25°C (0.604 mM) and a recombinant protein matrix at 14°C (0.143 mM).[1, 2] In deionized water at 25°C, Nereid had a CMC of 0.604 mM, but in the recombinant protein matrix tempered to 14°C it dropped to just 0.143 mM — a 4.2-fold difference. This dramatic shift demonstrates that proteins, salts, pH, and temperature collectively alter detergent self-assembly in ways that cannot be predicted from standard reference values alone. The plasma protein matrix experiments were conducted at 17°C, not 14°C, making temperature one of several contributing variables.
TX-100's CMC increased in biopharmaceutical matrices compared to pure water, while Nereid's CMC decreased — showing matrix effects are not uniform across detergents.[1, 2] TX-100 had a CMC of 0.274 mM in water, rising to 0.345 mM in the plasma protein matrix (17°C) and 0.432 mM in the recombinant protein matrix (14°C). Conversely, Nereid went from 0.604 mM in water down to 0.400 mM in the plasma matrix and 0.143 mM in the recombinant matrix. TX-100R, the reduced form, showed a CMC of 0.259 mM in water, 0.358 mM in the plasma protein matrix, and 0.321 mM in the recombinant protein matrix — also increasing in matrices, mirroring TX-100 behavior. This non-predictable, detergent-specific response means each candidate must be individually characterized in its actual application environment.
PS80's CMC in the recombinant protein matrix could not be determined because surface tension decreased continuously without the expected two-phase profile needed for CMC calculation.[1, 2] Force tensiometry measurements of PS80 in the recombinant protein matrix produced a constant decrease in surface tension across a wide concentration range rather than the characteristic two-phase curve (declining, then flat) needed for CMC determination. This anomalous behavior, noted in both the Results and Methods sections of the primary paper, highlights the complexity of detergent behavior in protein-rich environments. Critically, the paper also notes that CMC determination of the full S/D mixture — as actually used in manufacturing — yielded ambiguous measurement curves that precluded reliable CMC determination entirely, meaning all reported CMC data represents single detergents tested in isolation.
Sub-CMC Virus Killing
Even at concentrations where no micelles should form, these detergents still destroyed viruses — with an important caveat.
TX-100R achieved a 3.1 log₁₀ reduction factor against BVDV at approximately 0.25 mM — below its measured CMC of 0.358 mM in the plasma protein matrix.[1, 2] At approximately 0.25 mM (1.5% of the standard ~16 mM manufacturing target), TX-100R was below its measured CMC of 0.358 mM in the plasma protein matrix, yet it still inactivated BVDV (a Hepatitis C model virus) by 3.1 log₁₀ within the experiment. A 3.1 log₁₀ reduction means roughly 99.92% of virus was eliminated. Under the same sub-CMC conditions, TX-100 achieved 2.5 log₁₀ and Nereid achieved 2.2 log₁₀. However, the authors themselves note a critical caveat: when concentrated detergent stock solutions were added to the matrix before stirring was complete, there was a brief moment of locally high concentration during which micelles could have transiently formed, meaning it cannot be definitively concluded that monomeric detergent alone drove all observed inactivation.
At 3% of manufacturing concentration (above CMC), TX-100 achieved >6.1 log₁₀ reduction and Nereid achieved 4.6 log₁₀ — meaning less than one in a million viruses survived TX-100 treatment.[1, 2] When TX-100 was used at approximately 0.5 mM, just above its CMC in the plasma protein matrix, it delivered a log₁₀ reduction factor of >6.1 against BVDV, with no residual infectivity detected in any sample. Nereid at the same 3% manufacturing concentration achieved 4.6 log₁₀ reduction. The '>' notation for TX-100 indicates the true reduction may be even higher, limited only by assay sensitivity. Both results confirm that concentrations representing a tiny fraction of standard manufacturing levels provide overwhelming virus inactivation when in the micellar regime.
Molecular Shape Trumps CMC
A detergent's ability to kill viruses may depend more on its shape than on when it forms micelles.
PS80 and PS20 have CMC values 10–30 times lower than TX-100, yet TX-100 is far more potent at killing viruses — demonstrating CMC does not predict antiviral potency.[1, 5] PS80 had a CMC of just 0.013–0.022 mM and PS20 had 0.030–0.071 mM across matrices, compared to TX-100's 0.274–0.432 mM. Despite forming micelles at much lower concentrations, polysorbates are substantially less effective at virus inactivation even when applied well above their CMC. The Gooran et al. supported lipid bilayer study also confirmed that TX-100 induced rapid, irreversible, complete membrane solubilization while the alternative detergent caused only gradual, reversible membrane budding — pointing to mechanistic differences beyond simple micelle formation.
The authors propose that polysorbates' bulky, multi-branched molecular geometry physically prevents smooth intercalation into viral lipid bilayers, while the slender TX-100 family can slide between phospholipids.[1] Polysorbates have a sorbate core branched with multiple polyethylene glycol chains, creating a spherical, bulky three-dimensional shape. The primary paper's Discussion section proposes — using hedged language ('could be rationalized by,' 'might prevent') — that this bulkiness may physically prevent smooth intercalation into the tightly packed lipid bilayer of a viral envelope. In contrast, the TX-100 family has a comparatively slender linear structure. This geometric hypothesis is proposed rather than experimentally validated: no direct structural measurements of detergent insertion into viral membranes were performed in this study.
A single methylene group difference between Nereid and TX-100 caused a roughly 2.2-fold difference in CMC in pure water (0.604 mM vs. 0.274 mM).[1, 2] Nereid's CMC in deionized water was 0.604 mM versus TX-100's 0.274 mM — a 2.2-fold difference attributable to the addition of just one CH₂ group in Nereid's chemical structure compared to TX-100. This remarkable sensitivity of micelle formation to minimal structural changes underscores why CMC alone is insufficient for predicting antiviral behavior: tiny molecular modifications dramatically alter physical chemistry without necessarily changing biological activity in a proportional way.
The Triton X-100 Replacement Race
One of pharma's most trusted chemicals is effectively banned — and the industry is scrambling for alternatives.
ECHA listed Triton X-100 (octylphenol ethoxylate) as a Substance of Very High Concern under REACH in 2017, with a sunset date of January 4, 2021, prohibiting EU use without authorization because its metabolites act as endocrine disruptors.[1, 6, 8, 9] Triton X-100 (4-tert-OPnEO, octylphenol ethoxylate) was added to REACH Annex XIV by EU Commission Regulation (EU) 2017/999, with a sunset date of January 4, 2021. Post-sunset, the substance cannot be placed on the market or used without authorization. Its environmental degradation products, including octylphenol, are persistent endocrine disruptors that mimic estrogen and harm aquatic organisms. This has forced biopharmaceutical manufacturers — who rely on TX-100 for virus inactivation in plasma-derived and cell-derived medicines — to seek validated replacements, a complex process requiring new viral inactivation studies, process revalidation, and global regulatory approval.
Lonza screened 16 candidate detergents against criteria including solubility, virus inactivation feasibility, CMC, and storage conditions, identifying C16-AO and C11/15-sEO9 as top TX-100 replacements via Multi-Criteria Decision Analysis.[4] Yadav et al. (2025) conducted a systematic development program at Lonza Biologics, screening 16 potential alternatives. Using Multi-Criteria Decision Analysis (MCDA), four candidates advanced to a second-stage assessment, with C16-AO and C11/15-sEO9 emerging as the top practical alternatives. C13-EO8 also showed good viral inactivation capability but requires further investigation regarding detergent clearance. The study underscores that while low pH is the preferred inactivation method, detergent treatment remains essential for proteins with limited stability at low pH.
Deviron 13-S9, a non-ionic C11-15 secondary alcohol ethoxylate, achieved >5 log reduction against enveloped viruses across multiple biopharmaceutical matrices and is readily biodegradable per OECD 301B guidelines.[3, 10] Banerjee et al. (2025) demonstrated that Deviron 13-S9 matched or exceeded TX-100's virus inactivation performance in human IgG, clarified cell culture harvest, and fractionated plasma matrices. It is confirmed as readily biodegradable according to OECD 301B guidelines and exhibits no binding to typical downstream chromatography resins, with effective removal via Protein A resin. OECD 301B (the CO2 Evolution/Modified Sturm Test) requires achieving 60% biodegradation within 28 days, within a 10-day window, to qualify as readily biodegradable.
In standard manufacturing, TX-100 is used at approximately 1% w/w (~16 mM) — roughly 37–50 times higher than its measured CMC across matrices — providing a substantial safety margin even accounting for matrix-induced CMC shifts.[1, 7] The typical manufacturing concentration of TX-100 is 1% w/w (~16 mM), while its highest measured CMC across all matrices in the primary study was 0.432 mM (recombinant protein matrix at 14°C). This represents a ~37-fold excess at the matrix CMC peak. The paper's sub-CMC experiments used concentrations approximately 30–60 times lower than manufacturing standard, representing extreme worst-case scenarios. The Plasma Protein Therapeutics Association analysis of 308 viral inactivation studies found the TX-100/TNBP combination achieved log reductions ranging from >2.9 to >6.5, confirming robust efficacy at manufacturing concentrations.
Caveats and open questions
What this paper doesn't settle
The study tested only five detergents from the nonionic TX-100 and polysorbate families, using only one virus (BVDV) as a model system, in two specific biopharmaceutical matrices. It remains genuinely uncertain whether sub-CMC inactivation occurs with ionic detergents, alcohol ethoxylates, or against other enveloped viruses with different lipid envelope compositions. The molecular geometry hypothesis — while plausible and proposed in the Discussion — is not experimentally validated; no direct structural measurements of detergent insertion into viral membranes were performed. Additionally, CMC of the actual S/D mixture used in manufacturing could not be reliably determined, meaning all CMC data comes from single detergents tested in isolation — a further limitation not reflected in current regulatory frameworks.
The honest skeptical read
The authors themselves acknowledge a key caveat undermining the strongest interpretation of their sub-CMC findings: when concentrated detergent stock solutions were added to the matrix before stirring was complete, there was a brief moment of locally high concentration during which micelles could have transiently formed. This means it cannot be definitively concluded that monomeric detergent molecules alone drove all observed virus inactivation — some or all of the sub-CMC killing may have occurred in that initial high-concentration contact window. The Gooran et al. supported lipid bilayer study also found membrane activity only at or above CMC, though it used a different endpoint (physical disruption vs. infectivity).
Common misconception
A common misconception is that the critical micelle concentration is a fixed physical constant for a given detergent. In reality, CMC is highly dependent on temperature, ionic strength, pH, and the presence of other solutes — it can shift by more than 4-fold between laboratory reference conditions and actual biopharmaceutical manufacturing environments. Moreover, a lower CMC does not indicate a more potent virus-killing detergent; polysorbates form micelles at far lower concentrations than TX-100 yet are substantially less effective at virus inactivation.