Lipid nanoparticles are not manufactured in the conventional sense: they self-assemble spontaneously when an ethanolic solution of lipid components meets an aqueous nucleic acid solution under controlled mixing conditions. The particle size, polydispersity, and encapsulation efficiency of the resulting LNPs depend almost entirely on the mixing dynamics at the moment of assembly, making the choice of mixing technology one of the most consequential decisions in LNP process development.
Key Takeaways |
|
|
|
|
|
For the buffer management and in-line dilution context in which LNP formulation buffers are prepared and controlled, see Solving the Buffer Management Crisis in Biomanufacturing. For the downstream purification context covering the filtration and TFF steps that follow LNP formation, see Mastering Downstream Purification: Advanced Chromatography and Filtration in Biomanufacturing.
How LNPs Form: The Self-Assembly Mechanism That Mixing Controls
Lipid nanoparticles form through a mechanism of rapid precipitation and self-organization. A four-component lipid mixture consisting of an ionizable lipid, a phospholipid, cholesterol, and a PEG-lipid is dissolved in ethanol at defined molar ratios. This ethanolic lipid solution is then rapidly mixed with an aqueous buffer at low pH, typically between pH 3.5 and 5.5, under conditions that protonate the ionizable lipid and create a strong electrostatic interaction with the negatively charged phosphate backbone of the nucleic acid payload.
Working in analytical science?
Register for a FREE Separation Science account to subscribe to the Separation Science Newsletter.
When the two streams meet in the mixing device, the rapid dilution of ethanol into the aqueous phase creates a transient supersaturated condition in which the lipid components simultaneously interact with the nucleic acid and self-organize into nanoparticle structures. The kinetics of this self-assembly are highly sensitive to mixing speed: if mixing is too slow, the lipids aggregate into large heterogeneous structures; if mixing is sufficiently rapid, the result is uniform nanoparticles in the eighty to two hundred nanometer range with encapsulation efficiencies greater than eighty-five percent for mRNA payloads.
The ionizable lipid is uncharged at physiological pH but becomes protonated at the acidic pH of the formulation buffer, facilitating mRNA encapsulation. After formation, the pH of the LNP suspension is raised to physiological pH for purification and final formulation, at which point the ionizable lipid returns to its uncharged state and the particle stabilizes. This pH-dependent behavior means that the formulation buffer pH during mixing is a critical process parameter that must be controlled within tight limits for consistent encapsulation efficiency.
Why Does the Mixing Method Determine LNP Quality?
The quality attributes of an LNP preparation, specifically particle size, polydispersity index (PDI), encapsulation efficiency, and morphology, are determined by the competition between the timescale of nucleic acid-lipid interaction and the timescale of lipid self-assembly. If lipid molecules have too much time to encounter each other before the ethanol concentration drops sufficiently to drive precipitation, they will form large disordered aggregates. Rapid mixing compresses both timescales, ensuring that all lipid molecules experience the same transition from organic solvent to aqueous environment simultaneously, which produces uniformly sized particles.
A comprehensive 2025 review published in Pharmaceutics on microfluidic and turbulent mixing for mRNA LNP vaccines analyzed the full range of commercially available mixing devices used in LNP formulation, from staggered herringbone microfluidic chips to turbulent impingement jet mixers. The review confirmed that mixing speed, expressed as the Reynolds number of the flow at the point of stream contact, is the primary determinant of particle size and PDI, and that all successful LNP mixing approaches share the objective of achieving turbulent or chaotic mixing within milliseconds of stream contact.
Single-use mixing formats address a specific manufacturing challenge: lipid components, particularly PEG-lipids and ionizable lipids, form thin adhesive films on the surfaces of stainless steel tubing, glass capillaries, and other reusable materials. Over the course of a production run, this surface fouling progressively alters the effective channel geometry of the mixing device, changing the mixing dynamics and therefore the particle size distribution of the product. Single-use cartridges, discarded after each formulation run, eliminate this source of within-batch and between-batch variability.
Single-Use Mixing Technologies for LNP Formulation
Microfluidic Staggered Herringbone Mixers
The staggered herringbone micromixer (SHM) was the first microfluidic format to demonstrate precise, reproducible LNP formation at small scale and remains the reference format for development-stage LNP formulation. SHM channels contain a series of angled ridges on the channel floor that create chaotic advection, rapidly folding and stretching the two fluid streams into a laminar interdigitated structure from which diffusive mixing across the thin lamellae is rapid. The resulting mixing time of less than ten milliseconds produces LNPs with the smallest achievable sizes and narrowest size distributions, often below thirty nanometers for optimized formulations. Research published in 2025 demonstrated that parallelized SHM architectures incorporating 256 mixing channels with antifouling surface coatings achieved stable production rates at commercial liter-per-hour throughputs while maintaining the particle quality characteristics of single-channel devices, demonstrating that SHM parallelization can bridge the throughput gap between development and commercial manufacturing.
Impingement Jet Mixers
Impingement jet mixers (IJMs), also called T-jet mixers in some commercial implementations, direct the two fluid streams toward each other at high velocity through opposing nozzles, creating a zone of intense turbulent mixing at the impingement point. The turbulent kinetic energy at the impingement point is much higher than in laminar microfluidic mixing, providing rapid mixing without requiring the micron-scale channel geometry of SHM devices. IJMs are simpler to manufacture at scale, more tolerant of particulate matter in the feed streams, and available in single-use formats from multiple commercial suppliers.
The trade-off is that IJMs typically produce slightly larger particles with somewhat broader size distributions than optimized SHM devices, and the mixing characteristics are more sensitive to viscosity changes in the feed streams than to the flow geometry parameters. For most clinical and commercial LNP formulations where particle sizes of eighty to two hundred nanometers are acceptable, IJMs provide adequate control and are preferred at scales above ten milliliters per minute where SHM devices require parallelization.
Comparison of Single-Use LNP Mixing Technologies
Technology | Mixing Mechanism | Particle Size Range | Single-Use Format | Scale Application |
SHM microfluidic chip | Chaotic advection via herringbone ridges; laminar but geometrically induced chaotic mixing | Less than thirty to one hundred fifty nanometers; narrowest achievable PDI (less than 0.1) | Disposable polymer or PDMS chips; one-time use per run eliminates fouling carryover | Preclinical to clinical; parallelized formats extend to commercial |
Impingement jet mixer (IJM) | Turbulent mixing at high-velocity jet impingement point | Eighty to two hundred nanometers; PDI typically 0.1 to 0.2 | Single-use nozzle assemblies and flow paths; stainless steel reusable body with single-use fluid path | Clinical to commercial; standard format for most commercial mRNA LNP manufacturing |
Cross-flow/T-junction mixer | Controlled stream convergence in T or Y junction; relies on diffusion and low-level turbulence | Eighty to three hundred nanometers; broader size distribution than SHM or IJM | Single-use tubing and connectors; simple configuration with minimal specialized components | Early development and process scouting; limited to low flow rates for adequate mixing |
Parallelized microfluidic device (PMD) | Multiple SHM or chaotic channels in parallel on a single chip; uniform flow distribution across channels required | Equivalent to single-channel SHM; particle quality preserved at higher throughput | Single-use chip with integrated parallelization; flow resistors ensure uniform distribution | Clinical to commercial; bridges throughput gap between single-channel SHM and IJM |
What Critical Process Parameters Govern LNP Size and Polydispersity?
Four process parameters govern particle formation and must be defined and controlled within validated ranges for each LNP formulation:
Parameter | What It Controls | Typical Range | Quality Impact |
Flow rate ratio (FRR) | Ratio of aqueous volume to organic (ethanol) volume per unit time; determines the dilution rate of ethanol at the mixing point | One to one up to four to one (aqueous:organic); three to one commonly used for mRNA LNPs | Higher FRR produces smaller particles by increasing the rate of ethanol dilution and accelerating precipitation; FRR is the primary lever for particle size tuning |
Total flow rate (TFR) | Combined volumetric flow rate of both streams through the mixing device; determines mixing Reynolds number and mixing intensity | One to fifty milliliters per minute for microfluidic formats; up to hundreds of milliliters per minute for IJM | Higher TFR increases mixing intensity and generally reduces particle size at constant FRR; defines throughput capacity and must be matched to mixer design for optimal performance |
Ethanol concentration in organic stream | Solubility and aggregation behavior of lipid components; determines viscosity of organic stream | Seventy to ninety-five percent ethanol by volume; typically ninety percent for mRNA LNP formulations | Lower ethanol concentration reduces lipid solubility, increasing aggregation risk; higher concentrations affect aqueous phase pH at the mixing point |
Aqueous phase pH | Protonation state of the ionizable lipid during particle formation; directly controls encapsulation efficiency through electrostatic interaction with nucleic acid | pH 3.5 to 5.5 for most ionizable lipid formulations; controlled by citrate or acetate buffer | pH must be below the pKa of the ionizable lipid during mixing to drive encapsulation; deviations of more than 0.2 pH units can substantially reduce encapsulation efficiency |
Scaling from Bench to Commercial: The Parallelization Challenge
Scaling LNP production from milligrams to grams of nucleic acid payload per batch is not achieved by simply building larger microfluidic channels. The mixing quality achieved in a microfluidic SHM device is a function of the channel dimensions and the flow velocity within those channels; a larger channel at the same linear velocity produces different mixing dynamics than a smaller channel. Scaling instead requires either switching to a fundamentally different mixing technology (such as IJM) at higher throughput, or parallelizing the microfluidic mixing geometry across multiple identical channels operating simultaneously. A 2025 high-throughput aerofoil-structured microfluidic platform from University College Dublin demonstrated precisely this principle: by designing aerofoil-shaped channel structures optimized for performance across a wide flow rate range of 0.2 to fifty milliliters per minute, the platform achieved consistent LNP quality from screening volumes to liter-scale production without changing the mixing geometry.
Parallelized microfluidic devices address the throughput limitation of single-channel SHM designs by distributing flow across tens to hundreds of identical mixing channels on a single chip. Studies on scalable mRNA and siRNA LNP production using parallelized microfluidic devices confirmed that the critical requirement for quality-consistent parallelization is uniform flow distribution across all channels: any channel receiving a different flow rate than its neighbors produces LNPs with different size and encapsulation characteristics, degrading the overall batch quality. Individual flow resistors at the entrance of each channel, ensuring equal pressure drop and equal flow in each mixing unit, are the engineering solution that makes parallelized microfluidic LNP production feasible.
At commercial manufacturing scale, most facilities use impingement jet mixing as the primary technology for mRNA LNP production because it achieves the required throughput in a single mixing device without requiring the chip-level flow distribution engineering of parallelized microfluidics. Single-use IJM assemblies, in which the nozzles and fluid path are disposable, are the standard commercial format. The key critical quality consideration in scaling from SHM to IJM is verifying that the particle size and encapsulation efficiency of the product are equivalent at each scale, since the mixing mechanism differs between the two approaches.
What Downstream Processing Steps Follow LNP Formation?
LNPs emerging from the mixing step are suspended in a mixture of the aqueous formulation buffer and approximately twenty to thirty percent ethanol carried over from the organic stream. This post-formation intermediate requires downstream processing to remove the ethanol, exchange the low-pH formulation buffer for the final drug substance buffer, and concentrate the LNP suspension to the target drug substance concentration.
Tangential flow filtration is the primary technology for both ethanol removal and buffer exchange. Hollow fiber TFF modules, operating with a molecular weight cutoff in the one hundred to three hundred kilodalton range that retains the LNP particles while allowing buffer components, ethanol, and unencapsulated nucleic acid to pass into the permeate, can simultaneously remove ethanol, increase pH to physiological levels through diafiltration, and concentrate the LNP suspension to the final target concentration in a single integrated TFF operation.
Unencapsulated nucleic acid represents a critical quality attribute for LNP drug substances, since free nucleic acid in the final product can contribute to adverse immunostimulatory responses. TFF removes a substantial fraction of unencapsulated material through the permeate, but the ultimate encapsulation efficiency specification is determined by the combination of the formation step and the TFF polishing step together. For mRNA LNP formulations, encapsulation efficiencies above ninety percent after TFF processing are achievable with optimized formation parameters.
For the in-line buffer dilution systems used to prepare the LNP formulation buffers and the diafiltration buffers at precise pH and conductivity, see In-Line Buffer Dilution: Engineering and Implementation. For the facility-level design of buffer management infrastructure supporting LNP manufacturing, see Solving the Facility Bottleneck: Advanced Buffer Management.
This article was produced under Separation Science's AI Editorial Guidelines.
