Development of a measurement cell for AF4 SANS

Abstract (Swedish)

Asymmetrisk flödesfraktionering (AF4) är en separationsteknik som används för att separera större partiklar, såsom lipidnanopartiklar (LNP) eller makromolekylära polymerer. I kombination med småvinkelneutronspridning (SANS) ger detta möjlighet att undersöka partiklarnas inre struktur, istället för att endast mäta partiklarnas allmänna storlek. Att integrera SANS-mätningen som en kontinuerlig mätning medför dock vissa utmaningar. Eftersom neutronstrålen är större vid lägre flux än till exempel röntgenstrålning spelar geometrin en viktig roll för både uppehållstid, toppbreddning och signalhöjd. Med hjälp av beräkningsströmningsdynamik är det möjligt att digitalt förutsäga prestandan hos flera möjliga geometrier innan man tillverkar en kostsam fysisk cell.

Med hjälp av Ansys Fluent skapades en pipeline som genererar geometrier, beräknar flödes\-mönstret och sedan använder antingen en diskret fasmodellering (DPM) för att spåra partiklar genom fluidcellen eller en transient transportmodell för att simulera en spårämneskoncentration. Med hjälp av dessa partikelspår beräknades sedan ett simulerat fraktogram, från vilket halvbredden (FWHM) samt topphöjden beräknades. Dessa mått jämfördes sedan mellan tre olika serier av geometrier: en droppformad cell som tidigare använts i fysiska experiment, en rektangulär cell med parametriserade övergångar samt en serpentinformad flödesceller med konstant flödesbredd.

Resultaten visar att DPM-modellen och transportmodellen ger mycket likartade resultat, vilket bekräftar användningen av DPM-modellen. Dessutom visar det vissa avvägningar mellan den täckta strålytan och dess effekt på toppens bredd och höjd.

Abstract

Asymmetric flow field flow fractionation (AF4) is a prominent separation technique used to separate larger particles such as lipid nanoparticles or macromolecular polymers. When combined with Small angle neutron scattering (SANS), this provides the possibility to probe the inner structure of the particles, instead of only probing the general size of the particles. However, putting the SANS measurement in-line as a continuous measurement provides some challenges. Since the beam is bigger at a lower flux than for example X-rays, the geometry plays an important role in affecting both residence time, peak broadening and signal height. Using computational fluid dynamics, it is possible to predict the performance of multiple possible geometries digitally before manufacturing a costly physical cell.

Using Ansys Fluent, a pipeline was created which creates geometries, calculates the flow pattern and then uses either a discrete phase modeling (DPM) to track particles through the fluid cell or a transient species transport model to simulate a tracer concentration. Using these particle tracks, a simulated fractogram was then computed from which full width at half maximum (FWHM) as well as peak height was calculated. These metrics were then compared between three different series of geometries: a teardrop shaped cell used previously in physical experiments, a rectangular cell with parametrized transitions, as well as a serpentine flow cell with constant flow width.

The results show that the DPM model and the species model give highly similar results, verifying the use of the DPM model. Furthermore it shows some tradeoffs between the beam area covered and its effect on peak width and height.

Popular Abstract

How can we look into the core of microscopic particles that are used when delivering modern mRNA vaccines? By combining sorting with neutron beams! This masters thesis uses numerical simulations to design the perfect measurement chamber. These will ensure that enough data is collected while preventing previously separated particles from mixing again, enabling scientists to look even further into particles to understand their inner structure.

When the Covid-19 pandemic broke out, the broader public got introduced to the concept of lipid nanoparticles; microscopic fat particles engineered in such a way that they can be loaded with a payload such as a vaccine or other medicine. However, the design of these particles is far from simple, so to make sure that the particles work as intended, one need to look both at the size of the particle, but also at the inside. To achieve this, scientists combine two tools. Firstly a sorting technique to separate the particles by size and then a powerful microscope using neutrons instead of light. Unlike normal light, the neutrons are able to penetrate into the particles to image the inside without damaging the structure of the cell.

The combination of these two techniques is however far from simple. The neutron beam is wider than for example X-rays and is at the same time orders of magnitude dimmer. Compare the number of photons coming from a 15 watt lightbulb which in a conservative estimate emits 10^15 photons per second with the neutron beam which has about 10^9 neutrons per second, meaning that the normal lightbulb emits around 1 million more photons a second. As such, it becomes important to use as many of the neutrons as possible in the measurements.

The tube coming from the separation is less than 1 mm in diameter, while the beam is about 5 mm in diameter. Therefore, the tube needs to widen so that the whole beam will pass through the tube. If the tube widens too much, some particles will pass along the side of the beam, and never be hit by it, while if the tube is too narrow, some of the neutrons will not have the possibility to hit a particle. This is were this thesis comes in.

Instead of manufacturing many costly prototypes of quartz or sapphire glass, suitable for neutrons, this project uses computational fluid dynamics to essentially build a digital copy of the tube. This calculates how a fluid would behave if it would flow through the tube, which is then used to predict the path that particles would move.

This virtual screening gives the possibility to test more possible geometries, while at the same time cutting down on the number of physicals prototypes needed. And as computers continue to speed up, this process will get even faster than creating the physical model and running the actual samples.

Knowing the benefits and drawback of each cell will enable further researchers to select the cell that works best for their specific case. Maybe they absolutely need to make sure that their particles do not get mixed again, but don't care as much if all particles get hit? Or maybe it is more important that all particles get hit, but it is okay if they get mixed a bit during the process? This will be the task of future users of these cells.

• author: Tamm, Witko
• supervisor: Andreas Håkansson
• organization: Chemical Engineering (M.Sc.Eng.)
• course: KLGM16 20261
• year: 2026
• type: H2 - Master's Degree (Two Years)
• subject: Technology and Engineering; Chemistry
• keywords: AF4, SANS, Flowcell, CFD, Pharmaceutical formulation
• language: English
• language: Swedish
• id: 9231205
• date added to LUP: 2026-06-04 12:30:52
• date last changed: 2026-06-08 10:43:05