Cis-, trans-Perfluorodecalin 95% was ordered from Merck KGaA/Sigma-Aldrich (Darmstadt, Germany). PLGA was obtained as Resomer RG 503 H from Evonik Industries (Darmstadt, Germany). Polyvinyl alcohol (PVA), Type: Mowiol 4–88, was obtained from Kuraray Co., Ltd. (Tokyo, Japan).

As primary packaging 3 mL ISO glass cartridges (siliconized) were obtained pre-crimped from Schott AG (Mainz, Germany). The cartridges were crimped with Bi-Layer combiseals (6194) by West Pharma (Eschweiler, Germany). The plunger stoppers used were West NovaPure WP-457.

Glass cartridges of this type are a common primary packaging of injectables. The glass cartridges were used for the sedimentation experiment, while for the centrifugation experiment common Eppendorf-tube-like 2 mL plastic tubes were used.

For particle preparation we utilized an Ultra-Turrax T25 digital with S25N-G8 tool manufactured by IKA®-Werke GmbH & CO. KG (Staufen, Germany). For optical measurements a Tecan Reader infinite i200 was used (Tecan Trading AG, Switzerland). For particle size measurements a Zetasizer Ultra by Malvern Panalytical GmbH (Kassel, Germany) was employed.

A custom-made python script was used to calculate statistical tests. For differing between sample and control sets one-sided, two-sample welch-tests were used (99% confidence). For the outlier a two-sided, one-sample t-test was used (95% confidence). For consecutive tests a Bonferroni correction was applied. Mean values were calculated as arithmetic mean, standard deviations were calculated from unbiased sample variance.

PLGA particles were prepared with an emulsion solvent diffusion method: 50 mg of PLGA were dissolved in 1.5 mL of ethyl acetate. Then 2.5 mL of an aqueous PVA solution (2%) were added. The resulting ~ 4 mL were homogenized with the UltraTurrax for 30 s with 10,000 RPM in order to obtain an ethyl acetate in water emulsion. To precipitate the particles from the emulsion 16 mL of the PVA solution were added to the emulsion. This resulted in the precipitation of the particles from the emulsion droplets and the dissolution of the partially miscible ethyl acetate in water. The liquid mixture was left under a hood with constant stirring of ~ 400 RPM for evaporation of the ethyl acetate.

In order to obtain the final particle suspension, the process was conducted 5 times. All 5 suspensions were pooled and after ethyl acetate had evaporated the volume was adjusted to 200 mL with MilliQ water. The resulting PLGA solution was thus 250 mg PLGA particles in 200 mL liquid.

Eppendorf tubes were used for the centrifugation experiment. All tubes (6 samples and 6 controls) were filled with 1.5 mL of particle suspension. The 6 samples were additionally filled with 0.25 mL of PFD. Additionally, 3 tubes were filled with particle suspension to use the supernatant later as background measurement of the optical analysis.

All tubes were centrifuged with a Sigma 3-30KS centrifuge with 30,000 RCF for 30 min at 20 °C.

After centrifugation, the 12 tubes were taken from the centrifuge and inverted once for resuspension. 1 mL of the supernatant (after one inversion) was removed from the vial for measurement. For the measurement 200 µL of the removed supernatants were put into a 96-well plate. Furthermore, 200 µL of the 3 background controls were put into the plate. The plate was measured at 430 nm wavelength for absorption. The expectation here is that samples which resuspended better have a higher absorption because more particles float in the liquid.

The cartridges had to be filled nearly air bubble free. Remaining air was only permissible in a quantity where it would not contribute to a resuspension if the cartridge was inverted. In order to achieve this, cartridges were filled to the maximum possible point and a plunger stopper was put on the end of the cartridge while sliding it over the liquid, bulging out of the cartridge, sideways. It was then pressed a small part into the cartridge. Then a cannula was sticked through plunger stopper sideways, and the plunger stopper was pushed into the cartridge allowing the displaced liquid to leave through the cannula. The plunger stopper was inserted into the cartridge until the end of it met the end of the cartridge. The cannula which was now pinched between the glass of the cartridge and the plunger stopper was removed. The total volume inside the cartridge was thus ~ 3.5 mL. The amount of PFD in the samples was chosen to be 0.5 mL. With a relation of 6:1 of the two liquids to each other, the PFD amount was large enough to be more than ‘just a small bubble’ and small enough to be the minor part to the ‘formulation’ and an easy to use total volume.

Six sample cartridges were filled with 0.5 mL PFD and then with water-based particle suspension to the top (~ 3 mL). For the 6 controls the cartridges were filled first with 0.5 mL MilliQ water and then to the top with water-based particle suspension. This way the total amount of particles in both cartridge sets was be the same, which we deemed important when trying to get the experiment as close to conditions of a real medicine where the amount of particles correlates with the dose of the medicine. To compensate this dilution of 1:1.143 in the controls of the sedimentation experiment, all measured absorption values of the controls were multiplied with 1.143. The samples were stored at 2–8 °C in a refrigerator.

After roughly two months of sedimentation (55 days) the cartridges were retrieved from the refrigerator and brought to the laboratory bench (room temperature). They were taken out of the box in which they were stored without being tilted. Then they were inverted twice for resuspension. Following that a needle was pierced through the septum of the cartridge and the cartridge was pointed downwards. Then the plunger stopper at the back of the of the cartridge was pushed with plunger rod taken from a syringe in order to press liquid out through the needle into an Eppendorf tube. (Note: for the samples with PFD the needle was always pushed deep enough into the cartridge to bypass the PFD which would run to the tip of the cartridge because of its high density.) This way approximately 1.5 mL of potentially resuspended particles were withdrawn from the cartridges and put into Eppendorf tubes.

The retrieved liquid was then resuspended to homogeneity and 200 µL of each was pipetted into a well of a 96-well plate. The plate was measured analogously to the centrifugation experiment.