Assessing particle generation in the LevMixer™ 10 L single-use mixing system

Table 1. Equipment and materials used

Table 2. Experimental conditions and parameters

We set up the LevMixer™ 10 L single-use mixing system hardware, and the installed the single-use biocontainer, following procedure from Cytiva’s published Instructions For Use (IFU).

Before filling the biocontainer, we flushed the filling assembly, containing a Mini Kleenpak™ DJL 0.1 µm liquid filter inline with 2 L of ultrapure water to hydrate the membrane and thoroughly rinse the fluid line. The single use biocontainer was then filled and mixed with 5 L of ultrapure water twice for five minutes to rinse the bag, minimize initial particle levels, and improve the accuracy of particle detection, even though our biocontainers meet USP <788> compliance.

To verify system integrity, we collected two control samples:

• One directly from the ultrapure water source.
• A second sample from the filling manifold, after filtration to ensure filter integrity.

The biocontainer was first inflated to its final 3D geometry using a manual hand pump through a Supor™ EKV Kleenpak™ 0.1 µm vent filter. It was then filled by weight with 5.4 ± 0.05 kg of ultrapure water (assuming a density of 1000 kg/m³), using a pre-hydrated Mini Kleenpak™ DJL liquid filter and a quick-fit connector. Mixing was initiated at maximum impeller speed and maintained for 24 hours under controlled temperature conditions. For the data points at 40 °C and 4 °C, gentle agitation was started overnight prior to full-speed mixing.

Sampling

We collected 300 mL samples from the mixing biocontainer, at four time points during mixing:

• Sample 1: T = 0 h (after 1 minute of mixing)
• Sample 2: T = 2 h
• Sample 3: T = 6 h
• Sample 4: T = 24 h

Before each sample collection, we flushed the sampling line with 100 g of bulk water from the mixer (discarded) to clear the line and eliminate any residual content from the previous sample. We then collected a 300 g sample. All samples were drawn by gravity through the biocontainer’s drain port while mixing was active. The targeted initial working volume of 5 L was obtained after the first sample was collected (T=0h).

Data processing, analysis and particle counting

We analyzed all water samples, including blanks, for sub-visible and visible particles according to the USP <788> count Method 2 (Microscopic particle count test). We determined the total particle counts by summing the particles detected in the initial 300 mL sample with those observed in three subsequent 200 mL flushes of the biocontainer. We used a new single use biocontainer for each sample.

To account for potential background contamination, we collected blank samples prior to each experiment and processed them identically to the test samples. These included:

• An aliquot drawn directly from the ultrapure water source.
• Three 200 mL flushes using the same water.

For each particle size category, we subtracted the sum of the particles counted in the blank from the sum of particles in the samples. This meant we took into consideration the potential particle presence in the ultrapure water. This was calculated as follows:

Total adjusted particle count = (∑ sample + flushes) – average (∑ blanks + flushes)

To compare particle levels over time, we normalized the data obtained based on the maximum count for the T = 24 h sample, which was considered as being the 100%, This approach allowed us to track particle trends across different batches and temperature conditions, while accounting for baseline variability.

Normalization was carried out as shown in the calculation below:

Where the:

• Particle count in sample corresponds to the individual particle counting for each sample time per experiment (batches 1, 2, and 3 at room temperature and runs at 4 °C and 40 °C) and particle size (≥ 10, ≥ 25 and ≥ 100 µm).
• Max particle count (T=24h) represents the particle count at T = 24 hours for the batch that exhibited the highest total, for each particle size category (≥ 10 µm, ≥ 25 µm, and ≥ 100 µm).