Liquid-to-liquid mixing without vortex formation

In this study, we characterized mixing of miscible liquids with differing densities using conductivity sensors and visual observation in four of our single-use mixing systems covering a range of volumes:

• Allegro™ mixer (50 L cubical tank).
• LevMixer™ system (100 L cubical tank).
• LevMixer™ system (400 L cubical tank).
• Magnetic mixer (1500 L cubical tank).

Under the conditions tested, we recorded homogenization times ranging from 0.32 min to 4.58 min without vortex formation, satisfying our target requirement of ≤ 5 min.

Introduction

Mixing of miscible liquids of similar characteristics is a common requirement in bioprocessing and selecting appropriate mixing technologies and mixing conditions requires consideration. Typically, there is a requirement for the mixing impeller speed to be high enough to reach homogeneity in a generally acceptable timeframe (≤ 5 min), while being low enough to avoid vortex formation (which can damage products sensitive to air-liquid interface interactions).

In our experiments outlined below we describe liquid-liquid homogenization conditions which meet these requirements at various filling levels from 50 to 1500 L in the selected mixers. Our Magnetic mixer and LevMixer systems use the same tanks with mixing impellers with the same dimensions. Therefore, the results from the LevMixer system are also deemed applicable to the Magnetic mixer system and vice versa. To cover most typical biotech sensitive applications, we conducted the tests under stringent conditions in terms of density of the spiking solution and bulk.

Our mixing impellers are bottom mounted, so we considered addition from the top of a low-density liquid to a high-density bulk solution to be a worst-case scenario. We chose glucose (24.3% w/w aqueous solution, density of 1.098 kg/L) as the bulk solution, and NH4Cl (10% w/w aqueous solution, density of 1.029 kg/L) was used as the spiking solution.

Materials and methods

For each mixing tank, we installed the biocontainer, inflated, and then cut open at the top for better visibility of the mixing process.

As presented in Figure 1, two conductivity sensing probes were inserted into the biocontainers:

• At the bottom, through the front probe fitment.
• At the top, in the front corner (at a depth of 5 to 10 cm below the liquid surface).

We prepared the bulk solution, containing 24.3% glucose w/w and 0.05% polysorbate 80 w/w in advance. After starting the timer and data logging on the transmitter, the biocontainer was filled to various volumes with the bulk solution and stirred at various speeds. Each mixing test started at nominal maximum volume (e.g., 400 L in a 400 L biocontainer), and then we reduced the volume stepwise for successive tests. In each test, we evaluated the vortex by viewing it from the top and through the door window. After mixing for at least 1 min (to reach a steady state), spike additions of 0.1% of tank fill volume were made with NH4Cl 10% solution. The addition was made at the top corner opposite the top conductivity probe (see Fig 1) from a height of 5 to 20 cm above the liquid surface.

Mixing continued for at least 5 min before the next mixing conditions were set. If homogenization time or vortex formation were not adequate, we adapted the mixing speed accordingly.

Based on Dechema recommendations, 95% and 99% homogenization times were defined as the time elapsed after the addition so that all sensor readings plateaued within ± 5% and ± 1% respectively of their signal shifts (1).

Air-liquid interfaces were evaluated by visually confirming:

• the absence of significant vortex formation (< 3 cm depth).
• the absence of foaming.

Fig 1. Location of liquid addition and probe insertion in the biocontainer.

Results

We conducted tests under conditions which avoided significant vortex formation. The homogenization times are reported in Table 1.

Table 1. Homogenization times based on conductivity in different technologies and tank sizes at various mixing speeds and filling levels

Tank size (L)	Mixer technology	Filling level (%)	Mixing speed (rpm)	99% Homogenization time (min)	95% Homogenization time (min)
50	Allegro mixer	100	46	0.65	0.53
50	Allegro mixer	100	46	0.70	0.53
50	Allegro mixer	100	46	0.72	0.52
50	Allegro mixer	100	39	1.53	0.67
50	Allegro mixer	44	55	0.48	0.32
50	Allegro mixer	44	39	0.72	0.53
50	Allegro mixer	18	39	1.20	0.43
50	Allegro mixer	8	39	0.92*	0.28*
50	Allegro mixer	4	39	0.97†	0.88†
100	LevMixer system	100	50	0.77	0.60
100	LevMixer system	100	50	0.72	0.65
100	LevMixer system	100	42	0.85	0.63
100	LevMixer system	44	42	0.87	0.57
100	LevMixer system	44	42	0.73	0.52
100	LevMixer system	44	42	0.75	0.43
400	LevMixer system	100	101	1.20	0.90
400	LevMixer system	100	73	2.03	1.02
400	LevMixer system	100	73	1.97	1.50
400	LevMixer system	100	73	1.88	1.43
400	LevMixer system	60	73	1.55	0.78
400	LevMixer system	60	65	1.20	0.77
400	LevMixer system	60	65	1.43	1.03
400	LevMixer system	35	65	1.17‡	0.68‡
400	LevMixer system	35	58	1.02	0.78
400	LevMixer system	35	58	1.23	0.78
400	LevMixer system	35	58	1.23	0.78
400	LevMixer system	15	58	1.48	0.98
400	LevMixer system	15	58	1.23	0.73
400	LevMixer system	15	58	0.92	0.82
1500	Magnetic mixer	100	180	3.75	2.55
1500	Magnetic mixer	100	180	2.55	1.10
1500	Magnetic mixer	97	101	6.42	5.12
1500	Magnetic mixer	55	100	1.87	1.23
1500	Magnetic mixer	55	100	1.68	1.33
1500	Magnetic mixer	52	161	1.10	0.88
1500	Magnetic mixer	52	161	1.00	0.73
1500	Magnetic mixer	52	161	0.97	0.77
1500	Magnetic mixer	40	82	2.17	1.10
1500	Magnetic mixer	40	82	1.65	0.92
1500	Magnetic mixer	40	50	4.42	3.87
1500	Magnetic mixer	40	50	3.78	2.78
1500	Magnetic mixer	27	43	3.53	2.90
1500	Magnetic mixer	27	43	3.08	2.18
1500	Magnetic mixer	27	43	3.38	1.65
1500	Magnetic mixer	8	43	4.58	2.98
1500	Magnetic mixer	8	43	3.57	2.23

^*Only the bottom probe was used.

^†Only the top probe was used. Initially no stable conductivity signal could be collected so empirical homogenization times were chosen.

^‡Intermittent vortex formation > 3 cm < 50% of the time was observed.

Conclusion

We characterized mixing of miscible liquids with differing densities using conductivity sensors and visual observation in four of our single-use mixing systems: Allegro mixer 50 L cubical tank, LevMixer system 100 L cubical tank, LevMixer system 400 L cubical tank, and Magnetic mixer 1500 L cubical tank. Under the conditions tested, homogenization times ranged from 0.32 min to 4.58 min without vortex formation, satisfying the target requirement of ≤ 5 min.

Reference

1. W. Meusel et al, Recommendations for Process Engineering Characterisation of Single-Use Bioreactors and Mixing Systems by Using Experimental Methods Dechema, 2016. https://dechema.de/dechema_media/Downloads/Positionspapiere/SingleUse_ProcessEngineeringCaracterisation_2016.pdf  (accessed October 18, 2021).

CY54062