Key points
SINGLE-use (SU) systems have been used in industry for up to the last 15 years. They are used in manufacturing a range of products including pharmaceuticals, biopharmaceuticals and within fast moving consumer goods (FMCG). This article looks at how single-use mixers compare with stainless steel systems, in particular for difficult-to-dissolve powders.
Single Use?
Here we provide a brief overview of the pros and cons of using SU systems (see Reference 1 for a more detailed review).
Single Use System would appear not be good for the environment due to disposal issues but as discussed below, on balance they are environmentally neutral or close to neutral when compared with traditional fixed stainless-steel equipment.
One of the principal drivers for using SU systems is speed to market. Detailed information on single use systems is not readily available in standard texts and publications. Therefore, we took this opportunity to find more detailed information from the vendor. Patent protection is very important in the pharmaceutical industry and with limited patent lives it is important to place a product on the market swiftly to gain the maximum amount of patent life as possible. Speed to market (among other things) is improved by the following factors.
Cleaning and Sterilisation
Since SU systems do not need to be cleaned and sterilised (they are sterilised at manufacture using gamma irradiation), this means:
Reduced high purity water, steam and energy use.
Lower capital cost.
Lower space requirement.
No commissioning, qualification and validation (CQV) required, and hence shorter project times.
Modularity
SU systems are usually suitable for working in modular systems which means that production can be scaled out rather than scaled up and again, saving on CQV requirements.
Configurability
SU systems are simple and cheap to reconfigure. This means that process changes can be made rapidly and cheaply. Some SU systems are highly configurable and have many of the features of fixed stainless systems. However, these systems have a higher capital cost than standard less configurable SU systems and may have a capital cost similar to stainless steel systems.
Operating Costs
Although the usage of high purity water, steam and energy is reduced with SU systems compared to stainless steel systems it is worth noting that the SU components are specialist items and their purchase price can be significant and the overall running costs for SU systems may be no lower and can be even higher than for stainless steel systems depending on the setup of the SU System.
Extractables and Leachables
A potential disadvantage of SU systems is that materials can be extracted (under test conditions) or leached (under process conditions) from the system’s plastic film. This is considerably less of a problem with stainless steel systems. Manufacturers of SU systems have recognised this problem and minimising this risk.
| Comparison factor | Single use systems | Stainless steel systems |
| Cleaning and sterilisation | Not required | Requires the use of cleaning/sterilisation liquids and potentially steam |
| Modularity | Good for scale-out | Possible but with a high capital cost |
| Configurability | Good for reconfiguration | Possible but with a high capital cost |
| Operating costs: materials | High, due to the cost of the single use system components plus the process materials | Only includes process materials |
| Operating costs: utilities | Low, since cleaning and sterilisation is not required | High, since cleaning and sterlisation is required |
| Operating costs: labour | Depends on system configuration | Depends on system configuration |
| Extractables and leachables | Potentially a problem but largely addressed by manufacturers | Not usually a problem |
SU system case study
We have been working with SU systems for some time and along with many other chemical engineers, we have found that SU systems are particularly useful in biopharmaceutical applications. During a recent project we were able to take a detailed look at the performance of a standard SU system in more detail. We have been involved with a project where the Flexsafe® Pro Mixer System is being used and this equipment was selected by the equipment users project team based on prior experience.
The SU system that we reviewed was the Flexsafe® Pro mixer system from Sartorius (Figure 2), which, as is the case with most standard SU systems, is available in sizes from 50L to 3000L. 2000L. It consists of a portable drive unit and stainless steel Palletank casing, into which the SU bag is inserted. In this application the mixing system also included weighing system which is a common feature in my SU mixing systems.
The drive unit couples to the mag drive impeller mixer, which is located inside the SU bag. The impeller (see Figure 3) can operate between 20–750 rpm mixing speed.
Mixing
A detailed study has been carried out Sartorius on the Flexsafe® Pro Mixers mixing capability[ref2] for sodium CMC (carboxymethylcellulose) dissolution, using a 650L system.
Addition of the sodium CMC in small amounts of 1.5–2.5 kg mixed into the vortex created by the impeller helped create a slow and homogenous mixture in the bag and helped increase mixing efficiency.
The results of the Sodium CMC dissolution are shown in Table 2.
| Run | Impeller speed | Mixing time confirmed by visual inspection | Observation |
| 1 | 700 rpm | 1 h 30 | Sodium CMC in solution with no remaining clumps prior to final water addition |
| 2 | 700 rpm | 1 h 50 | Few clumps still observed after 1 h 35 when the filling to 100% started |
| 3 | 400 rpm (for 1 h 30) then 700 rpm | 2 h 12 | Still a great number of clumps observed after 1 h 30 at 400 rpm. Impeller speed raised to 700 rpm. Only a few clumps visible after 1 h 45 when the filling to 100% started |
In runs 1 and 2, complete dissolution of the sodium CMC was achieved due to the slow addition of the powder into the water while the impeller mixed at a high speed of 700 rpm. Complete dissolution was achieved prior to final water addition. In run 3 a slower mixing speed of 400 rpm was used to begin with, and this showed that clumps formed prior to final water addition to bring to final volume. The speed was increased to 700 rpm and complete dissolution was achieved within 42 minutes at this speed (see Figures 4 and 5).
| Run | Impeller Speed | Testing Time | Mixing status after visual inspection | Observation |
|---|---|---|---|---|
| 1 | Max speed | 3h 03 | Few clumps left | Clumps still observed after 2h30 when the filling to 100% started |
| 2 | Max speed | 3h 50 | Full dissolution confirmed | Few clumps still observed after 2h45 when the filling to 100% started |
| 3 | Max speed | 4h | Few clumps left | Clumps till observed after 3h when the filling to 100% started |
| Volume | ||||||
| 50L | 100L | 200L | 400L | 650L | 1000L | |
| Reynolds Number | 3.2E+05 | 3.2E+05 | 3.2E+05 | 3.2E+05 | 3.2E+05 | |
| Regime | Turbulent | Turbulent | Turbulent | Turbulent | Turbulent | Turbulent |
| Power number N | 1.82 | 1.92 | 2.14 | 2.28 | 226 | 2.21 |
| Power(W) | 372.7 | 393.2 | 438.3 | 466.9 | 462.8 | 452.6 |
| Flow number N | 0.54 | 0.52 | 0.48 | 0.48 | 0.45 | 0.44 |
| Q(m^3/min) | 1.66 | 1.60 | 1.47 | 1.47 | 1.38 | 1.35 |
| Circulation time tc (sec) | 1.8 | 3.8 | 8.1 | 16.3 | 28.2 | 44.4 |
| P/V (kW/m^3) | 7.45 | 3.93 | 2.19 | 1.17 | 0.71 | 0.45 |
Table 4 lists performance data across a range of system sizes, from 50–1,000 L; notably, high Reynolds number of 3.20 x 105 and good turbulent flow are recorded in this range. From the author’s previous experience, the power numbers and the power per unit volume is comparable to other standard single use systems but lower than that expected for comparable stainless steel systems. The power per unit volume is also lower than for the configurable SU system shown in the box.
A couple of speed profile studies were completed. Figure 7 shows side view velocity analysis using computational fluid dynamics (CFD). In most of the trials, vortex formation was confirmed, characterised by the yellow and red colours. The green colour observed near the top surface of the system indicates a fluid speed of 0.5 m/s.
The impeller in all standard SU systems is close to the base of the bag and this impacts on the mixing flow pattern. As Figure 3 shows the Pro Mixer impeller is raised about the bottom of the bag and this aids with the mixing. The flow near the top surface is important when mixing solids that tend to float on the surface since it ensures that the solids can be drawn down into the vortex created by the impeller. It is worth noting the impeller for the configurable SU system shown in the box is mounted off the bottom of the vessel which is one of the reasons the mixing is more comparable with stainless steel systems.
This study confirms that turbulent flow and good mixing velocities are observed throughout the bag. In the 1,000 L system, mixing speeds of 2 and 3 m/s were observed. The turbulence is important and is generated in a radial flow. When the turbulent flow reaches the walls of the bag, an axial circulation is also observed throughout the bag which helps avoid dead zones or low velocity areas. This creates a re-circulation pattern above the stirrer which allows more efficient mixing. It has been highlighted that a moderate vortex is required for rapid inclusion and dispersion of floating powders into the liquid.
In traditional cylindrical stainless steel systems, the use of baffles removes vortex formation but including floating powders is difficult, as the surface velocity slows down and leads to powders floating on the surface with slow mixing into the liquid. It is also noted that a rapid dissolution of powder into the liquid is important when using products that are susceptible to foaming.
The Flexsafe ® bag s’ square corners create a partial baffle effect in the bag which helps to create a moderate vortex.
The Pro Mixer was confirmed as the best solution for the mixing application in this project.
Heat Transfer
“U” Values
The SU systems purchased were required to heat and cool the mixer being manufacture so we requested detailed heat transfer information from Sartorius as part of the selection process. As part of their own product development work Sartorius have performed a heat transfer study on their SU system [Reference 3].].
In the study, four mixer bag sizes were used: 50 L, 100 L, 200 L and 1,000 L. T two set-points that this study was looking to achieve were 1) a heating set-point of 38°C +/-0.5°C and 2) a cooling set-point of 3°C +/- 0.5°C
Observed overall heat transfer coefficient (U) was plotted against predicted U values for 200 L MLR (Multiple Linear Regression). See Figure 8. For the different volume systems, the R2 value was above 0.75. This is a good result as a R2 around 1 shows complete correlation between the dependent and the independent variables.
Heating/Cooling Times
Comparable heating and cooling times are shown for the studies performed at 50 L (two heating studies only), 200 L, 1,000 L and 2,000 L. See Figure 9. This is important and shows that heating from 4°C to 20°C in a 2,000 L system and heating from 20°C to 37°C is the same even up to 2,000L, and takes approx. 5 hours in both systems.
The average heating and cooling rates are recorded in Table 5.
| Palletank® volume | Average heating rate (°C/h) | Average colling rate | ||
|---|---|---|---|---|
| 4°C to 20°C | 20°C to 37°C | 37°C to 20°C | 20°C to 4°C | |
| 50L | 16.68 | 17.24 | – | – |
| 200L | 9.56 | 9.89 | 11.21 | 10.85 |
| 1000L | 4.98 | 5.14 | 5.01 | 4.85 |
| 2000L | 3.24 | 3.13 | 3.13 | 3.03 |
The system’s impeller can mix from 20–750 rpm. The cube-shaped bag has features which allow improved heat transfer and help achieve turbulent flow . This cube shape creates a partial baffle effect which helps mix floating powder residues into the liquid. Also, the design of the impeller allows a radial flow pattern to be generated, and when it reaches the walls of the bag it changes to an axial flow. As a result of this motion a recirculation loop is achieved above the bottom-mounted impeller, which gives improved mixing compared to unbaffled cylindrical SU and unbaffled stainless steel systems. In addition, the corners of the bag create a moderate vortex from the square-shape design.
Jacket Pressure Drop
As part of the Sartorius development work, the jacket pressure drop was also recorded and monitored for each of the systems. Jacket pressure loss data is required to allow the design of the heat transfer media supply system. See a summary of the pressure drop results in Figure 10.
The pressure drop increases with the jacket flow rate and with the Palletank volume. It is interesting to note that the pressure drop is higher during cooling than in the heating phase for the same volume system. The pressure drop is changing due to the impact of temperature on the viscosity. For the cooling phase the viscosity is higher, so the pressure drop is more, and for the heating phase the viscosity is lower therefore a lower pressure drop is observed.
This study also examined the flow rate of coolant in the jacket versus time to reach the target temperature.
Three jacket flow rates were assessed in the 200 L system. These were 20 L/min, 35 L/min and 50 L/min.
The results of this study are recorded in Figure 11 below.
The pressure drops for each of the studies are recorded in Table 5 below.
| Jacket flowrate (L/min) | Time needed to reach target temperature | Pressure drop (psig) |
|---|---|---|
| 20 | 2H19 | 0.76 |
| 35 | 1H58 | 2 |
| 50 | 1H43 | 4.1 |
It was shown that the slowest jacket flowrate gave the smallest pressure drop. Contrary to this, the fastest flowrate of cooling solution through the jacket gave the highest pressure drop. The data shows that increasing the flowrate of glycol through the jacket by 2.5 times leads to a reduction of 26% in time. Again, this information enhances the design of the heat transfer and feeds into defining the batch time requirements.
Conclusions
Difficult-to-dissolve powders such as NaCMC can be successfully mixed to form a true solution using the SU system. From the studies performed, a true solution was achieved at 700 rpm in 1 hour 30 minutes.
Heat transfer was also performed in the same SU system; heating and cooling rates were determined for system sizes of 50–2,000 L. The average heating rates are much faster in the 50 L system and slowest in the largest 2,000 L system. A similar trend was obtained for cooling where cooling was the fastest in the 200 L system and the slowest in the 2,000 L system.
Pressure drop was also an important parameter that was monitored. The pressure drop increases with the jacket flow rate and with the Palletank volume. The pressure drop is highest during cooling than in the heating phase for the same jacket volume.
The detailed information provided by Sartorius as part of their product development programme allowed the project team to confirm that the chosen SU system complied with all of the function requirements of the relevant part of the manufacturing process. Also, this information illustrates the typical performance of a standard SU mixing systems but the performance of SU systems from other vendors will differ in a number of ways and therefore we recommend that others considering SU systems should ensure that they request similar performance data from the vendors that they are reviewing of part of their vendor pre-qualification process.
Bioreactors provide a useful comparison between highly configurable SU systems (as opposed to a standard SU systems) and stainless steel vessels. A comparison can be made using the mass transfer coefficient which is an indication of the efficiency of the bioreactor configuration.
Equation 11–19 from the Handbook of Industrial Mixing gives:
Where kL is mass transfer coefficient and a is the gas-liquid interfacial are per unit volume. kLa – the mass transfer factor is frequently taken as indication of how good a vessel is at promoting mass transfer. For vessels using the same gas liquid mixture, this formular shows that kLa is proportional to power per unit volume.
If you compare the power input of a 2,000 L highly configurable SU system with a 2,000 L stainless steel vessel (complete with baffles), the power per unit volume for each system is of the same order and hence the mass transfer will be similar. The exact mass transfer will depend on the detailed configuration of each system.
References
- https://www.pharmaceuticalonline.com/doc/single-use-systems-the-benefits-the-challenges-and-selection-considerations-0001
- Audubey, Delphine et al, “Efficiency of NaCMC Powder Dissolution Using the Flexsafe® Pro Mixer”, December 2020, 1-6
- Audubey, Delphine et al Flexsafe® Pro Mixer, “The Fast, Flexible and Intelligent Solution for Heat Transfer Applications”, July 2020, 1-9
- “Flexsafe® Pro Mixer Computational Fluid Dynamic Studies”, 1-11
Bioreactor Agitation Parameters’ from “A rapid low risk approach for Process Transfer of Biologics from Development to Manufacturing Scale” - “Biostat STR® Generation 3 and Biobrain® Automation Platform”, 1-4
- “Validation Guide”, Flexsafe® Pro Mixer Bag, 1-41
- Handbook of Industrial Mixing: Science and Practice, 2004, Wiley
BIOSTAT® STR2000 – Corpus-C Design Agentur – Product Design, User Interface Design
Authors:
Mary Carty
Mary Carty, MSAT (Manufacturing, Science and Technology) Senior Engineer, currently working in biopharmaceuticals. She has 11 years experience working in the pharmaceuticals, biopharmaceuticals and cosmetics goods industries. She has been a Process Engineer for most of her career to date.
Mary has been directly involved in creating and implementing change to improve manufacturing processes, along with the introduction of lean six sigma problem solutions and the establishment of root causes for deviations that occur. She has also held leading roles in continuous improvement projects in production and also in Safety and Environmental Projects.
