Downstream Industrial Biotechnology: Recovery and Purification

By Michael C. Flickinger

An affordable, easily accessible desk reference on biomanufacturing, focused on downstream recovery and purification

Advances in the fundamental knowledge surrounding biotechnology, novel materials, and advanced engineering approaches continue to be translated into bioprocesses that bring new products to market at a significantly faster pace than most other industries. Industrial scale biotechnology and new manufacturing methods are revolutionizing medicine, environmental monitoring and remediation, consumer products, food production, agriculture, and forestry, and continue to be a major area of research.

The downstream stage in industrial biotechnology refers to recovery, isolation, and purification of the microbial products from cell debris, processing medium and contaminating biomolecules from the upstream process into a finished product such as biopharmaceuticals and vaccines.

Downstream process design has the greatest impact on overall biomanufacturing cost because not only does the biochemistry of different products ( e.g., peptides, proteins, hormones, antibiotics, and complex antigens) dictate different methods for the isolation and purification of these products, but contaminating byproducts can also reduce overall process yield, and may have serious consequences on clinical safety and efficacy. Therefore downstream separation scientists and engineers are continually seeking to eliminate, or combine, unit operations to minimize the number of process steps in order to maximize product recovery at a specified concentration and purity.

Based on Wiley's Encyclopedia of Industrial Biotechnology: Bioprocess, Bioseparation, and Cell Technology, this volume features fifty articles that provide information on down- stream recovery of cells and protein capture; process development and facility design; equipment; PAT in downstream processes; downstream cGMP operations; and regulatory compliance.

It covers:

• Cell wall disruption and lysis
• Cell recovery by centrifugation and filtration
• Large-scale protein chromatography
• Scale down of biopharmaceutical purification operations
• Lipopolysaccharide removal
• Porous media in biotechnology
• Equipment used in industrial protein purification
• Affinity chromatography
• Antibody purification, monoclonal and polyclonal
• Protein aggregation, precipitation and crystallization
• Freeze-drying of biopharmaceuticals
• Biopharmaceutical facility design and validation
• Pharmaceutical bioburden testing
• Regulatory requirements

Ideal for graduate and advanced undergraduate courses on biomanufacturing, biochemical engineering, biophar- maceutical facility design, biochemistry, industrial microbiology, gene expression technology, and cell culture technology, Downstream Industrial Biotechnology is also a highly recommended resource for industry professionals and libraries.

• Language: English
• Publisher: Wiley
• Release date: Jul 17, 2013
• ISBN: 9781118618981

Book preview

Chapter 1

Bioprocess Design, Computer-Aided

Victor Papavasileiou

Intelligen Europe, Leiden, The Netherlands

Charles Siletti

Intelligen, Inc., Mt. Laurel, New Jersey

Alexandros Koulouris

Intelligen Europe, Thermi, Greece

Demetri Petrides

Intelligen, Inc., Scotch Plains, New Jersey

1.1 Introduction

Bioprocess design is the conceptual work done prior to commercialization of a biological product. Given information on the potential market demand for a new product, bioprocess design endeavors to answer the following questions: What are the required amounts of raw materials and utilities for manufacturing a certain amount of product per year? What is the required size of process equipment and supporting utilities? Can the product be manufactured in an existing facility or is a new plant required? What is the total capital investment for a new facility? What is the manufacturing cost? How long does a single batch take? What is the minimum time between consecutive batches? During the course of a batch, what is the demand for various resources (e.g. raw materials, labor, and utilities)? Which process steps or resources are the likely production bottlenecks? What process and equipment changes can increase throughput? What is the environmental impact of the process? Which design is the best among several plausible alternatives?

Bioprocess design and project economic evaluation require the integration of knowledge from many different scientific and engineering disciplines. Design and evaluation are also carried out at various levels of detail. Table 1.1 presents a common classification of design and cost estimates and typical engineering costs for a $50 million capital investment project (1).

Table 1.1 Types of Design Estimates

NumberTable

Order-of-magnitude estimates are usually practiced by experienced engineers who have worked on similar projects in the past. They take minutes or hours to complete, but the error in the estimate can be as high as 50%. Table 1.2 provides a good example of information typically employed for order-of-magnitude estimates of the capital investment for cell culture facilities. It lists capital investment for cell culture facilities of various sizes built in the last 10 years. The last column displays unit cost of capital investment expressed in millions of US dollars per cubic meter of production bioreactor capacity. The numbers range between 2.5 and 6.2 and for the more recent facilities the numbers are in the 5–6.2 range. Consequently, using the data of Table 1.2, one can safely estimate the capital investment for a new cell culture facility with production bioreactor capacity of 100 m³ to be in the range of $500–650 million.

Table 1.2 Capital Investment for Cell Culture Facilities

NumberTable

Engineers employed by operating companies usually perform level 2 and 3 studies. Such studies take days or weeks to complete using appropriate computer aids. The main objective of such a study is to evaluate alternatives and pinpoint areas of high cost and low yield. The results are used to plan future research and development and to generate project budgets.

Level 4 and 5 studies are usually performed by engineering and construction companies that are hired to build new plants for promising new products that are at an advanced stage of development. These detailed estimates are beyond the scope of this chapter. Instead, the rest of this chapter will focus on level 2 and 3 studies. It should also be noted that opportunities for creative process design work are usually limited to preliminary studies. By the time detailed engineering work is initiated, a process is more than 80% fixed. Furthermore, most of the important decisions for capital expenditure and product commercialization are based on results of preliminary process design and cost analysis. This is why it is so important for a new engineer to master the skills of preliminary process design and cost analysis.

1.2 Benefits From the Use of Computer Aids

Process design calculations are greatly facilitated by the use of computer aids, such as spreadsheets, process simulators, finite capacity scheduling (FCS), and other specialized tools. Use of appropriate computer aids allows the process design team to quickly and accurately redo the entire series of calculations with a different set of assumptions and other input data. The benefits from the use of such tools depend on the type of product, the stage of development, and the size of the investment. For commodity biological products such as biofuels, minimization of capital and operating costs are the primary benefits. For high-value biopharmaceuticals, systematic process development that shortens the time to commercialization is the primary motivation. Figure 1.1 shows a pictorial representation of the benefits from the use of computer aids at the various stages of the commercialization process.

Figure 1.1 Benefits from the use of computer aids.

1.1

1.2.1 Idea Generation

When product and process ideas are first conceived, process modeling tools are used for project screening, selection, and strategic planning on the basis of preliminary economic analyses.

1.2.2 Process Development

During this phase, the company's process development groups look into the various options available for synthesizing, purifying, characterizing, and formulating the final product. At this stage, the process undergoes constant change. Typically, a large number of scientists and engineers are involved in the improvement and optimization of individual processing steps. The use of process simulation tools at this stage can introduce a common language of communication and facilitate team interaction. A computer model of the entire process can provide a common reference and evaluation framework to facilitate process development. The impact of process changes can be readily evaluated and documented in a systematic way. Once a reliable model is available, it can be used to pinpoint the cost-sensitive areas of a complex process. These are usually steps of high capital and operating cost or low yield and production throughput. The findings from such analyses can be used to focus further lab and pilot plant studies to optimize those portions of the process. The ability to experiment on the computer with alternative process setups and operating conditions reduces the costly and time-consuming laboratory and pilot plant effort.

The environmental impact of a process is another issue that can be readily evaluated with computer models. Material balances calculated for the projected large-scale manufacturing reveal the environmental hot spots. These are usually process steps that use organic solvents and other regulated materials of high disposal costs. Environmental issues not addressed during process development may lead to serious headaches during manufacturing. This is especially true for biopharmaceuticals because after a process has been approved by the regulatory agencies, it is extremely costly and time consuming to implement process changes.

1.2.3 Facility Design and/or Selection

With process development near completion at the pilot plant level, simulation tools are used to systematically design and optimize the process for commercial production. Availability of a good computer model can greatly facilitate the transfer of a new process from the pilot plant to the large-scale facility. If a new facility needs to be built, process simulators can be used to size process equipment and supporting utilities, and estimate the required capital investment. In transferring production to existing manufacturing sites, process simulators can be used to evaluate the various sites from a capacity and cost point of view and select the most appropriate one. The same can apply to outsourcing of manufacturing to contract manufacturers.

1.2.4 Manufacturing

In large-scale manufacturing, simulation tools are mainly used for on-going process optimization and debottlenecking studies. Other computer aids that play an important role in manufacturing include FCS, manufacturing resource planning (MRP), and enterprise resource planning (ERP) tools. FCS tools play an important role in batch chemical manufacturing. They are used to generate production schedules on an on-going basis in a way that does not violate constraints related to the limited availability of equipment, labor resources, utilities, inventories of materials, and so on. FCS tools close the gap between ERP/MRP tools and the plant floor (2). Production schedules generated by ERP/MRP tools are typically based on coarse process representations and approximate plant capacities and, as a result, solutions generated by those tools may not be feasible, especially for multiproduct facilities that operate at high capacity utilization. This can often lead to late orders that require expediting and/or to large inventories in order to maintain customer responsiveness. Lean manufacturing principles, such as just-in-time production, low work-in-progress (WIP), and low product inventories cannot be implemented without good production scheduling tools that can accurately estimate capacity (3, 4).

1.3 Commercially Available Tools

Process simulation programs, also known as process simulators, have been in use in the chemical and petrochemical industries since the early 1960s. Established simulators for those industries include: Aspen Plus and HYSYS from Aspen Technology, Inc. (Cambridge, MA), ChemCAD from Chemstations, Inc. (Houston, TX), and PRO/II from SimSci-Esscor, Inc. (Lake Forest, CA).

The above simulators have been designed to model primarily continuous processes and their transient behavior. Most biological products, however, are produced in batch and semicontinuous mode (5, 6). Such processes are best modeled with batch process simulators that account for time-dependency and sequencing of events. Batches from Batch Process Technologies, Inc. (West Lafayette, IN) was the first simulator specific to batch processes. It was commercialized in the mid-1980s. All of its operation models are dynamic and simulation always involves integration of differential equations over a period of time. In the mid-1990s, Aspen Technology (Cambridge, MA) introduced Batch Plus, a recipe-driven simulator that targeted batch pharmaceutical processes. Around the same time, Intelligen, Inc. (Scotch Plains, NJ) introduced SuperPro Designer. A unique feature of SuperPro is its ability to model batch as well as continuous processes (7).

Discrete-event simulators have also found applications in the bioprocessing industries. Established tools of this type include ProModel from ProModel Corporation (Orem, UT), Arena and Witness from Rockwell Automation, Inc. (Milwaukee, WI), Extend from Imagine That, Inc. (San Jose, CA), and FlexSim from FlexSim Software Products, Inc. (Orem, UT). The focus of models developed with such tools is usually on the minute-by-minute time-dependency of events and the animation of the process. Material balances, equipment sizing, and cost analysis tasks are usually out of the scope of such models. Some of these tools are quite customizable and third-party companies occasionally use them as platforms to create industry-specific modules. For instance, BioPharm Services, Ltd. (Bucks, UK) have created a module that runs on top of Extend and focuses on biopharmaceuticals.

MS Excel from Microsoft is another common platform for creating models for integrated processes that focus on material balances, equipment sizing, and cost analysis. Some companies have even developed models in Excel that capture the time-dependency of batch processes. This is typically done by writing extensive code (in the form of macros and subroutines) in VBA (Visual Basic for Applications) that comes with Excel. K-TOPS from Alfa Laval Biokinetics, Inc. (Philadelphia, PA) belongs to this category.

In terms of production scheduling, established tools include Infor SCM from Infor Global Solutions (Alpharetta, GA), Optiflex from i2 Technologies, Inc. (Irving, TX), SAP APO from SAP AG (Walldorf, Germany), ILOG Plant PowerOps from ILOG SA (Gentilly, France), Aspen SCM (formerly Aspen MIMI) from Aspen Technology, Inc. (Cambridge, MA), and so on. Their success in the biochemical industries, however, has been rather limited so far. Their primary focus on discrete manufacturing (as opposed to batch chemical manufacturing) and their approach to scheduling from a mathematical optimization viewpoint are some of the reasons for the limited market penetration.

SchedulePro from Intelligen, Inc. (Scotch Plains, NJ) is a new FCS tool that focuses on scheduling of batch and semicontinuous biochemical and related processes. It is a recipe-driven tool with emphasis on generation of feasible solutions that can be readily improved by the user in an interactive manner.

The rest of this chapter will address, through an illustrative example, the use of simulation and scheduling tools for evaluating and optimizing integrated biochemical processes. Analysis and assessment of additional bioprocesses can be found in the literature (8).

1.4 Monoclonal Antibody Example

Monoclonal antibodies (Mabs) are the fastest growing segment within the biopharmaceutical industry (9). More than 20 Mabs and Fc fusion proteins are approved for sale in the United States and Europe and approximately 200 Mabs are in clinical trials for a wide variety of indications (2). The market is predicted to grow by around 20% per year and reach $17 billion in 2008 (10).

The high-dose demand for several Mabs translates into annual production requirement for purified product in the metric ton range. Such a process is modeled and analyzed with SuperPro Designer in the rest of this chapter. Figure 1.2 displays the flow sheet of the overall process. The generation of the flow sheet was based on information available in the patent and technical literature combined with our engineering judgment and experience with such processes. The computer files for this example are available as part of the evaluation version of SuperPro Designer at the website www.intelligen.com/literature. Additional examples dealing with other biopharmaceuticals as well as commodity biological products are available at the same website.

Figure 1.2 Monoclonal antibody production flow sheet.

1.2

To model an integrated process on the computer using SuperPro Designer, the user starts by developing a flow sheet that represents the overall process. The flow sheet is developed by putting together the required unit procedures (see the next paragraph for an explanation), and joining them with material flow streams. Next, the user initializes the flow sheet by registering the various materials that are used in the process and specifying operating conditions and performance parameters for the various operations.

Most biopharmaceutical processes operate in batch mode. This is in contrast to petrochemical and other high-throughput industries that use continuous processes. In continuous production, a piece of equipment performs the same action all the time. In batch processing, on the other hand, a piece of equipment goes through a cycle of operations. For instance, an inoculum preparation step (P-5 in SBR1) includes the following operations (Fig. 1.3): SIP, SET UP, TRANSFER IN-1(media), TRANSFER IN-2 (inoculum), FERMENT (fermentation operation), TRANSFER OUT (emptying vessel), CIP (cleaning in place). In SuperPro, the set of operations that compose a processing step is called a unit procedure (as opposed to a unit operation). The individual tasks contained in a procedure (e.g. transfer in, Ferment, and CIP) are called operations.

A unit procedure is represented on the screen with a single equipment icon. In essence, a unit procedure is the recipe that describes the sequence of actions required to complete a single processing step. Figure 1.3 displays the dialog through which the recipe of a vessel unit procedure is specified. On the left-hand side of that dialog, the program displays the operations that are available in a vessel procedure; on the right-hand side, it displays the registered operations. The hierarchical representation of batch processes (also known as recipes) using unit procedures and operations is an approach that is recommended by the Instrument Society of America (ISA) because it facilitates modeling, control, and scheduling of batch operations (11).

Figure 1.3 The operations associated with the P-5 unit procedure of Fig. 1.2. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.3

For every operation within a unit procedure, the simulator includes a mathematical model that performs material and energy balance calculations. On the basis of the material balances, it performs equipment-sizing calculations. If multiple operations within a unit procedure dictate different sizes for a certain piece of equipment, the software reconciles the different demands and selects an equipment size that is appropriate for all operations. The equipment is sized so that it is large enough that it will not be overfilled during any operation, but it is no larger than necessary (to minimize capital costs). If the equipment size is specified by the user, the simulator checks to make sure that the vessel is not overfilled. In addition, the tool checks to ensure that the vessel contents will not fall below a user-specified minimum volume (e.g. a minimum stir volume) for applicable operations.

1.4.1 Process Description

1.4.1.1 Upstream

The upstream part is split in two sections: the inoculum preparation section and the bioreaction section. The inoculum is initially prepared in 225-mL T-flasks. The material is first moved to 2.2-L roller bottles, then to 20-L and subsequently to 100-L disposable bag bioreactors. Sterilized media is fed at the appropriate amount in all of these four initial steps (3.6, 11.4, 43.6, 175.4 kg/batch, respectively). The broth is then moved to the first (1000 L) and second (4000 L) seed bioreactor. For the seed bioreactors the media powder is dissolved in water for injection (WFI) in two prep tanks (MP-101 & MP-102) and then sterilized/fed to the reactors through 0.2-µm dead-end filters (DE-101 and DE-102). In the bioreaction section, serum-free low-protein media powder is dissolved in WFI in a stainless-steel tank (MP-103). The solution is sterilized using a 0.2-µm dead-end polishing filter (DE-103). A stirred-tank bioreactor (production bioreactor, PBR1) is used to grow the cells, which produce the therapeutic Mab. The production bioreactor operates under a fed batch mode. High media concentrations are inhibitory to the cells, so half of the media is added at the start of the process and the rest is fed at a variable rate during fermentation. The concentration of media powder in the initial feed solution is 17 g/L. The fermentation time is 12 days. The volume of broth generated per bioreactor batch is approximately 15,000 L, which contains roughly 22.6 kg of product (the product titer is approximately 1.5 g/L).

1.4.1.2 Downstream

Between the downstream unit procedures there are 0.2-µm dead-end filters to ensure sterility. The generated biomass and other suspended compounds are removed using a Disc-Stack centrifuge (DS-101). During this step, roughly 2% of Mab is lost in the solids waste stream resulting in a product yield of 98%. The bulk of the contaminant proteins are removed using a Protein-A affinity chromatography column (C-101). The following operating assumptions were made: (i) resin binding capacity is 15 g of product per liter of resin, (ii) the eluant or elution buffer is a 0.6% w/w solution of acetic acid and its volume is equal to 5 column volumes (CVs), (iii) the product is recovered in 2 CVs of eluant with a recovery yield of 90%, and (iv) the total volume of the solution for column equilibration, wash, and regeneration is 14 CVs. The entire procedure takes approximately 27 h and requires a resin volume of 362 L. The protein solution is then concentrated fivefold and diafiltered 2 times (in P-21/DF-101) using WFI as diluent. This step takes approximately 5 h and requires a membrane of 15 m². The product yield is 97%. The concentrated protein solution is then chemically treated for 1.5 h with Polysorbate 80 to inactivate viruses (in P-22/V-111). An ion exchange chromatography step follows (P-24/C-102). The following operating assumptions were made: (i) the resin's binding capacity is 40 g of product per liter of resin, (ii) a gradient elution step is used with a sodium chloride concentration ranging from 0.0 to 0.1 M and a volume of 5 CVs, (iii) the product is recovered in 2 CVs of eluant buffer with a yield on Mab of 90%, and (iv) the total volume of the solutions for column equilibration, wash, regeneration and rinse is 16 CVs. The step takes approximately 22.3 h and requires a resin volume of 158 L. Ammonium sulfate is then added to the ion exchange (IEX) eluate (in P-25/V-109) to a concentration of 0.75 M to increase the ionic strength in preparation for the hydrophobic interaction chromatography (HIC; P-26/C-103) that follows. The following operating assumptions were made for the HIC step: (i) the resin binding capacity is 40 g of product per liter of resin, (ii) the eluant is a sodium chloride (4% w/w) sodium di-hydrophosphate (0.3% w/w) solution and its volume is equal to 5 CVs, (iii) the product is recovered in 2 CVs of eluant buffer with a recovery yield of 90%, and (iv) the total volume of the solution for column equilibration, wash, and regeneration is 12 CVs. The step takes approximately 22 h and requires a resin volume of 142 L. A viral exclusion step (DE-105) follows. It is a dead-end type of filter with a pore size of 0.02 µm. This step takes approximately 2.3 h and requires a membrane of 1.45 m². Finally, the HIC elution buffer is exchanged for the product bulk storage (PBS) buffer and concentrated 1.5-fold (in DF-102). This step takes approximately 4 h and requires a membrane of 7 m². The approximately 580 L of final protein solution is stored in fifteen 50-L disposable storage bags (DCS-101). Approximately, 14.6 kg of Mab are produced per batch. The overall yield of the downstream operations is approximately 64.5%.

1.4.2 Material Balance

Table 1.3 provides a summary of the overall material balance of the process. Note the large amount of WFI utilized per batch. A major part of WFI is consumed for cleaning and buffer preparation. Approximately, 14.6 kg of Mab are produced per batch.

Table 1.3 Raw Material Requirements

NumberTable

1.4.3 Scheduling and Cycle Time Reduction

Figure 1.4 displays the Gantt chart of the process for four consecutive batches. The schedule represents a plant that has a single production train. The cleaning-in-place (CIP) skids can be seen at the top of the graph. The batch time is approximately 50 days. This is the time required from the start of inoculum preparation to the final product purification of a single batch. A new batch is initiated every 2 weeks (14 days). The production bioreactor (PBR1) is the time (scheduling) bottleneck. On an annual basis the plant processes 20 batches and produces approximately 292 kg of purified Mab. It is clear from the chart that under these conditions the downstream train is underutilized and the cycle time of the process—the time between consecutive batch starts—is relatively long. The cycle time of the process can be reduced and the plant throughput increased by installing multiple bioreactor trains that operate in staggered mode (out of phase) and feed the same purification train. Figure 1.5 represents a case where four bioreactor trains feed the same purification train. The new cycle time is 3.5 days, which is one-fourth of the original. Under these conditions, the plant processes 80 batches per year and produces approximately 1167 kg of Mab per year. Some biopharmaceutical companies have installed more than four bioreactor trains per purification train aiming at cycle times as low as 2 days.

Figure 1.4 One bioreactor train feeding one purification train. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.4

Figure 1.5 Four bioreactor trains feeding one purification train. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.5

1.4.4 Sizing of Batch Utilities

Another characteristic of batch processing is the variable demand of resources such as labor, utilities and raw materials as a function of time. Sizing of WFI systems is a common challenge during the design of new facilities and the retrofit of existing ones. WFI is used for preparing media and buffer solutions, for cleaning equipment, for generating clean steam, and so on. A WFI system consists of a distillation unit that generates the distilled water, a surge tank, and a circulation loop for delivering the material around the plant. The capacity may be limited by any of the following:

The process can not, on average, consume more water than the still can generate.

The process peak demand can not exceed the capacity of the circulation system.

The surge vessel must be large enough to maintain capacity during peak operation.

In some plants, periodic sanitization cycles may interrupt all purified water draws.

Process modeling can provide reasonable estimates for the sizes of the still, the surge tank, and the pumping capacity of the circulation loop. Figure 1.6 displays the demand for WFI of the Mab process over time. The plots show the instantaneous and the 12-h average (heavy-line) demands. The chart also shows the 12-h cumulative amount that corresponds to the y-axis on the right. The peak instantaneous demand indicates the minimum pumping capacity for the system (11,500 kg/h or 50.7 gpm). The peak 12-h average rate provides an estimate for the capacity of still (1800 kg/h or 8 gpm), and the corresponding peak 12-h accumulation is an estimate of the surge tank capacity of 25,000 L. The trade-off between still rate and surge capacity can be examined by changing the averaging time. Selecting a longer period predicts a larger surge tank and a lower still rate.

Figure 1.6 WFI demand as a function of time. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.6

Figure 1.7 displays the inventory profile of WFI in the surge tank for a tank size of 25,000 L and a still rate of 3500 L/h. The generation still is turned on when the level in the tank falls below 30% and it remains on until the tank is full. The operation rate of the still is depicted by the blue step-function lines. (The reader is requested to refer to the online version of this chapter for color indication.)

Figure 1.7 WFI inventory profile in the surge tank. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.7

1.4.5 Economic Evaluation

Cost analysis and project economic evaluation are important for a number of reasons. For a new product, if the company lacks a suitable manufacturing facility with available capacity, it must decide whether to build a new plant or outsource the production. Building a new plant is a major capital expenditure (Table 1.2) and a lengthy process. To make a decision, management must have information on capital investment required and time to complete the facility. When production is outsourced, a cost-of-goods analysis serves as a basis for negotiation with contract manufacturers. A sufficiently detailed computer model can be used as the basis for the discussion and negotiation of the terms. Contract manufacturers usually base their estimates on requirements of facility/equipment utilization and labor per batch, which is information that is provided by a good model. SuperPro performs thorough cost analysis and project economic evaluation calculations. It estimates capital as well as operating cost. The cost of equipment is estimated using built-in cost correlations that are based on data derived from a number of vendors and literature sources. The fixed capital investment is estimated based on equipment cost and using various multipliers, some of which are equipment specific (e.g. installation cost) while others are process specific (e.g. cost of piping and buildings). The approach is described in detail in the literature (12–14). The rest of this section provides a summary of the cost analysis results for this example process.

Table 1.4 provides a list of major equipment items along with their purchase costs (generated by SuperPro Designer). The total equipment cost for a plant of this capacity (four production bioreactors each having a working volume of 15,000 L) is around $24 million. Approximately, a quarter of the equipment cost is associated with the four production bioreactors. The cost of vessels and filters that are seen in Fig. 1.2 but are missing from the table are accounted for under the Cost of Unlisted Equipment item. The economic evaluation also takes into account the vessels required for buffer preparation and holding that are not included in Fig. 1.2. A full model that includes all buffer preparation and holding activities and other advanced process modeling features can be downloaded from www.intelligen.com/literature.

Table 1.4 Major Equipment Specification and Purchase Costs (Year 2007, Prices in USD)

NumberTable

Table 1.5 displays the various items included in the direct fixed capital (DFC) investment. The total DFC for a plant of this capacity is around $240 million or approximately 10 times the total equipment cost. The total capital investment that includes the cost of start-up and validation is around $300 million.

Table 1.5 Fixed Capital Estimate Summary (Year 2007, Prices in USD)

Table 1.6 provides a summary of the operating cost. The total annual operating cost is $111.6 million resulting in a unit production cost of around $95.6/g (1167 kg of purified product are produced annually). The facility-dependent cost is the most important item, accounting for around 40% of the overall operating cost. This is common for high-value biopharmaceuticals. Depreciation of the fixed capital investment and maintenance of the facility are the main contributors to this cost. Raw materials account for around 18% of the overall cost. Serum-free media accounts for 90% of the raw materials cost (Table 1.7). A price of $500/kg was assumed for serum-free media in dry powder form. Labor and consumables come third and fourth, respectively, each accounting for around 17% of the overall manufacturing cost. Consumables include the cost of chromatography resins and membrane filters that need to be replaced on a regular basis. The replacement of the Protein-A resin accounts for 62.3% of the total consumables cost (Table 1.8). A unit cost of $6000/L and a replacement frequency of 60 cycles were assumed for the Protein-A resin. Approximately 63% of the manufacturing cost is associated with the upstream section (inoculum preparation and fermentation) and 37% with the downstream section (product recovery and purification).

Table 1.6 Operating Cost Summary (Year 2007, Prices in USD)

Table 1.7 Raw Materials Cost Breakdown (Year 2007, Prices in USD)

NumberTable

Table 1.8 Consumables Cost Breakdown (Year 2007, Prices in USD)

NumberTable

1.4.6 Sensitivity Analysis

After a model of the entire process has been developed on the computer, tools like SuperPro Designer can be used to ask and readily answer what if questions and to carry out sensitivity analysis with respect to key design variables. In this example, we looked at the impact of the number of bioreactor trains, product titer, and bioreactor volume on unit production cost.

Figure 1.8 displays the impact of bioreactor trains on the unit production cost. The cost analysis calculations of the previous section correspond to the case of four production bioreactors (each having a working volume of 15,000 L) feeding a single purification train and resulting in a unit cost of around $96/g. If just a single bioreactor train feeds the purification train, then, the manufacturing cost increases by 50%. For more than four

production bioreactors per purification train, the unit cost drops a bit and asymptotically approaches a value of around $90/g. Multiple-production bioreactors that feed a single purification train lead to reduced manufacturing cost because the plant throughput is increased (it is proportional to the number of bioreactors) without the need for additional capital investments in the purification train. Four to six bioreactor trains per purification train is probably the optimum number for cell culture processes that have a fermentation time of around 12 days. Such processes operate with cycle times ranging between 3.5 and 2.5 days.

Figure 1.8 Production cost of Mab as a function of the number of bioreaction trains.

1.8

Figure 1.9 displays the impact of product titer and bioreactor volume on the unit production cost. All points correspond to four production bioreactors feeding a single purification train. For low product titers, the bioreactor volume has a considerable effect on the unit production cost. For instance, for a bioreactor product titer of 0.5 g/L, going from 10,000 L to 15,000 L and finally to 20,000 L of production bioreactor volume, the unit cost is reduced from approximately $330/g to $250/g and $210/g, respectively. For high product titers (e.g. around 2.5 g/L), on the other hand, the impact of bioreactor scale is not as important. This can be explained by the fact that at high product titers, a major part of the manufacturing cost is associated with the purification train. It is therefore wise to shift R&D efforts from cell culture to product purification as the product titer in the bioreactor increases. A key assumption for the results of the sensitivity analysis is that the composition and cost of the cell culture media is independent of product titer.

Figure 1.9 Production cost of Mab as a function of product titer and production bioreactor volume.

1.9

1.4.7 Variability and Uncertainty Analysis

Process simulation tools typically used for batch process design, cycle time reduction, and cost estimation employ deterministic (cause and effect) models. They model the average or expected situation commonly referred to as the base case or most likely scenario. However, variability occurs in all bioprocesses despite best efforts to ameliorate their effects. Modeling many cases can help determine the range of performance with respect to key process parameters. However, such an approach does not account for the relative likelihood of the various cases. Monte Carlo simulation is a practical means of quantifying the variability and uncertainty in process parameters (15). In Monte Carlo simulation, uncertain input variables are represented with probability distributions. A simulation calculates numerous scenarios of a model by repeatedly picking values from a user-defined probability distribution for the uncertain variables and using those values for the model to calculate and analyze the outputs in a statistical way to quantify risk. For models developed in SuperPro, Monte Carlo simulation can be performed by combining SuperPro with Crystal Ball from Decissioneering, Inc. (Denver, Colorado). Crystal Ball is an Excel add-in application that facilitates Monte Carlo simulation. It enables the user to designate the uncertain input variables, specify their probability distributions, and select the output (decision) variables whose values are recorded and analyzed during the simulation. For each simulation trial (scenario), Crystal Ball generates random values for the uncertain input variables selected in frequency dictated by their probability distributions using the Monte Carlo method. Crystal Ball also calculates the uncertainty involved in the outputs in terms of their statistical properties, mean, median, mode, variance, standard deviation, and frequency distribution.

For the base case of this example we estimated an annual throughput of 1167 kg of purified Mab and a unit production cost of $95.6/g. Let us assume that our objective is to reliably manufacture at least 1100 kg/year of purified product at a cost of no more than $100/g. If there is variability and uncertainty in some key process and market parameters, how confident are we about the values of our objective?

To illustrate the benefits of such an exercise, we assigned probability distributions to some parameters that affect the annual throughput of the plant and the unit production cost. Since the production bioreactors are the time bottleneck of the process (i.e. have the longest cycle time), delays in the harvesting of production bioreactors will impact the annual number of batches and, consequently, the annual plant throughput and the unit production cost. For the base case scenario, a fermentation time of 12 days was assumed. To investigate the impact of variability in the harvesting of production bioreactors, a Weibull distribution was assumed for the fermentation time as part of this exercise (Fig. 1.10a) with values varying from 11.5 to 14 days and a mean value of approximately 12 days. The variability in the fermentation time accounts for the combined variability of the entire inoculum preparation and fermentation line because any delays in inoculum preparation affect the start and, consequently, the harvesting of production bioreactor batches.

Figure 1.10 Assumed probability distribution for (a) fermentation time in days, (b) serum-free media price in dollars per kilogram, (c) Protein-A resin price in dollars per liter, and (d) Protein-A resin replacement frequency in cycles. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.10

The unit cost of the serum-free media, which accounts for 90% of the raw materials cost, was the second parameter that was investigated. As part of this exercise, its cost was represented with a normal distribution (Fig. 1.10b) whose mean is $500/kg (equal to the base case value) and the standard deviation is 50.

Finally, the Protein-A resin, which dominates the cost of consumables, was also investigated. Figure 1.10c represents the probability distribution for the unit cost of Protein-A. The mean of $6000/L corresponds to the value of the base case. Figure 1.10d represents the probability distribution for the replacement of the Protein-A resin. The mean of 60 cycles corresponds to the base case value. If this type of analysis is done for an existing facility, historical data should be used to derive the probability distributions. Crystal Ball has the capability to fit experimental data.

The decision (output) variables considered in this example are the production cost and the annual throughput of the facility. Figures 1.11 and 1.12 display the results of the Monte Carlo simulation. The analysis reveals that the production cost will be less that $100/g (or $100,000/kg) with a certainty of 91% (dark area of Fig. 1.11). Similarly, the annual throughput of the facility will be higher than 1100 kg of purified product with a certainty of 87% (Fig. 1.12). Such findings constitute a quantification of the risk associated with a process and can assist the management of a company in making decisions on whether to proceed or not with a project idea. Additional information on Monte Carlo simulation and risk assessment can be found in the literature (16, 17).

Figure 1.11 Calculated probability distribution for the production cost (in $/kg). (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.11

Figure 1.12 Calculated probability distribution for the annual throughput of the process (in kg/year). (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.12

1.5 Design and Operation of Multiproduct Facilities

Many biopharmaceutical facilities produce more than one product in parallel. The multiple production lines in such facilities typically share utilities (e.g. WFI and clean steam) and labor resources. They may also share auxiliary equipment (e.g. CIP skids and transfer panels), buffer preparation and holding tanks, and even main processing equipment. However, the sharing of resources across multiple lines renders the design and operation of such facilities more challenging. Computer models developed for such environments must capture the interaction among production lines at the facility level. During the design of such facilities, appropriate computer models assist engineers in sizing the shared utilities and figuring out equipment requirements. During operation, similar models are used for generating feasible production schedules that respect all the major constraints. Owing to the inherent variability of biological processes, scheduling tools employed in the biopharmaceutical industry must be able to handle rescheduling easily. Production scheduling results are communicated through Gantt charts and reports that provide information on tasks that need to be executed during a certain time period. The simple example that follows illustrates some of the challenges associated with the design and operation of multiproduct facilities.

Figure 1.13 displays a simplified flow diagram of a two-line Mab facility. Each line has its own bioreactor suites and purification train. However, the two lines share a centrifugal separator for biomass removal (broth clarification). They also share a buffer preparation and holding

area for the first two chromatography steps. A single CIP skid, CIP-2, is available in that area for cleaning all the buffer preparation and holding tanks.

Figure 1.13 Two-product Mab facility with common buffer preparation and holding area. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.13

The facility and the two recipes were modeled using SchedulePro (Intelligen, Scotch Plains, NJ). Figure 1.14 displays a typical production schedule for this facility with a 3.5-day cycle time for each product line (represented by the different colors). The top line of the chart displays the occupancy of CIP-2, the single CIP skid used for cleaning the tanks of the shared buffer preparation and holding area. Under normal conditions, CIP-2 can clean all the tanks. However, its utilization is quite high and that may lead to conflicts in cases where the schedule deviates from its normal one.

Figure 1.14 Production schedule for the two-line Mab facility. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.14

For instance, what if there was a 24-h delay in the harvesting of the bioreactor of the fourth batch? That would lead to the delay of the start of the downstream operations. When such a delay is introduced to the model, the algorithm warns the user about the conflicts that it will create and offers to reschedule. Rescheduling typically involves delaying of activities that have not started yet. If the user does not authorize rescheduling, the tool simply displays the conflicts created by the delay (Fig. 1.15). SchedulePro creates multiple lines for conflicting equipment, draws a red frame around the conflicting activities, and displays an exclamation mark on the y axis. (The reader is requested to refer to the online version of this chapter for color indication.) The user can manually resolve the conflicts through drag-and-drop of tasks and local rescheduling and evaluate whether a solution with the current available resources is possible.

Figure 1.15 Introducing a delay and viewing conflicts. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

1.15

Elimination of conflicts related to CIP-2 through the installation of a second CIP skid can be readily evaluated with tools like SchedulePro. Other issues that can be easily investigated include potential conflicts related to the limited availability of labor, utilities, inventories of raw materials, storage and treatment capacity for waste materials, and so on.

1.6 Summary and Conclusions

Process simulation and production scheduling tools can play an important role throughout the life cycle of process development and product commercialization. In process development, simulation tools are becoming increasingly useful as a means to analyze and evaluate process alternatives. During the transition from development to manufacturing, they facilitate technology transfer and process fitting. Production scheduling tools play a valuable role in manufacturing. They are used to generate feasible production schedules and enable manufacturing personnel to efficiently handle process delays and equipment failures. Such tools also facilitate capacity analysis and debottlenecking tasks.

The biopharmaceutical industry has only recently begun making significant use of process simulation and scheduling tools. Increasingly, universities are incorporating the use of such tools in their curricula. In the future, we can expect to see increased use of these technologies and tighter integration with other enabling information technologies, such as supply chain tools, manufacturing execution systems (MES), batch process control systems, and process analytics tools (PAT). The result will be more robust processes and efficient manufacturing leading to more affordable biological products.

References

1. Douglas JM. Conceptual design of chemical processes. New York: McGraw-Hill; 1988.

2. Pavlou AK, Belsey MJ. Eur J Pharm Biopharm 2005; 59: 389–396.

3. Plenert G, Kirchmier B. Finite capacity scheduling–management, selection, and implementation. New York: John Wiley & Sons; 2000.

4. Pinedo ML. Planning and scheduling in manufacturing and services. New York: Springer Science; 2005.

5. Korovessi E, Linningerr AA. Batch processes. Boca Raton, FL: Taylor & Francis; 2006.

6. Hwang F. Pharm Eng 1997; January/February 28–43.

7. Petrides DP, Koulouris A, Lagonikos PT. Pharm Eng 2002; 22: 56–64.

8. Heinzle E, Biwer A, Cooney C. Development of sustainable bioprocesses. West Sussex: John Wiley & Sons; 2006.

9. Walsh G. Nat Biotechnol 2006; 24: 769–775.

10. Langer ES, Junker B, editors. Advances in large scale biopharmaceutical manufacturing and scale-up production. Rockville, IN: ASM Press & ISTM; 2004. pp. 152–190.

11. Parshall J, Lamb L. Applying S88–batch control from a user's perspective. Research Triangle Park, NC: ISA; 2000.

12. Harrison RG, Todd P, Rudge SR, Petrides DP. Bioseparations science and engineering. New York: Oxford University Press; 2003.

13. Peters MS, Timmerhaus KD. Plant design and economics for chemical engineers. 4th ed. New York; McGraw-Hill; 1991.

14. Valle-Riestra JF. Project evaluation in the chemical process industries. New York: McGraw-Hill; 1983.

15. Mun J. Applied risk analysis. Hoboken, NJ: John Wiley & Sons; 2004.

16. Achilleos EC, Calandranis JC, Petrides DP. Pharm Eng 2006; 26(4): 34–40.

17. Papavasileiou V, Koulouris A, Siletti C, Petrides D. Pharm TechnolInnovations 2006; 22(1): s28–s38.

Part II

Downstream Recovery of Cells and Protein Capture

Chapter 2

Cell Separation, Centrifugation

Hans Axelsson

AlfaLaval AB, Tumba, Sweden

2.1 Introduction

The first application of centrifugal separation outside milk processing was harvesting of cells in baker's yeast production; centrifugation is still the only choice for cell separation and protein recovery in many fermentation processes. In many such processes, the most optimal phase separation is obtained by combining centrifugation with some type of fine filtration.

2.2 Centrifugal Separation

Centrifugal separation is a sedimentation operation accelerated by centrifugal force. Thus, a prerequisite for the separation is a difference in density between the phases. This applies to both solid–liquid and liquid–liquid separation.

The settling velocity Vg of a solids particle (or droplet) under the influence of gravity alone is given by Stokes' law (refer to section on nomenclature for abbreviations and symbols):

2.1 2.1

In the centrifuge field, the settling velocity becomes:

2.2 2.2

where

2.3 2.3

is called relative centrifugal force (RCF) or G-number.

It is however, not the RCF alone that determines the performance of a given centrifugal separator. The residence time of the particle or droplet in the centrifugal force field is also important, so that a larger volume of the rotor will increase the possible flow rate. In the most common centrifugal separators, the disk bowl machines, a given bowl material can cope with a certain peripheral speed; the stronger the bowl material, the higher the peripheral speed. This leads to the fact that the maximal RCF is inversely proportional to the bowl diameter, but not necessarily that the lower RCF machine has a lower performance.

2.3 Types of Centrifugal Separators

2.3.1 General

What is determining the design of a centrifuge is the method by which the solids phase in the process material is handled. The first two types developed were applied to duties where the resulting products were in liquid form, cream and skimmed milk, and yeast cream and cell-free wort. In the beginning of this century, centrifuges were applied to processes with solids, amorphous or crystalline. In the biotechnology industry, the solids phase is often recovered in a diluted suspension or is, even at high solids concentrations, aslurry with good flow properties.

Types of machines are described in the work by Rushton et al. (1). The most comprehensive description of centrifuges can be found in the book by Sokolov (2). Brochures and technical publications from sedimenting centrifuge manufacturers such as Alfa Laval AB, also provide information on machine types and sizes.

In this chapter, centrifugal filtration will not be dealt with because of the limited use in this industry, with the exception of crystal recovery in antibiotic production. For a description of various centrifugal filters, see Rushton et al. (1).

2.3.2 Disk Bowl Machines

2.3.2.1 Solid Bowl

A solid bowl machine is shown in Fig. 2.1a. The feed enters from above through a stationary pipe that leads into a volume formed by the distributor, the component that carries a multitude of thin conical disks, less than 1-mm thick, separated by spacers, usually 0.3–1.5 mm thick. On the inside of the distributor are 4–16 radial wings; however, a large bowl needs more wings. The purpose of the wings is to start the rotation of the incoming liquid. This design principle is the same for all disk bowls with feed from above. Like all other disk bowl machines, it may have one or two liquid outlets. The outlet(s) may be open or equipped with paring disks. The paring disk is a stationary device, and has the shape of a pump wheel. Liquid fills the chamber around the paring disk. Part of the inner layer of the rotating liquid in the chamber is pared off, enters the pump wheel and is forced out by the static pressure created when the kinetic energy of the rotating liquid is converted to static pressure. The version in Fig. 2.1a has two outlets, each equipped with a paring disk. The solids accumulate at the periphery and must be removed manually. In the separation of two nonmiscible liquids, a coalescence of the dispersed droplets takes place at an internal interface in the bowl. The positioning of this interface, its diameter Di is very important. The nearer it is placed to the light-liquid outlet, the more contaminated the light liquid will be with droplets of the heavy phase. The position is determined by a gravity disk (Fig. 2.1a), over which the heavy liquid flows. With every machine a set of gravity disks with different inner diameters is supplied. The gravity disk's inner diameter is given by the relation:

Figure 2.1 Vertical cuts of various disc centrifuge bowls. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

2.1

2.4 2.4

where Di and Dl are the diameter of the interface and the outflow diameter of the light phase, respectively.

If properly designed, and if the feed flow rate is high enough to prevent sedimentation in the bowl, this type of machine can be used to recover a dilute cell concentrate in the heavy-liquid outlet. This is also valid for three-phase solids-ejecting machines. Solid bowl machines are made in a large number of sizes with bowl diameters ranging from 140 to 750 mm.

2.3.2.2 Solids-Ejecting Separators, Radial Discharge

Of all types of disk bowl machines, the solids-ejecting separators with radial discharge (Fig. 2.1b, c) are the most common. The feed zone in the machine in Fig. 2.1c resembles that of the machine in Fig. 2.1a. With an automated, periodic partial discharge of the sediment, it is possible to obtain a considerably higher concentration of the solids than in the peripheral nozzle machine (Fig. 2.1f). In its upper position, the sliding bowl bottom (cf. Fig. 2.1b) keeps the bowl closed. At discharge it is pressed downwards during about 0.1 sec, making it possible for the solids to escape through large ports in the outer bowl wall. This is illustrated in a video clip (the reader is requested to refer to the online version for Videoclip1.mov). To make the discharge at the right moment, the discharge can be initiated by a self-triggering mechanism that hydraulically senses the solids level in the solids space (3). Three-phase separation units are available. The solids-ejecting machines can be equipped with hermetic seals and a hermetic feed from below in a hollow spindle (Fig. 2.1c). In these true hermetic machines the split between the exiting light and heavy liquids can be made by controlling the back pressure in the outlet pipes. The heavy liquid can contain large amounts of solids (cells or proteins) if the machine is properly designed. The solids-ejecting separators are today available in about 10 sizes with bowl diameters between 180 and 1000 mm. Several small- and medium-sized units are available in versions equipped for BL2-LS requirements. A pilot-scale sterilizable separator module is seen in Fig. 2.2. A large-scale unit is seen in Fig. 2.3. A machine with radial intermittent discharge has been developed primarily for recovery of protein precipitates (4). When the solids space is full of solids, the liquid in the bowl is decanted and a discharge takes place by a pneumatic mechanism.

Figure 2.2 Solids-ejecting disc bowl centrifuge of pilot size for sterilizable contained installation. Type Culturefuge™ 100. The height of the separrator only is 1.3 m. Courtesy Alfa Laval AB. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

2.2

Figure 2.3 Radially solids-ejecting disk bowl centrifuge in production size. Capacity in breweries up to 90 m³/h. Type BREW 3000. The height of the separator is 2.2 m. Courtesy Alfa Laval AB. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

2.3

2.3.2.3 Solids-Ejecting Separators, Axial Discharge

When comparing axially and radially discharging separators (Fig. 2.1b, d) one finds that the former has a different geometry. The lock ring that holds together the upper and lower parts of the rotor (Fig. 2.1d) can be placed at a smaller diameter, because no sliding bowl bottom needs to be fitted. This reduces the stresses in the thread of the lock ring. Also, the solids discharge takes place through a few small axial channels, which do not induce large stresses in the bowl shell. These two factors make it possible to increase the bowl speed considerably above that of the radially discharging machines, so that the RCF is at least doubled. The discharge system is operated by pressurized air, which forces a ring slide, equipped with valve seats, downward, opening the axial channels. To facilitate the transport of solids toward the channels, the

bowl is machined to have a star-shaped inner contour. This type of machine exists at present in three sizes with bowl diameters ranging from 500 to 900 mm. BL-2LS versions are available. A medium-sized unit is seen in Fig. 2.4.

Figure 2.4 Axially solids-ejecting disk bowl centrifuge in production size for bacteria and mammal cell recovery. Capacity up to 12 m³/h. Type BTAX 215S. The height of the separator including piping is 1.8 m.

Courtesy Alfa Laval AB. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

2.4

2.3.2.4 Nozzle Machines with Pressurized Discharge of Concentrate

The apparent viscosity of bacteria and yeast suspensions as well as some protein precipitates decreases with increasing velocity gradient. Therefore, they are able to flow even at high concentrations in channels or tubes from the periphery of the bowl in nozzle machines with pressurized discharge of concentrate (Fig. 2.1e). The tubes empty into a central chamber, where a centripetal pump, a paring tube, picks up the concentrate and pumps it out of the bowl. The paring tube is a stationary radial tube acting in the same way as a paring disk (Fig. 2.1a). The nozzles are placed at the end of the concentrate tubes, just in front of the chamber. This machine type is suitable only for applications where the solids have the specific flow properties mentioned above, and in the absence of other type of solids. In some machines of this type, a vortex chamber is placed in front of the nozzle (5). Liquid enters the chamber tangentially, creating a viscosity-dependent whirl, and leaves through the nozzle in the center of the chamber. The viscosity effect gives a self-regulating function so that varying solids flow rates (within limits) to the machine will result in a constant concentration of the concentrate,

thereby reducing the risk of clogging the machine. The machine type was originally developed for bakers’ yeast production. The yeast, being grown on molasses as the carbon source, needs to be washed clean from substrate residues by water. A three-stage countercurrent yeast harvest and washing system is shown in Fig. 2.5. One of the units of this type (Fig. 2.6) is available in a sterilizable, contained and aseptic version, conforming to BL2-LS. In order to make this possible, the cleaning-in-place (CIP) has to be efficient. Therefore, the sterilizable machine uses the discharge system of the machine in Fig. 2.1d during CIP. This system is also used in noncontained versions. Bowl diameters between 500 and 900 mm are manufactured.

Figure 2.5 (a) Flow sheet for a countercurrent washing system with two washing stages for yeast. Centrifuges with pressurized yeast outlet. (b) Nozzle machines with pressurized solids discharge for yeast separation, installed for countercurrent washing. Capacity on bakers’ yeast up to 100 m³/h. Type FEUX 214. The height of the separator including piping is 2.3 m. [Courtesy of Alfa Laval AB.] (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

2.5

Figure 2.6 Nozzle machine with pressurized solids discharge for yeast and bacteria separation. Capacity on E. coli about 3000 l/h. Type BTUX 510. The height of the separator is 1.8 m. Sterilizable and contained installation. Courtesy Alfa Laval AB. (This figure is available in full color at http://onlinelibrary.wiley.com/book/10.1002/9780470054581.)

2.6

2.3.2.5 Nozzle Machines with Peripheral Nozzles

In machines with peripheral nozzles (Fig. 2.1f), nozzles with a diameter of 0.5–3 mm are situated at the periphery of the bowl, which has sloping walls toward the nozzles. The number of nozzles varies between 4 in small bowls and 20 in the largest bowls. It can be equipped with a recirculation device for the sediment built as follows: Inside the feed tube is a separate central tube for recirculated liquid. The central tube enters into a chamber at the bottom of the bowl, from which tubes lead the recirculated liquid along the bowl bottom. Each tube ends just in front of the nozzle. This makes it possible to increase the concentration of the solids phase without decreasing the nozzle size which leads to increased risk of nozzle clogging. The nozzle flow rate is proportional to number of nozzles, distance from center of rotation, bowl speed, and square of the nozzle diameter. The peripheral nozzle machines are used for the largest flow rates; the bowl diameters are up to 1050 mm.

2.3.3 Decanter Centrifuges

In the 1940s, the solid bowl separator with scroll discharge of solids, the decanter centrifuge (Fig. 2.7), was developed into the design it has today. It is intended for use with process liquids containing a high percentage of suspended solids. It is equipped with a screw conveyor that rotates at a speed slightly higher or lower than the bowl speed (the reader is requested to refer to the online version for Videoclip2.mov).

Figure 2.7 Decanter centrifuge. Vertical cut.

2.7

The differential speed between the conveyor and bowl can be varied. Careful control of this parameter, torque control, and some newer design features (e.g. the baffle disk), have recently made the decanter interesting to use for biological sludges as well. However, the conventional mechanical design of these machines does not allow more than 5000 G for a medium-diameter machine (6).

For three-phase decanters the