Determining key process parameters of continuous virus inactivation reactors and systems and methods for designing and manufacturing the same

Abstract

A virus inactivation device includes at least one experimental continuous virus inactivation reactor, the reactor having at least an inlet, an outlet and a tubular flow path, and a computer system that can design, select, make and / or manufacture a scaled actual reactor based on the experimental continuous virus inactivation reactor. The tubular flow path includes a set of alternating turns that form a serpentine or interweaving pattern between the inlet and the outlet.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of U.S. Provisional Application No. 62 / 742,506, filed on October 8, 2018, the contents of which are expressly incorporated herein by reference. Technical Field

[0003] The present disclosure generally relates to a system and method for determining key process parameters of a continuous viral inactivation reactor and designing and manufacturing the same. Background Art

[0004] Currently, defining the residence time of virus particles in a plug flow reactor (PFR) is difficult to quantify due to the fluid dynamics that occur during the flow of the process stream in the circulating pipe, where the flow rate of the process stream in the center of the pipe can be twice the average flow rate of the process stream and it becomes almost stagnant near the pipe wall. Therefore, the only way to determine the correct PFR parameters for viral inactivation is currently through experimentation. This trial-and-error approach is inefficient and time-consuming. Summary of the invention

[0005] In one aspect, a method for designing, selecting, making and / or manufacturing an actual piston flow reactor is described. The method includes introducing a process stream including a detectable particle / tracer into an experimental reactor having a known radius of curvature and a known inner diameter, wherein the experimental reactor is connected to at least one of a first detector and a second detector; detecting a flow rate of the process stream in the experimental reactor by at least one of the first detector and the second detector; detecting a fluid phase parameter of the process stream by at least one of the first detector and the second detector; detecting the detectable particle / tracer leaving the experimental reactor by the second detector; determining an empirical value associated with at least one of the experimental reactor parameter and the fluid phase parameter based on the introduced process stream including the detectable particle / tracer; determining a non-empirical value associated with at least one of the experimental reactor parameter and the fluid phase parameter; and designing, selecting, making and / or manufacturing an actual reactor based on the determined empirical value and the determined non-empirical value.

[0006] In one aspect, a system for determining, selecting, making and / or manufacturing actual reactor dimensions is described. The system includes a processor; and a non-transitory machine-readable storage medium storing machine-readable instructions executable by the processor to: receive parameters of an experimental reactor in communication with at least one of a first detector and a second detector; detect a flow rate of a process stream including a detectable particle / tracer in the experimental reactor by at least one of the first detector and the second detector; detect a fluid phase parameter of the process stream in the experimental reactor by at least one of the first detector and the second detector; detect the detectable particle / tracer leaving the experimental reactor by the second detector; determine an empirical value of at least one of a reactor parameter of the experimental reactor and a fluid parameter of the process stream; determine a non-empirical value of at least one of a reactor parameter of the experimental reactor and a fluid parameter of the process stream; design, select, make and / or manufacture an actual reactor for an actual process stream having a predetermined volume of fluid, wherein the fluid includes parameters substantially similar to the fluid parameters in the experimental reactor.

[0007] In another aspect, a system for designing, selecting, making and / or manufacturing an actual reactor for continuous inactivation of viruses during the manufacture of a biological product is described. The system includes an experimental reactor having a known radius of curvature and a known diameter and designed to receive a process stream; at least one of a first detector and a second detector in communication with the experimental reactor, wherein at least one of the first detector and the second detector detects a fluid phase parameter of the process stream, and wherein the second detector detects detectable particles / tracers exiting the experimental reactor; a processor; and a non-transitory machine-readable storage medium storing machine-readable instructions executable by the processor to: determine an empirical value of a parameter corresponding to at least one of the fluids of the experimental reactor and the process stream; determine a non-empirical value of a parameter corresponding to at least one of the fluids of the experimental reactor and the process stream; and design, select, make and / or manufacture an actual reactor for an actual process stream having a predetermined volume of fluid based on the determined empirical value and the determined non-empirical value.

[0008] In another aspect, a system for designing, selecting, fabricating and / or manufacturing a practical reactor for continuous inactivation of viruses during bioproduct manufacturing is provided. The system includes: an experimental reactor having an inlet, an outlet, and a tubular flow path, the tubular flow path including a set of alternating turns forming a serpentine pattern between the inlet and the outlet, wherein the serpentine pattern includes a predetermined radius of curvature and a predetermined diameter, and wherein the experimental reactor is designed to receive a process material flow; at least one of a first detector and a second detector connected to the experimental reactor, wherein at least one of the first detector and the second detector detects a fluid phase parameter of the process material flow, and wherein the second detector detects detectable particles / tracers leaving the experimental reactor; a processor; and a non-transitory machine-readable storage medium storing machine-readable instructions executable by the processor to: determine empirical values ​​of at least one of the experimental reactor parameters and the fluid parameters of the process material flow; determine non-empirical values ​​of at least one of the reactor parameters of the experimental reactor and the fluid parameters of the process material flow; and design, select, make and / or manufacture an actual reactor based on the determined empirical values ​​and the determined non-empirical values, wherein the actual reactor includes a serpentine pattern substantially similar to the serpentine pattern of the experimental reactor, but is configured to accommodate an actual process material flow having a predetermined volume of fluid.

[0009] In yet another aspect, a system for designing, selecting, making and / or manufacturing an actual reactor for continuous inactivation of viruses during the manufacture of a biological product is provided. The system comprises an experimental reactor having an inlet, an outlet and at least one interwoven tubular flow path, the tubular flow path comprising a plurality of turns located on different, non-parallel planes; at least one of a first detector and a second detector in communication with the experimental reactor, wherein at least one of the first detector and the second detector detects a fluid phase parameter of a process stream, and wherein the second detector detects detectable particles / tracers exiting the experimental reactor; a processor; and a non-transitory machine-readable storage medium storing machine-readable instructions executable by the processor to: determine empirical values ​​of at least one of an experimental reactor parameter and a fluid parameter of a process stream; determine non-empirical values ​​of at least one of a reactor parameter of the experimental reactor and a fluid parameter of the process stream; and design, select, make and / or manufacture an actual reactor based on the determined empirical values ​​and the determined non-empirical values, wherein the actual reactor comprises an interwoven tubular flow path substantially similar to the interwoven tubular flow path of the experimental reactor, but designed and configured to accommodate an actual process stream having a predetermined volume of fluid.

[0010] Other features and advantages of various embodiments will be explained in part in the following description, and will become apparent in part from the description, or can be known by the practice of various embodiments. Through the elements and combinations specifically pointed out in the description here, the purposes and other advantages of different embodiments will be recognized and obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Several aspects and embodiments of the present disclosure may be more fully understood from the detailed description and accompanying drawings, in which:

[0012] Figure 1A is a top view of a tubular flow path of a continuous tubular reactor according to an example of the present disclosure;

[0013] Figure 1B is a partial perspective view of a continuous tubular reactor according to an example of the present disclosure;

[0014] Figure 1C are side and isometric views of a continuous tubular reactor according to an example of the present disclosure;

[0015] Figure 1D illustrates a plurality of continuous tubular reactors connected to each other according to examples of the present disclosure;

[0016] Figure 1E is an isometric view of an exemplary continuous flow reactor tube having a single tube according to examples of the present disclosure;

[0017] Figure 1F The example according to the present disclosure is shown Figure 1E A continuous flow reactor tube having turns about a single longitudinal axis but located in different non-parallel planes;

[0018] Figure 1G is along the lines of the examples according to the present disclosure Figure 1E A cross-sectional view of a longitudinal passage of a tube;

[0019] Figure 1H According to the example of the present disclosure Figure 1E A side view of an exemplary continuous flow tube;

[0020] Fig. 1I According to the example of the present disclosure Figure 1H Detailed view of area A of a side view of an exemplary continuous flow tube.

[0021] Figure 1J is an isometric view of an exemplary continuous flow tube having four tubes according to an example of the present disclosure;

[0022] Figure 1K An overview of a system according to an example of the present disclosure is illustrated;

[0023] Figure 2A An exemplary system according to an example of the present disclosure is illustrated;

[0024] Figure 2B is a graph showing experimental raw data detected by a second detector according to an example of the present disclosure;

[0025] Figure 2C is a graph showing the inlet and outlet of volume and time of a detectable particle or detectable tracer according to an example of the present disclosure;

[0026] Figure 2D is a graph showing a fitted curve and raw data according to an example of the present disclosure;

[0027] Figure 2E The relationship between the Dean Number and HETP in experimental data and fitting data according to an example of the present disclosure is illustrated;

[0028] Figure 3 Another exemplary system according to examples of the present disclosure is illustrated;

[0029] Figure 4 illustrates an overview of the relationship between a comparator and a non-empirical variable according to an example of the present disclosure;

[0030] Figure 5A illustrates an overview of how a system according to an example of the present disclosure is derived with reactor tube path length and flow rate;

[0031] Figure 5B illustrates details of how a system according to an example of the present disclosure is derived with reactor tube path length and flow rate;

[0032] Figure 5C is a graph illustrating the relationship between the reactor volume and the flow rate of an actual reactor according to an example of the present disclosure;

[0033] Fig. 6A The system according to the example of the present disclosure is illustrated as follows: min and T max Overview of derivation;

[0034] Figure 6B The system according to the example of the present disclosure is illustrated as follows: min and T max Details of the derivation;

[0035] Figure 6C is a diagram illustrating the T of an actual reactor according to an example of the present disclosure. min and T max Graph of

[0036] Fig. 7A illustrates an overview of how a system according to an example of the present disclosure may be derived with an actual reactor path length;

[0037] Figure 7B illustrates details of how a system according to an example of the present disclosure is derived with an actual reactor path length;

[0038] Figure 7C is a graph illustrating the relationship of flow rate and reactor volume for designing, selecting, fabricating and / or manufacturing a practical reactor according to examples of the present disclosure;

[0039] Fig. 8A illustrates an overview of how a system according to an example of the present disclosure is derived with actual reactor path lengths and inner diameters of reactor tubes;

[0040] Figure 8B illustrates details of how a system according to an example of the present disclosure is derived with actual reactor path lengths and inner diameters of reactor tubes;

[0041] Figure 8C is a graph of HETP versus linear flow rate according to an example of the present disclosure;

[0042] Fig.8D is a graph of predicted HETP versus actual HETP according to an example of the present disclosure;

[0043] Fig. 8E is a diagram showing a method for satisfying a predetermined flow rate and T according to an example of the present disclosure. min A graph of reactor volume and inner diameter;

[0044] Fig. 9 Another exemplary system according to examples of the present disclosure is illustrated;

[0045] Fig. 10A illustrates details of how a system according to an example of the present disclosure is derived with actual reactor path lengths and inner diameters of reactor tubes;

[0046] Fig. 10B is a graph illustrating a relationship between HETP and Dean number by using a small JIB, a medium JIB, and a large JIB according to an example of the present disclosure;

[0047] Fig. 10C is the value according to an example of the present disclosure when HETP is normalized by the inner diameter Fig. 10B Graph of

[0048] Fig. 10D is a diagram illustrating an example of the present disclosure. min is 60 and T maxA graph showing the relationship between path length and inner diameter at a flow rate of 500 mL / min of 75;

[0049] Fig.11A is a graph illustrating volume diffusion response of an exit pulse injection of variable flow rate and path length from a JIB operation according to an example of the present disclosure;

[0050] Fig. 11B is a graph illustrating normalization of path length by HETP according to an example of the present disclosure;

[0051] Fig.12 is a graph illustrating the correlation of the Dean number required to generate a transition in Taylor-Couette flow with two inflection points and asymptotes produced in the expansion of the RTD distribution of the JIB according to an example of the present disclosure;

[0052] Fig.13A is a graph illustrating the response of an exit pulse injection of variable flow rate and viscosity from a JIB operation according to an example of the present disclosure;

[0053] Fig. 13B is a graph illustrating normalization of various viscosity experimental HETPs by Dean number according to an example of the present disclosure;

[0054] Fig. 13C is a graph illustrating a third order polynomial fit to a lumped data set over all tested flow rates and viscosities according to an example of the present disclosure;

[0055] Fig.13D and 13E are the standard deviation and T according to the examples of the present disclosure, respectively. min Resulting parity plot of the lumped data set of (5σ);

[0056] Fig.14A is a graph of the response of exit pulse injection from JIB operation for variable flow and viscosity, except for the data point where De<100, according to an example of the present disclosure.

[0057] Fig. 14B is a graph illustrating a third order polynomial fit to the lumped data set over all flow rates and viscosities tested, except for data points where De<100, according to an example of the present disclosure;

[0058] Fig. 14C and 14D are the standard deviation and T respectively according to the example of the present disclosure. min Resulting parity plot of the lumped data set of (5σ), where data points with De < 100 are excluded;

[0059] Fig.15A and15B illustrates a selection of standard deviations selected from Table 2A and Table 2B on an RTD distribution according to an example of the present disclosure;

[0060] Fig.16 The example of the present disclosure satisfies the arbitrary but strict requirement T min (5σ)>60min and T max Two RTD distributions for different combinations of path length and flow rate for (3σ) < 79 min;

[0061] Fig.17A The results of applying pulse injection to a simulated protein elution peak occurring before the addition of glucose, the 2-peak mid-height and the maximum peak are illustrated according to an example of the present disclosure.

[0062] Fig. 17B is the predicted T based on isocratic data according to an example of the present disclosure min (5σ) Comparison graph with results from dynamic peak results;

[0063] Fig.18 is a comparison graph of a peak value generated by pulse injection according to an example of the present disclosure and a peak value after correction;

[0064] Fig.19 is a comparative graph of time (5σ) calculated from an original trajectory and a modified trajectory according to an example of the present disclosure;

[0065] Fig. 20 This is a comparative curve diagram of the low-viscosity phage pulse injection experiment, where (+) virus represents the first time point when the virus is detected positive, and (-) virus represents the last time point when the virus is detected negative;

[0066] Fig.21 This is a comparative curve diagram of the high-viscosity phage pulse injection experiment, where (+) virus represents the first time point when the virus is detected positive, and (-) virus represents the last time point when the virus is detected negative;

[0067] Fig. 22 is a graph illustrating the relationship between the reactor volume and the flow rate of an actual reactor according to an example of the present disclosure; and

[0068] Fig.23 is a graph illustrating the solution of enlarging the inner diameter.

[0069] Like reference numerals refer to like elements throughout the specification and drawings. DETAILED DESCRIPTION

[0070] It is to be understood that both the foregoing summary and the following detailed description are exemplary and explanatory only and are intended to provide explanations of various embodiments of the present teachings.

[0071] In the description below, the phrase "experimental reactor" refers to a reactor used for non-commercial purposes, such as low-capacity operations, simulated operations, initial data collection, etc. In addition, the phrase "actual reactor" refers to a reactor used for any other purpose not used by the experimental reactor. These purposes include, for example, commercial purposes. In addition, the phrase "hypothetical reactor" refers to an experimental reactor, where data was previously collected and there is no reason to conduct the experiment again. The phrases "first detector" and "second detector" refer to one or more detectors capable of detecting different characteristics of a process stream and its contents.

[0072] Overview

[0073] In one example, a method for manufacturing a biological product may include using a PFR to continuously perform virus inactivation by continuously adding a virus inactivator to a product containing a feed stream (process material stream) and homogenizing it. From then on, the process material stream can be pumped through and / or introduced into the PFR, and the virus inactivation state is maintained for a predetermined time. It is difficult to quantify the residence time of virus particles in the PFR because the flow rate of the process material stream in the center of the pipeline can be twice the average flow rate of the process material stream and is almost stagnant near the pipeline wall. In order to determine the optimal residence time of virus particles, a newly designed or manufactured PFR may include a set of alternating turns forming a serpentine pattern between the inlet and outlet, thereby forming a serpentine flow path, an interlaced path, and / or a boxed fixture (JIB) design that generates Dean vortexes to promote radial mixing can be utilized, as described in the co-owned and co-applied U.S. patent applications entitled "New Continuous Flow Reactor for Low pH Virus Inactivation" and "Continuous Flow Reactor for Virus Inactivation", the specifications of which are incorporated herein by reference. In this newly designed or manufactured PFR, the ability to predict when the first virus particle leaves the reactor is crucial. In general, three alternative approaches can be used to determine or estimate the minimum residence time experienced by discrete detectable particles or detectable tracers, which is equivalent to the incubation time (T) for bulk viral inactivation. min). One approach assumes ideal uniformity in the flow path (i.e., plug flow) and simply divides the reactor volume by the flow rate. However, this idealized approach can lead to an underestimation of the required residence time, even as reactors with serpentine flow paths or interwoven flow paths continue to improve in efficiency. Another approach is to assume that the center of the flow path remains at twice the average velocity of the process stream. However, due to the increased efficiency of serpentine or interwoven reactors, this approach can lead to an overestimation of the required residence time for the process stream. Another approach is to use process development to determine efficiency factors by which the ideal approach can be modified to account for non-ideal conditions. However, this approach requires testing of large-scale JIBs and does not account for potential anomalies in viscosity and flow rate. Therefore, the systems and methods of the present invention use a creative technical solution to accurately infer the performance of JIBs with respect to different path lengths, flow rates, reactor designs, inner diameters, and viscosities.

[0074] Able to predict T min Allows estimation and / or configuration of required flow rates and reactor sizes to suit process needs.

[0075] This is particularly useful when submitting data to regulatory agencies, such as the Food and Drug Administration (FDA), where data validation is required for compliance with applicable regulations. The system described below will allow the user to run a small-scale process stream for data validation and then scale the reactor for large-scale production without affecting the process stream or the final outcome of the process stream during scale-up. Given that viral inactivation conditions can also degrade the target product as a function of residence time, the maximum residence time (T) spent in the reactor is max ) (i.e., the maximum residence time experienced by the last large amount of target product leaving the reactor) should also be determined. The present invention is the first known method to address the stability of the target product. The first application of the system and method allows the use of T min and T max The second application uses the opposite method, that is, using the working flow rate and path length to predict T min and T max The third application is to approximate a certain T when scaling the reactor size. minA method for determining the required internal diameter and path length. In these applications, the effects of viscosity, Dean number, and reactor volume on the residence time distribution all need to be determined and quantified. In these applications, the user can select whether the user wants to use reactors with essentially the same diameter or reactors with different diameters (i.e., scaled reactors). Alternatively or in addition, the system can recommend whether a reaction tube reactor of the same size as the experimental reactor should be used in the actual process, or whether the experimental reactor should be scaled up or down. In this optional or additional example, when scaling up or down the reactor, the system or user can select the experimental reactor and the actual reactor to have the same aspect ratio, such as Figure 2A As shown, or the experimental reactor and the actual reactor are selected to have different aspect ratios, such as Figure 3 shown.

[0076] A continuous flow reactor having a tubular flow path having a set of alternating turns forming a serpentine pattern

[0077] As described above, the system can design, select, fabricate and / or manufacture an actual reactor having a reactor tube having a serpentine pattern. Figures 1A to 1D The details of the serpentine pattern are described in reference Figure 1A , each curve 10 in the tubular flow path 12 may include a vertical (L1) center-to-center distance of about 1.5 cm (e.g., about 1.479 cm) between turns. In addition, each curve 10 may include a horizontal (L2) center-to-center distance of about 1.375 cm between turns. In addition, the radius of each curve 10 in the tubular flow path 12 may be substantially constant. In one example, when the ROC is about 0.85 cm to about 0.99 cm, the angle of curvature in each curve 10 may be about 270°. In another example, the ROC is greater than or equal to about 0.99 cm, then the angle of curvature of each curve may be the same or the angle of curvature of the first curve 10 may be about 270°, and the angle of curvature of the second curve 10 adjacent to the first curve may be equal to or greater than 270°. In each of the exemplary scenarios described above, R1 and R2 may differ from each other within 0.05 cm to prevent significant differences in the Dean number between alternating turns.

[0078] In one example, each curve 10 in the tubular flow path 12 may include the same radius, such as a radius of 1 cm. In another example, each curve 10 in the tubular flow path 12 may include a different radius. For example, the first curve 10 may include a radius R1, which may be 1 cm, and the second curve 10 may include a radius R2, which may be 1.02 cm. In this example, the curvature angle corresponding to the radius R1 may be approximately 270°, and the curvature angle corresponding to the radius R2 may be approximately 278.27°. In another example, the first half of the curve 10 in each tubular flow path 12 may include a first radius R1, which may be 1 cm, and the second half of the curve 10 in each tubular flow path 12 may include a second radius R2, which may be 1.02 cm.

[0079] refer to Figure 1B and 1C In order to accommodate about 325 alternating 270° turns in a compact design, the tubular flow path 12 in the shell-and-tube CVI reactor 10 can be vertically divided into a plurality of stacked layers 14, such as from 2 layers 14 to 26 layers 14 or more, such as 26 layers 14, as shown in FIG. Figure 1C As shown in . Each layer in the plurality of stacked layers 14 may include a thickness of about 0.5 cm or less to about 2 cm or greater, such as about 0.7 cm to about 1.2 cm. In one example, each layer 14 (a)-14 (z) in the stacked layers 14 may include about 10.5 turns or less to about 15.5 turns or more in a single plane. For example, each layer 14 (a)-14 (z) in the stacked layers 14 may include 12.5 turns. In one example, each layer 14 may be connected to its adjacent lower layer 14 by a 180° vertical turn 16. Alternatively, the second half of the last turn 10 of the flow path 12 in each layer 14 may be vertically rotated 180° to connect the tubular flow path 12 in the first layer (e.g., layer 14 (a) is connected to the tubular flow path 12 in the second layer 14 (b)). In one example, the shell-and-tube CVI reactor 10 includes 26 layers 14, and the 26 layers 14 may be connected to each other by 25 180° vertical turns 16.

[0080] refer to Figure 1BIn one example, each layer 14 in the shell-and-tube CVI reactor 100 may include a depth L3. The depth L3 may be the distance from the center of the tubular flow path 12 in the first layer 14 to the center of the tubular flow path 12 in the second layer 14 directly below the first layer 14. The depth L3 may be about 0.7 cm or less to about 1.2 cm or more, for example, about 0.8 cm to about 0.9 cm, such as a depth of about 0.835 cm. In one example, the distance from the bottom of the tubular flow path 12 in the first layer to the top of the tubular flow path 12 in the second layer directly below the first layer may be about 0.15 cm (1.5 mm) to about 0.4 cm (4 mm), for example, it may be about 0.17 cm (1.7 mm) to about 0.255 cm (2.55 mm), such as about 0.2 cm (2 mm).

[0081] In one example, if Figure 1D As shown, in order to allow for changes in path length and incubation time, in addition to the shell-and-tube CVI reactor 10 having multiple layers 14, multiple shell-and-tube CVI reactors 10 can be connected in series with each other. This can be achieved by one or more flange connectors 18. In one example, at least 2 shell-and-tube CVI reactors 10 can be connected to each other, for example, there are at least 6 shell-and-tube CVI reactors 10 or more. In this particular example, the tubular flow path 12 at each end of the shell-and-tube CVI reactor 10 can partially extend out of the shell-and-tube CVI reactor 100 (extension portion 15). The extension portion 15 can also include a flange 20, such as Figure 1B As shown. The connector 18 may include a horizontal 180° rotation and / or be in a "U" shape. One end of the connector 18 may be connected to the tubular flow path 12 or the flange 20 of the first tubular CVI reactor 10, and the second end of the connector 18 may be connected to the tubular flow path 12 or the flange 20 of the adjacent tubular CVI reactor 10.

[0082] The connector 18 can be connected by a clamp 22 (such as Figure 1D ) or other fastener means (such as screws, adhesives, etc.) connected to each tubular flow path 12 or flange 20. In one example, a gasket can be placed between the end of the tubular flow path 12 or flange 20 and each end of the connector 18.

[0083] In one example, the shell-and-tube CVI reactor 10 may include a body or footprint of 20×4.9×23 cm and may include a flow path 12 having a length of approximately 16.43 m, thereby generating a flow rate of approximately 520 ml. The body of the shell-and-tube CVI reactor 10 may include a first side 24 and a second side 26, such as Figure 1CAs shown. In one example, the first side 24 may include at least one groove or notch 24A and the second side 26 may include at least one protrusion 26A. The at least one notch 24A and the at least one protrusion 26A may be arranged so that when two shell-and-tube CVI reactors 10 face each other, they are aligned and can secure one shell-and-tube CVI reactor 10 to an adjacent shell-and-tube CVI reactor 10. Continuous Flow Reactor with Interwoven Tubular Flow Paths

[0084] As described above, the system may design, select, fabricate and / or manufacture an actual reactor having interwoven reactor tubes. Figures 1E to 1J Details of interwoven reactor tubes are described. Figures 1E to 1J An exemplary continuous flow reactor tube 10Z that can operate at low Re is illustrated. The continuous flow reactor tube 10Z may include a tubular flow path 12Z that includes turns or curves 14Z and bends 16Z. At least two of the turns or curves 14Z are disposed on a single longitudinal axis LX, but are located on different non-parallel planes, such as plane A and plane B, such that the turns or curves may form an angle of about 25° to about 60° about the longitudinal axis LX. Depending on the number of paths used to create the continuous flow reactor tube 10Z, the turns may be in two or more different non-parallel planes, such as about 6 to about 13 different planes, for example, 8 different planes. In addition, at least two turns are arranged so that the planes corresponding to the at least two turns may intersect with each other, for example, as Figure 1F Turns can also create patterns that may or may not repeat after a predetermined number of turns. For example, Figure 1G , which shows a cross section of a single path along its longitudinal axis, the single path may include a pattern 50Z repeated at least twice (see also Figure 1E ). Each flow path may include about 4 turns to about 128 turns or more alternating turns, such as about 16 to about 32 turns, for example, 12.5 alternating turns. Each turn 14Z may include an angle of about 110° to about 280°, such as about 135° to about 140°. In one example, a first turn may include an angle (such as an angle of about 135°) that is less than an angle of a second turn (such as an angle of about 140°). In addition, as Figure 1E , 1F As shown in Figures 1 and 1J, each flow path may also include about 8 to about 64 or more bends 16Z, such as about 8 to about 16 bends 16Z. Each bend 16Z may include an angle of about 15° to less than about 135°, such as an angle of about 30° to about 90°, such as an angle of about 45°. In one example, each pattern 50Z may be repeated after about 4 or more bends, such as after about 8 bends.

[0085] Additionally or alternatively, when the continuous flow reactor comprises an interwoven tubular flow path, the interwoven tubular flow path may comprise a plurality of interwoven tubular flow paths. 3 From about 6 to about 100 turns, such as from about 6.5 to about 93.2 turns.

[0086] In another example not shown in the figure, the path of the continuous flow reactor tube 10Z may include two or more different patterns that may or may not be repeated. When the continuous flow reactor tube 10Z includes a plurality of interlaced flow paths, each path of the continuous flow reactor tube 10Z may include a substantially similar pattern. Alternatively or in addition, each path of the continuous flow reactor tube 10Z may include different patterns. In addition, each path of the continuous flow reactor tube 10Z may include a similar number of repeating patterns (e.g., two similar repeating patterns), or may include more than or less than two repeating patterns. For example, the second path may include two similar repeating patterns, or may include three similar repeating patterns.

[0087] refer to Figure 1H and 1I , each turn 14Z in the continuous flow reactor tube 10Z can include a vertical L1 center-to-center distance between turns of about 1 cm to about 2 cm (e.g., about 1.5 cm). In addition, each turn 14Z can include a horizontal L2 center-to-center distance between turns of about 1 cm to about 2 cm (e.g., about 1.63 cm). In addition, each turn 14Z can include an end-to-end distance L3 of about 3 cm to about 4 cm, (e.g., about 3.85 cm). The radius of each turn 14Z in the continuous flow reactor tube 10Z can be substantially constant. For example, referring to Fig. 1I , radii R1 and R2 may differ from each other within a range of 0.05 cm or less, such as about 0.02 cm, to prevent significant differences in Dean numbers between alternating turns. For example, R1 may be about 1.10 cm and R2 may be about 1.12 cm. In one example, turn 14Z may be arranged to generate vortices to induce mixing of a process stream having a laminar flow with a Reynolds number of about 187.7 to about 375.5.

[0088] In one example, the plurality of turns 14Z may be along a three-dimensional path that may include a flow direction change of about 45° at the center of the turn. Additionally, each turn in the plurality of turns 14Z may include an angle of about 125° to about 180°.

[0089] Systems using serpentine and interwoven reactor tubes

[0090] refer to Figure 1K, in order for the system 1000 to design, select, fabricate, and / or manufacture an actual reactor 400 that can manage an actual process stream for commercial purposes and have the same results as the experimental reactor 100, the system 1000 may include a processor 1001 and a non-transitory machine-readable storage medium 1002 storing machine-readable instructions executable by the processor 1001 to estimate and / or determine desired parameters of the actual reactor 400. In one example, the processor 1001 may receive known values ​​and empirical values ​​of reactor parameters of the experimental reactor 100 and / or fluid parameters of the process stream entering the experimental reactor 100. The instructions stored on the non-transitory machine-readable storage medium 1002 store machine-readable instructions executable by the processor 1001 to determine non-empirical values ​​of reactor parameters of the experimental reactor 100 and / or fluid parameters of the process stream entering or introduced into the experimental reactor 100 using the received known values ​​and empirical values ​​of reactor parameters and / or fluid parameters. Instructions stored on the non-transitory machine-readable storage medium 1002 store machine-readable instructions that are executable by the processor 1001 to forward empirical and non-empirical values ​​and request the processor 1001 to determine, design, select, fabricate, manufacture and / or recommend an actual reactor 400 for an actual process material flow.

[0091] In one example, the experimental reactor 100 can be a hypothetical reactor with certain known parameters. Typically, the experimental reactor 100 can be a fixed reactor including a constant inner diameter id and a radius of curvature Rc (cm). The known empirical and non-empirical values ​​depend on the task. That is, depending on the goals that the user wants to achieve, the values ​​of at least some reactor parameters and / or fluid parameters can be empirical or non-empirical. In one example, the main variable associated with the empirical and non-empirical values ​​related to the reactor parameters and / or fluid parameters can be T min 、T max , the inner diameter id, the volume flow rate Q (mL / min) of the process stream, the path length L (cm) of the flow path, the radius of curvature Rc (cm) of the reactor tube, the density (ρ) of the fluid in the process stream, and the dynamic viscosity μ (mPa*s) of the fluid in the process stream, as well as the variance σ 2 time (min 2 ).

[0092] Use a pilot reactor to determine empirical values

[0093] In one example, the empirical value may be a value corresponding to an experimental reactor parameter and / or a fluid phase parameter that is linearly related to the experimental data set. Figure 2A and 2BAs shown, by introducing the process stream into the experimental reactor 100 using pulse injection at 10, a first detector 30 that can be in communication with the experimental reactor 100 can determine and / or have process stream fluid phase parameters such as density (ρ), dynamic viscosity (μ) and Dean number (De). Furthermore, as Figure 2C As shown, the first detector 30 can detect the volume / number of detectable particles or detectable tracers injected into the process material flow, as well as the time taken to inject the detectable particles or detectable tracers into the process material flow and / or the time taken to introduce the detectable particles or detectable tracers into the experimental reactor 100. In addition, the first detector 30 can determine or have the type of experimental reactor (e.g., JIB), the flow rate Q of the process material flow entering the experimental reactor 100, the path length L of the experimental reactor 100, and the volume of the experimental reactor 100. In one example, some of the above values ​​may be known to the user, so the detector does not need to detect these values. In one example, the detectable particle or detectable tracer can be, but is not limited to, viral / bacteriophage particles, riboflavin, salts, dyes, proteins, and / or sugars. For example, as Figure 2B As shown, the first detector 30 can determine that the process fluid is water, the experimental reactor type is JIB, the flow rate Q of the process material flow is 50 mL / min, the path length L of the experimental reactor is 1644 cm, the reactor volume is 520 mL, and the Dean number is 118.94.

[0094] See also Figure 2B The second detector 40 may also be in communication with the experimental reactor 100 and may detect and measure the time when the first detectable particle or detectable tracer exits or leaves the experimental reactor 100 and the time when the last large number of detectable particles or detectable tracers exit or leave the experimental reactor 100. In one example, the second detector 40 may also detect a parameter that may be detected or measured by the first detector 30. Based on Figure 2B The experimental raw data shown, the second detector 40 can generate a fitting curve, such as Figure 2D Based on this fitting curve, the variance σ is determined 2 time and standard deviation σ time For example, for a flow rate of 50 mL / min in the experimental reactor 100 with a path length of 1644 cm, based on the fitted curve, the variance σ 2 time is 0.5115min 2 , and the standard deviation σ time It is 0.7152min.

[0095] like Figure 2E As shown, given the variance σ2 time is 0.5115min2, the experimental reactor path length L is 1644cm, the Dean number De is 118.94, and the average residence time T Ave 4 min, the comparator 50 can derive a theoretical plate equivalent (HETP) value of 7.87 cm. The comparator 50 can then create a graph of experimental data HETP and Dean number corresponding to multiple flow rates and process stream viscosities (e.g., 0 g / L glucose, 50 g / L glucose, 100 g / L glucose, and 200 g / L glucose). In one example, based on the experimental data graph, the comparator 50 can also create an experimental data fit graph, such as Figure 2E As shown. Figure 2A As shown, at 200, the empirical value / variable may be forwarded to the system 1000 or received by the system 1000 to determine and / or create the actual reactor 400 of interest. In another example, the empirical value may have been derived or determined in advance, and therefore, the first detector 30 or the comparator 50 may not be required, such as Figure 3 As shown, or for that matter, no experimental reactor 100, first detector 30 or second detector 40 (not shown) is required.

[0096] Limiting factors in designing, selecting, fabricating, manufacturing and / or determining the actual reactor 400 may be product stability, desired viral latency, process parameters and / or lack of operational or kinetic considerations.

[0097] When the limiting factor is product stability, a target protein that is highly sensitive to viral inactivation chemicals is provided. In this example, the system 1000 can be made to determine acceptable reactor lengths and flow rates based on the minimum residence time required for viral inactivation kinetics and the maximum residence time required for product stability. When the limiting factor is a process parameter, the downstream process is limited by the volume flow rate Q and the path length L. In this example, the system 1000 can be made to determine the minimum residence time of the process material flow required for viral inactivation and the maximum residence time required for product stability. If there are no operational or kinetic considerations, the target protein may not include stability considerations. In this example, the system can be made to determine the resulting T min Appropriate Q and L.

[0098] Limiting Factors Product Stability - Use T min and T max System and method for determining actual reactor operating flow rates and path lengths

[0099] In one example, a user may use the system 1000 to determine the desired min and Tmax to develop or create the actual reactor. For example, the user requires T min For 60 minutes, T max In addition, the user can input the desired inner diameter of the reaction tube of the experimental reactor 100 and the radius of curvature of the experimental reactor 100. In this particular example, the experimental reactor 100 and the actual reactor 400 may include the same inner diameter id and radius of curvature Rc. Figure 2A , the density ρ and dynamic viscosity μ of the fluid in the process material flow are either known or detected by the first detector 30 and can be provided to the system at 200. Therefore, the system 1000 can provide the design and manufacturing specifications of the actual reactor 400 based on the above-mentioned input parameters. The new design and manufacturing specifications of the actual reactor 400 (i.e., the path length of the reaction tube of the actual reactor 400) can be changed based on the operating volume flow rate Q. For the operation volume flow rate Q at σ 2 time <σ 2 max This is particularly true for systems operating under T min and T max However, when the system is in σ 2 time =σ 2 max When operating under these conditions, only one combination of reactor path length L and flow rate Q satisfies T min and T max Require.

[0100] refer to Figure 4 , 5A At 50, the system 1000 may determine or the user may input known values ​​210 / 210A, 250 / 250A and empirical values ​​and / or variables 260 / 260A. These empirical values ​​and / or variables 260 / 260A may be classified as reactor parameters and / or fluid phase parameters of the experimental reactor 100. For example, Figure 5A and 5B As shown, at 212, the desired predetermined T of the experimental reactor 100 may be min is input into the comparator 50. At 214, the desired predetermined T of the experimental reactor 100 may be max is input into the comparator 50. Also, at 216, the inner diameter id of the experimental reactor 100 may be input into the comparator 50. At 218, the radius of curvature Rc of the experimental reactor 100 may be input into the comparator 50. Thus, in this particular example, the input known value corresponding to the reactor parameter may be T min 、T max, inner diameter id and radius of curvature Rc, such as Figure 5B As shown in 210A.

[0101] In addition, reference Figure 5B At 252, the density (ρ) of the fluid in the process stream from the first detector 30 may be input into the comparator 50. At 254, the dynamic viscosity (μ) may be input into the comparator 50. Thus, in this particular example, the input known values ​​corresponding to the fluid phase parameters may be density ρ and dynamic viscosity μ, as shown at 250 / 250A.

[0102] As described above, empirical values ​​can then be determined using pulsed injection of the process stream into the experimental reactor 100. For example, Figure 5A and 5B , at 260A, the comparator 50 can use T min (e.g., 60 min) and T max (e.g., 75 min) to predict and / or determine the empirical value. In this particular example, given T min and T max , the comparator 50 can determine the maximum variance σ 2 time (i.e. σ 2 max ) and the corresponding average residence time T Ave , as shown below:

[0103] (1)T Ave =T min +(n*σ max ), where n may be 5;

[0104] (2)T Ave =T max -(m*σ max ), where m may be 3;

[0105] (3) ΔT = 8σ max

[0106] (4) 15 = 8σ max

[0107] σ time =1.875min,σ 2 time =3.52min 2 ,T Ave –69.375min

[0108] Given a certain variance σ 2 time , T Ave and standard deviation σtime The comparator 50 uses the data from the first detector 30 and the second detector 40 respectively, and can derive the cm at the empirical value. 2 and / or the theoretical plate (HETP) of the Dean number De, as described above. For example, HETP can be defined as follows:

[0109] (5)HETP=(aDe 3 +bDe 2 +cDe+d),

[0110] where De is the Dean number, and a, b, c, and d are fitted based on empirical data and are only valid when the Dean number is ≥100.

[0111] By applying the panel with Dean number (De) ≥ 100, each De fixes a flow rate Q and returns a HETP as shown below.

[0112] (6)

[0113] (7)

[0114] (8)

[0115] In one example, if Figure 4 As shown, non-empirical values ​​associated with reactor parameters can be derived at 302, or non-empirical values ​​associated with fluid phase parameters can be derived at 304. For example, referring to Figure 5A and 5B , the empirical equation above and the determined σ 2 time , σ time and T Ave It may be forwarded to the system at 200 so that the system at 302A can simultaneously solve the equations shown below related to the empirical values ​​to determine the path length (L) and flow rate (Q) of the actual reactor 400.

[0116] (9)

[0117] (10)T Ave =T min =+(n*σ time ), where n can be 5

[0118] (11)T Ave =T max –(m*σ time ), where m can be 3

[0119] For a constant T Ave, a fixed Q value will fix the corresponding LT Ave Each Q value and LT Ave The value combination returns a resulting σ 2 time Values ​​as shown below.

[0120] (12)

[0121] Solving for σ 2 time The minimum reactor volume is affected by T Ave =T min +(5*σ max ) and T Ave =T max -(3*σ max )constraint. Figure 5C Different flow rates and their respective corresponding reactor tube length paths L are shown to have a desired predetermined T min and T max Additionally or alternatively, such as Figure 5B As shown in steps 312 and 314, the outputs may include reactor tube path length L and flow rate Q, respectively.

[0122] Since the experimental reactor in this example is the same as the actual reactor, based on the selected flow rate Q and the corresponding reactor tube path length, a series of reactors can be connected to each other to achieve the corresponding reactor tube path length.

[0123] The limiting factors are the process parameters - reactor volume RV and flow rate Q determine T min and T max System and method

[0124] In one example, a user using the system 1000 can determine the T of a reactor with known reactor parameters and fluid phase parameters. min and T max For example, Fig. 6A As shown, at 210B, the known reactor parameters may include the reactor volume RV (e.g., 3120 mL), the flow rate Q (e.g., 50 mL), the inner diameter id, and the radius of curvature Rc. Fig. 6A As shown, at 250B, the known fluid phase parameters may include the density (ρ) and dynamic viscosity (μ) of the fluid in the process stream. In this example, the reactor volume RV, flow rate Q, inner diameter id, and radius of curvature Rc may be input into the comparator 50. The comparator 50 may then determine empirical values, such as the variance σ 2 time , average residence time T Ave and HETP.

[0125] refer to Figure 6B , in this example, the reactor volume RV of the reactor can be input into the system at 222, the flow rate Q of the process stream can be input into the system at 224, the inner diameter id of the reaction tube of the reactor can be input into the system at 216, and the radius of curvature Rc of the reaction tube of the reactor can be input into the system at 218. Therefore, in this particular example, the input values ​​corresponding to the reactor parameters at 210B can be the reactor volume RV, the flow rate Q of the process stream, the inner diameter id of the reaction tube of the actual reactor, and the radius of curvature Rc of the reaction tube of the actual reactor.

[0126] Additionally, at 252, the density (ρ) of the fluid in the process stream may be input into the system 1000. At 254, the dynamic viscosity (μ) may be input into the system 1000. Thus, in this particular example, the input known values ​​corresponding to the fluid phase parameters at 250B may be the density p and the dynamic viscosity μ.

[0127] Based on the known process material flow rate Q (e.g., 50 mL / min), the known reactor volume RV (e.g., 3120 mL), the known reactor tube inner diameter id, and the known radius of curvature Rc, the processor 1001 of the system 1000 can predict and determine T min and T max .

[0128] In order to predict and / or determine T min and T max , the known value can be input into the comparator 50. The comparator 50 at 260B, using the experimental reactor as described above, can use the process material flow rate Q and the reactor volume RV to predict and / or determine the average residence time (T Ave ), as shown in the following equation.

[0129] (13)

[0130] (14)

[0131] (15)T Ave =62.4min

[0132] Once the comparator 50 derives T Ave value, the comparator 50 can then utilize T Ave , Q, De, Rv and L are used to predict and / or determine the variance σ using the experimental reactor 100 and the following equation: 2 time .

[0133] HETP=(aDe3 +bDe 2 +cDe+d), where De is the Dean number, and a, b, c, and d are based on empirical data fitting and are only valid when the Dean number is ≥100.

[0134] For a Q of 50 mL / min and a De of 118.94, HETP may be equal to 7.464 cm. Based on the derived HETP value, the comparator 50 may predict or determine the variance σ using the following equation: 2 time .

[0135] (16)

[0136] (17)

[0137] (18)

[0138] (19)σ 2 time =2.95min 2

[0139] The above empirical and known values ​​can then be forwarded to the system at 200. The variance σ has been derived 2 time The processor 1001 can use the variance σ 2 time , standard deviation σ time , reactor tube length L, radius of curvature Rc, Dean number De, flow rate Q and T Ave , to estimate and / or determine the actual reactor T min and T max , as shown in the equation below.

[0140] (20)T Ave =T min +(n*σ max ), where n may be 5;

[0141] (21)T Ave =T max -(m*σ max ), where m may be 3;

[0142] (22)T min =53.81min

[0143] (23)max=67.55min

[0144] In one example, if Figure 6B As shown, at 316, the display may display T minvalue, and at 318, the display may display T max Additionally or alternatively, the display may display T min and T max A graphical representation of Figure 6C shown.

[0145] No operational or kinetic considerations - use T min System and method for determining the path length of an actual reactor based on the operating flow rate (Q)

[0146] In one example, a user may use the system 1000 to determine the desired min (60min), process material flow rate Q (50mL / min), the inner diameter id of the reaction tube (0.635cm), the radius of curvature Rc, the density ρ and the dynamic viscosity μ to develop or create an actual reactor. For example, Fig. 7A As shown, at 210C, known reactor parameters may include T min , process material flow rate Q, the inner diameter id of the reaction tube, and the radius of curvature Rc. Fig. 7A As shown, at 250C, the known fluid phase parameters may include the density (ρ) and dynamic viscosity (μ) of the fluid in the process stream. In this example, these known values ​​may be input into the comparator 50 so that the empirical value, such as the variance σ, may be determined. 2 time and the average residence time T Ave .

[0147] refer to Figure 7B , in this example, T min The system may be input at 212, the flow rate Q of the process stream may be input at 224, the inner diameter id of the reactor tube may be input at 216, and the radius of curvature Rc of the reactor tube may be input at 218. Thus, in this particular example, the input values ​​corresponding to the reactor parameters at 210C may be T min , process material flow rate Q, the inner diameter id of the reaction tube of the actual reactor, and the curvature radius Rc of the reaction tube of the actual reactor.

[0148] Additionally, at 252, the density (ρ) of the fluid in the process stream may be input into the system 1000. At 254, the dynamic viscosity (μ) may be input into the system 1000. Thus, in this particular example, the input known values ​​corresponding to the fluid phase parameters at 250C may be the density (ρ) and the dynamic viscosity (μ).

[0149] Based on the known T min, process material flow rate Q (e.g., 50 mL / min), known reaction tube inner diameter id and known curvature radius Rc, the processor 1001 of the system 1000 can predict and determine the reaction tube flow length L of the actual reactor 400.

[0150] In order to predict and / or determine the reaction tube flow length L of the actual reactor 400, a known value can be input into the comparator 50. The comparator 50 is at 260C, using the experimental reactor described above, and can use T min , process material flow rate Q and Dean number De to predict and / or determine T Ave and variance σ 2 time , as shown in the equation below.

[0151] (twenty four)

[0152] (25)

[0153] a, b, c, and d are based on empirical data fitting and are only valid when the Dean number is ≥100.

[0154] (26)T Ave =Tmin+(n*σ time ), where n can be 5

[0155] (27) Where n can be 5.

[0156] By rearranging the equation

[0157] (28)

[0158] (29)

[0159] refer to Fig. 7A and 7B , the above empirical equation and known values ​​can be forwarded to the system at 200 so that the system can determine the path length (L) at 302C given a T min of 60 minutes, a flow rate Q of 50 mL / min and an id of 0.635 m in the following equation.

[0160]

[0161] Solving the above equation, L will equal 108.99m or a reactor volume of 3.46L. Figure 7C A graphical representation may also be displayed at 300 showing all data points corresponding to combinations of flow rate Q and reactor volume RV that result in a Tmin of 60 minutes.

[0162] In all of the above examples, the inner diameter id and radius of curvature of the experimental reactor 100 and the actual reactor 400 remain the same. However, in one example, as described below, the system can also design, select, make, manufacture, and recommend a reactor that includes an inner diameter id of a reaction tube that is different from the inner diameter id of the experimental reactor. This is particularly useful when submitting data to regulatory agencies, such as the FDA, because applicable regulations require data to demonstrate compliance. The system described below will allow a user to run a small-scale process material flow for data purposes and then scale it up for large-scale production without making any significant changes to the process material flow or the final results of the process material flow during scale-up production.

[0163] Zoom in or out using aspect ratio

[0164] In one example, a user may use the system 1000 to determine the desired min (60 min), process stream flow rate QExit (100 mL / min), density ρ and dynamic viscosity μ, and a known aspect ratio to develop or create an actual reactor 400. For example, Fig. 8A As shown, at 210D, known reactor parameters may include T min , process material flow rate Q and aspect ratio. The aspect ratio can be defined as the ratio of the radius of the reaction tube of the experimental reactor to the radius of curvature Rc of the experimental reactor. Moreover, the aspect ratio can be about 0.01 to about 10, such as about 0.05 to about 5, for example about 0.1 to about 0.5. In addition, Fig. 8A As shown, at 250D, the known fluid phase parameters may include the density ρ and dynamic viscosity μ of the fluid in the process stream. In this example, these known values ​​may be input into the comparator 50 so that the empirical value, such as the variance σ, may be determined. 2 time and the average residence time T Ave .

[0165] refer to Figure 8B , in this example, T min (e.g., 60 min) can be input into the system at 212, and the flow rate Q of the process stream (e.g., 100 mL / min) can be input into the system at 224. Thus, in this particular example, the input value corresponding to the reactor parameter at 210D can be T min and process material flow rate Q.

[0166] Additionally, at 252, the density (ρ) of the fluid in the process stream may be input into the system 1000. At 254, the dynamic viscosity μ may be input into the system 1000. Thus, in this particular example, the input known values ​​corresponding to the fluid phase parameters at 250D may be the density ρ and the dynamic viscosity μ.

[0167] To predict and / or determine the reaction tube flow length L and the actual inner diameter id of the reactor 400, known values ​​may be input into the comparator 50. The comparator 50, at 260D, using the experimental reactor 100 as described above, may utilize T min and process stream flow rate Q to predict and / or determine T Ave and variance σ 2 time , as shown in the following equation.

[0168] (31)

[0169] (32)HETP=f(v)=(av 3 +bv 2 +cv+d),

[0170] where a, b, c and d are fitted based on empirical data and are applicable to all Dean numbers

[0171] (33)T Ave =Tmin+(n*σ time ), where n can be 5.

[0172] (34)T Ave =T max -(m*σ time ), where m can be 3.

[0173] (35)

[0174] (36)Q=v*CA

[0175] (37)

[0176] (38) By rearranging the equation (39)

[0178]

[0179] refer to Fig. 8A and 8B, the empirical equation and known values ​​above can be forwarded to the system at 200. In this particular example, the system at 500 can ask the user if he would like to use an actual reactor having a substantially similar diameter to the experimental reactor. If the user answers yes, then the system can Fig. 7A and 7B Determine the length. However, if the user answers no or selects scaling of the reactor, the system can obtain the aspect ratio and the processor 1001 of the system 1000 can solve for the reactor length L in the equation below.

[0180] (40)

[0181] In one example, scaling of the reactor may include at least one of: (i) scaling the size of the experimental reactor to an actual reactor having the same aspect ratio as the experimental reactor but a different inner diameter; (ii) scaling the size of the experimental reactor to an actual reactor having the same aspect ratio as the experimental reactor and the same inner diameter as the experimental reactor; (iii) scaling the size of the experimental reactor to an actual reactor having a different aspect ratio than the experimental reactor and a different diameter than the experimental reactor; and (iv) scaling the size of the experimental reactor to an actual reactor having a different aspect ratio than the experimental reactor but the same diameter as the experimental reactor.

[0182] Once L is determined, the system 1000 calculates the reactor length L based on the standard deviation σ time and the derived value of the mean linear flow velocity (cm / min) can be determined internally using the following equation:

[0183] (41)

[0184] (42)

[0185] (43)

[0186] (44)

[0187] For this particular example, in order to design, select, fabricate and / or manufacture a practical reactor with a fixed aspect ratio, a graph between HETP and linear flow rate can be derived. For example, Figure 8C It can be derived from an experimental small-scale reactor with a fixed aspect ratio and an inner diameter of about 0.1 cm to about 0.2 cm, such as 0.156 cm tested with water, and an experimental small-scale reactor with the same fixed aspect ratio and an inner diameter of about 0.6 cm to about 0.7 cm, such as 0.635 cm tested with water and glucose.

[0188] Fig.8D is a plot between experimental HETP and predicted HETP, which shows the Figure 8C Parity between predicted and experimental HETP.

[0189] like Figure 8B As shown in 312 and 320 of FIG. 1 , the reactor path length and inner diameter of the actual reactor can be derived and / or determined. Fig. 8E As shown, based on the derived reactor path length, derived inner diameter, and predicted HETP value, a model can be created that shows that Q = 100 mL / min and T min = 60 for all solution values ​​of reactor volume and inner diameter.

[0190] Fig. 9 An exemplary embodiment is shown in which the system 1000 can determine an ideal reactor for a given fluid and reactor volume. In this example, as in the previous example above, a user can input known parameters into the comparator 50. The comparator 50 can determine empirical values ​​based on the known values, which can be input into the system 1000 at 200. Non-empirical values, such as reactor tube length L, can then be determined at 300 as described above. Given the reactor length, the system can determine that the reactor volume is 700. At 750, based on known fluid properties, detectable particles / tracers, and the determined reactor volume, the system can communicate with a database 770. The database 770 can include previously designed or manufactured reactors based on reactor volume and known fluid and detectable particle / tracer parameters. The database 770 can then provide the system 1000 with a list of different actual reactors, each having a substantially similar volume, for use in achieving the same desired end result using similar fluids and detectable particles / tracers. The processor 1001 of the system 1000 may then review the provided list of actual reactors to select the best actual reactor 400 for the intended purpose.

[0191] refer to Fig. 10A , in another example, T min (e.g. 60 min) can be entered into the system at 212, T max (e.g., 75 min) can be input into the system at 213, and the flow rate Q of the process stream (e.g., 500 mL / min) can be input into the system at 224. Thus, in this particular example, the input values ​​corresponding to the reactor parameters at 210D can be two T min and the process stream flow rate Q. Additionally, the aspect ratio at 600 may be entered at this time or later, as described below.

[0192] Additionally, at 252, the density (ρ) of the fluid in the process stream may be input into the system 1000. At 254, the dynamic viscosity μ may be input into the system 1000. Thus, in this particular example, the input known values ​​corresponding to the fluid phase parameters at 250D may be the density ρ and the dynamic viscosity μ.

[0193] To predict and / or determine the reaction tube flow length L and the actual inner diameter id of the reactor 400, known values ​​may be input into the comparator 50. The comparator 50, at 260D, using the experimental reactor 100 as described above, may utilize T min , T max and process stream flow rate Q to predict and / or determine T Ave and variance σ 2 time , as shown in the following equation.

[0194] (45)T Ave =T min +(n*σ max ), where n is 5

[0195] (46)T Ave =T max -(m*(σ max )), where m can be 3

[0196] (47)ΔT=8σ max

[0197] 15=8σ max

[0198] σ time =1.875min;σ 2 time =3.52min 2 ; T Ave =69.375min

[0199] refer to Fig. 10A , the above empirical value equation and / or corresponding values ​​and known values ​​can be forwarded to the system at 200. In this particular example, the system at 500 can ask the user if he is willing to use an actual reactor with a substantially similar diameter to the experimental reactor. If the user answers yes, then the system can be based on the above example and Fig. 7A and 7B However, if the user answers no or selects scaling of the reactors, the system can obtain the aspect ratio and the processor 1001 of the system 1000 can divide the experimental HETP experimental data points by the JIB inner diameters of the small, medium, and large reactors, such as Fig. 10B The system can then fit a curve to the data set, as Fig. 10C Using the equation below, the system can then be plotted as a function of path length versus internal diameter for a flow rate of 500 mL / min, T min is 60, T max is 75.

[0200] (48)

[0201] (49)

[0202] a, b, c, and d are based on fits to empirical data and are applicable to all Dean numbers.

[0203] (50)

[0204] Fixed id returns the path length item, such as Fig. 10D shown.

[0205] Example 1

[0206] Dwell time distribution generation.

[0207] The JIB was designed based on a previous development project by Boehringer Ingelheim and 3D printed using SLA technology by 3D Systems (Rock Hill, SC). Riboflavin and glucose used to create the mobile phase and pulse tracer were purchased from Thermo Fisher Scientific (Suwanee, GA). The viscosity of the solution was measured by a microVISC S viscometer using an A05 chip (San Ramon, CA). The density of the solution was measured by a Mettler-Toledo Densito density meter (Columbus, OH).

[0208] The medium-sized 3D-printed JIBs were tested using an Akta Avant 150, while the large-sized JIBs were tested using an Akta Pilot 600 from GE Healthcare (Uppsala, Sweden). The JIBs were first flushed with 1 reactor volume of mobile phase. Next, a fixed volume of riboflavin dissolved in the mobile phase was pulsed and discharged with the mobile phase. This produced a residence time distribution (RTD) profile upon exiting the reactor, which was detected and quantified by UV-Vis absorption at the riboflavin absorption maxima (i.e., 267, 372, and 445 nm). The RTD peak was then analyzed by fitting a Gaussian distribution. Based on this fit, the peak variance of the measured values, which is the spread of the RTD, was determined. The method was tested over a range of flow rates and viscosities that varied with glucose concentration. These values ​​were converted to HETP. A plot of HETP versus Dean number was created and a third-order polynomial was fitted. The following series of equations were then used:

[0209] 1. Start taking control of the equation

[0210] a)

[0211] b)

[0212] a, b, c, and d are based on fits to empirical data and are valid only when the Dean number is ≥100.

[0213] c)

[0214] d)

[0215] 2. Rearrange the equation

[0216] e)

[0217] f)

[0218] 3. Solve for L:

[0219]

[0220] 4. Fill in variables

[0221] mAb concentration (g / L) <![CDATA[Approximate kinematic viscosity (m 2 / s)]]> 10 <![CDATA[1.2x10 -6 <!-- 17 -->]]> 20 <![CDATA[1.4x10 -6 ]]> 50 <![CDATA[2.0x10 -6 ]]>

[0222] The following table shows the T values ​​of the medium and large reactors at 15 min, 30 min, and 60 min. min .

[0223]

[0224]

[0225]

[0226] Evaluating the impact of flow dynamics on key process parameters in a continuous viral inactivation reactor

[0227] The JIB was designed based on a previous development project and 3D printed by 3D Systems (Rock Hill, SC) using SLA technology. Riboflavin and glucose used to create the mobile phase and pulse tracer were purchased from ThermoFisher Scientific (Suwanee, GA). The viscosity of the solution was measured using a microVISC S viscometer with an A05 chip (San Ramon, CA). The density of the solution was measured by a Mettler-Toledo Densito density meter (Columbus, OH).

[0228] Small and medium 3D printed JIBs were tested using an Akta Avant 150, while large JIBs were tested using an Akta Pilot 600 from GE Healthcare (Uppsala, Sweden). The JIB was first flushed with 1 reactor volume of mobile phase. Next, a fixed volume of riboflavin dissolved in the mobile phase was pulsed and emptied along with the mobile phase. This produced an RTD profile upon exiting the reactor, which was detected and quantified by UV-Vis absorption at the riboflavin absorption maxima (i.e., 267, 372, and 445 nm). Table 1 below lists the inner diameters, flow rates, mobile phases, and multiple serially connected JIBs tested in this study.

[0229] Table 1 Overview of all combinations of mobile phase, inner diameter, path length, and flow rate tested by pulse injection experiments using JIB

[0230]

[0231] The peak value is then analyzed by fitting a Gaussian distribution. Based on this fit, the peak value of the measured value as the RTD expansion is determined, To better understand the impact of the quantitative value of the variance, equation 1 was calculated, where Q is the volume flow rate. In addition, the data set was transformed using equations 2 and 3, where HETP is the height equivalent to a theoretical plate, T Ave is the average residence time, RV is the reactor volume, and L is the length of the JIB process.

[0232] (1)

[0233] (2)

[0234] (3)

[0235] Flow rate and path length

[0236] like Fig.11A As shown, for all three reactor sizes, the slowest flow rates produce the broadest peaks. This is shown in the graph as having a relatively high standard deviation (e.g., about 82 mL for the JIB(1) data set at 5 mL / min), which describes the distribution of the peaks from the center (i.e., T Ave) to 34% of the total mass to the left or right. As the flow rate increases, the peak narrows and the standard deviation decreases at an exponential rate. However, when the flow rate increases above 20mL / min, an inflection point occurs. The peak becomes wider until another inflection point is reached at the 30mL / min set point. As the flow rate increases, the peak narrows significantly until the 55mL / min set point is reached. Overall, the progression from slower flow rates to faster flow rates shows an initial asymptote, two inflection points, and finally a second asymptote. View Fig.11A For the JIB(1), JIB(2), and JIB(6) data in , the phenomenon of two asymptotes and two inflection points maintains the same flow velocity over all path lengths.

[0237] Comparing path lengths, the peak broadens as path length increases. This is a well-characterized observation and a reproducible phenomenon of PFR. When converted using Equation 3 Fig.11A When the data points are taken and converted to HETP equivalent heights, the three path length datasets are overlaid to create Fig. 11B .

[0238] To understand the driving forces of this transition (i.e., the two inflection points), previously published work on Dean Vortices was further explored. A suspension in water was used to visualize the flow pattern when between two rotating drums (i.e., Taylor-Couette flow). As the flow rate in the flow cell was increased, Aider made observations for specific Dean numbers where the flow pattern transitioned from laminar to chaotic. Similar experiments were performed in JIB using suspended mica in water and the same laminar to chaotic flow transition was found. These specific Dean numbers and corresponding observations are outlined by Aider et al., and together plotted in Fig.12 The Dean number is defined by Equation 4, where ρ is the fluid density, u is the average linear flow velocity, D is the flow path inner diameter, μ is the dynamic viscosity, and Rc is the radius of curvature of the serpentine pattern.

[0239]

[0240] The two inflection points and the faster flow velocity asymptote correspond to the visually observable manifestations of the flow transitioning from unstable flow, wavy waves, and fully turbulent flow, respectively. Considering the flow in a circular straight pipe, the onset of turbulence is usually observed at a Reynolds number of about 2000. JIB is able to simulate turbulent behavior at Reynolds numbers of about 174. Due to this large difference, the term "weakly turbulent" is used.

[0241] To demonstrate the validity of the transition of the 2 asymptotes and 2 inflection points behaviors found in the above section is controlled by the Dean number, the viscosity of the mobile phase was increased with three concentrations of glucose. Fig.13AThe response to the addition of glucose is shown. For all four mobile phases, the standard deviation starts at the highest value at the slowest flow rate and narrows as the flow rate increases. However, the flow rate required to reach inflection point 1 increases with increasing glucose concentration and increasing viscosity. The same is true for inflection point 2 and the flow rate required to reach the second asymptote.

[0242] To illustrate this clear transition in the inflection point and asymptote, the flow rate is converted to that described in Equation 4 and Fig. 13B The x-axis is normalized to the Dean number shown in . Once normalized to the Dean number, the asymptote and the inflection point align. Although the shift of the data set returns to normal with correction, the magnitude of the spread appears to be corrected only for higher Dean number (i.e., De>100) operation. This is due to the Dean vortex taking over radial mass transfer above De=70. For lower Dean numbers (De<70), other dispersion mechanisms dominate and the mobile phase grades from widest to narrowest at 200, 0, 50, and 100 g / L glucose do not follow a concentration-based trend.

[0243] To inform the operation of the JIB, two predictive modeling approaches can be generated. The first uses a lumped data pool approach, utilizing all experimental data from various glucose mobile phase experiments and normalizing the data to HETP and Dean number ( Fig. 13B ). Based on this data set, a third-order polynomial can be fitted to this data pool, resulting in Fig. 13C and Equation 5, where a, b, c, and d are based on polynomial fits. Equation 5 allows the prediction of the width of the peak based on the Dean number. This model can be Fig.13D ) to generate a parity plot that describes how well the model predicts the volume expansion of the pulse injection in the JIB. The R2 value of the fit is relatively low, 0.8405, which is expected due to the large amount of variability in the De<100 dataset. Using Equations 2 and 6, T can be approximately estimated min (5σ) (i.e., an estimate of the time when the first detectable particle, such as a virus particle or surrogate tracer, exits the reactor) is expected to appear, e.g. Fig.13E As shown. min The significance of (5σ) will be explained later. The fit is improved, with an R2 value of 0.9175, however, all data points below the y=x line represent cases where the model predicted an incubation time that was longer than what actually occurred. Most of the error occurs at larger T corresponding to experiments with lower Dean numbers. min Time point. This is an important question because a key unit operation specification is residence time.

[0244] (5)HETP=aDe 3 +bDe 2 +cDe+d

[0245] The second approach is applicable if the criterion for JIB unit operation is to maintain a Dean value greater than 100. When this condition is true, an exclusion criterion is implemented into the model that only allows data points collected at a set point of De>100 ( Fig.14A ). This approach aims to reduce the noise in the model by removing the highly variable low Dean number data sets. When this change is applied, the variability is reduced, which allows the peak spread in the JIB to be predictable for fluids of variable viscosity and density operating at different flow rates ( Fig. 14B ). Fig. 14C Parity plots showing pulse volume expansion and R for the second model approach compared to the concentrated dataset approach are shown. 2 The value increases to 0.9201. When T is calculated and plotted min (5σ) Fig.14D ), resulting in a better prediction model, R 2 The value is 0.9812.

[0246] This model has two main applications when determining the design and operating conditions of a JIB-based CVI installation.

[0247] 1. In view of the requirements of product flow (i.e. T min and T max ), determine the length of the reactor and the operating flow rate

[0248] 2. Given the size of the reactor and the operating flow rate, predict the output of the product stream

[0249] Starting with an understanding of the minimum residence time required for viral inactivation and the maximum amount of time that the target molecule can be in acidic conditions before affecting product quality, Equations 6 and 7 can be applied to help determine the flow rate and path length required to meet these specifications.

[0250] (6)T Ave =T min +(n*σ time )

[0251] (7)T Ave =T max -(m*σ time )

[0252] (8)

[0253] Table 2A below summarizes the selection of σ time Quantitative aspects of the "n" and "m" values.

[0254] Table 2A. Tangible quantification of standard deviation parameters relevant to viral clearance risk assessment

[0255]

[0256]

[0257] Fig.15A The final determination of the RTD profile is illustrated. The theoretical starting point for viral and target product breakthrough is defined by T in Equation 8 min (2v). This value was derived by Hagen and Poiseuille who stated that for flow in a circular pipe, the fastest part of the flow occurs at the geometric center of the cross-sectional area of ​​the flow path and runs at 2 times the mean flow velocity. This is considered our "speed of light". Due to the influence of mass transfer phenomena (i.e., convection, diffusion, and Dean vortices), this condition is almost impossible to achieve in real practice and requires highly specific and extreme conditions to achieve operation in a JIB.

[0258] The exit pulse injection is considered as a Gaussian peak, so its spread is considered to be calculated as σ. For example, n = 5 (i.e., T min (5σ)) is understood to represent 0.00003% of the product at T min (2v) and T min (5σ) leaves the reactor. T min (2v) and T min This difference between (5σ) can be found in Fig.15A This is considered a conservative ideology that overvalues ​​under-incubated groups. min (2v) and T min The intersection of (5σ) is derived from the ideology of estimation, ignoring the possibility of the real world. min (2v) is the “speed of light” discussed above, σ time There is no limit to the value of "n". Choosing a sufficiently large value of "n" allows calculation of T occurring in negative time min .

[0259] In a similar manner, Table 2B and Fig. 15B Corresponding to T max Determine, where m = 3 (ie T max (3σ)) It is expected that about 99.865% of the product pool will be less than or equal to T max Exit the reactor.

[0260] Table 2B Tangible quantification of standard deviation parameters relevant to product quality risk assessment

[0261]

[0262] This decision will depend on the stability data of the product or the acceptable yield loss under acidic or any other virus inactivation conditions. If we combine equations 6 and 7, we get equation 9.

[0263] (9)T max -T min =ΔT = (n + m) * σ time

[0264] With a confirmed T min and T max , we can calculate T Ave and σ time Using Equations 2, 3, and 5, it is possible to find multiple combinations of path lengths and flow rates (i.e., starting with a flow rate that produces De>100) that satisfy Figure 5C The calculated T Ave and σ time The minimum flow rate and the minimum reactor volume can be calculated. This is an undesirable operating position because any increase or decrease in flow rate will result in a residence time that is not subject to T. min or T max As the reactor volume increases, the volume flow rate also increases to maintain the target T max As the flow rate increases, the efficiency of the JIB also increases. Figure 5C The JIB operation shown as error bars in the figure is more flexible. In addition to increasing efficiency, changing the operating window can be done by changing n, m, T in equations 6 and 7. min or T max value to modulate.

[0265] In order to intuitively understand this idea and phenomenon, and considering T min (5σ) and T max (3σ) are strictly defined for processes >60 min and <79 min, respectively, generating Fig.16 For smaller reactors operating at slower flow rates, σ time The value is larger, which can be seen from the difference in the width of the two peaks. Operating at the target flow rate, both designs comply with T min and T max As the reactor size increases and the corresponding flow rate increases, σ time Reduce, allowing flexibility in deviations during operation.

[0266] Variable flow rates and viscosities are unavoidable when CVI is actually operated in a GMP environment. Through the work performed in this series of experiments, this variability can be addressed by understanding the extremes of process operation and the corresponding worst-case scenarios, and predicting how they propagate into the process output. Based on these isovosity experiments, the worst-case scenario for viral incubation time appears to be a high viscosity solution. Given that the viscosity of the chromatographic elution peak will experience a peak, this should be considered the worst-case scenario. All other parts of the peak (i.e., the front and tail) have lower viscosities relative to the maximum peak, larger Dean numbers, and therefore better mixing.

[0267] To test this claim, a simulated protein peak was generated using glucose to increase viscosity and NaCl to generate the conductivity trace. In the theoretical case of a Protein A column, the mAb would elute from the column in a Gaussian-like shape with some tail. This general peak shape was generated using the gradient function of the Akta Avant 150, where the A1 line contained deionized water, A2 contained riboflavin dissolved in water, and B1 contained 200 g / L glucose and approximately 150 mM NaCl. To evaluate different viscosity gradient positions, four pulse injection positions were selected, the first occurring before the addition of glucose (0 g / L glucose), 2 peak mid-heights (50 g / L glucose), and the peak maximum (100 g / L glucose). Fig.17A All four injections are shown. To insert a peak without disturbing the density gradient, the A pump was switched from A1 to A2 while maintaining the gradient slope. This explains why the pulse injections are at different heights under the glucose curve. The pulse injection positions were evaluated in their own experiments because the injections overlapped.

[0268] The phenomenon of peak diffusion as a function of viscosity occurs in a dynamic composition setting. Fig. 17B It shows that the prediction model is effective in predicting T min The difference between the isocratic prediction and the dynamic experimental results is less than 1 minute, while the difference between the first water pulse injection and the maximum peak pulse is about 4 minutes, making it a worst-case scenario for virus incubation.

[0269] Example 2

[0270] Mobile phase and flow chamber.

[0271] The JIB was designed based on a previous development project, which is described in pending U.S. Patent Serial No. 62 / 742534 (incorporated herein by reference in its entirety), and was 3D printed using SLA technology from 3D Systems (Rock Hill, SC). Riboflavin, Tris buffer (TSB), and glucose used to make the mobile phase were purchased from ThermoFisher Scientific (Suwanee, GA). The viscosity of the solution was measured by a microVISC S viscometer (San Ramon, CA) using an A05 chip. The density of the solution was determined by a calibrated pipette and balance.

[0272] Phage selection:

[0273] ΦX174 and the corresponding host bacteria E. coli C were purchased from ATCC (ATCC catalog numbers: 13706-B1 and 13706, respectively). The concentration of ΦX174 was quantified by using a standard plaque formation assay, which requires co-inoculation of the liquid and the host bacteria E. coli C onto a tryptic soy agar plate with plating agar (i.e., tryptic broth containing 0.7% agarose). E. coli C was inoculated onto a tryptic soy agar plate with plating agar (i.e., tryptic broth containing 0.7% agarose). Bacteriophage ΦX174 was selected as a suitable tracer for this experiment due to some of its inherent properties. ΦX174 is a relatively resilient phage with a low chance of losing infectivity when suspended in appropriate mobile phase conditions, but can also be easily sterilized using 0.1 M NaOH and a reasonable contact time. The surface characteristics of this phage are relatively inert compared to other viruses. Previous experience has shown that this phage has significantly lower surface adsorption to positively charged, hydrophobic, and multimodal chromatography resins relative to other virus models. Therefore, in slightly alkaline solutions with low ionic strength (i.e., pH 7.5 and 150 mM NaCl), the probability of non-specific adsorption of the virus to 3D printed plastics is low. The plaque morphology of ΦX174 is also advantageous. The plaque forming units (pfu) of ΦX174 produce very large bull's eye-shaped plaques that are easily identifiable.

[0274] Preliminary work

[0275] To determine the validity of the experimental protocol, some preliminary experiments were performed. First, a pulse injection of ΦX174 was introduced into the JIB at the highest Dean number (i.e., high flow rate and low viscosity), which corresponds to the most chaotic gas flow caused by Dean vortices. The discharge from the JIB was then collected and statistically analyzed, which enabled the determination of the mass balance of the injection. The results showed that the recovered phage titer was within the typical error range of the titration method (i.e., (+ / -) 0.5 logs). The dead volume remaining in the outlet valve was sampled, and a dead volume of approximately 300 pfu / mL was persistent. A disinfection procedure was then created to thoroughly disinfect the injection valve, JIB, and outlet valve with 0.1 M NaOH for a contact time of ≥15 minutes. At the end of the disinfection cycle, no infectious particles remained in the outlet line.

[0276] Determine the minimum dwell time (ΦX174).

[0277] 3D printed JIBs with 0.32 and 0.64 cm inner diameter were tested using an Akta Explorer 100 from GE Healthcare (Uppsala, Sweden). To prepare for the experiment, a 30 mL aliquot of the mobile phase was taken and reserved for spiking. The sample was then spiked at 0.06% (v / v) with a target mobile phase concentration of 106.5 pfu / mL. The spike was intentionally made at a fairly low level to ensure that the fluid properties of the experimental injections were not altered by the ΦX174 spike. 25 mL of the spiked sample was loaded into a 50 mL capacity Superloop from GE Healthcare (Uppsala, Sweden) using a syringe, while the remainder was placed on the bench as a reserve sample to determine whether significant ΦX174 death was a function of mobile phase conditions and not related to flow through the JIB. ΦX174 never deviated from the target mobile phase concentration within the typical error of the titration assay (i.e., (+ / -) 0.5 logs). Finally, the empty fraction collection container was weighed to determine the tare weight.

[0278] To begin the experiment, Akta started the injection valve in the "Inject" position, pushing mobile phase into the Superloop, introducing ΦX174 spiking buffer into the JIB at 3% of the total reactor volume, with the effluent directed into a large volume container. The injection valve was then switched to the "Load" position, stopping the flow of mobile phase through the Superloop and redirecting it directly into the JIB for flush injection, with the effluent remaining in the same large volume container. After a predetermined time, the outlet valve switched to direct flow from the large volume container to the small volume container. Subsequently, the outlet valve was switched two more times, generating two more fractions. min3, 4, and 5σ, the time and volume of three small and one large fractions were determined using a modified peak analysis method. The outlet flow path for the three small volume fractions was a 1 mm capillary PEEK tube from GE Healthcare (Uppsala, Sweden). After the experiment was completed, the three small containers and one large container were weighed again. To sample the virus, the dead volume of the capillary was drained into a sterile tube with a sample volume of approximately 100uL. The sample remaining in the capillary will be the last drop of that fraction and can be considered an instantaneous grab sample. The entire volume of this grab sample is then titrated without dilution. Therefore, the question of "probability of detecting virus at low concentrations" outlined in ICH Q5A does not apply.

[0279] After all Akta post-experiment activities were completed, the remaining spiked mobile phase was drained from the Superloop and the Superloop was taken offline. The injection loop position was then replaced with a PEEK capillary and the above-mentioned sanitization steps were completed, followed by quenching with TSB.

[0280] result:

[0281] From preliminary results on bacteriophages, we found that the computational model discussed above provides a very conservative minimum residence time (T min ). The original peaks have been modified to account for this. Fig.18 A comparison of the peak generated by the Akta detector during dye pulse injection (i.e., original trace, dashed line) and the corrected peak (i.e., modified trace, solid line) is shown. The corrected trace is created by determining the maximum absorbance of the peak and mirroring the left trace to the right of the peak maximum. Fig.19 The calculated T min (5σ) as a function of the original trajectory and the modified version.

[0282] Fig. 20 Results from low viscosity phage experiments using TSB as the mobile phase are shown. The sampling strategy for these experiments was to min Instantaneous grab samples were collected during JIB discharge at 3, 4, and 5σ. For the three flow rates tested, 4σ and 5σ tested negative for viruses, while 3σ tested positive for viruses. From the equation, it is known that increasing viscosity will reduce the efficiency of the JIB. Since 4σ and 5σ are negative for viruses at low viscosities, the same volumetric sampling strategy was utilized, assuming that there would be enough volume space to capture the reduced efficacy. When viscosity was increased by adding glucose, viral breakthrough occurred faster for higher viscosities. However, as Fig.21 As shown, the 75 mL / min data point has missing (-) virus results. This is due to the sampling strategy and not the efficiency of the reactor or the calculations.

[0283] Example 3

[0284] Determining parameters to scale up the reactor five-fold

[0285] In one example, a user uses Figure 1C The number of reactors in series can be calculated (e.g. Figure 1D ) to provide a target residence time distribution for the process. In addition, the user can use the work done to calculate the size of a larger scale reactor. In this experiment, the Figure 2A In the system shown Figure 1C The flow path inner diameter (id) and radius of curvature of the reactor are shown. Figure 2A In the system shown, determine Figure 1C The inner diameter of the flow path in the reactor is 0.635 cm and the radius of curvature is 0.6825 cm. To obtain these parameters, a tracer pulse is injected into Figure 1C The results of these experiments are shown in Figure 2B After the pulse injection experiment is completed, the peak value generated is analyzed by fitting the discharge peak value with a Gaussian curve, such as Figure 2D This value is then converted to HETP (using Equation 1 below), as Figure 2E shown and plotted against the Dean number (using Equation 2 below).

[0286] (1)

[0287] (2)

[0288] Apply the best fit line to Figure 2E The data set is shown and expressed in Formula 3 below.

[0289] (3) HETP = f(De) = (aDe 3 +bDe 2 +cDe+d)

[0290] a, b, c, and d are based on fits to empirical data and are applicable to all Dean numbers.

[0291] Formulas 1 and 3 are then combined and rearranged to form Formula 4 below.

[0292] (4)

[0293] It is required that 99.99997% of the product stays in the reactor for ≥ 60 minutes (i.e. T min), so that n = 5 (in Formula 5 below). In addition, it is required that 99.865% of the product leaves the reactor for processing ≤ 90 minutes (i.e., T max ), so that m=3 (in Formula 6 below). The maximum kinematic viscosity allowed in the reactor is 1.5*10-6m2 / s. Based on this viscosity and the reactor dimensions mentioned above, to meet the requirement of Dean number ≥100, the flow rate in the reactor must be ≥65mL / min, and due to any process limitations, the flow rate must be ≤95mL / min.

[0294] (5)T Ave =T min +(n*σ max ), where n can be 5 or 60 minutes

[0295] (6)T Ave =T max -(m*(σ max )), where m can be 3 or 90 minutes

[0296] From the constraints above, a locus of solutions that satisfy the constraints can be generated. Four flow rates were solved as shown in the table below. The target flow rate, reactor design (i.e., the inner diameter and radius of curvature of the flow path), and the maximum kinematic viscosity calculate the worst-case Dean number. The output Dean number is then entered into Equation 3, and this value is entered into Equation 4. Using a guess-and-check approach, different reactor volumes are calculated using a step-by-step cascade of equations. The average residence time is then calculated by dividing the suggested reactor volume by the flow rate (see Equation 7 below), which is then divided by the cross-sectional area of ​​the inner diameter of the flow path to obtain the path length. The path length and average residence time are also entered into Equation 4 to obtain a value. Equations 5 and 6 are then used to solve for T min and T max .

[0297] (7)

[0298] Table 1 shows the min 60 minutes or T max The maximum and minimum reactor volumes for each flow rate are shown for the 90 min case. The specifications for these reactors are shown in Fig. 22 In, with Figure 5C Any flow rate and reactor volume selected between these two lines will satisfy the selected target residence time distribution (i.e., T min and T max ). Fig. 22 and Figure 5C The difference lies in the viscosity term used in the two figures. Fig. 22The higher viscosity of the reactor reduces the efficiency of the reactor and is suitable for faster flow rates and larger reactor volumes. Table 2 shows the decision for the reactor operating specifications. The flow rate was chosen relatively arbitrarily to keep the flow rate low. With the choice of flow rate, the reactor volume was chosen as the midpoint between the maximum and minimum reactor volumes.

[0299] Table 1 In T min 60 minutes or T max Maximum and minimum reactor volume solutions for each flow rate for the 90 min case

[0300]

[0301] Table 2. Determine the operating specifications of the reactor

[0302]

[0303] In addition, the design specifications of the reactor in terms of inner diameter, radius of curvature, flow rate and path length are determined to meet large-scale operations. In this example, the user requires the size to operate at 5 times the process volume flow rate (i.e., 350 mL / min) and requires that the ratio between the inner diameter and the radius of curvature be maintained constant. Figure 2E The data set shown is divided by the inner diameter of the reactor used to generate the data set (i.e., 0.635 cm), yielding a value similar to Fig. 10C The best fit line was then applied to the data set. Table 3 shows the known and unknown details of the new reactor. With the target flow rate and T in Table 2 Ave , the reactor volume can be calculated according to Formula 7. From the above constraints, a solution trajectory that satisfies the constraints can be generated. As shown in Table 4 and Fig.23 As shown, three inner diameters are solved. Any inner diameter selected along this plotted line will provide an appropriate residence time distribution. The final choice of inner diameter will be determined by a trade-off between residence time distribution and pressure drop in the reactor, with smaller inner diameters and longer path lengths resulting in greater pressure. Fixed flow rate and average residence time fix the volume of the reactor (Equation 7). The variable inner diameter is divided by the reactor volume to obtain the path length of the reactor. The output Dean number is then entered into Equation 8 below and multiplied by the inner diameter to obtain the HETP. The path length and average residence time are also entered into Equation 4 to obtain a value. Equations 5 and 6 are then used to solve for T min and T max As the inner diameter increases, the Dean number decreases, thereby reducing the efficiency of the reactor. Any inner diameter selected between 1.5-1.7 cm will provide an appropriate residence time distribution.

[0304]

[0305] Table 3 Known and unknown details of the new reactor

[0306]

[0307] Table 4 Derived inner diameter

[0308]

[0309] From the above description, it will be appreciated by those skilled in the art that this teaching can be realized in various forms. Therefore, although these teachings have been described in conjunction with their specific embodiments and examples, the true scope of this teaching should not be so limited. Different changes and modifications can be made without departing from the scope of the present invention.

[0310] The scope of this disclosure is to be interpreted broadly. This disclosure is intended to disclose equivalents, means, systems and methods for implementing the devices, activities and mechanical actions disclosed herein. For each device, article, method, means, mechanical element or mechanism disclosed, it is the purpose of this disclosure to also include and teach in its disclosure the equivalents, means, systems and methods for practicing the many aspects, mechanisms and devices disclosed herein. In addition, this disclosure relates to a coating and many aspects, features and elements thereof. Such a device can be dynamic in use and operation, and this disclosure is intended to include equivalents, means, systems and methods for using the device and / or manufacture, as well as many aspects that conform to the description and spirit of the operation and function of this disclosure. The claims of this application should also be interpreted broadly.

[0311] The description of many embodiments of the present invention is merely exemplary, therefore, changes that do not deviate from the gist of the present invention are within the scope of the present invention. These changes should not be considered as departing from the spirit and scope of the present invention.

Claims

1. A method for designing a practical reactor for virus inactivation, comprising: introducing a process stream including a detectable particle / tracer into a test reactor having a known radius of curvature and a known inner diameter, wherein the test reactor is in communication with at least one of a first detector and a second detector; detecting a flow rate of a process stream in a test reactor by at least one of a first detector and a second detector; detecting a fluid phase parameter of a process stream in a test reactor by at least one of a first detector and a second detector; detecting detectable particles exiting the experimental reactor via a second detector; determining an empirical value associated with at least one of an experimental reactor parameter and a fluid phase parameter based on an introduced process stream including detectable particles; determining a non-empirical value related to at least one of an experimental reactor parameter and a fluid phase parameter; as well as Designing the actual reactor based on the determined empirical values ​​and the determined non-empirical values, wherein designing the actual reactor comprises at least one of the following: scaling the size of the experimental reactor to an actual reactor having the same aspect ratio as the experimental reactor but a different internal diameter; scaling the size of the experimental reactor to an actual reactor having the same aspect ratio and the same internal diameter as the experimental reactor; scaling the size of the experimental reactor to an actual reactor having a different aspect ratio than the experimental reactor and a different diameter than the experimental reactor; scaling the size of the experimental reactor to an actual reactor having a different aspect ratio than the experimental reactor but the same diameter as the experimental reactor, wherein (1) When the actual reactor includes scaling the experimental reactor to have the same aspect ratio as the experimental reactor, the method requires the following equations (a)-(g) to derive the HETP, reactor volume, and internal diameter based on the average flow rate: (a) (b) HETP = f(v) = (av 3 +bv 2 +cv+d), where a, b, c, and d are based on empirical data for all Dean numbers (c) (d) (e) (f) (g) or (2) When the actual reactor includes scaling the dimensions of the experimental reactor to an actual reactor having the same aspect ratio as the experimental reactor, the method requires the use of the following equations (a)-(e) to derive the HETP, reactor volume, and internal diameter based on the reduced HETP and Dean number: (a) (b) (c)T Ave =T min +(n*σ max ) (d)T Ave =T max –(m*(σ max )) (e) or (3) When the actual reactor comprises scaling the experimental reactor to have the same aspect ratio and the same internal diameter as the experimental reactor, the method requires the derivation of at least the HETP and the path length using the following formula: (a) (b) a, b, c, and d are based on empirical data fits and are only valid when the Dean number is ≥ 100 (c)T Ave =Tmin+(n*σ time ); (d)T Ave =Tmax-(m*σ time ); (e) as well as (f) in Π = constant π, σ max = Maximum allowable standard deviation set by the user, σ time = standard deviation in time units, σ 2 time = variance in time units, A = cross-sectional area of ​​the flow path, CA = cross-sectional area of ​​the flow path, De = Dean number, The function f(……)=(……), h = reduced plate height, id=inner diameter, L = path length, L TAve = reactor length required to provide the target mean residence time, n = number of modifications, m = number of modifications, Q = flow rate, r = radius of the flow path, RV = reactor volume, T Ave = average residence time, Tmax = maximum residence time, Tmin = minimum residence time, v = speed.

2. The method of claim 1, wherein the detectable particle is at least one of a viral particle and a surrogate tracer. The method of claim 1 , wherein the empirical value and the non-empirical value are task dependent.

4. The method of claim 1 , wherein the experimental reactor parameters and fluid phase parameters corresponding to the empirical and non-empirical values ​​are theoretical or estimated minimum residence time, theoretical or estimated maximum residence time, inner diameter, volume flow rate, path length of the flow path, radius of curvature, density of the process stream, dynamic viscosity and variance.

5. The method of claim 1, wherein the empirical value is a value corresponding to at least one of an experimental reactor parameter and a fluid phase parameter that is linearly related to a set of experimental data, and wherein the non-empirical value corresponds to at least one of a theoretical or estimated minimum residence time, a theoretical or estimated maximum residence time, an inner diameter, a volume flow rate, a flow path length, and a radius of curvature.

6. The method according to claim 1, in, Detecting detectable particles exiting the experimental reactor also includes determining the minimum residence time experienced by one of the detectable particles that is equivalent to the incubation time for bulk viral inactivation, and the maximum residence time experienced by the last significant amount of detectable particles exiting the experimental reactor.

7. The method of claim 1, wherein the experimental reactor parameters include determining at least one of the following: a minimum residence time experienced by one of the detectable particles that is equivalent to an incubation time for bulk virus inactivation, a maximum residence time experienced by a significant number of detectable particles before they leave the experimental reactor, an inner diameter of the reaction tube, a volume flow rate, a length of the reaction tube, a radius of curvature, and a volume of the experimental reactor.

8. The method of claim 1, wherein the fluid phase parameter comprises at least one of density and dynamic viscosity.

9. The method of claim 1, wherein the experimental reactor comprises a reaction tube comprising at least one of: (i) a set of alternating turns forming a serpentine pattern between an inlet and an outlet and (ii) an interwoven path.

Images