Automated double emulsion droplet library generator

Abstract

Methods and systems for automated double emulsion droplet library generators are described herein. Optimizing the geometry and composition of double emulsion droplets is essential for maximizing their utility in various scientific and industrial applications. The microfluidic method is effective to generate a droplet library, which is useful to optimize droplets. However, monitoring and intervention by a human are necessary to generate droplet library. The disclosure provided herein presents an automated double-droplet library generator that utilizes a fine-tuned object detector and feedback control. The system autonomously produces libraries of single-core double emulsion droplets tailored to user-defined sizes, shell thicknesses, and solute concentrations.

Description

103241.007507 / 25-10932AUTOMATED DOUBLE EMULSION DROPLET LIBRARY GENERATORCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of and priority to U.S. Provisional Patent Application No. 63 / 708,032, “Automated Double Emulsion Droplet Library Generator” (filed October 16, 2024), the entirety of which application is hereby incorporated herein by reference for any and all purposes.TECHNICAL FIELD

[0002] The present disclosure relates to the field of process control and to the field of microfluidics.BACKGROUND

[0003] Double emulsions are a liquid dispersion system in which the inner phase droplets are dispersed again in the middle phase, resulting in core-shell structured liquid droplets. Thanks to their ability to encapsulate different materials within the core and shell, double emulsions have been widely utilized in science and industry.

[0004] To effectively apply these droplets in specialized fields, optimizing their geometry and composition is essential. For instance, the shell thickness of drug delivery capsules derived from double emulsions influences the rate and timing of drug release.

[0005] Among the methods for generating double emulsions, microfluidic droplet generation stands out as it produces monodisperse products. By adjusting flow rates, one can control the geometry and composition of the double emulsion droplets. Utilizing the droplet generator, it is effective to generate set of droplets with various properties called the droplet library, which is useful to find double emulsion droplet with optimized properties for a specific application.

[0006] But double emulsion generation is sensitive to disturbances that can destabilize the system. In one case, the middle phase surrounding the inner phase delaminates and the droplet generator fails to produce double emulsion droplets and impossible to be restored without an operator’s intervention. Accordingly, there is a long-felt need in the field for improved techniques for controlling microfluidic production, particular for controlling production of double emulsions.- 1 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932SUMMARY

[0007] Methods and systems for automated double emulsion droplet library generators are described herein. In one aspect, a method can include: verifying production of a predetermined amount of a first type of double emulsion droplets selected from a library of types of double emulsion droplets, the first type of double emulsion droplets corresponding to a first set of double emulsion droplet criteria; operating a double emulsion droplet generator so as to effect production of further double emulsion droplets, inputting image data of the further double emulsion droplets into a trained classification model; and based on output of the trained classification model, adjusting a parameter of the double emulsion droplet generator so as to effect production of a second type of double emulsion droplets selected from the library of types of double emulsion droplets, the second type of double emulsion droplets corresponding to a second set of double emulsion droplet criteria.BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In the drawings, which are not necessarily drawn to scale, like numerals may describe similar components in different views. Like numerals having different letter suffixes may represent different instances of similar components. The drawings illustrate generally, by way of example, but not by way of limitation, various aspects discussed in the present document. In the drawings:

[0009] FIG. 1. Automated double emulsion droplet library generator, (a) Schematic illustration of the automated double emulsion droplet library generator. The system includes a computer (1), three syringe pumps (2-4), a pressure controller with a reservoir (5), a high-speed camera attached to an optical microscope (6), a microfluidic device comprising a droplet generator and a micro mixer, a selector (8), and a rotating collector (9). (b) Functions of the automated droplet library generator program.

[0010] FIG. 2. Selective collection and recovery of single-core double emulsion (SCDE) mode, (a) Images of droplet generation and detection results during the SCDE recovery process, to indicates the moment when only the middle phase (MP) mode is detected. Temporal changes in (b) the outer diameter of double emulsion droplets and (c) the detected droplet generation mode during the SCDE recovery process.- 2 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932

[0011] FIG. 3. 5*5 double emulsion droplet library of the shell thickness and the solute concentration, (a) Programmed flow rates to generate 5*5 double emulsion droplet library of the concentration and the shell thickness, (b) Temporal variation of the outer phase pressure. The "collect" and "waste" data represent instances when droplets are selectively collected and discarded, respectively, (c) A set of optical micrographs of double emulsion droplets from the 5x5 droplet library, illustrating variations in shell thickness and dye concentration. Scale bar: 100 pm. (d) The outer diameter and (e) the thickness of double emulsion droplets in the library as a function of the flow rate ratio of the middle phase to the inner phase.

[0012] FIG. 4. Automated double emulsion droplet library generator for size-controlled microcapsules production. A schematic illustration of the automated double emulsion droplet library generator, consisting of: (1) a user interface program on a computer (2); (3 and 4) two syringe pumps for the inner phase (IP) and middle phase (MP); (5) a pressure controller and reservoir for the outer phase (OP); (6) a high-speed camera attached to an optical microscope; (7) a microfluidic double emulsion generator; (8) a “Everything is done under water” (EDW) collector.

[0013] FIG. 5. “Everything is done under” (EDW) collector, (a) A schematic showing the components of the EDW collector, (b) Arrangement of an outlet tubing and the EDW collector to i) collect and ii) reject droplet generation products, (c) five different PLGA double emulsions collected separately on the vessels of the EDW collector, (d) Initiation of the solvent evaporation process by lowering the level of the collection fluid.

[0014] FIG. 6. Automated generation and collection of PLGA double emulsions resulting microcapsules, (a) Flow rates of the inner (QIP) and middle (QMP) phases during the production of PLGA double emulsions, (b) Position of the EDW collector between the collection (Clt) and rejection (Rjt). (c) temporal change outer phase pressure (POP) for adjusting outer diameter (DO) of double emulsions to the target specifications, (d) temporal variation of the outer diameters (DO) and coefficient of variation (CV) of PLGA double emulsions during automated production. The pale green dotted lines represent target outer diameters, (e) Optical micrographs of five distinct PLGA microcapsules varying consolidated PLGA shell thicknesses (tshelT). The scale bar is 100 pm.- 3 -103241.00750714908-6808-9715.1103241.007507 / 25-10932

[0015] FIG. 7. Automated droplet library generator (ADLib) for tailored microcapsule fabrication, (a) A schematic illustration of the ADLib, consisting of: (1) a personal computer; (2 and 3) two syringe pumps for the inner (blue) and middle phase (red); (4) a pressure controller and reservoir for the outer phase (light blue) ; (5 and 6) a high-speed camera attached to an optical microscope; (7) a microfluidic double emulsion generator; (8) a dip collector for in-liquid droplet collection, (b) Workflow of the ADLib for tailoring microcapsules as drug delivery system.

[0016] FIG. 8. Graphical user interface (GUI) of the automated droplet library generator for tailoring microcapsules (ADLib for T-MCs). (a) In the Design tab of the ADLib for T- MCs program, users can specify the geometrical parameters of double emulsion droplets, such as outer or inner diameters and middle-phase shell thickness, as well as postprocessing options including osmotic annealing and solvent evaporation, (b) The program visualizes the predicted characteristics of double emulsion droplets and resulting microcapsules in a pop-up window based on the user’s input. Example microcapsule designs generated using the GUI with (i) solvent evaporation only and (ii) combined osmotic annealing and solvent evaporation options.

[0017] FIG. 9. Dip collector, (a and b) Comparison of emulsion collection methods, (a) Dropwise collection: double emulsions exit from the outlet tip into air and fall into the collection vessel, resulting in defective single emulsions and double emulsions with thin shells (red arrows and circles), (b) Dip collection: the outlet is submerged in collection fluid, significantly reducing the number of defective droplets, (c and d) Schematic and photography of the dip collector setup, (e) Selective collection using the dip collector. In the ‘collect’ position (i), monodisperse single-core double emulsions are captured in a cell strainer. In the ‘reject’ position (iii, iv), defective droplets are discarded at the bottom. In this way, a module can move between (1) a first state or location to dispense conforming droplets for collection and (2) a second state or location to dispense non-conforming droplets for discarding.

[0018] FIG. 10. Droplet post-processing using a dip collector, (a) Regulation of solvent evaporation rate by adjusting the height of the collection fluid, (b) Transition of double emulsion droplets into microcapsules via i) solvent evaporation only (DE ■=> SE) in isotonic collection fluid, and ii) osmotic annealing followed by solvent evaporation (DE ■=> OA ■=> SE) in hypotonic collection fluids. Osmotic annealing induces additional change in- 4 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932 droplet geometry, (c) Transformation of double emulsion droplets (top raw) with varying middle-phase thickness (zW = 3, 5, 7, 9, and 11 pm) into microcapsules via osmotic annealing (middle raw) and solvent evaporation (bottom raw), (d) Inner diameter (Dr), middle-phase thickness (zw), and average shell thickness (ts,avg) of double emulsion droplets (DE) and microcapsules (MC) as functions of the flow rate ratio of the middle phase to the inner phase QMP / QIP .

[0019] FIG. 11. Automated production of double emulsion for tailored PLGA microcapsule fabrication, (a-c) Temporal changes of key parameters during automated double emulsion generation; (a) flow rates (0), (b) outer diameters (Do) of double emulsion droplets and pressure of the outer phase POP , and (c) droplet generation modes identified by the automated droplet library generator (ADLib). (d) Automatically generated double emulsions varying shell thicknesses with the constant inner diameter of 54 pm. The scale bar is 100 pm (e) Failure detection and double emulsion generation recovery process by ADLib. The width of micrographs is 640 pm.

[0020] FIG. 12. Characterization of tailored PLGA microcapsules by the automated droplet library generator (ADLib). (a) optical micrographs of microcapsules varying consolidated PLGA shell thickness, zw is middle phase layer thickness of double emulsion before a solvent evaporation. ts,avgis the average PLGA membrane thickness microcapsules. The scalebars represents 200 pm (b) Confocal micrographs of microcapsules encapsulating model drugs (yellow; Evans blue) with PLGA shell (red; Nile red). The scale bar is 50 pm. (c) Scanning electron microscopy (SEM) images of freeze- dried microcapsules. The scale bar is 100 pm. (d) Normalized red fluorescence intensity over PLGA shell of the microcapsules. The intensity is measured on a line through the center of the microcapsules as shown in the inset. The blank circles indicate the peak of the intensity, r is an inner radius of microcapsules. Scalebar of the inset is 30 pm. (e) 3D reconstructed confocal micrograph of the PLGA shell with ts,avg= 830 pm and 130 pm, respectively. The insets depict SEM images of cross-sections of the thinnest parts of the shell. The scalebar represents 200 nm.

[0021] FIG. 13. Drug release of tailored PLGA microcapsules, (a) Schematic illustration drug release from microcapsules with varying PLGA shell thickness under hypotonic physiological conditions, (b) Cumulative release profile of encapsulant from microcapsules triggered by osmotic stress. ts, avg is the average PLGA membrane thickness- 5 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932 microcapsules. Error bars represent standard deviations, and solid lines indicate multiple exponential fits to the experimental data, (c) Time-lapse optical micrographs of microcapsules during incubation in physiological conditions. The time labels represent an incubation duration, (d) Temporal changes in microcapsule diameter (DIDo) and relative dye intensity for samples with different ts,avg. The centers and perimeters of ovals represent the mean and standard deviation diameter and intensity, respectively, while dots correspond to individual microcapsules.DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS

[0022] The present disclosure may be understood more readily by reference to the following detailed description of desired embodiments and the examples included therein.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. In case of conflict, the present document, including definitions, will control. Preferred methods and materials are described below, although methods and materials similar or equivalent to those described herein can be used in practice or testing. All publications, patent applications, patents and other references mentioned herein are incorporated by reference in their entirety. The materials, methods, and examples disclosed herein are illustrative only and not intended to be limiting.

[0024] The singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise.

[0025] As used in the specification and in the claims, the term "comprising" can include the embodiments "consisting of' and "consisting essentially of.” The terms “comprise(s),” “include(s),” “having,” “has,” “can,” “contain(s),” and variants thereof, as used herein, are intended to be open-ended transitional phrases, terms, or words that require the presence of the named ingredients / steps and permit the presence of other ingredients / steps.However, such description should be construed as also describing compositions or processes as "consisting of' and "consisting essentially of' the enumerated ingredients / steps, which allows the presence of only the named ingredients / steps, along with any impurities that might result therefrom, and excludes other ingredients / steps.

[0026] As used herein, the terms “about” and “at or about” mean that the amount or value in question can be the value designated some other value approximately or about the- 6 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932 same. It is generally understood, as used herein, that it is the nominal value indicated ±10% variation unless otherwise indicated or inferred. The term is intended to convey that similar values promote equivalent results or effects recited in the claims. That is, it is understood that amounts, sizes, formulations, parameters, and other quantities and characteristics are not and need not be exact, but can be approximate and / or larger or smaller, as desired, reflecting tolerances, conversion factors, rounding off, measurement error and the like, and other factors known to those of skill in the art. In general, an amount, size, formulation, parameter or other quantity or characteristic is “about” or “approximate” whether or not expressly stated to be such. It is understood that where “about” is used before a quantitative value, the parameter also includes the specific quantitative value itself, unless specifically stated otherwise.

[0027] Unless indicated to the contrary, the numerical values should be understood to include numerical values which are the same when reduced to the same number of significant FIGs. and numerical values which differ from the stated value by less than the experimental error of conventional measurement technique of the type described in the present application to determine the value.

[0028] All ranges disclosed herein are inclusive of the recited endpoint and independently of the endpoints. The endpoints of the ranges and any values disclosed herein are not limited to the precise range or value; they are sufficiently imprecise to include values approximating these ranges and / or values.

[0029] As used herein, approximating language can be applied to modify any quantitative representation that can vary without resulting in a change in the basic function to which it is related. Accordingly, a value modified by a term or terms, such as “about” and “substantially,” may not be limited to the precise value specified, in some cases. In at least some instances, the approximating language can correspond to the precision of an instrument for measuring the value. The modifier “about” should also be considered as disclosing the range defined by the absolute values of the two endpoints. For example, the expression “from about 2 to about 4” also discloses the range “from 2 to 4.” The term “about” can refer to plus or minus 10% of the indicated number. For example, “about 10%” can indicate a range of 9% to 11%, and “about 1” can mean from 0.9-1.1. Other meanings of “about” can be apparent from the context, such as rounding off, so, for example “about 1” can also mean from 0.5 to 1.4.- 7 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932

[0030] Further, the term “comprising” should be understood as having its open-ended meaning of “including,” but the term also includes the closed meaning of the term “consisting.” For example, a composition that comprises components A and B can be a composition that includes A, B, and other components, but can also be a composition made of A and B only. Any documents cited herein are incorporated by reference in their entireties for any and all purposes.

[0031] Any embodiment or aspect provided herein is illustrative only and does not limit the scope of the present disclosure or the appended claims. Any part or parts of any one or more embodiments or aspects can be combined with any part or parts of any one or more other embodiments or aspects.

[0032] Methods and systems for automated double emulsion droplet library generators are described herein. The disclosure provided herein discusses an automated doubledroplet library generator. The generator can accept user-defined properties for the desired double emulsion droplets and generates a library of 25 distinct types of double droplets (FIG. la). The droplet generation mode and the size of droplets can be monitored in realtime using a fine-tunned object detector (FIG. lb i). Based on detection results, the library generator can selectively collect monodispersed single-core double emulsion droplets (FIG. lb ii). If the generator fails to produce double emulsions, the generator can automatically restore the generation mode, allowing production to continue autonomously. The library generator can also regulate droplet size to a targeted value through feedback control (FIG. lb iv). Droplets can be collected in microtubes attached to a rotating collector, and by sequentially altering the flow rate and microtubes, a library of 25 different properties can be created (FIG. lb v). This automated generator may enhance research and industrial applications involving double emulsion droplets.Materials and Methods

[0033] The automated droplet library generator can include three syringe pumps corresponding to the inner phases 1 and 2(IP1 and IP2), and the middle phase (MP), respectively, a presser controller for driving the outer phase (OP), a droplet generator, a micro mixer, a high-speed camera attached to an optical microscope, a selector, and a rotating collector (FIG. 1). The droplet generator can be fabricated using glass capillaries. HFE 7500 with 2% krytox FSH 157 surfactant can be used as MP. IP1 and OP is a 2% PVA aqueous solution., while 5 mg / mL of Evan's blue can be added to the PVA solution- 8 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932 for IP2 to visualize the concentration difference of the inner phase. The two inner phase solutions can be thoroughly mixed before being injected into the droplet generator. To monitor the droplet generation, a fine-tuned YOLOvlOn object detector can be used. Results and Discussion

[0034] A soft sensor, including the highspeed camera and objective detector, can detect the droplet generation mode and can approximate droplet size with a latency of around 170 ms. Based on the object detector's results, the library generator can collect products when the detected mode is a single-core double emulsion (SCDE), the approximated droplet size is within 10% tolerance of the previous size, the coefficient of variation (CV) is below a critical value, and the previous position of the selector is set to collection (FIG. 2a i, vi, and b). Otherwise, the product can be directed to a waste chamber. To maintain SCDE production, the library generator can recover from a failure mode, referred to as MP mode, where the library generator is unable to produce double emulsion. The recovery step can be initiated upon detection of MP mode (FIG. 2a ii, iii, and c). Once double emulsion generation resumes, the library generator can wait for the droplet generation to settle down and defects, such as multicore droplets double emulsion droplets outside the target size, to exit the droplet generator (FIG. 2a iv and v, and b). The library generator can then resume collection (FIG. 2a v).

[0035] To generate a double droplet library with varying solute concentrations in the core and shell thickness, the library generator can calculate a matrix of flow rates for IP 1, IP2, and MP based on user inputs regarding target size, shell thickness, solute concentration range, and dispersed phase flow rates (FIG. 3a). The generator can automatically adjust the outer phase pressure (POP) to regulate droplet diameter to the target size, which in some cases can be independent of the flowrate ratio and composition (FIG. 3b). Consequently, a 5x5 double emulsion droplet library can be produced, varying in shell thickness and core solute concentration (FIG. 3c). The outer diameter of the droplets can be controlled to the target size of 53 pm, with the shell thickness closely matching input values of 1.5, 2, 3, 4, and 5 pm (FIG. 3d and e). In conclusion, the library generator can produce a droplet library on user’s demand while precisely controlling the droplet geometry and composition.ADLib 2: automated production of size-controlled biodegradable microcapsules- 9 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932

[0036] Uniformity of microcapsules (MCs) is crucial for drug delivery applications, ensuring precise drug transport, consistent release profiles, and controlled simultaneous activation. Double emulsions serve as ideal templates for monodisperse MCs; however, their generation process is highly sensitive, often leading to unstable droplet generation producing defects and even leading to break down microfluidic devices. Moreover, the slow production rate of droplet generators (0.1-1 mL / hr for dispersed phase in general) necessitates hours of manual supervision to obtain sufficient sample quantities. Controlling the droplet characteristics — such as core size and shell thickness — requires expertise in fluid dynamics and experiences posing a significant barrier to entry.

[0037] To address these challenges, we previously introduced Automated droplet library (ADLib) system, an Al-driven automated droplet library generator that integrates object detection, decision-making algorithms, and handling systems to autonomously monitor, assess, and control droplet generation. In this study, we upgraded the ADLib system to automate the generation of poly(lactic-co-glycolic acid) (PLGA) microcapsules (FIG. 4). The newly developed graphical user interface (GUI) allows users to specify target properties, including core size (DI) and up to five consolidated PLGA shell thicknesses (tshelT), as well as other essential parameters such as PLGA concentration and the dispersed phase flow rate. The system then determines the droplet dimensions and calculates the required flow rate matrix (QIP and QMP) for their formation.

[0038] A YOLOvlOn object detection model, fine-tuned on a custom dataset, monitors W / O / W double emulsion formation in real time. The model classifies five droplet generation modes — single-core double emulsions, multi-core droplets, jetting, satellite droplets, and intermediate states — while rapidly estimating droplet sizes. Based on the detection result, the decision-making algorithm adjusts the outer phase pressure to ensure the production of single-core double emulsions in target specification. If the failure mode occurs, the recovery protocol restores the single-core double emulsion generation mode within a few seconds. Following the algorithm, the ADLib system selectively collects or rejects droplets.

[0039] It requires minimizing air exposure and impact during collection process to maintain the core-shell structure and uniformity of PLGA double emulsions. The newly designed “Everything is done under water” (EDW) collector enables all collection processes to occur within a collection fluidic environment (FIG. 5). The collector consists- 10 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932 of a glass dish, five cell strainers as collection vessels, and a 3D- printed cell strainer holder. A stepper motor adjusts the collector’s angle to collect different droplets in separated collection vessels while rejecting defects. The five different PLGA droplet types are stored in the cell strainers, while defects pass through openings in the holder and are discarded onto the bottom of the glass dish.

[0040] After collection, the continuous phase level is lowered to activate solvent (here, CHCU is used) evaporation, consolidating PLGA. During solvent evaporation, some double emulsions rupture, forming single PLGA droplets that transition into PLGA particles. Since these single particles are generally smaller than half the size of PLGA MCs, they pass through the cell strainer mesh, which is a bit smaller than the diameter of MCs (i.e. DMCs is ~ 48 pm while the mesh size of the cell strainer is 40 pm).

[0041] The resulting five PLGA MCs exhibit uniform core sizes and distinct shell thicknesses from 320 to 1370 nm. These size-controlled PLGA MCs demonstrate varying release profiles based on natural degradation and responses to activation stimuli such as pressure, shear forces, and osmotic pressure, making them highly valuable for drug delivery system design. The user-friendly GUI and automated generation process eliminate the need for prior experience or fluid dynamics expertise, enabling seamless operation of microfluidic devices and broadening ADLib’s applicability across various fields.ADLib 2 Part 2: automated production of size-controlled biodegradable microcapsules

[0042] Uniformity of microcapsules (MCs) is crucial for drug delivery applications, ensuring precise drug transport, consistent release profiles, and controlled simultaneous activation. Double emulsions serve as ideal templates for monodisperse MCs; however, their generation process is highly sensitive, often leading to unstable droplet generation producing defects and even leading to break down microfluidic devices. Moreover, the slow production rate of droplet generators (0.1-1 mL / hr for dispersed phase in general) necessitates hours of manual supervision to obtain sufficient sample quantities. Controlling the droplet characteristics — such as core size and shell thickness — requires expertise in fluid dynamics and experiences posing a significant barrier to entry.

[0043] To address these challenges, we previously introduced Automated droplet library (ADLib) system, an Al-driven automated droplet library generator that integrates object detection, decision-making algorithms, and handling systems to autonomously monitor,- 11 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932 assess, and control droplet generation. In this study, we upgraded the ADLib system to automate the generation of biodegradable poly(lactic-co-gly colic acid) (PLGA) microcapsules (FIG. 7). The newly developed graphical user interface (GUI) allows users to design microcapsules for tailoring their characteristics such as size and PLGA membrane thickness, customizing drug release profile (FIG. 8).

[0044] A YOLOvlOn object detection model, fine-tuned on a custom dataset, monitors W / O / W double emulsion formation in real time. The model classifies five droplet generation modes — single-core double emulsions, multi-core droplets, jetting, satellite droplets, and intermediate states — while rapidly estimating droplet sizes. Based on the detection result, the decision-making algorithm adjusts the outer phase pressure to ensure the production of single-core double emulsions in target specification. If the failure mode occurs, the recovery protocol restores the single-core double emulsion generation mode within a few seconds. Following the algorithm, the ADLib system selectively collects or rejects droplets.

[0045] The Dip collector, a submerged droplet handing system, enables in-liquid droplet collection without air exposure and defect formation, and controlled environment for reproducible post processing of droplets (FIG. 9). By dynamically controlling the fluid level and osmotic conditions, this collector ensures that droplets experience uniform masstransfer and solvent-removal environments throughout post-processing. As a result, the osmotic annealing and solvent evaporation steps proceed under consistent and reproducible conditions, leading to highly uniform and reproducible microcapsule structures (FIG. 10).

[0046] The automated system generated five monodisperse PLGA double emulsions with a fixed core diameter of ~54 pm and variable middle-phase thicknesses of 2-10 pm (FIG. 11). Using the Dip collector, the droplets underwent controlled osmotic annealing followed by solvent evaporation, forming PLGA microcapsules with average shell thicknesses of 130, 280, 440, 630, and 830 nm, respectively (FIG. 11).Optical and confocal microscopy confirmed structural uniformity and stable encapsulation of model drugs, while SEM analysis showed smooth consolidated shells (FIG. 12). Osmotic stress-triggered drug release tests revealed systematically tunable release kinetics — 24-hour cumulative release ranging from 98.8% to 67.7% depending on shell thickness(FIG. 13). Importantly, the consistency of osmotic annealing and evaporation- 12 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932 within the Dip collector substantially reduced inter-sample variation, validating the improved reproducibility of microcapsule fabrication.

[0047] The ADLib 2 platform achieves fully automated, reproducible, and userindependent fabrication of tailored PLGA microcapsules. By coupling Al-driven droplet generation with uniform in-liquid post-processing in the Dip collector, it resolves key limitations of manual microfluidic systems and provides a standardized route to reproducible encapsulation and release control. This technology offers broad applicability for precision drug delivery, bioactive material encapsulation, and programmable microreactor fabrication, where control and reproducibility at the microscale are critical. Exemplary Embodiments

[0048] The following embodiments are exemplary only and do not serve to limit the scope of the present disclosure of the appended claims. It should be understood that any part of any one or more Embodiments can be combined with any part of any other one or more Embodiments.Embodiment 1

[0049] A method, comprising: receiving image data related to droplets output from a double emulsion droplet generator; inputting the image data into a trained classification model; collecting those droplets whose image data satisfies a set of double emulsion droplet criteria corresponding to a double emulsion droplet type of a plurality of double emulsion droplet types of a droplet library generator.Embodiment 2

[0050] The method of Embodiment 1, further comprising diverting from collection those droplets whose image data do not satisfy the set of double emulsion droplet criteria.Embodiment 3

[0051] The method of any one of Embodiments 1-2, further comprising segregating droplets whose image data satisfies a first set of double emulsion droplet criteria corresponding to a double emulsion droplet type from droplets whose image data satisfies a second set of double emulsion droplet criteria corresponding to a double emulsion droplet type.Embodiment 4- 13 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932

[0052] The method of any one of Embodiments 1-3, further comprising receiving input corresponding to one or more double emulsion droplet types, the input optionally being from a library of inputs.Embodiment 5

[0053] The method of Embodiment 4, wherein the input comprises a set of double emulsion droplet criteria corresponding to a double emulsion droplet type.Embodiment 6

[0054] The method of any one of Embodiments 1-5, wherein a criterion comprises any one or more of shell thickness, droplet contents concentration, or cross-sectional dimension.Embodiment 7

[0055] The method of any one of Embodiments 1-6, further comprising adjusting a parameter of the double emulsion droplet generator in response to image data. Embodiment 8

[0056] The method of any one of Embodiments 1-7, wherein the trained classification model comprises an object detection model.Embodiment 9

[0057] A method, comprising: verifying production of a predetermined amount of a first type of double emulsion droplets selected from a library of types of double emulsion droplets, the first type of double emulsion droplets corresponding to a first set of double emulsion droplet criteria; operating a double emulsion droplet generator so as to effect production of further double emulsion droplets, inputting image data of the further double emulsion droplets into a trained classification model; and based on output of the trained classification model, adjusting a parameter of the double emulsion droplet generator so as to effect production of a second type of double emulsion droplets selected from the library of types of double emulsion droplets, the second type of double emulsion droplets corresponding to a second set of double emulsion droplet criteria.Embodiment 10

[0058] The method of Embodiment 9, further comprising: segregating droplets satisfying the first set of double emulsion droplet criteria from droplets satisfying the second set of double emulsion droplet criteria.Embodiment 11- 14 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932

[0059] The method of any of Embodiments 9 and 10, further comprising: determining a subset of droplets produced by the double emulsion droplet generator fail to satisfy the second set of double emulsion droplet criteria subsequent to the adjusting of the parameter; and refraining from collecting the subset of droplets based on the determining. Embodiment 12

[0060] The method of Embodiment 11, further comprising: based on the determining, readjusting the parameter, or adjusting a second parameter, of the double emulsion droplet generator so as to effect production of the second type of double emulsion droplets. Embodiment 13

[0061] The method of any of Embodiments 9-12, further comprising receiving input corresponding to one or more double emulsion droplet types.Embodiment 14

[0062] The method of Embodiment 13, wherein the input comprises a set of double emulsion droplet criteria corresponding to a double emulsion droplet type.Embodiment 15

[0063] The method of any one of Embodiments 9-14, wherein a criterion comprises any one or more of shell thickness, droplet contents concentration, or cross-sectional dimension.Embodiment 16

[0064] The method of any one of Embodiments 9-15, wherein the parameter comprises any one or more of a flow rate, a pressure, a capillary number, a Weber number, or an amount of a surfactant.Embodiment 17

[0065] The method of any one of Embodiments 9-16, wherein the trained classification model comprises a object detection model.Embodiment 18

[0066] The method of any one of Embodiments 9-17, wherein a criterion comprises a characteristic of an inner phase of a double emulsion droplet, a middle phase of a double emulsion droplet, or an outer phase of a double emulsion droplet.Embodiment 19

[0067] The method of any one of Embodiments 9-18, wherein verifying production of the predetermined amount of the first type of double emulsion droplets by monitoring a- 15 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932 flow rate of the emulsion droplet generator and a production time for the first type of double emulsion droplet.Embodiment 20

[0068] A system configured to perform the method of any of Embodiments 9-19, the system comprising: an imager configured to collect the image data of an output of the droplet generator; the trained classification model configured to classify the image data; and a controller configured to adjust the parameter of the droplet generator.- 16 -103241.007507\4908-6808-9715.1

Claims

103241.007507 / 25-10932What is Claimed:

1. A method, comprising: receiving image data related to droplets output from a double emulsion droplet generator; inputting the image data into a trained classification model; collecting those droplets whose image data satisfies a set of double emulsion droplet criteria corresponding to a double emulsion droplet type of a plurality of double emulsion droplet types of a droplet library generator.

2. The method of claim 1, further comprising diverting from collection those droplets whose image data do not satisfy the set of double emulsion droplet criteria.

3. The method of any one of claims 1-2, further comprising segregating droplets whose image data satisfies a first set of double emulsion droplet criteria corresponding to a double emulsion droplet type from droplets whose image data satisfies a second set of double emulsion droplet criteria corresponding to a double emulsion droplet type.

4. The method of any one of claims 1-2, further comprising receiving input corresponding to one or more double emulsion droplet types, the input optionally being from a library of inputs.

5. The method of claim 4, wherein the input comprises a set of double emulsion droplet criteria corresponding to a double emulsion droplet type.

6. The method of any one of claims 1-2, wherein a criterion comprises any one or more of shell thickness, droplet contents concentration, or cross-sectional dimension.

7. The method of any one of claims 1-2, further comprising adjusting a parameter of the double emulsion droplet generator in response to image data.- 17 -103241.007507\4908-6808-9715.1103241.007507 / 25-109328. The method of any one of claims 1-2, wherein the trained classification model comprises an object detection model.

9. A method, comprising: verifying production of a predetermined amount of a first type of double emulsion droplets selected from a library of types of double emulsion droplets, the first type of double emulsion droplets corresponding to a first set of double emulsion droplet criteria; operating a double emulsion droplet generator so as to effect production of further double emulsion droplets, inputting image data of the further double emulsion droplets into a trained classification model; and based on output of the trained classification model, adjusting a parameter of the double emulsion droplet generator so as to effect production of a second type of double emulsion droplets selected from the library of types of double emulsion droplets, the second type of double emulsion droplets corresponding to a second set of double emulsion droplet criteria.

10. The method of claim 9, further comprising: segregating droplets satisfying the first set of double emulsion droplet criteria from droplets satisfying the second set of double emulsion droplet criteria.

11. The method of any of claims 9-10, further comprising: determining a subset of droplets produced by the double emulsion droplet generator fail to satisfy the second set of double emulsion droplet criteria subsequent to the adjusting of the parameter; and refraining from collecting the subset of droplets based on the determining.- 18 -103241.00750714908-6808-9715.1103241.007507 / 25-1093212. The method of claim 11, further comprising: based on the determining, readjusting the parameter, or adjusting a second parameter, of the double emulsion droplet generator so as to effect production of the second type of double emulsion droplets.

13. The method of any of claims 9-10, further comprising receiving input corresponding to one or more double emulsion droplet types.

14. The method of claim 13, wherein the input comprises a set of double emulsion droplet criteria corresponding to a double emulsion droplet type.

15. The method of any one of claims 9-10, wherein a criterion comprises any one or more of shell thickness, droplet contents concentration, or cross-sectional dimension.

16. The method of any one of claims 9-10, wherein the parameter comprises any one or more of a flow rate, a pressure, a capillary number, a Weber number, or an amount of a surfactant.

17. The method of any one of claims 9-10, wherein the trained classification model comprises a object detection model.

18. The method of any one of claims 9-10, wherein a criterion comprises a characteristic of an inner phase of a double emulsion droplet, a middle phase of a double emulsion droplet, or an outer phase of a double emulsion droplet.

19. The method of any one of claims 9-10, wherein verifying production of the predetermined amount of the first type of double emulsion droplets by monitoring a flow rate of the emulsion droplet generator and a production time for the first type of double emulsion droplet.

20. A system configured to perform the method of any of claims 9-10, the system comprising: an imager configured to collect the image data of an output of the droplet generator;- 19 -103241.007507\4908-6808-9715.1103241.007507 / 25-10932 the trained classification model configured to classify the image data; and a controller configured to adjust the parameter of the droplet generator.- 20 -103241.007507\4908-6808-9715.1

Images