In this report you get to read about the goals, problems and end result, how the group took on the different areas and what boundaries we had. You will get an understanding of our background and what research we had to do.Furthermore the implementation of Artificial Intelligence combined with the optical spectrum and what obstacles we faced during the project. Finally you will read about our results, the improvement points we had on the project and what we would have done differently if we redid the project. How we succeeded in transforming the script and what could have been done to finish the AI. The group has worked together with Uppsala University and Peafowl Plasmonics to transform, optimize and implement Artificial Intelligence to regulate a pump to a microfluidic reactor. The microfluidic reactor is automated by a genetic algorithm which is the core of the project. The main goals of the project was to transform the code from 32-bit to 64-bit, fit the optical spectrum and implement AI to the GA. To reach the goals the project was divided into different stages; data collection, implementation, verification and presentation. The implementation started with the transformation which was successful when all the correct files were given, and the code could run in 64-bit in the laboratory in Ångström. Furthermore, when the TOOLBOX for the optical spectrum in MATLAB were working for both spherical and nonspherical particles, this part of the project was concluded. The last part about the AI implementation was not successful due to time and problems with lack of knowledge about the subject of AI. However, a conclusion could be drawn about the optimal way to implement the AI to the GA, where the strategy was to first run the GA to collect data, and then run the GA and artificial Neural Network in parallel.

This project encompasses the implementation of a computational model to simulate the microfluidic separation of like-charged particles in a continuous flow environment. By accomplishing this task the model can be used to optimize future fractionations by tailoring the process parameters to the properties of the target particles. The primary goal of this project is to develop a vectorized code within Matlab® that captures a sufficient quantity of the physics in separations to assist with the optimization and design of microfluidic systems. This project differs from other computational models in that it utilizes a personal computer to run the simulation in an optimized format rather than utilizing a highly parallelized system for the computing. Based on previous literature from computational models of fluid-particle systems a model was developed to simulate the separation process. Computational experiments of separation processes were conducted with this model to validate the simulation and to investigate the impacts of microfluidic fractionation parameters on the purity and yield of like charged particles in a continuous flow environment. By adapting the input parameters the separation results can be customized for the particles in the sample. The implementation and use of this this model can improve the efficiency of separation processes.

This project encompasses the design and development of a pretreatment microfluidic device for samples of physiological conductivity, namely a saline solution. The conductivity was reduced through the combination of dilution and ion removal using electric fields to enable downstream electro kinetic manipulations. The two major parts of this project include (1) designing a pretreatment protocol to reduce the conductivity of the sample solution to an acceptable level and (2) designing /fabricating a microchip that will effectively allow aim (1) to be performed on chip. This project is one of the first to observe the effects of an electric field, used in the application of ion removal, to reduce the conductivity of a sample. Through the combination of sample and low conductivity buffer, as well as the presence of an electric field, a conductivity pretreatment chip is created. Since biomarkers and analytes of interest are difficult to detect in complex raw samples, such as blood, this chip is a necessary preliminary step that allows for successive separations. Using previous literature from the field of capillary electrophoresis, a design and pretreatment protocol was developed to pretreat a sample into a target conductivity range. A PMMA device was fabricated using a laser photoablation system located on the Cal Poly campus. Off-chip electrodes were used to induce electrophoretic movement of ions across a membrane and out of the sample. The combination of dilution and electrical fields yielded samples that had their conductivity reduced 80%. Dilution was found to be more effective in a chip designed with a short process time and continuous flow. Ultimately, we wish to incorporate this device with other pre-fabricated pretreatment and electrokinetic devices to optimize certain bioseparations.

Colorectal cancer is the third most common cancer in the United States with a 5-year late-stage survival rate of only 14%. Due to the lack of translation between animal models and clinical trials as well as the inefficacy of many chemotherapeutics in initial clinical trials, researchers are turning to in vitro drug screening models in an effort to mimic the conditions in vivo. This research project aimed to validate an in vitro tumor culture model within a microfluidic device using a clinically relevant chemotherapy drug. The first experiment consisted of a cell density and drug concentration study to determine the correct cell density and oxaliplatin concentration combinations that would result in a spectrum of quantifiable effects on the tumor cells. This experiment was then converted from a monolayer cell culture on glass into a 2D culture on top of a fibrin extracellular matrix (ECM) to ensure that the cells would respond in a similar way to the drug in the presence of an ECM as they did in the first experiment. The third experiment involved SW620 cells cultured within the fibrin hydrogel to create a 3D tumor model that better mimics the growing conditions in vivo. The goal of this experiment was again to ensure that the cells would respond in the same way to the oxaliplatin treatments as the previous experiments when adding complexity to the model. The final experiment was then to convert this 3D experiment performed in chamber slides into a 3D culture within a microfluidic device with media and oxaliplatin treatments perfused through the chamber using a syringe pump. The purpose of this experiment was to assess whether tumor cells could grow and survive within a microfluidic device with interstitial flow as well as determining if they responded as expected to the oxaliplatin treatment. The first three experiments performed within chamber slides showed that tumor count and average tumor size decreased with increasing oxaliplatin concentrations as expected, which is comparable to the in vivo tumor response to the drug. The fourth experiment demonstrated that, although cells are able to grow within the microfluidic device, this model did not accurately replicate the in vivo condition and future work needs to be aimed at improving the design of the device as well as optimizing parameters within the experiment.

Respiratory distress syndrome is a major cause of mortality among infants. Current therapies are limited in terms of invasiveness, cost, infrastructure, and leads to long term morbidities such as bronchopulmonary dysplasia. As a result a form of respiratory support termed as “artificial placenta” has been developed that allows natural development of lungs and avoids long term morbidities. The artificial placenta is connected via the umbilical vessels and provide pumpless respiratory support and is characterized by non-invasiveness, low cost and low infrastructure. Our group previously reported on a development of porous PDMS membrane artificial placenta. To build upon its development, one of the objectives of this thesis was to reduce the variation in the oxygen saturation of the input blood for testing the oxygenator. Another objective was to setup a mathematical model to predict the oxygen uptake in an oxygenating unit and use the model to optimize the geometric parameters of a design. The final objective was to improve the oxygen uptake of the oxygenating unit of the artificial placenta by redesigning the blood flow path and the membrane material. The experimental setup was improved to employ an active controller that actively maintained the oxygen saturation of the input blood for testing the oxygenator within a variation of ±3% of the set point for at least an hour. As compared to previous experimental setup the blood deviated from the set point by 9%. Later, the blood flow path in the oxygenator was redesigned from a flat height profile to a sloping height profile; and the PDMS membrane was reinforced with a thin steel mesh. Such changes improved the oxygen uptake at the operating pressure of 30 mmHg from 16 µL/min in case of an oxygenator with flat height profile and PDMS membrane to 26 µL/min in case of an oxygenator with flat profile and composite membrane. Finally, a mathematical model was developed that coupled oxygen uptake, pressure drop and membrane expansion. The model was validated against experimental results and was later used to optimize the configuration of the oxygenator with sloping profile and composite membrane. The predicted oxygen uptake of the optimized configuration at the operating pressure of 30 mmHg was 78.8 µL/min. / Thesis / Master of Applied Science (MASc)

Preterm infants are at risk for respiratory distress syndrome (RDS) due to immature lungs, leading to notable neonatal mortality. About 10% of US births are premature. While mechanical ventilation is a common RDS treatment, it can cause complications. If it fails, extracorporeal membrane oxygenation (ECMO) is employed, but standard ECMO devices are not suited for preterm babies. The limitations of hollow fiber membrane oxygenators used in ECMO have spurred interest in an artificial placenta that connects to the umbilical cord and supports lung growth. Microfluidic blood oxygenators, with their biomimetic designs, have being explored for this purpose. This thesis advances microfabrication techniques for Lung Assist Devices (LADs), focusing on two main objectives: I. Improving Throughput for Elevated Blood Flow Rates: This section delves into refining Microfluidic Blood Oxygenators (MBOs) to accommodate greater blood flow rates. By combining parallel units, we increased throughput and optimized LAD designs. Newly designed MBOs, with an expanded gas exchange surface area, can manage blood flow rates up to 60 mL/min. Using these enhanced MBOs, we constructed a novel LAD achieving superior oxygenation compared to predecessors. Our in vitro tests confirmed that this LAD can sustain blood flow rates of up to 150 ml/min, elevating oxygen saturation by approximately 20%—equivalent to an oxygen transfer of 7.48 mL/min, a leading figure for AP-type devices. II. Hierarchically Designed Microchannels: The second objective revolves around developing microchannels with a hierarchical layout to mitigate stagnation and high shear stress regions. Traditional photolithography poses challenges at channel intersections, inducing clotting risks. We pioneered alternative microfabrication methods, yielding diverse microchannels and intricate hierarchical designs that emulate natural vascular networks devoid of dead zones. These advancements have propelled the microfabrication domain for artificial placenta-like LADs. Utilizing our method, we produced channels varying from hundreds to a few microns in height with a single exposure and an opal diffuser. Thin membranes (~60 µm top and ~45 µm bottom) were amalgamated, culminating in a total depth of about 200 µm. Such oxygenators excel in oxygenating blood even at intense flow rates of up to 15 mL/min per unit. Leveraging these hierarchically designed MBOs, we crafted a LAD supporting a flow rate of 100 mL/min, offering an oxygen transfer of 5.21 mL/min. Both LADs developed in this research proficiently support premature neonates weighing up to 2 kg. Notably, the priming volume of the LAD using the enhanced MBOs has been substantially minimized, underscoring its advancements over earlier models. Realizing these objectives can transform neonatal care, addressing respiratory challenges in premature neonates and bolstering their chances for a healthier life. / Thesis / Master of Science (MSc)

No description available.

No description available.

No description available.

Colonoscopy is a necessary procedure to diagnose diseases in the lower gastrointestinal tract. Nevertheless, there exists a risk of bleeding during the colonoscopy, caused by perforations, diverticuli, post-biopsy complications, and polyps, which could go undetected as the camera is only equipped on the tip of colonoscope. A soft sensor, capable of detecting blood and distinguishing it from other GI fluids, has been developed using optical fibers for blood detection and data transmission. However, the sensor’s numerous optical fibers make it harder for the surgeon to hold and maneuver the colonoscope. In addition, the fibers are fragile and sensitive to external forces. This makes the sensor’s fabrication difficult and signal interpretation less reliable. Presented in this thesis is a wireless soft blood sensor utilizing deformable polymeric materials and microelectronic technologies. Opto-electronic components and a microcontroller installed on the flexible PCB allow the sensors to sense blood, recognize fluid types, account for the external forces, comply to bending of colonoscope, and wirelessly transmit data. The wireless data transmission is implemented by a millimeter-scale transmitter-receiver module. A Lithium ion battery powers the sensor. Without optical fibers, multiple blood sensors can be installed along the length of colonoscope. Consequently, this increases efficiency and reliability of blood detection while remaining safe for patients without interruption on the clinical workflow. / 2027-05-31T00:00:00Z