ELECTROCHMICAL IMPEDANCE SPECTROSCOPY BIOSENSOR PLATFORM FOR EVALUATION OF BIOFILM by Matthew Connor Dusenbery McGlennen A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Mechanical Engineering MONTANA STATE UNIVERSITY Bozeman, Montana May 2023 ©COPYRIGHT by Matthew Connor Dusenbery McGlennen 2023 All Rights Reserved ii ACKNOWLEDGEMENTS First and foremost, I would like to thank my advisors, Dr. Stephan Warnat, and Dr. Christine Foreman for their continued support, ideas, problem solving, and benevolence. Their passion for interdisciplinary science is inspiring. Thank you both for allowing me to pursue my own ideas and for your commitment to my professional and personal development throughout my PhD. I would also like to thank my committee members Dr. Cecily Ryan and Dr. Stephen Sofie for accepting the invitation to participate in my PhD work. A special thank you to Dr. Markus Dieser for teaching me, a mechanical engineer, on how to be a microbiologist. I would like to thank the Center for Biofilm Engineering for providing a collaborative multi-disciplinary space to conduct research. Thanks to all the researchers and fellow graduate students who have participated in experiments, troubleshooting, discussions, and camaraderie – Michael Neubauer, Daniel Collins, Haley Kettler, Seth Kane, Jesse Arroyo, Bruce Boles, Madelyn Wilis, Dr. Heidi Smith, Dr. Kristin Brileya, Dr. Andy Lingley, Dr. Joshua Heinemann. I would also like to thank all the undergraduate students that I have had the privilege of mentoring during my time as a graduate student – Ruby Jackson, Lauren Jackson, and Amelia Burda. And most of all, I thank Taylor Boucher for being my best friend and partner in life’s adventures. iii TABLE OF CONTENTS 1. INTRODUCTION .......................................................................................................................1 Motivation ...................................................................................................................................1 Dissertation Outline .....................................................................................................................3 Background .................................................................................................................................5 Biofilms ..............................................................................................................................5 Microfabricated Systems for Real-time Biofilm Detection ...............................................8 Electrochemical Impedance Spectroscopy (EIS) .............................................................10 Importance of the Electrochemical Interface for Electrochemical Sensing .............................................................................................16 Effects of Biofilm on Electrochemical Systems ...............................................................18 µIDEs for Biological Recognition ....................................................................................20 Knowledge Gaps .......................................................................................................................23 Research Goals, Hypotheses, and Objectives ...........................................................................24 2. EFFECTS OF ESCHERICHIA COLI K12 BIOFILM ON SENSOR THIN FILM MATERIALS ................................................................................29 Contribution of Authors and Co-Authors ..................................................................................29 Manuscript Information .............................................................................................................30 Abstract .....................................................................................................................................31 Introduction ...............................................................................................................................31 Materials and Methods ..............................................................................................................32 Sample Preparation ........................................................................................................32 Culture conditions and Biofilm Reactors ......................................................................34 Materials Characterization ............................................................................................35 Results .......................................................................................................................................36 Conclusions ...............................................................................................................................39 3. USING ELECTROCHEMICAL IMPEDANCE SPECTROSOPY TO STUDY BIOFILM GROWTH IN A 3D-PRINTED FLOW CELL SYSTEM ..........................................................................41 Contribution of Authors and Co-Authors ..................................................................................41 Manuscript Information .............................................................................................................42 Abstract .....................................................................................................................................43 Introduction ...............................................................................................................................44 Material and Methods ................................................................................................................46 Sensor Fabrication .........................................................................................................46 Sensor Electropolymerization .......................................................................................48 Design and Fabrication of CLSM Flow Cell System ....................................................49 iv TABLE OF CONTENTS CONTINUED EIS Measurements and Data Analysis ..........................................................................51 Characterization of PPy:PSS Coated Sensors ...............................................................52 Cell Culturing Conditions .............................................................................................52 Flow Cell System Operation .........................................................................................53 CLSM Imaging and Data Analysis ...............................................................................54 Results .......................................................................................................................................54 Characterization of Electropolymerized µIDEs ............................................................54 Sensor Variability and Stability in Abiotic Media ........................................................55 Flow Chamber Fluid Flow Simulation ..........................................................................57 EIS Characterization of Biofilm Growth .......................................................................58 Real-time Single Frequency Impedance Analysis and CLSM of Biofilm .....................................................................................60 Discussion .................................................................................................................................63 Conclusions ...............................................................................................................................66 4. APPLICATION OF IMPEDANCE SENSORS FOR BIOFILM CONTROL IN REAL-TIME ...........................................................................68 Contribution of Authors and Co-Authors ..................................................................................68 Manuscript Information .............................................................................................................69 Abstract .....................................................................................................................................70 Introduction ...............................................................................................................................71 Material and Methods ................................................................................................................74 System Overview ..........................................................................................................74 Sensor Design ................................................................................................................74 Flow Cell Operation ......................................................................................................76 Bacterial Strain Selection and Culturing Conditions ....................................................76 Biofilm Treatment .........................................................................................................77 Electrochemical Impedance Spectroscopy Measurement .............................................78 Biofilm Assessment with Confocal Laser Scanning Microscopy and Image Analysis ...................................................................79 Statistical Analyses ........................................................................................................80 Results .......................................................................................................................................80 Real-time Impedance Measurements of Biofilm Growth .............................................80 Real-time Impedance Measurements of Biofilm Treatment .........................................81 Real-time Impedance Measurements of Delayed Biofilm Development ......................84 Validation of Biofilm Removal and Delay by CLSM Imaging ....................................85 Discussion .................................................................................................................................87 Conclusions ...............................................................................................................................91 v TABLE OF CONTENTS CONTINUED 5. MICROSENSORS IN ICY ENVIRONMENT TO DETECT MICROBIAL ACTIVITIES ...............................................................................93 Contribution of Authors and Co-Authors ..................................................................................93 Manuscript Information .............................................................................................................94 Abstract .....................................................................................................................................95 Introduction ...............................................................................................................................95 Materials & Methods .................................................................................................................97 Sensor and Probe Manufacturing ..................................................................................97 Culture Conditions for Laboratory Experiments .........................................................100 Laboratory Tests ..........................................................................................................101 Field LCR Meter .........................................................................................................101 Field Tests ...................................................................................................................102 Results and Discussion ............................................................................................................103 Laboratory Impedance Measurements ........................................................................103 Comparison of the Laboratory and Handheld LCR Meters ........................................107 Field Tests ...................................................................................................................108 Conclusion ...............................................................................................................................110 6. CONCLUSIONS AND FUTURE OUTLOOK .......................................................................112 REFERENCES CITED ................................................................................................................118 APPENDICES .............................................................................................................................136 APPENDIX A: Supplemental material for: Using Electrochemical Impedance Spectroscopy to Study Biofilm Growth in a 3D Printed Flow Cell System .............................................137 APPENDIX B: Supplemental Material for: Application of Impedance Spectroscopy Sensors for Assessing Biofilm Control in Real Time ...........................................................................................142 vi LIST OF TABLES Table Page 1.1 Summary of reports using EIS µIDE sensors for biofilm detection ............................22 2.1 Previous Work on Material and Sensor Reliability in Aqueous Environments ..........................................................................33 vii LIST OF FIGURES Figure Page 1.1 Graphical representation of an EIS measurement ........................................................11 1.2 Graphical Representation of EIS spectra .....................................................................15 1.3 Schematic of the electrochemical double layer ...........................................................17 1.4 Example electrical equivalent circuit and contribution of biofilm to the electrochemical system .........................................................................20 1.5 Graphical representation of a µIDE .............................................................................21 2.1 CDC biofilm reactor containing E. coli K12 biofilm growing on aluminum and silicon nitride substrate ........................................................................35 2.2 Collinear four-point-prove sheet resistance of aluminum thin film measured at time intervals of 0, 3, 5, and 7 weeks in CDC biofilm reactor ................36 2.3 SEM images of aluminum PVD thin film surfaces .....................................................37 2.4 FTIR spectra of a-SixNy:H thin films ...........................................................................38 2.5 White light interferometry image of a-SixNy:H thin films ...........................................39 3.1 Optical microscope images of microfabricated sensors and their features .............................................................................................48 3.2 Design of the flow cell system with integrated microfabricated sensors used for biofilm growth experiments ..............................................................50 3.3 Evaluation of sensor variability in abiotic 1:10X TSB from sensors coated with PPy:PSS with deposition charges ranging from 0 to 450 µC ...................56 3.4 Finite element analysis using COMSOL Multiphysics of the flow chamber design ....................................................................58 3.5 Bar chart comparing average impedance changes at select frequencies ......................60 3.6 Representative CLSM images of biofilm growth ........................................................62 viii LIST OF FIGURES CONTINUED Figure Page 4.1 Graphical representation of platform design and experimental setup for analyzing biofilm .............................................................75 4.2 Impedance changes for untreated abiotic controls and biofilm growth .......................81 4.3 Impedance changes for treated abiotic controls and biofilm .......................................83 4.4 Magnitude of impedance change between pre-treated and post-treated abiotic controls and biofilm ............................................84 4.5 CLSM images of biofilm at pre-treatment and post-treatment ....................................86 4.6 CLSM images of untreated and QSI exposed biofilm .................................................87 5.1 Optical microscope image of µIDE sensor ..................................................................99 5.2 Impedance sensors for laboratory and field applications ...........................................100 5.3 Impedance sensors deployed in natural low temperature environments ...................................................................................103 5.4 Nyquist and Bode plots of impedance for Flavobacterium sp. ANT 11 in frozen and liquid conditions ..................................................................................104 5.5 Equivalent circuit representation of experimental data for frozen cell suspension ...................................................................................105 5.6 Calibration curve for frozen cell concentration of Flavobacterium sp. ANT 11 ......................................................................................107 5.7 Comparison measurements of dummy cells between handheld and benchtop impedance analyzers dummy cell ........................................108 5.8 Bode plot comparing the measured samples in the field ...........................................109 A.1 Microfabricated sensor design and fabrication flow .................................................138 A.2 Summary of experimental procedure for operating flow-cell for biofilm growth .....................................................................139 ix LIST OF FIGURES CONTINUED Figure Page A.3 Bode plot of impedance and phase angle for different electropolymerization coatings ..................................................................139 A.4 Evaluation of sensors drift from EIS sensor in sterile 1:10X TSB for sensors coated with different thicknesses of PPy:PSS .............................................................................................140 A.5 Relative impedance response of entire measured frequency range at selected timepoints .....................................................................141 B.1 Relative impedance response of entire measured frequency range at select timepoints of 5% MWF flow in flow-cell .........................................144 x ABSTRACT Microbial biofilms are organized communities of surface-attached microorganisms encased in a self-produced extracellular matrix that pose significant challenges in medicine, the environment, and industry. Biofilms can cause chronic infections, biofouling, and equipment failure, while existing methods for biofilm detection are slow, costly, and labor-intensive. Recently, the use of microfabricated electrochemical impedance spectroscopy (EIS) biosensors has emerged as a promising technique for evaluating biofilm growth in real-time with advantages of small-size, adaptability, low-cost, and high-sensitivity. In this work, EIS biosensors featuring gold micro-interdigitated electrodes were produced using standard microfabrication techniques. Sensors were integrated into a custom 3D-printable flow cell system, enabling EIS measurements and confocal laser scanning microscopy (CLSM) imaging simultaneously. Green fluorescently labeled Pseudomonas aeruginosa PA01, a model biofilm forming bacteria, was introduced into flow chambers and subsequent growth was monitored by EIS, CLSM, and biomass enumeration. Using the system, biofilm growth, dispersal, and the effects of cell-signaling suppression were evaluated. The sensors were also tested in an oil-water emulsion and field-tested on an alpine snow-patch and pond. Improved stability of EIS measurements was achieved by coating the sensors’ counter and reference electrodes with an electrically conductive polymer. Biofilm growth was successfully detected using EIS biosensors at an optimized single-frequency, with average decreases in impedance of ~22% by 24 hours. Likewise, biofilm dispersal via chemical treatments were successfully detected with average increases in impedance of ~14% over the ensuing 12 hours. When cells were exposed to a quorum sensing inhibition agent, impedance did not decrease for 18 hours. Impedance changes due to biofilm growth, dispersal, and effects of quorum sensing inhibition were validated by CLSM images and biofilm enumeration. Similarly, in an oil-water emulsion the biosensors successfully detected biofilm growth, dispersal, and effects of quorum sensing inhibition. In an alpine field-test, samples containing varying concentrations of microbes could be detected using the EIS biosensors. This work demonstrates that EIS biosensors are a promising tool for real-time monitoring of biofilm dynamics in a variety of aqueous environments. Overall, EIS biosensing holds great potential for in situ and real-time data regarding biofilm colonization that is not possible with existing techniques. 1 INTRODUCTION Motivation Microfabricated biosensors offer great potential for scientific and engineering innovations, and they have found widespread applications in various fields such as medicine, industry, and environmental conservation. The benefits of these sensors are vast and compelling, including their compact size, cost-effectiveness, high sensitivity, versatility, and ability to provide real-time data. Additionally, these biosensors can be customized and designed to work in diverse settings, enabling them to collect data from otherwise unreachable areas. With their miniaturization, they can be integrated into various devices and systems, allowing for continuous monitoring and analysis. Among the many types of sensors, microfabricated electrochemical impedance spectroscopy (EIS) sensors are gaining adoption due to their numerous advantages, including simple design, low-cost fabrication, and adaptability to a wide variety of detection targets. EIS is an informative method of characterizing materials that can be used to analyze bulk and interfacial properties of liquids, solids, or gasses (Lvovich, 2012; Vadhva et al., 2021). When the technique of EIS is applied to microfabricated sensors and is used to measure biological or chemical reactions, the resulting devices are termed EIS biosensors. EIS biosensors are non-invasive, sensitive, and cost-effective devices, which makes them preferable tools for rapid and portable in situ or point-of-care detection (Cesewski and Johnson, 2020). EIS biosensors are also label-free, meaning they do not require any tagging/labeling agents which can affect the activity of the targets of interest (Mazzaracchio et al., 2019; Queiros et al., 2013). The affinity of EIS biosensors can be enhanced by adding ligands (e.g., DNA, RNA, antibodies, or peptides) to the sensor surface, which 2 allows conversion of bio-recognition events such as antibody-antigen, substrate-enzyme, or whole cell binding, into meaningful signals (Trunzo and Hong, 2020). EIS biosensors are suitable for various applications in areas such as medical diagnostics and environmental monitoring, as they can detect trace amounts of target molecules. Given that microorganisms are abundant in many aqueous environments, EIS biosensors could be deployed for their detection. Microorganisms in aqueous environments form biofilms, which are defined as proliferating communities of microbes attached to surfaces, embedded within a self-produced extracellular polymeric substance (EPS) matrix (Hall-Stoodley et al., 2004). In nature, the vast majority of microorganisms exist as biofilms (Mazza, 2016). Analysis of biofilms can be particularly difficult due to their tendency to develop in inaccessible locations and their constantly evolving nature. Biofilms often exist in polymicrobial communities that interact with each other in intricate ways (Roy et al., 2018). Depending on microbial species, environmental conditions, and nutrient availability, biofilm structure, and composition can vary. Combined, these factors make the study of biofilm challenging. An inexpensive microfabricated EIS biosensor that could accurately and promptly detect biofilm contamination in real-time would be highly desirable in a range of industrial and scientific contexts. Although EIS biosensors show promise in studying biofilm, challenges remain to fully realize their capabilities and establish the technology as a reliable tool for biofilm research. EIS signals must be correctly interpreted in the context of the testing conditions. For instance, EIS biosensor signals are influenced by fluid types, environmental conditions, and types of organisms present (Sopoušek et al., 2021). Standardized procedures for EIS biosensing need to be established to ensure consistent and reproducible results across different studies and applications. The research 3 presented in this dissertation aims to address some of the gaps by improving design and integration, exploring the limits, and expanding on the capabilities of EIS biosensors. Ultimately, these advances will lead to better adoption of microfabricated biosensors into mainstream analysis of biofilm. Dissertation Outline Chapter 1: Introduction. An overview of biofilm physiology is provided. This is followed by a discussion of microfabricated systems for characterizing biofilms, electrochemical impedance spectroscopy, the importance of electrochemical interfaces for biosensing, the effects of biofilm on electrochemical interfaces, and micro-interdigitated electrodes for biofilm detection. Key research gaps and research objectives are presented. The main body of this dissertation (Chapter 2-5), address the outlined objectives. Chapter 2: “Effects of Escherichia coli K12 Biofilm on Sensor Thin Film Materials” published in the Journal of The Institute of Electrical and Electronics Engineers Sensors, 2019. In this study, the long-term effects of biofilm formation on the properties of aluminum (metallic thin- film material) and a-SixNy:H (dielectric thin-film material) are evaluated. Material degradation caused by Escherichia coli K12 biofilm growth is determined by electrical sheet resistance measurements (collinear four-point-probe) and Fourier-transform infrared spectroscopy (FTIR) absorption spectra over a time period of 7 weeks. The results suggest that sensor materials (e.g., gold, polycrystalline silicon, and nickel) should be investigated to determine their susceptibility to microbially influenced corrosion, and encapsulation with a-SixNy:H can minimize degradation. From this study, the corrosive nature of biofilm is evident which underscores the need for corrosive resistant, reliable, sensor design in future studies. 4 Chapter 3: “Using Electrochemical Impedance Spectroscopy to Study Biofilm Growth in a 3D Printed Flow Cell System”, published in the Journal Biosensors and Bioelectronics: X, March 2023. This work demonstrates that microfabricated sensors modified with PPy:PSS coatings enable highly stable time-resolved EIS measurements under both abiotic (free of microbes) conditions and when subjected to Pseudomonas aeruginosa biofilm growth. Electrochemical impedance data from sensors integrated into a novel 3D-printed flow cell corroborates distinct biofilm growth stages that is confirmed using confocal laser scanning microscopy (CLSM). Given the success of the sensor in monitoring biofilm growth, the subsequent logical progression is to evaluate the sensor's efficacy in detecting biofilm dispersal maneuvers. Chapter 4: “Application of Impedance Sensor for Assessing Biofilm Control in Real Time”, prepared for submission to the Journal of Industrial Microbiology and Biotechnology, April, 2023. In this study, real-time impedance measurements of biofilm growth under flow conditions were made possible through a custom flow cell system equipped with integrated sensors, proving to be effective at assessing the growth and dispersal (using either chlorine or biocide) of P. aeruginosa biofilm in growth medium and an oil-water emulsion. Additionally, the sensors successfully detect delayed biofilm growth in the presence of a bacterial quorum sensing inhibitor. To broaden the potential uses of the sensor, the next logical step is to assess its effectiveness in other fluid environments containing microorganisms. Chapter 5: “Microsensors in Icy Environments to Detect Microbial Activities”, published in the Journal of Microelectromechanical Systems, 2020. In this study, data is presented on the use of microfabricated EIS sensors for the detection of microbes in both laboratory and field-based 5 liquid, semi-solid, and icy environments. The results demonstrate that samples that vary in phase (solid vs. liquid), temperature, and microbial concentration generate discrete impedance spectra in laboratory conditions and subsequently in an environmental field-test. Chapter 6: Conclusions and future direction of EIS biosensing for biofilm monitoring. This chapter highlights the key findings and conclusions of this work. Remaining challenges, and possible future directions of this technology are discussed. Background The following sections provide essential background about biofilm physiology, and the current analytical methods used to investigate biofilms. This is followed by sections that elucidate the concepts of electrochemistry, electrochemical impedance spectroscopy (EIS), and EIS biosensing, which are crucial to understanding the operation of this sensor technology. Biofilms A biofilm is a cooperative aggregate of microorganisms associated with a surface, enclosed in self-produced extracellular polymeric substance (EPS) composed of polysaccharides, proteins, lipids, and extracellular DNA (Azeredo et al., 2016; Lappin‐Scott and Costerton, 1989; Wimpenny et al., 2000). It is estimated that up to 80% of all microbial biomass exists as biofilm (Bar-On and Milo, 2019). Biofilm-forming microorganisms attach to surfaces, communicate with each other in a coordinated manner, and carry out regulated processes leading to maturation. Cell-to-cell communication occurs via signaling molecules called autoinducers in a process referred to as quorum sensing (QS), which allows cells to sense and respond to population density and regulate gene expression (Davies et al., 1998). Biofilms maintain a protected mode of growth that allows 6 them to survive for long durations with minimal nutrients and endure in high-stress conditions or harsh environments (Costerton et al., 1999). Biofilm contamination is a ubiquitous problem in many aqueous environments. For instance, industry contamination of precision machining and grinding fluids causes premature degradation of fluids and workplace pathogenic exposure (Simpson et al., 2003). In the medical field, biofilms form on surgical implants and indwelling catheters, leading to infections and detrimental health effects (Lisoń et al., 2022). In environmental settings, biofilms contaminate municipal water supplies or pipe networks, potentially leading to widespread illness (Wingender and Flemming, 2011). Collectively, the economic cost of biofilm related problems worldwide is in excess of $5 trillion a year (Cámara et al., 2022). The success of removing biofilm is greatest when targeting microorganims at certain life stages. For example, during the early stages of biofilm development, free-floating cells are more susceptible to antimicrobials and biocides (Gu et al., 2019). However, mature biofilms exhibit high resistance to antimicrobials making treatment more difficult (Davies, 2003; Hall-Stoodley et al., 2004). This resistance is possibly due to the microbes being encased and protected by EPS. This barrier acts to deter physical removal and provides enhanced resilience against chemical disruption by blocking the diffusion of antimicrobials into the bulk of the biofilm (Stewart, 2003). Once surfaces and fluids are contaminated with biofilm decontamination is difficult, even after meticulous cleaning, since even a few residual cells can quickly repopulate the system. The problem is further exacerbated as biofilms often reside in hard-to-reach locations. Currently, techniques that are available to measure biofilm formation and growth generally require time for culture and/or involve labor-intensive handling steps (Pantanella et al., 2013). 7 Harvested sessile bacteria can be enumerated using colony-forming unit plate-counts and staining methods, like crystal violet (CV), safranin, or trypan blue (Christensen et al., 1982; Pitts et al., 2003). The reproducibility can vary based on the skill and training of the operator. Biofilms can be analyzed with quantitative Polymerase Chain Reaction (qPCR), various microscopy techniques (e.g., epifluorescence and confocal laser scanning microscopy (CLSM), transmission electron microscopy (TEM), or scanning electron microscopy (SEM) (Hassan et al., 2011; Peeters et al., 2008; Thurnheer et al., 2004). However, these techniques are either labor-intensive, expensive, or require labeling, and generally are destructive to the biofilm providing only a single time point for each sample. Furthermore, different techniques are deployed for different biofilm types, making comparability and consistency challenging across studies. As a result, there is a lack of consensus on how best to evaluate biofilms (Azeredo et al., 2016). Pseudomonas aeruginosa is a model biofilm forming organism (Hall-stoodley et al., 2004) and has been the focus of intense study because of its prominent role in disease and widespread environmental presence. The ability of P. aeruginosa to readily form biofilms, hinders eradication (Moradali et al., 2017). P. aeruginosa grows readily in many diverse environments and is an especially virulent opportunistic infection when normal immune defenses are disabled. For instance, patients with cystic fibrosis whose airways become colonized with P. aeruginosa, develop chronic infections that are almost impossible to eradicate (Killough et al., 2022; Tuon et al., 2022;). P. aeruginosa also is a major pathogen in burns and other chronic wounds, on implanted biomaterials such as knees, hips, or valves, and within hospital surfaces and water supplies (Breathnach et al., 2012). 8 P. aeruginosa is especially difficult to eradicate because it is naturally resistant to many antibiotics. According to the world health organization (WHO), P. aeruginosa is recognized as a priority pathogen for new antibiotic development since number of multidrug-resistant strains are increasing (World Heath, 2017). Pseudomonas is a common genus of bacteria found in industry, such as in metalworking fluids (MWFs) which are complex mixtures of chemicals used as cooling and lubricating agents in a variety of machining processes (Saha and Donofrio 2012). The quality of MWFs is degraded by microbial contaminates which not only pose potential workplace health concerns, such as dermatitis or hypersensitivity pneumonitis, but may also lead to premature deterioration of the costly fluid (Koch, 2023). For the above reasons P. aeruginosa is chosen as a model biofilm producing bacteria for development of an EIS biosensor. The research aim is to use EIS to detect all stages of P. aeruginosa biofilm development including initial attachment, proliferation, and mature biofilm production. Microfabricated Systems for Real-time Biofilm Detection In the past decade, microfabricated systems have expanded into the field of microbiology and have been increasingly used to evaluate/detect biofilm (Tiozzo-Lyon et al., 2023). Microfabricated system design has brought about advantages such as manufacturing precision, ability to control feature size, potential for high reproducibility, ease of modification for specific applications, and low-cost for large quantity fabrication (Baracu and Dinu Gugoasa, 2021). Consequently, microfabrication has advantages that can be applied for the creation accurate biofilm detection devices. 9 Broadly, microfabricated detection systems can be categorized as having optical, mechanical, or electrochemical transduction mechanisms. Each technique has strengths and weaknesses for evaluating biological phenomenon including biofilm monitoring (Subramanian et al., 2019). Briefly, optical systems detect changes in light properties, such as intensity, wavelength, polarization, or phase shift, and turn the signal into a measurable quantity, such as voltage, current, or frequency (Zhang et al., 2011). However, optical systems are susceptible to interference from non-biofilm components and are unsuitable for opaque fluids. Mechanical microsystems operate by physically measuring the change of mechanical properties such as mass, stiffness, or viscoelasticity, translating the change into a signal (Chalklen et al., 2020). Mechanical sensors are limited by susceptibility to ambient noise and vibrations. These limitations can lead to inaccurate measurements making mechanical sensors less suitable for biofilm analysis, especially in industrial locations where vibrations may be encountered. Electrochemical sensors detect changes in electric potential, charge buildup, or current resulting from chemical reactions between the analyte and electrodes (Grieshaber et al., 2008). These changes are then converted into a signal that can be utilized. Electrochemical sensors offer several advantages, including immunity to ambient vibrations, non-invasive analysis with high sensitivity and selectivity, real-time response, suitability for use in various fluid types (including opaque fluids), and low power consumption. Electrochemical sensors can be classified into three categories: amperometric, potentiometric, and impedimetric (EIS). Amperometric sensors require an electrical current and often require the addition of redox mediators which are intermediate electron carriers or reservoirs that allow reversible charge transfer through the electrode/electrolyte interface (Zhang et al., 10 2022). Potentiometric sensors require auxiliary reference electrodes to obtain valid data complicating the design and operation (Zdrachek and Bakker, 2019). In contrast, EIS measurements use a sinusoidal electrical perturbation, which has less impact on the electrochemical system compared to amperometric measurements, where the applied potential could alter the system. Additionally, EIS sensors do not require the addition redox mediators (Dorledo de Faria et al., 2019). Lastly, EIS measurements can be carried out over a broad range of applied frequencies, enabling collection of additional electrochemical data that encompasses both the system's surface and bulk properties. Collectively, these advantages make EIS a desirable technique for studying biofilms. Electrochemical Impedance Spectroscopy (EIS) EIS is a technique in which a sine wave voltage or current is applied to an electrochemical system, and the corresponding current or voltage is measured (Figure 1.1a). Fundamentally, impedance measures the opposition of electrical flow in an alternating current (AC) system. The quotient of the time dependent voltage to the time dependent current is termed impedance (Figure 1.1b; Lvovich, 2012). EIS is used in a wide range of application because of the informative outputs it provides. Examples where EIS is applied include the characterization of fuel cell reactions, the development of coating materials, the evaluation of battery performance, and the monitoring of metal and alloy corrosion (Lasia, 2014; Wang et al., 2021). 11 Figure 1.1: Schematic representation of EIS. a) The ratio of the potential (V) to the current (I) is the impedance (|Z|), and the phase shift difference between the voltage and current (ϕ). b) The magnitude of impedance can be represented in both real (Zre) and imaginary (Zim) parts and are related mathematically with the phase lag. EIS measurements are carried out in either potentiostatic (sine wave voltage) or galvanostatic (sine wave current) conditions. Several conditions must be satisfied to obtain valid EIS data. First, the measured impedance must be causal, meaning that the impedance response must come only from the applied perturbation. Second, to ensure no alterations to the electrochemical system under test, and to ensure the measurements are “pseudo-linear”, the electrical perturbations in EIS measurements are kept small (i.e., ≤50 mV or ≤100 µA). Third, while the measurement is taking place, the system remains stable. In the process of performing an EIS measurement, the impedance is measured at various perturbation frequencies, typically ranging from 0.1 Hz to 1 MHz. This procedure produces an impedance spectrum, which is a graphical representation of the system's impedance as a function of applied frequency. By examining this spectrum, insights into the electrical characteristics of the system are obtained (Vivier and Orazem, 2022). 12 Impedance extends from the concepts of Ohm’s law, which describes the relationship between voltage and current passing through a resistor, shown in equation 1.1: 𝑉 = 𝐼𝑅 (1.1) In an alternating current (AC) system, impedance is represented as Z, as shown in equation 1.2: 𝑉 = 𝐼𝒁 (1.2) However, in a standard EIS experiment applying a sinusoidal voltage to a material or system, the resulting current response can be expressed as a sinusoidal function with the same frequency, but with a phase shift and possibly a different amplitude, as shown in equation 1.3: 𝑉(𝑡) |𝑉|𝑠𝑖𝑛(𝜔𝑡) 𝒁 = = (1.3) 𝐼(𝑡) |𝐼| sin(𝜔𝑡+ 𝜙) where t is time, ω is the applied angular frequency in radian s-1 related to frequency, f, in Hz by ω=2πf, and ϕ is the phase shift difference between the voltage and current. Using Euler's identity, which simplifies and converts trigonometric functions into exponential functions in polar coordinates, equation 1.3 is converted to equation 1.4: |𝑉|𝑒𝑗𝜔𝑡 𝒁 = (1.4) |𝐼|𝑒𝑗𝜔𝑡+𝜔 where j is the complex imaginary square root of -1. Impedance spectroscopy measurements are typically presented in the Cartesian complex plane, where the complex impedance is divided into its real and imaginary components, as shown in equation 1.5: 𝒁 = 𝑍𝑟𝑒 + 𝑗𝑍𝑖𝑚 (1.5) where Zre is the real part of impedance, and Zim is the imaginary part of impedance. The impedance of an ideal resistor, R, is purely real, as described in Equation 1.6: 𝑍𝑟𝑒 = 𝑅 (1.6) 13 whereas both inductors and capacitors have purely imaginary impedance and contribute to the imaginary portion of impedance, as described in Equation 1.7: 𝑍𝑖𝑚 = 𝑋𝐿 − 𝑋𝐶 (1.7) where XL is the reactive inductance, and XC is the reactive capacitance. Impedance of both inductors and capacitors varies with frequency. The reactive inductance of an inductor increases with the applied angular frequency as shown in Equation 1.8: 𝑋𝐿 = 𝑗𝜔𝐿 (1.8) where L is the inductance. Furthermore, the impedance of a capacitor decreases as the applied frequency increases, as shown in Equation 1.9: −1 𝑋𝐶 = (1.9) 𝑗𝜔𝐶 where C is the capacitance. The overall impedance of a system is often expressed as the magnitude of impedance using equation 1.10: |𝑍| = √(𝑍 2 2 𝑟𝑒) + (𝑍𝑖𝑚) = √𝑅2 + (𝑋 − 𝑋 )2 𝐿 𝐶 (1.10) where |Z| is the magnitude, or modulus of impedance. Finally, the phase angle shift difference between the voltage and current is represented in Equation 1.11: 𝑍 𝑋 𝜙 = tan−1 𝑟𝑒 = tan−1 (1.11) 𝑍𝑖𝑚 𝑅 Since EIS involves complex numbers, there are two common ways to represent impedance spectra. The first method involves plotting either the magnitude of impedance (|Z|), or the calculated phase angle shift (ϕ) with respect to the applied frequency. This is called a Bode plot (Fig 1.2a). The second method involves plotting real components of impedance (Zre) against the imaginary components of impedance (Zim) across the impedance spectrum. This is called a Nyquist plot (Figures 1.2b and 1.2c). 14 One approach to elucidate impedance spectra is by fitting impedance data to electrochemical equivalent circuits (EECs) (Furst and Francis, 2019). EECs are used to fit measured EIS spectra with known circuit elements such as resistors (R), constant phase elements (CPEs), capacitors (C), and Warburg elements (W) combined in either parallel or series (Furst and Francis, 2019). Each of the circuit elements represent unique electrochemical components in an electrochemical system. The Randle’s circuit is frequently used to represent electrode-electrolyte systems because it provides a simple and effective way to model the complex electrochemical processes that occur at the electrode-electrolyte interface (MacDonald and Andreas, 2014). The circuit is composed of Rs (solution resistance), CDL (double layer capacitance), RCT (charge transfer resistance), and W (Warburg element). In a Randal’s circuit (Figure 1.2b), resistance of the bulk electrolyte solution is Rs. The CDL is the capacitance that arises from the separation of charges that occurs at the interface, due to the adsorption of ions or molecules on the electrode surface. The RCT is the resistance to the electrochemical reaction occurring at the electrode surface. Lastly, W represents the diffusion of ions in the electrolyte solution (Bahadır and Sezgintürk, 2016). Impedance measurements can be categorized into two types based on their surface characteristics and how measurements are taken (Zhang et al., 2022). The differences are defined by the charge transfer capability, which can be either faradic or non-faradaic. In electrochemical faradic systems, charge is transferred across an interface (Figure 1.2b). Faradic impedance systems allow charge to transfer through the electrode surface and often incorporate a redox mediator that can undergo cyclical oxidation and reduction with the electrode. Moreover, a Direct Current (DC) offset is sometimes required. Non-faradic measurements do not allow charge to transfer through 15 the electrode surface, meaning that the electrode behaves as a pure insulator. No redox-active molecule is required with this type of measurement. The Randal's circuit assumes a faradaic electrochemical system but can be modified to account for non-faradaic circumstances. If there are no redox-active molecules present, the electrode cannot facilitate charge transfer, resulting in the replacement of both RCT and W elements with RLEAK (Figure 1.2c). This parameter represents the ohmic losses that occur within the system. Non-faradaic systems are advantageous for remote sensing applications since they do not require the addition of redox mediators, which could alter the system under test. As discussed in the following section, the use of an EIS biosensor is designed to operate under non-faradaic conditions, but biofilm potentially contributes redox mediators resulting in signals from both faradaic and non-faradaic processes (Bellin et al., 2014; Bellin et al., 2016). Figure 1.2: Graphical representation of EIS spectra: a) Bode plot of impedance, b) Nyquist plot and Randle’s equivalent circuit for faradaic impedance, c) Nyquist plot and modified Randle’s equivalent circuit for non-faradaic impedance. 16 Importance of the Electrochemical Interface for Aqueous Electrochemical Biosensing The electrochemical double layer (EDL) is an important concept in electrochemistry and in understanding how EIS biosensors operate. The EDL occurs at the interface between a solid electrode and a liquid electrolyte and is formed when electrons from the electrode surface interact with ions in the surrounding electrolyte solution (Kosri et al., 2022). In the commonly cited Gouy– Chapman–Stern (GCS) EDL model (Kosri et al., 2022), the electrochemical interface is divided into five regions (Fig 1.3a). The first layer is the electrode itself, where the applied current or electrical potential dictates the surface charge. The second layer, called the inner Helmholtz plane, consists of species specifically adsorbed on the electrode surface, such as reactants and solvent molecules (Dunwell et al., 2018). Just beyond the inner Helmholtz plane is the outer Helmholtz plane, which consists of stationary solvent ions and molecules with opposite charges to the electrodes drawn to the electrode by electrostatic forces (Wu, 2022). Beyond the outer Helmholtz plane is the diffuse layer consisting of ions of both signs distributed in the electrolyte solution (Wu, 2022). Lastly, the bulk layer is composed of the electrolyte solution beyond the influence of the electrode’s charge. Taken together, these components describe the charge distribution in an electrochemical system. The potential difference across the EDL is known as the electrode potential, which is defined as the difference between the electrochemical potential of the electrode and that of the bulk electrolyte (Brown et al., 1999; Markovic, 2013). The magnitude of the electrode potential is influenced by the nature of the electrode, the electrolyte composition, temperature, and the presence of other ions or molecules at the electrode surface (Dunwell et al., 2018). The magnitude of electrode portential decay in a double layer is shown in Figure 1.3b. 17 Figure 1.3 a) Schematic of the electrochemical double layer and its components, b) the electrode potential as a function of distance from the electrode surface. A biosensor electrode in contact with a target liquid analyte forms a capacitive double- layer between the electrode surface and the bulk fluid of the analyte, and collectively comprise the electrochemical system. During an electrochemical biosensing event, biochemical reactions in the system, occurring near the electrode, alter the charge distribution and can be detected. Specifically, when target molecules interact with the biosensor electrode, changes to charge distribution in the double layer and/or changes in relative permittivity occur. These changes can be detected with impedance measurements (Bahadır and Sezgintürk, 2016). This research focuses on impedimetric biosensing, but it should be noted that electrochemical biosensors also include potentiometric and amperometric sensors, which measure either changes in electrode potential, or current responses due to biochemical reactions, respectively. 18 Effects of Biofilm on Electrochemical Systems When an EIS biosensor is deployed in a fluid seeded with microorganisms, microbial growth and EPS production are initiated, leading to biofilm development. The biofilm (e.g., microorganisms, EPS matrix, and nutrients), and the surrounding liquid can collectively be considered an electrochemical system (Wang et al., 2009). Growth, attachment, or detachment of biofilm on the sensor can correspond with distinct changes to the electrochemical signals (Marsili et al., 2010; van Duuren et al., 2017). Detecting biofilm with EIS biosensors is based on the principle that both the electrochemical interface and bulk properties of an aqueous system change when a biofilm proliferates. It is unclear what causes the electrochemical changes, but several mechanisms have been proposed. For instance, electrochemically active biofilms such as Pseudomonas aeruginosa produce at least four redox-active molecules, such as phenazines, including pyocyanin (PYO), phenazine-1-carboxamide (PCN), 5-methylphenazine-1-carboxylic acid (5-MCA), and phenazine-1-carboxylic acid (PCA), all of which have strong redox activity (Bellin et al., 2016, Qiao et al., 2015). These substances can increase the concentration of charge carriers near the electrode interface, and/or become integrated into the biofilm matrix, creating conductive pathways for electron transfer which influence impedance spectra. EPS components such as polysaccharides, proteins and extracellular DNA may also alter the electron transfer characteristics in biofilms (Gula et al., 2020; Li et al., 2016; Malvankar et al., 2011; Tan et al., 2019; Wang et al., 2022). Another possible mechanism by which biofilms might generate changes in electrochemical properties is through the extension of electrically conductive pili or nanowires, which extend from the surface of some bacteria and could facilitate electron transfer to an electrode (Malvankar and Lovley, 2014; Maruthupandy et al., 2015). Additionally, bacterial metabolism involves the conversion of large organic compounds (i.e., carbohydrates and sugars) into organic 19 acids and carbon dioxide, releasing the ionic metabolites into the medium which could also alter conductivity (Brosel-Oliu et al., 2019a). Figures 1.4B-4E are schematic representations of the electrochemical processes that can affect the electrochemical properties of a biofilm including attachment of cells to the substrate (Figure 1.4B), the production of EPS matrix (Figure 1.4C), presence of metabolically active planktonic cells (Figure 1.4D), and secretion of proteins, macromolecules, and redox metabolites onto the electrode surface (Figure 1.4E). EECs can be used to model electrochemical processes that occur when a sensor is deployed in biofilm, depicted in Figure 1.4A. As previously described, the Randle’s circuit is frequently used to represent electrode-electrolyte systems, but when applied to EIS biosensors, modifications to the circuit components should be considered. To take into consideration the non-idealities at the electrochemical interface, a CPEDL is often implemented (Figure 1.4A). The system resistance including electrical connection and the resistance of the thin film electrodes is modeled as Rconnection. The system's overall capacitance including parasitic losses and the baseline empty cell capacitance is modeled as Cgeometry. Several additional equivalent circuit models for characterizing biofilm formation have been considered. For instance, bacteria and biofilm matrix characteristics have been modeled with resistance and capacitance elements in both parallel and series configurations (Dominguez-Benetton et al., 2012; Furst and Francis, 2019; Kim et al., 2012; Kumar et al., 2023; Paredes, J. et al., 2014; Paredes et al., 2013; Subramanian et al., 2019; Zheng et al., 2013). Eventually EEC fitting may lead to a deeper understanding of the various biofilm processes. 20 Figure 1.4: (A) Equivalent complex circuit model of an electrode-electrolyte electrochemical impedimetric cell containing microorganisms and biofilm. Bacteria contribute to different mechanisms at the interface and bulk solution in an electrochemical cell. (B) bacteria in close contact with the surface contributing to the double-layer capacitance. (C) extracellular polymeric substance (EPS) production contributing to the double-layer capacitance. (D) planktonic growth contributing to the bulk solution resistance. (E) proteins and macromolecules adsorption onto the electrode surface contribute to changes in the double-layer capacitance (Image Credit: Matt McGlennen, CBE. Figure used with permission from Elsevier, Vélez Justiniano et al., (2023)). µIDEs for Biological Recognition This study focuses on the utilization of micro interdigitated electrodes (µIDEs), a specific type of EIS biosensor, which is a microfabricated electrode design with a unique structure consisting of a patterned array of "fingers", usually between 1-50 µm that are separated by small gaps, usually between 1-50 µm (Dizon and Orazem, 2020). The length and spacing of the electrodes can be individualized depending on the application of interest. Figure 1.5 illustrates the structure of the µIDEs. µIDEs operate under conventional two-electrode system configurations and have been used for sensing biological samples via impedance (Gomez et al., 2002; Radke and Alocilja, 2004; Yang et al., 2004). A significant advantage of µIDEs in biological applications is the maximized sensitivity achieved by the increased surface area in contact with the media under test (Ibrahim et al., 2013). The narrow gap between electrodes increases the signal-to-noise ratio (Kosri et al., 2022). In addition, µIDEs are easily fabricated at a low-cost, do not require a reference 21 electrode, and have a small footprint making them ideal for high-throughput manufacturing (Brosel-Oliu et al., 2019a; Maduraiveeran et al., 2018). Figure 1.5: Graphical representation of a µIDE The use of µIDEs for EIS biochemical sensing applications has gained interest in recent years. In the last two decades, µIDEs have been used to detect planktonic cells (Kim et al., 2012; Mallén-Alberdi et al., 2016; Tang et al., 2011; Xu et al., 2020) and to monitor biofilm formation (Estrada-Leypon et al., 2015; Liu et al., 2018; Paredes, J. et al., 2014; Paredes, Jacobo et al., 2014; Paredes et al., 2012, 2013; Subramanian et al., 2017). For instance, Yang was able to discern Salmonella Typhimurium concentrations down to 106 colony-forming units (CFUs) ml-1 with a simple µIDE setup (Yang, 2008) and could also detect foodborne pathogens using the same device (Yang and Bashir, 2008). Brosel-Oliu et al., developed a novel label-free µIDE biosensor functionalized with DNA aptamers that could detect Escherichia coli O157:H7 with a detection limit down to around 102 cfu ml−1. Moreover, no response of the sensor was registered in the presence of other bacterial strains (E. coli K12, Salmonella typhimurium, Staphylococcus aureus), confirming the selectivity (Brosel-Oliu et al., 2019b; Brosel-Oliu et al., 2018). Lastly, Paredes et al., showed that µIDEs on silicon substrates were an accurate and dependable way to detect Staphylococcus aureus biofilms in aqueous media (Paredes et al., 2012). 22 Comparison of past studies using EIS biosensors for detecting biofilm is limited by high variation between experimental setups, and methods of data analysis leading to discordant results (Table 1.1). Table 1.1 Summary of reports using EIS µIDE sensors for biofilm detection Reactor Duration; Target Electrode Abiotic Confirmational Statistical Type; Fluid Measurement Ref (year) Organism(s) modification control? technique Analysis Conditions Interval Candida albincans, Strep Well-plates; Abrantes et al. 48 hrs; 15 min NR Yes NR NR sanguinis, Step Static (2020) mutans Pseudomonas Yes, CLSM and aeruginosa PA01 Microfluidic; pearson’s Blanco-Cabra et al. 85 hrs; 12 hr NR NR Biofilm and PAET1, Staph Flow correlation (2021) Enumeration aureus MRSA test Pseudomonas Yes, Epifluoresence aeruginosa and Microfluidic; pearson’s Bruchmann et al. 3 days; <10 min NR Yes and Crystal Stenotrophomonas Flow correlation (2015) Violet maltophilia test Pseudomonas Crystal violet Well-plates; Chabowski et al. aeruginosa 168 hrs; 24 min NR Yes staining and NR Static (2017) ATCC14454 SEM Microfluidic; Optical and Estada-Leypon et al. Staph aureus 24 hrs; 10 min NR NR NR Flow SEM (2015) CFU Simulated Yes, Enumeration Huiszoon (2019, E. coli Catheter; 24 hr; 30 min NR Yes Endpoint T- and Optical 2021) Flow test Absorbance Pseudomonas NR 1 hr; 2 min NR Yes SEM NR Kim et al. (2012) aeruginosa PA01 E. coli, PDMS* 48hrs; 3hrs NR NR NR NR Liu et al. (2018) Salmonella Wells; Static Well-plates, CDCBGR, Parades et al. (2012, Staph epidermidis 25 hrs; 30min NR No Optical NR Petri Dish; 2013, 2014a 2014b) Varies Microfluidic; Subramanian et al. E. coli K12 60 hrs; 10-15 min NR Yes Epifluoresence NR Flow (2017) Pseudomonas Well-plates; Van Duuren et al. aeruginosa PA14 72 hrs; 5 min NR Yes CLSM NR Static (2017) and mutants CLSM and Pseudomonas Microfluidic; Yes; PPy:PSS McGlennen et al. 36-72 hrs; 30 min Yes Biofilm Yes, T-test aeruginosa PA01 Flow coating (This Study) Enumeration NR= Not Reported SEM = Scanning Electron Microscopy CLSM = Confocal Laser Scanning *PDMS=Polydimethylsiloxane Microscopy CDCBGR= CDC Bioflm Growth Reactor CV = Crystal Violet Staining 23 As evident in the literature, a variety of biofilm forming microorganisms have been evaluated by EIS, which may explain the discrepancy in impedance measurements, possibly due to difference in the electrochemical properties of different biofilms. A variety of experimental formats (i.e., static vs. continuous flow, batch vs. flow reactor), and sensor electrode geometries may also influence results. Most studies utilized bare gold electrodes, which are known to experience sensor drift. The duration of testing ranges from as little as 1 hr to 168 hrs. Many do not report on abiotic controls or lack the statistical assessment of differences in signals over time. Lastly, correlation of impedance data to standard biofilm techniques, such as CLSM imaging is often absent. Acknowledging the limitations in previous endeavors to implement EIS biosensing for biofilm monitoring, the research undertaken in this dissertation endeavors to mitigate these shortcomings and contribute to the advancement of the field. Knowledge Gaps There are significant deficits in the understanding of EIS biosensing for in situ monitoring of biofilms which need further study. One critical area that requires exploration is the reliability of sensors over extended periods of exposure to microbial contamination. Microbial colonization on surfaces has been shown to cause corrosion and material degradation, which can also occur on a sensor surface leading to compromised reliability and shortened lifespan, in a process called microbially influenced corrosion (MIC) (Beech and Sunner, 2004; McGlennen et al., 2019). The impacts of microbial fouling on microfabricated sensors must be understood further to optimize designs and coatings, and to minimize the effects of MIC. Although highly sensitive, EIS biosensors are susceptible to sensor drift, which refers to the gradual deviation of a sensor's output from its expected value or baseline over time (Roberts 24 and Sombers, 2018). Sensor drift can occur even in the absence of biologic material, causing shifts in the EIS spectra over time. This drift results from changes in the physical and chemical properties of the sensing elements, such as from surface imperfections or defects in the electrode structures resulting in non-uniform current distribution or alterations in the local electrochemical environment. Overall, minimizing drift in EIS sensors requires careful consideration of the sensor's design, fabrication, operation, and appropriate attention to signal processing and data analysis techniques. Further study is required to discern and reduce the impacts of sensor drift. An additional overlooked component of EIS biosensing studies is that majority of testing has been conducted under static conditions (no-flowrate) that might not reflect the realities of most naturally occurring biofilms. Opportunities to deploy a sensor in an actual fluid-flow environment would be a notable advancement in biofilm monitoring and research. Ultimately, the usefulness of an EIS biosensor lies in its ability to be used in a practical, real-world scenario to evaluate biofilm control and generate dependable data that can be utilized to take appropriate actions. This goal is especially challenging since biofilms flourish in harsh environments that are often difficult to access. An assessment on the capabilities of EIS biosensors applied in a real-world or simulated environment would drive the advancement of this innovative technology. Research Goals, Hypotheses, and Objectives Broadly, this work aims to develop and evaluate the use of microfabricated sensors as a novel methodology for real-time characterization of biological phenomenon. Specifically, this work aims to evaluate liquid environments that support microbial contamination and biofilm growth. The overarching goal of this work is to develop a highly-sensitive microfabricated sensor 25 and platform for in situ microbial analysis of biofilm-forming bacteria. The central hypothesis is that microfabricated EIS biosensors can reliably monitor biofilm in situ. To test this hypothesis, an optimized EIS biosensor is designed and fabricated, and a platform to validate sensor data is developed. Using a model biofilm-forming bacterial strain, the sensor is tested against biofilm growth, suppression, and dispersal in a variety of fluids. To support this overarching goal of developing a microfabricated sensor system, four objectives are considered: Objective 1: Investigating and mitigating the negative effects of biofilm on commonly used sensor materials. This objective aims to evaluate the severity of degradation of commonly used sensor thin film materials due to abiotic and biofilm-induced corrosion. The study seeks to quantify how microfabricated thin films degrade over time in a standard biofilm growth reactor, assess changes to the thin films’ material properties, and consider potential solutions for improving sensor design. To address this objective, material degradation caused by Escherichia coli K12 biofilm growth is determined by monitoring changes to electrical sheet resistance measurements (collinear four-point-probe) and Fourier-transform infrared spectroscopy (FTIR) absorption spectra over a time period of 7 weeks. Aluminum (metallic thin-film material) corroded significantly, whereas a- SixNy:H (dielectric thin-film material) remained unaffected by MIC. It is concluded that to mitigate the negative effects of biofilm induced corrosion on sensors, an encapsulation material such as a- SixNy:H can be utilized to protect the underlying conductive sensor material, and thus improve sensor durability. This objective is investigated in Chapter 2. Once the corrosive nature of biofilm is characterized, the next task is to develop corrosive resistant, reliable, sensor design. 26 Objective 2: Optimizing/enhancing a novel sensor design and developing an integrated platform to evaluate planktonic and biofilm growth under continuous flow conditions. This objective aims to improve sensor design by enhancing measurement stability and reducing sensor drift. An additional aim is to design a custom biofilm growth reactor flow cell system which integrates microfabricated sensors with CLSM imaging, a standard biofilm microscopy technique. To address this objective, unmodified microfabricated sensors are evaluated under abiotic conditions resulting in unpredictable drift. Sensors modified with PPy:PSS coatings mitigate drift and result in highly stable time-resolved EIS measurements. An optimal PPy:PSS coating thickness is selected. A novel 3D-printed flow cell system that integrates sensors and enables real- time EIS measurements with simultaneous CLSM imaging during biofilm growth is designed. Furthermore, the system operates under continuous flow conditions, mimicking naturally occurring biofilm. It is concluded that during biofilm growth, this system successfully links continuous EIS data with CLSM imaging and will serve as the experimental platform for subsequent objectives. These aims are investigated in Chapter 3. Given the successful development of a continuous flow platform with seamless CLSM and EIS integration, the subsequent logical progression is to evaluate the sensor's responses to biofilm growth. Objective 3: Understanding the relationship between biofilm development and sensor responses by correlating EIS data with a standard biofilm microscopy technique. This objective aims at collecting sensor signals throughout the stages of biofilm development. Using the flow cell system, EIS data is captured, and correlated with qualitative and quantitative CLSM assessments of biofilm. 27 To address this objective, real-time impedance measurements of biofilm growth under flow conditions were carried out using the novel 3D-printed flow cell system developed in objective 2. Pseudomonas aeruginosa PA01 served as the model biofilm-forming bacteria. Simultaneous EIS and CLSM were collected during biofilm growth from planktonic cell attachment through maturation. It is concluded that EIS microsensor data collection, in parallel with CLSM visualization, can reproducibly detect increasing levels of biofilm development. This objective is investigated in Chapter 3. Given the success of the sensor platform in correlating biofilm growth to EIS and CLSM, the subsequent logical progression is to evaluate the sensor's efficacy in detecting biofilm dispersal maneuvers. Objective 4: Demonstrating the utility of EIS biosensors during biofilm dispersal and suppression. The aim of this objective is to test the utility of the biosensors to monitor biofilm control interventions. Within this aim, sensors are tested during biofilm removal treatments with commercially available disinfectants. Additionally, sensors are tested to assess the effects of cell- signaling inhibition on biofilm. To address this objective, the novel 3D-printed flow cell system developed in objective 2 is used to monitor biofilm responses to antimicrobial treatments and a bacterial cell signaling suppressive agent. It is concluded that EIS microsensor data collection, in parallel with CLSM visualization, can reproducibly detect biofilm dispersal and suppression of cell signaling. This objective is investigated in Chapter 4. Given the success at detecting growth and dispersal in standard growth solution, the next logical step is to assess its utility in other biofilm relevant fluid environments. 28 Objective 5: Demonstrating the utility of EIS biosensors in an industrial fluid and in a natural setting. This objective aims at expanding the utility of the sensors outside of the laboratory. One aim is to test the sensors in an industrially relevant fluid prone to biofilm contamination. Sensors are tested for their ability to detect growth, dispersal, and inhibition in this fluid. A second aim involves testing the sensors to detect microbes in nature, specifically on a snow-patch and in an alpine pond. To address the first part of this objective, the novel 3D-printed flow cell system developed in objective 2 is used to monitor biofilm growth, dispersal, and suppression responses in an industrially relevant oil-water emulsion fluid. It is concluded that in an oil-water emulsion, EIS microsensor data collection can reproducibly detect biofilm dispersal and suppression of cell signaling. To address the second part of this objective, a field-test of the sensor is conducted in an icy environment. It is concluded that under icy condition in a field test, EIS microsensors can differentiate varying concentrations of microbes. This objective is investigated in Chapters 4 and 5. 29 EFFECTS OF ESCHERICHIA COLI K12 BIOFILM ON SENSOR THIN FILM MATERIALS Contribution of Authors and Co-Authors Manuscript in Chapter 2 Author: Matthew McGlennen Contributions: Investigation, Conceptualization, Methodology, Formal analysis, Data curation, Writing – original draft. Co-Author: Markus Dieser Contributions: Conceptualization, Formal analysis, Writing – review & editing. Co-Author: Christine M. Foreman Contributions: Conceptualization, Supervision, Funding acquisition, Writing – review & editing. Co-Author: Stephan Warnat Contributions: Conceptualization, Supervision, Funding acquisition, Writing – review & editing. 30 Manuscript Information Matthew McGlennen, Markus Dieser, Christine M. Foreman, Stephan Warnat 2019 IEEE Sensors Status of Manuscript: ____ Prepared for submission to a peer-reviewed journal ____ Officially submitted to a peer-reviewed journal ____ Accepted by a peer-reviewed journal x Published in a peer-reviewed journal Publisher: IEEE Published: Jan 14, 2020 DOI:10.1109/SENSORS43011.2019.8956741 31 Abstract Micro-fabricated sensors enable the study of chemical and physical dynamics in aqueous environments such as rivers, lakes or oceans at low cost. Sensors must work reliably in these environments, which include both biological and chemical challenges. However, sensor thin films have not been studied in detail for aqueous applications, and more specifically how biotic interactions may change sensor material properties. In this study, the long-term effects of biofilm formation on the properties of aluminum (electric conductor) and a-SixNy:H (insulating material) were investigated. Material degradation caused by Escherichia coli K12 biofilm growth was determined by electrical sheet resistance measurements (collinear four-point-probe) and Fourier- transform infrared spectroscopy (FTIR) absorption spectra over a time period of 7 weeks. Changes of the surface topography were tested using scanning electron microscopy (SEM) and white light interferometry. Aluminum was found to be heavily degraded at three weeks, whereas a-SixNy:H was inert during the entire investigation period. As differences between thin film sensor materials are evident, more detailed investigations including a broader range of materials should be explored. Introduction Biofilms form on many types of substrata, including natural and engineered surfaces. When conditions allow, bacterial cells colonize surfaces and become embedded in extracellular polymeric substance (EPS) (Azeredo et al., 2017). Within minutes of substrate submersion in aqueous environments, microbial growth and EPS production are initiated, leading to the development of biofilm (Videla & Herrera, 2005). Biofilms can inhibit the performance of sensors used in studying the dynamics of aqueous environments. Microbial activity leading to deterioration 32 of surfaces containing biofilm is termed bio-corrosion or microbially influenced corrosion (MIC) (Beech and Sunner, 2004). For instance, Pseudomonas aeruginosa, a bacterial strain that forms corrosive biofilms (Li et al., 2016), has been shown to alter surface and material properties of nickel-copper and nickel-zinc thin films (San et al., 2014). Escherichia coli has been shown to accelerate corrosion in marine like environments on aluminum alloys (Pratikno & Titah, 2017), likely due to the secretion of metabolic organic acids (Ashton et al., 1973). Surface micromachined sensors are based on the deposition of metallic and insulating materials. An inert interface between the active sensor surface and the aqueous media is required for reliable operation (Warnat et al., 2015). TiN and SiN have been reported to be inert in aqueous media, and still allow fully functional device operation. However, bio-corrosion/degradation of these materials has yet to be determined. Moreover, mechanisms to protect electrically conductive materials from environmental effects, and to improve sensor material reliability awaits investigation. Table 2.1 summarizes previous work on material and sensor reliability in aqueous environments. The present study examines the effects of biofilm growth on the material properties of two thin film materials. The focus was given to electronic conductivity, surface morphology, and bulk stoichiometric properties, and how they change as a function of exposure to biofilms. Materials and Methods Sample preparation Samples of aluminum were deposited using physical vapor deposition (PVD) with thermal evaporation using Modu-Lab PVD. Amorphous silicon nitride with incorporated hydrogen (a- SixNy:H) was deposited using PECVD with a chuck temperature of 100ºC. This low deposition temperature allows a-SixNy:H to be used as an inert encapsulation layer for most electrically 33 conductive materials, comparable to the use of aluminum herein. Higher deposition temperatures could cause chemical reactions between the material and the substrate (Sedky et al., 2001). Also, lower deposition temperatures increase the hydrogen concentration in the a-SixNy:H film through the use of NH3 and SiH4 reaction gasses. As an extreme example, non-stoichiometric a-SixNy:H films made from PECVD were studied for their ability to be used as an encapsulation layer for sensor applications in aqueous media by studying long term reliability under exposure to biofilm. PECVD a-SixNy:H films were deposited at the Utah Nanofabrication Facility. Samples were cleaved from the wafers and characterized prior to experimentation (see material characterization). Table 2.1: Previous work on material and sensor reliability in aqueous environments Subject Methods Results Ref Atomic layer deposition (ALD) Thermal actuation in Performance of (Warnat et post-processing of Al203 and aqueous MEMS structures al., 2015) TiO2 to encapsulate poly multi- media/seawater, and enhanced with user MEMS processing systems displacement ALD coating for (PolyMUMPs) devices for measurements long term use aqueous environments Dissolution kinetics of e-beam, AFM, ellipsometry, Combination (Kang et al., low-pressure chemical vapor and resistance changes encapsulation 2014) deposition (LPCVD) and layers plasma-enhanced chemical demonstrate vapor deposition (PECVD) SiO2 increased and a-SixNy:H thin films in electronic device aqueous media lifetime Biocorrosion of Al 6063 by E. Corrosion rates, and E. coli accelerates (Pratikno coli in seawater microscopy aqueous corrosion and Titah, rates 2017) MIC of Ni-Zn and Ni-Cu thin SEM, polarization, and P. aeruginosa (San et al., films from P. aeruginosa EIS increased 2014) corrosion of Ni- Cu alloy but protected Ni-Zn 34 Culture Conditions and Biofilm Reactors Center for disease control (CDC) biofilm reactors (Biosurface Technologies, Bozeman, MT) were used to grow biofilm following standardized methods (Goeres et al., 2005). Reactors were autoclaved for 30 min. Each reactor contained 350 mL of 1X tryptic soy broth (TSB; BD Bacto) media. Aluminum and a-SixNy:H coated wafers were cut into approximately 1x1 cm samples, and sterilized in a 30% hydrogen peroxide bath for 10 min, followed by 30 min in 70% ethanol. Sample wafers were then glued with a drop of silicone adhesive to polycarbonate sample holders, mounted to CDC reactor rods, and allowed to cure for 24 hours. To ensure sample wafers and rods were thoroughly sterilized, and to guarantee abiotic conditions for control experiments, final sterilization of the CDC reactor rod assemblies involved exposure to UV light (254 nm) for two hours. All steps were performed inside a biosafety cabinet. Samples were prepared in triplicates per time point. E.coli K12 was grown in TSB at 37°C while shaking at 125 rpm for 16 h, following standard methods (Sezonov et al., 2007). Subsequently, CDC reactors filled with 350 mL of autoclaved 1X TSB media were inoculated with 2 ml of bacterial enrichment. Inoculated reactors were run for 24 hours in batch mode at 22 °C on a stir plate at 125 rpm, followed by continuous flow operation of a 1:10 dilution of the TSB medium at a flow rate of 2 ml/min for 7 weeks (Figure 2.1). Abiotic control reactors were included using the same CDC reactor setup filled with either autoclaved deionized water (DI) or 1:10 diluted TSB media both operated in batch mode for 7 weeks. Triplicate sample wafers were collected at the beginning and 3, 5, and 7 weeks of operation. At each time point, sample wafers were removed from the polycarbonate sample holder with a sterile razor blade and transferred into 50 ml Falcon tubes containing 5 ml of 1X phosphate buffer saline solution (PBS). Sample wafers were placed into 50 ml Falcon tubes filled with 4.5g/L Tween 35 80 in 1X Dulbecco’s phosphate-buffered saline (DPBS; Thermo Fisher), 5 ml of lab grade acetone, and 5 ml of isopropanol to remove biofilms. Sample wafers were sonicated at 60 W for 10 min at 22 °C, followed by a deionized water rinse. After biofilm removal, the samples were air-dried and stored in sterile Petri dishes at 22 °C until further analyses. Figure 2.1: CDC biofilm reactor containing E. coli K12 biofilm growing on aluminum and silicon nitride substrate samples. Materials Characterization Techniques used to determine changes of the coated aluminum wafers included collinear four-point-probe sheet resistance measurement with a Keithley 2450 SourceMeter, scanning electron microscopy (SEM; Integrated Auger nanoprobe based on Physical Electronics 710), and energy-dispersive x-ray analysis (EDX; Bruker X-Flash 6I10). Wafers coated with a-SixNy:H were analyzed with Fourier transform infrared spectroscopy (FTIR) attenuated total reflectance absorbance (Thermo Fischer Scientific Nicolet iS10 with a Harrick VariGATR ATR attachment) as well as with Filmetrics Profilm 3D white light interferometry. 36 Results Collinear four-point-probe sheet resistance measurements of the aluminum sample wafers demonstrated a statistically significant increase in the sheet resistance after exposure to biofilm over the 7-week period (one way ANOVA: F(3, 236)=1185.3, P<0.001; Figure 2.2). At each time point investigated, an increase in sheet resistance was noticed, likely due to the corrosion of the film. Abiotic control studies show only slight increases in sheet resistance and showed no visual signs of degradation. Due to the high resistance and instability, a-SixNy:H sheet resistance values are not reported. Figure 2.2: Collinear four-point-probe sheet resistance of aluminum thin film measured at time intervals of 0, 3, 5, and 7 weeks in CDC biofilm reactor. 37 Apparent differences were observed between aluminum and a-SixNy:H coated samples exposed to E. coli biofilm. The original microstructure of the aluminum thin film (i.e., at the beginning of the experiment) was imaged using SEM (Figure 2.3a) and shows evidence of hillock formation on the surface (Bordo and Rubahn, 2012). After three weeks of exposure to biofilm, topographical features such as crevices and pits were abundant on the aluminum surface (Figure 2.3b). By week 7 much of the aluminum thin film was removed (Figure 2.3d) with only small amounts of aluminum remaining. EDX analysis confirmed the corrosion of aluminum film, showing the exposed underlying silicon substrate within a pit (Figure 2.3c). Figure 2.3: SEM images of aluminum PVD thin film surface a) as deposited showing aluminum hillocks on surface b) after 3 weeks in CDC biofilm reactor c) EDX analysis map of aluminum thin film surface after 3 weeks in biofilm reactor revealing underlying silicon substrate and d) SEM image of aluminum surface after 7 weeks in CDC biofilm reactor. 38 To study bulk structural changes on the a-SixNy:H samples, FTIR absorption spectra were measured over 500 to 3500 cm-1 (Figure 2.4). Spectra show both symmetric and antisymmetric Si-N stretching modes in all samples, regardless of the duration of exposure to biofilm. Relative absorbances remained the same throughout the specified time intervals and suggested no significant change in the film stoichiometry. These results suggest that a-SixNy:H is a promising material for encapsulating active sensor structures in aqueous media. Figure 2.4: FTIR spectra of a-SixNy:H thin films, measured at timepoints of 0, 3, and 7 weeks. White light interferometry data was collected on wafer samples from an as-deposited a- SixNy:H, and 7 weeks biofilm exposed sample (Figure 2.5). Area maps of sample surfaces were collected and show surface morphology and average roughness values. The as-deposited a- SixNy:H sample had an average roughness of 2.2 nm while the 7-week exposed sample had an average roughness of 2.3 nm. No visual differences were noticed between samples, further 39 suggesting that a-SixNy:H is a promising encapsulation material to protect metallic thin films used for aqueous sensing. Figure 2.5: White light interferometry image of A) as deposited a-SixNy:H thin film and B) after exposure to biofilm for 7 weeks. Conclusion Biofilm adhesion and growth affects material properties when attached to surfaces and allowed to propagate. In aqueous environments biofilm formation occurs on a wide variety of substrates and can lead to biofouling (Lappin‐Scott and Costerton, 1989). Conductive materials are very susceptible to corrosive biofilms and precautions should be made to mitigate damage that can occur due to prolonged exposure in aqueous environments. In the present study, biofilm adhesion and growth was shown to impact aluminum thin films. Aluminum thin films used in aqueous environments showed an apparent increase in sheet resistance, suggesting that they are not reliable for use in aqueous environments by themselves. Therefore, a protective barrier should be considered. Conversely, low deposition temperature a-SixNy:H coatings remained inert throughout the course of the experiment. The results show that this coating has the potential to protect conductive sensor materials from fouling and biological damage. These results suggest that 40 other sensor materials (e.g., gold, polycrystalline silicon, and nickel) should be investigated to determine their susceptibility to microbially influenced corrosion, and the ability of a-SixNy:H to encapsulate the materials to minimize material degradation. Future work should focus on higher resolution time-series investigations to determine when the materials are no longer viable for optimal sensor performance. 41 USING ELECTROCHEMICAL IMPEDANCE SPECTROSCOPY TO STUDY BIOFILM IN A 3D PRINTED FLOW CELL SYSTEM Contribution of Authors and Co-Authors Manuscript in Chapter 3 Author: Matthew McGlennen Contributions: Investigation, Conceptualization, Methodology, Formal analysis, Data curation, Writing – original draft. Co-Author: Markus Dieser Contributions: Conceptualization, Formal analysis, Writing – review & editing. Co-Author: Christine M. Foreman Contributions: Conceptualization, Supervision, Funding acquisition, Writing – review & editing. Co-Author: Stephan Warnat Contributions: Conceptualization, Supervision, Funding acquisition, Writing – review & editing. 42 Manuscript Information Matthew McGlennen, Markus Dieser, Christine M. Foreman, Stephan Warnat Biosensors and Bioelectronics: X Status of Manuscript: ____ Prepared for submission to a peer-reviewed journal ____ Officially submitted to a peer-reviewed journal ____ Accepted by a peer-reviewed journal x Published in a peer-reviewed journal Publisher: Elsevier Published: March 5, 2023 DOI: https://doi.org/10.1016/j.biosx.2023.100326 43 Abstract Biofilm contamination is a widespread issue that can occur anywhere when organisms attach to surfaces in the presence of water. In industrial environments, formation of biofilms can lead to component failure, material degradation, and biofouling or spoilage, which collectively come with significant economic costs. Microfabricated electrochemical impedance spectroscopy (EIS) sensors have emerged as a promising tool for monitoring biofilm as EIS sensors capture information about biofilm growth autonomously in real-time; however, sensors suffer from drift, and the technique lacks temporal interpretation of dynamic biofilm processes. In this work, microfabricated sensors featuring gold micro-interdigitated electrodes (µIDEs) were modified with an electrically conductive polymer layer resulting in EIS measurement variability that was significantly reduced compared to unmodified sensors, and enabled highly stable, time-resolved EIS measurements. EIS characterization of Pseudomonas aeruginosa biofilm in parallel with high- resolution confocal laser scanning microscopy confocal laser scanning microscopy (CLSM) was performed using a novel 3D-printed flow cell system, resulting in distinct changes to EIS data corresponding with consistent biofilm growth. We have shown that EIS microsensors can detect four stages of biofilm: (i) initial biofilm attachment to the sensor substrate, (ii) early-stage irreversible biofilm proliferation characterized by sparse biofilm coverage, (iii) mature biofilm detection characterized by uniform biofilm coverage, and (iv) changes due to detachment and regrowth of biofilm. 44 Introduction Most microorganisms exist as organized communities attached to each-other or surfaces embedded in a matrix of extracellular polymeric substance (EPS); creating structures known as biofilms (Azeredo et al., 2017; Costerton et al., 1999). Biofilms can profoundly affect human health, industrial processes, and natural and engineered systems. Collectively, biofilm associated costs have been estimated at around $4 trillion per year globally (Cámara et al., 2022). Biofilms are more resilient against mechanical and chemical treatments compared to planktonic cells, rendering common treatment options ineffective. Strategic solutions for in situ sensing of biofilms in environmental, medical, and industrial settings are needed to implement treatments more effectively so that biofilms can be reliably monitored. Measurement and analysis techniques using quartz crystal microbalance (QCM), open circuit potentiometry (OCP), voltammetry, or amperometry sensors allow for nondestructive monitoring and in situ analysis of biofilms at various stages of growth (Becerro et al., 2016; Poma et al., 2020; Ripa et al., 2020). However, these applications are susceptible to ambient vibrations and noise, or are limited to analyzing electroactive analytes, respectively (Ding et al., 2020; Grieshaber et al., 2008). Recently, electrochemical impedance spectroscopy (EIS) has shown potential in detecting biofilm growth in situ and has significant advantages over the previously mentioned techniques. EIS has the sensitivity to gather information occurring at the solid-liquid interface and in the bulk solution in real-time without the need of complex reagents and bulky instrumentation. EIS devices can be miniaturized using microfabrication techniques, which increase signal-to-noise, and improve sensitivity compared to conventional electrode geometries. In general, EIS applies a sinusoidal electrical perturbation (i.e., voltage or current) to an 45 electrochemical system across a range of frequencies and measures the electrical response (i.e., voltage or current), termed complex impedance (Vivier and Orazem, 2022). EIS spectra can be divided into several domains where certain regions embody electrochemical parameters of a system under test. For instance, EIS spectra may show impedance artifacts due to double-layer capacitances, diffusional impedances, solution resistances, and others (Magar et al., 2021). During an EIS measurement containing microorganisms, changes in EIS spectra over time may indicate the attachment, presence, and growth of microorganisms on the electrode surface. Among EIS techniques, the use of micro interdigitated electrodes (µIDEs) for biosensing applications have stood out for their simplistic design and fabrication, high sensitivity, high signal- to-noise ratio, and their use without the need of a reference electrode during measurements (Dorledo de Faria et al., 2019; Furst and Francis, 2019; Park et al., 2018). For microfabricated EIS sensors, gold is commonly used as an electrode material due to having good chemical stability, low reactivity, good biocompatibility, and high electrical conductivity. While gold µIDEs have been used for the detection of planktonic bacteria (Kim et al., 2012; Simic et al., 2020; Yang et al., 2004), biofilm (van Duuren et al., 2017; Ward et al., 2018), and microbes in icy-conditions (McGlennen et al., 2020), several research limitations can be noted. Most of these studies were performed under static conditions with no fluid shear and may have reached nutrient limitations. Further, confirming the presence of biofilm on µIDEs has largely relied on biofilm-destructive techniques at a finite time point and lack a temporal interpretation of EIS during cell attachment and biofilm development (Goikoetxea et al., 2018; Paredes et al., 2012, 2013, 2014a, 2014b). Electrochemical measurements, including the use of gold µIDEs, suffer from nonlinear, long-term baseline instability, (i.e., signal drift), which decreases measurement reproducibility and leads to 46 diminished quantitative confidence (Roberts and Sombers, 2018). Lastly, a clear differentiation and interpretation between biotic and abiotic effects on EIS measurements through continuous measurements has been ambiguous or lacking (Blanco-Cabra et al., 2021). Owing to the complex behavior of biofilms, a single analytical technique is often not sufficient to fully characterize the dynamics of biofilm growth. Herein, microfabricated EIS sensors were combined with live-cell confocal laser scanning microscopy (CLSM) imaging to achieve the analytical depth necessary for interpreting EIS data from biofilms over different stages of growth. The objectives of this work were to: (i) minimize EIS signal drift and improve the measurement stability of microfabricated EIS sensors, (ii) develop a novel 3D printed flow cell system for capturing EIS and CLSM images during biofilm growth, (iii) identify the optimal single-frequency impedance measurement that best characterizes biofilm growth, and (iv) validate the suitability of EIS sensors for the real-time detection of biofilm growth. Material and methods Sensor fabrication Sensors were fabricated using standard microfabrication techniques of physical vapor deposition, lithography, and wet etching processes with minor modifications (Neubauer et al., 2019). A 10 nm titanium adhesion promoter was used in lieu of chromium because of its increased resistance to dissolution, and limited oxidation potential (Hoogvliet and van Bennekom, 2001). The titanium film was deposited using electron beam physical vapor deposition in an Angstrom Thermal Evaporator (AMOD, Oxford Instruments, USA). Proceeding gold etching, the titanium adhesion promoter was etched by immersion in 40:1:1 H20:HF:H202 solution for 12 s, followed by deionized water (DIW) rinse. The µIDEs were designed to be 15 µm in width with a spacing of 10 47 µm, for a total of 50 electrode pairs. To allow only the µIDEs region of the sensors to participate in EIS, the electrical connection leading up to the µIDEs were encapsulated with SU-8 3010 (Kayaku Advanced Materials, Boston, USA) and were patterned with lithography (Figure 3.1b). The gold-patterned wafers were cleaned with O2 and Ar+ plasma for 5 min at 600 W in Ion Wave 10 Batch Plasma System (PVA Tepla, USA), and were immediately spun coated with SU-8 3010 photoresist at 500 RPM for 10 s, 4000 RPM for 30 s, soft baked at 65⁰C for 60 s and then 95⁰C for 600 s. The wafers were exposed on the contact aligner (ABM inc., USA) for 360 mJ cm-2 to a mylar photolithography mask (Cad Art Services; Bandon, OR). Post-exposure bake was performed at 65⁰C for 60 s, 95⁰C for 180 s, and 65⁰C for 30 s, respectively. The wafers were developed in SU-8 developer (Kayaku Advanced Materials, Boston, USA) for 360 s while shaking on an orbital shaker at 60 RPM, followed by a DIW rinse. Finally, the wafers were hard baked at 125⁰C for 1200 s on a hotplate. Additionally, sensors were equipped with a crosshair feature (Figure 3.1a-2) to streamline locating the center of the µIDE during CLSM analysis (see CLSM imaging and data analysis below). The wafers were diced into sensors measuring approximately 8.7 mm wide × 26 mm long using a Disco DAD3221 Dicing Saw (Disco Corp., Japan). The summary of the fabrication process can be found in Supplemental Figure A.1. The microfabricated sensors in this work featured three transducer elements within the same footprint (Figure 3.1), however the primary focus of this study was to use µIDEs. 48 Figure 3.1: Optical microscope images of microfabricated sensors and their features. (a) Layout of the microfabricated sensor. (a- 1) RTD sensor, (a-2) “crosshair” feature used for finding the center of IDE structures (a-3) μIDE structures, (a-4) ring impedance sensor. (b) μIDE sensor before electropolymerization. (c) μIDE sensor showing both bare gold WE and PPy:PSS coated CE/ RE regime. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Sensor electropolymerization Sensors were cleaned by placing a 100 µL drop of 7:2:1 H2O:NH4OH:H2O2 on the µIDEs for 1 hr, followed by DIW rinse (Kim et al., 2008). Electropolymerization of one side of the µIDEs was adapted from (El Hasni et al., 2017). Briefly, 0.2 M poly (4-styrenesulfonic acid) solution (PSS; Mw ~75,000, 18 wt% in water, Sigma-Aldrich, USA) and 0.2 M pyrrole (PPy; 98%, reagent grade, Sigma-Aldrich, USA) were freshly prepared in DIW. A 100 µL drop of 0.2 M PSS:0.2 M PPy was placed onto the surface of the µIDEs. Two different electrode modification procedures (i.e., one thin, and one thick) were evaluated to optimize the coating parameter that best minimizes measurement variability and drift. Electrodeposition was carried out under galvanostatic conditions with a current density of 0.1 mA cm2 applied to one side of the µIDEs using a Keithley 2450 source measure unit (SMU). The surface area of one side of the µIDEs was calculated as 49 0.015 cm2, which led to deposition times of either 113.33 s or 300.00 s. Resulting deposition charges were 170 µC and 450 µC, respectively (Figure 3.1c). Design and fabrication of CLSM flow cell system A flow cell system was designed for monitoring biofilm growth on top of the surfaces of the µIDEs so that simultaneous high-resolution live CLSM images and EIS data could be collected. The design of the flow cell system consisted of four components: (i) a base to mount the flow cell system to the CLSM microscope, (ii) the flow cell body, (iii) three microfabricated EIS sensors, and (iv) three electrical connectors (Figure 3.2). 3D CAD of the flow cell system and its associated components were made in Autodesk Inventor 2021 and Eagle (Autodesk, USA). The base was milled out of a 6061-aluminum sheet to fit the clamp width of the CLSM stage. The flow cell body and the electrical connection holders were 3D printed using an Epax X1 4k printer with biocompatible poly lactic acid (PLA) liquid resin (eSUN PLA pro, China). Structurally, the flow cell body was 100/40/10 mm (W/H/D) and housed three independent flow chambers (~35 µL each). Inlet and exit ports were vertically mounted and set at differing heights to minimize bubble formation. A 25 × 75 × 1 mm glass microscope slide (Fisher Scientific, USA) served as a viewing window and was adhered to the flow cell body using silicone (DAP, USA). The planar profile of the flow cell body enabled uninterrupted movement along the X-Y direction on the microscopy stage. 50 Figure 3.2: Design of the flow cell system with integrated microfabricated sensors used for biofilm growth experiments. (a) Picture of flow cell system prior to experiment. (b) Disassembled view of flow cell system revealing its four major components: (top piece) 3D printed flow cell body with luer-lock and barb-connectors, (middle pieces) silicone gaskets, sensors, and support blocks, (bottom piece) aluminum base for mounting to CLSM, and (electrical connection) PCB boards and holders. (c) The flow cell mounted beneath a wet immersion objective on CLSM, (c-1) influent barb connectors and silicone tubing, (c-2) injection ports, (c-3) effluent barb connectors and silicone tubing. The flow cell body was fitted with three microfabricated EIS sensors and sealed with laser- cut silicone gaskets (0.01” High-Purity High-Temperature Silicone Rubber, McMaster Carr, USA) pressed between laser cut acrylic blocks and the flow cell body. Plastic threaded barb connectors were used to connect 0.8 mm I.D. silicone tubing (P/S #13, Masterflex, USA) both for the inlet and outlet (Figure 3.2c-1 and 3.2c-3). Cell injection ports (Fisher Scientific, USA) were equipped 51 with luer-lock connectors and placed in line with the influent tubing (Figure 3.2c-2). Electrical connection was achieved using custom-made printed circuit boards (PCBs) and pogo-pins (Mill- Max, USA), which allowed reliable and non-permanent electrical connection to each sensor (Figure 3.3a). The PCBs were equipped with ultra-miniature connectors (UMC) and were connected to a Hioki IM3533 LCR meter with UMC to subminiature version A (SMA) cables (Cinch Connectors; USA). The entire assembly fastens to the custom-manufactured aluminum base using six M3 socket stainless steel cap screws. Computational fluid dynamics (CFD) analysis was performed to visualize the flow profiles and wall shear stresses within the flow chambers containing a microfabricated sensor using COMSOL Multiphysics Version 5.3. A physics controlled fine mesh was used for the simulation. Incompressible and steady-state flow conditions were simulated using the experimental flowrate of 1 µL min-1. EIS measurements and data analysis EIS measurements were carried out with a Hioki IM3533 LCR meter connected to a Keysight 970A data acquisition system (DAQ) with a DAQ905A RF 2 GHz Dual 1:4 RF Multiplexer, 50 Ohm switching module. The instruments were controlled by a custom-built virtual interface (VI; LabView 2017), providing multiplexed four channel EIS measurements. EIS was collected every 30 min across a frequency range of 200 kHz to 100 Hz and 25 logarithmically spaced data points. The signal source was a sine wave with an amplitude of 10 mV RMS and no DC bias. During EIS data collection sensors coated with PPy:PSS were connected to the LCR meter such that the coated side of the µIDEs served as both the counter electrode (CE) and 52 reference electrode (RE; Figure 3.1c).Time-dependent changes to impedance were reported given by equation (3.1): |𝑍(𝑓,𝑡)|−|𝑍(𝑓,𝑡0)| Impedance Change (%) = ∗ 100% (3.1) |𝑍(𝑓,𝑡0)| Where |Z(f,t0)| is the impedance measurement at the frequency of interest (f), measured at the initial timepoint (t0), and |Z(f,t)| is every subsequent impedance measurement. All EIS data were analyzed in Matlab 2021a and statistical differences were determined by Student’s T-test. Characterization of PPy:PSS coated sensors To evaluate sensor variability and stability, four µIDEs of each of three different coatings (i.e., 0, 170, and 450 µC terminal deposition charges of PPy:PSS) were submerged in sterile 1:10X tryptic soy both (TSB; Fisher Scientific, USA), inside of 1.5 mL microcentrifuge tubes for 18 hrs while logging EIS every 30 min (see above EIS measurements and data analysis). Impedance changes were measured at 100 Hz to monitor time-dependent changes to low-frequency impedance dominated by the double-layer capacitance. EIS sensor variability and stability at 100 Hz were evaluated statistically using Levene’s Test for equality of variances between the three coating parameters at t0 and between t18 and t0. Cell culturing conditions A green fluorescent protein (GFP) labelled Pseudomonas aeruginosa PA01 was used for all biofilm related experiments (Nivens et al., 2001). Cell cultures were grown in the presence of 50 µg mL-1 carbenicillin (Fisher Scientific, USA) plates. Freezer stocks were transferred to tryptic soy agar (TSA; Fischer Scientific, USA) and were incubated at 37⁰C for 24 h prior to experimentation. Cell culture enrichments were made from single colonies and were grown in 1X 53 TSB at 37⁰C while shaking at 150 rpm for 18 h. Subsequently, enrichments were harvested at 10,000×g for 5 min. The resulting cell pellets were resuspended in 1:10 X TSB and incubated at 22⁰C while shaking at 150 rpm for 6 hrs to allow cells to adjust to growth conditions selected for the flow cell. Dilute media was used because biofilm growth and EPS production are more substantial under depleted nutrient environments (Zhang et al., 2014). Enrichments were harvested at an optical density of OD600=0.2, resulting in a cell concentration of ~1x108 colony forming units (CFUs) mL-1. Enrichments were transferred to 10 mL sterile luer lock syringes (Becton Dickinson, USA), equipped with 20 ½ ga. needles (Fisher Scientific, USA) prior to injection into the flow chambers. Flow cell system operation The three flow chambers were sterilized by infusing 70% ethanol for 20 min followed by a DIW rinse. Biofilm growth was initiated by first filling each flow chamber with 1:10X TSB from individual 10 mL syringes. A continuous flowrate of 1 µL min-1 was provided to each flow chamber for 2 h. Next, the flow chambers were inoculated with 1 mL of cell culture enrichments (~1x108 CFUs) by aseptically inserting the syringe needle through the injection ports (Figure 3.2c) (seeding phase). Flow was paused during seeding for 2 h to allow cell attachment onto µIDEs. Subsequently, sterile 1:10X TSB was supplied to the flow chambers at a continuous flowrate of 1 µL min-1 for 48 h to remove excess planktonic cells in the bulk solution and to promote biofilm formation (growth phase). All experiments were performed in triplicate at 22⁰C. Abiotic controls were performed following the same procedure, minus the injection of cell culture enrichments. Before inoculation of the flow chambers with cell culture enrichments, an initial EIS measurement of the cell free culture media (1:10X TSB) was collected, which served as the 54 background impedance (|Z(f,t0)|). After infusion of cell culture enrichment into the flow chambers, EIS data were recorded every 30 min (|Z(f,t)|). A summary of the operation of the flow cell for biofilm growth is shown in Supplemental Figure A.2. CLSM imaging and data analysis Biofilm images were acquired using CLSM (SP5 upright confocal, Leica, USA) and analyzed in Imaris 9.8 (Bitplane, Switzerland). To minimize quantitative subjectivity and bias, biofilms were imaged at the same X-Y coordinates for each timepoint. Representative 3D CLSM images were captured using 40X (Leica, 3.3 mm WD, 0.80 NA) magnification. Total biovolume from each CLSM Z-stack was calculated using Imaris 9.8 “surfaces” analysis feature with absolute fluorescence intensity above 20.0, voxel absolute fluorescence intensity above 4.0, and eliminating particle sizes smaller than 2 µm in diameter as filter criteria. Data were normalized to the field of view by using equation (3.2) similarly to (Lim et al., 2016), creating a biofilm index. ∑ 𝐵𝑖𝑜𝑣𝑜𝑙𝑢𝑚e (µ𝑚3) Biofilm Index = (3.2) Field of View (µ𝑚2) Differences in biofilm indices were analyzed using a Student’s T-test to compare differences in means. Results Characterization of electropolymerized µIDEs A bode plot of impedance and phase angle for all three coatings parameters (i.e., 0, 170, and 450 µC PPy:PSS) of electrodeposited µIDEs displayed typical resistive and capacitive behaviors for electrode/electrolyte systems. The spectra were characterized by three distinct regions: a low-frequency capacitive region (i.e., <2 kHz), a middle frequency capacitive and 55 resistive region (i.e., 2 kHz20 kHz) (Supplemental Figure A.3). Below 2 kHz, phase angle reached its minimum, and impedance reached its maximum. In the middle-frequency region impedance and phase angle were reversed with higher frequencies. Above 20 kHz, the phase angle was highest, and the magnitude of impedance reached its minimum. Differences in EIS spectra were evident between the three coatings. Sensors coated with deposition charges of 170 µC and 450 µC PPy:PSS resulted in a clear reduction in impedance from 100 Hz to ~150 kHz. Conversely, 170 µC and 450 µC PPy:PSS coated sensors' phase angles were collectively higher than the 0 µC PPy:PSS coated sensors across all frequencies (Supplemental Figure A.3). Of relevance, the phase angle for each of the three coating parameters approached - 90⁰ below 2 kHz, where capacitive effects (i.e., electric double-layer capacitance and electrode polarization effect) were the dominating signal. For this reason, 100 Hz was chosen to further characterize sensor variability and stability. Sensor variability and stability in abiotic media EIS measurements of µIDEs with and without electrically conductive PPy:PSS coatings were recorded over 18 h in sterile 1:10X TSB to determine measurement stability over time at 100 Hz (Figure 3.3; Supplemental Figure A.4). Impedance signals for uncoated sensors (i.e., 0 µC PPy:PSS) either increased or decreased relative to t0 with no predictable trend. Sensors with PPy:PSS coatings behaved unidirectionally, with a reduction in total sensor drift variability related to coating thickness. 56 Figure 3.3: Evaluation of sensor variability in abiotic 1:10X TSB from sensors coated with PPy:PSS with deposition charges ranging from 0 to 450 μC. (a) Normalized impedance at t0 for 100 Hz from sensor batches coated with deposition charges of 0 μC, 170 μC, and 450 μC PPy:PSS. (b) Normalized impedance at t18 with respect to t0 for 100 Hz from sensor batches coated deposition charges of 0 μC, 170 μC, and 450 μC PPy:PSS. Values for bare gold are indicated by “0 μC” deposition charge. Please note different scaling on y-axes. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Figure 3.3a shows normalized impedance at 100 Hz for each of the three coatings immediately (i.e., t0) after sensors were submerged in 1:10X TSB. Differences in the variability of 57 the three conditions of the sensors at t0 were evident. At t0, variability of impedance for both 0 µC PPy:PSS and the 170 µC PPy:PSS coated sensors varied largely from 62.1 to 181.8%, and 82.1 to 114.9%, respectively. When coated with 450 µC PPy:PSS, impedance variability at 100 Hz extended from only 97.3 to 101.9%. The variation of normalized impedance of 170 µC PPy:PSS coated sensors was not significantly different from the 0 µC PPy:PSS coated sensors (Levene’s Test, F=3.2259, DF=7, P value=0.123). Conversely, the 450 µC PPy:PSS coated sensors were significantly different from the 0 µC PPy:PSS (Levene’s Test, F=7.4285, DF=7, P value=0.034) and the 170 µC PPy:PSS coated sensors (Levene’s Test, F=89.7715, DF=7, P value<0.001). Impedance variability at 100 Hz for the three differently coated sensors submerged in 1:10X TSB for 18 h (i.e., t18 with respect to t0) mirrored the initial response (Figure 3.3b). At t18, impedance drift for the coated sensors were bidirectional. Further, impedance drift for both the 0 µC PPy:PSS and 170 µC PPy:PSS coated sensors varied largely over a range of 58.4% to 28.0%, respectively. Conversely, when coated with 450 µC PPy:PSS, the variability of impedance drift was narrow, spanning 2.0% across a slight upward trajectory. The variability of sensors coated with 450 µC PPy:PSS were significantly different from 0 µC PPy:PSS coated sensors (Levene’s Test, F=8.3456, DF=7, P value=0.028). Therefore, each sensor was functionalized with 450 µC PPy:PSS for the remainder of our study. Flow chamber fluid simulation Fluid flow velocity-profiles and wall shear stresses within the flow chambers were simulated at a steady-state flow rate of 1 µL min-1 (Figure 3.4). Cross-sections through the longitudinal plane of the transverse flow profile showed a homogeneous distribution of flow within the active sensor region (Figure 3.4a and 3.4b). The flow velocities along the sensor surface did 58 not exceed 5 µm s-1, and no eddies were apparent at any position within the boundaries of the flow chambers. Simulated wall shear at the interface of the sensor surface was found to be evenly distributed in the active sensor area averaging ~20 µPa (Figure 3.4c). Higher wall shear at the inlet and outlet of the flow chambers were negligible as shear dissipated prior to reaching the sensor. Collectively, low-flow velocity and wall shear stress provided ideal conditions to validate the EIS response to biofilm formation (Araujo et al., 2016). Figure 3.4: Finite element analysis using COMSOL multiphysics of the flow chamber design. (a) flow trajectories indicated by directional arrows within the flow chamber, (b) Z-X cross section of flow profile, (c) wall shear stress of X–Y plane parallel to integrated microfabricated sensor showing homogenous stress distribution across surface. EIS characterization of biofilm growth Specific frequency regions of interest were defined at 100 Hz