Bubble-driven cell detachment

Abstract

On-demand cell detachment is of great importance in various applications in biosensitive environments. Existing methods such as enzymatic treatments and mechanical scraping are often time-consuming, labor-intensive, and harmful to cells. In this work, we demonstrate a method of detaching cells from substrates using electrochemical bubble generation without biocide generation. We demonstrate that shear stress generated by fluid flow beneath a rising bubble is the primary mechanism for cell detachment. This strategy, relying solely on physical forces and independent of cell or surface chemistry, can therefore work with a large range of the media, surfaces, and cells. We successfully implement this discovery at the lab-scale by designing a prototype for on-demand cell detachment that maintains high cell viability. The developed principle could find applications in high-throughput culture settings, such as algae photobioreactors or cell culture environments.

On-demand, electrochemical bubble generation efficiently detaches cells while preserving their viability.

INTRODUCTION

Controlling cell adhesion is an important challenge for a wide range of industries and applications including pharmaceuticals and biomedicine (1, 2), biosensors and other implants (3–5), food processing (6–8), and biofuel production (9). For two-dimensional (2D) cell cultures, passaging requires the viable removal of adherent cells from the surface for subsequent re-culturing. Current cell detachment strategies for mammalian cell cultures rely heavily on enzymatic approaches, a time-intensive method that generates large volumes of waste; a process largely unchanged for over 100 years (10). Another example is algae adhesion to the walls of photobioreactors. Photobioreactors are a common tool used to cultivate algae for applications such as cosmetics, biofuel, dietary supplements, and medical drugs among others. Currently their operation is limited due to fouling related downtimes for cleaning, thereby increasing operational costs (9).

Over the last several decades, many technologies have emerged as potential solutions for unwanted cell adhesion in biosensitive environments, such as surface coatings with biocidal compounds or functionalization inhibiting protein adsorption (11–16). However, the rapid depletion of molecules in such coatings and the negative impact on cell health have limited their practical implementation (17–19).

Recent advancements in antifouling and cell-detachment strategies have used active and on-demand methods of local biocide formation, such as sodium hypochlorite (bleach) and hypochlorous acid, generated on the fouled surface through electrochemical means (electrolysis) (20–25). While highly effective in detaching cells and foulants, the formation of biocides limits its use to applications where the survival of detached cells and surrounding environment are not critical concerns, such as with underwater sensors or desalination membranes (20, 24, 25).

To illustrate these challenges, we perform an electrolysis-driven cell detachment experiment as shown in Fig. 1A. Chlorella vulgaris freshwater algae are allowed to settle on and adhere to a metallic electrode surface. Electrolysis is then initiated to effectively detach algae adhered to surfaces (Fig. 1B). However, in chloride-containing environments, electrolysis also triggers unwanted side-reactions, such as the generation of sodium hypochlorite (NaClO), a potent biocide (Fig. 1C). This is evidenced by a marked loss of cell viability, where the algae solution, initially vibrant green, becomes clear following exposure (Fig. 1D). More info can be found in section S1.

Fig. 1. Electrolysis-driven algae detachment leads to biocide formation and cell death.

(A) Schematic of adhered algae electrolytic-detachment concept. (B) Schematic of electrolysis side-reactions in media containing chloride, including electrochlorination and the generation of NaClO. (C) Experiment showing C. vulgaris algae detachment before (i) and after (ii) bubbling at 200 mA/cm² for 1 min. (D) Images of 50-ml C. vulgaris algae solution before (i) and after (ii) bubbling at 200 mA for 1 min.

This electrochemical process also generates bubbles directly on the electrode surfaces. Until now, these bubbles have been viewed as by-products that potentially enhance the function of the biocide. Namely, it has been stipulated that effects like triple-contact-line movement (21) and growing bubble dynamics (20, 22, 23) could play a secondary role as well in detaching the cells in these systems. However, their exact contribution in the foulant release mechanism from the surface remains unclear. In this work we demonstrate that, even when eliminating biocide formation, bubbles alone can efficiently detach adhered cells and foulants. This provides a purely physical mechanism for cell detachment, enabling the on-demand, repeatable, and viable detachment of cells. We use C. vulgaris microalgae as a model cell, a commonly grown alga in photobioreactors.

To demonstrate this concept, we generate bubbles via electrolysis on a biofouled surface composed of a 10-nm transparent gold film deposited on glass. We first experimentally investigate the dynamic interactions between bubbles detaching from a surface and the algae adhered to the electrode. This is carried out using a specially designed millifluidic imaging platform, in a chlorine-free medium to underscore the importance of hydrodynamic factors in both bubble and cell detachment processes. We also use an inverted fluorescence microscopy setup allowing for the imaging of both the bubble dynamics and the attached algae on the surface. We then delve into experiments using individual bubble departures on surfaces populated with adhered cells to ascertain the effective radius within which a departing bubble influences its surroundings. To further understand the interaction, we evaluate the adhesion strength of the algae to the electrode surface using a microfluidic device that can systematically vary the shear force imposed on the cells, as previously demonstrated (26–28). By comparing these experimental results with an analytical model, we show that the shear stress generated upon bubble departure is sufficient to detach attached cells. Therefore, unlike existing strategies that prevent adhesion through toxic coatings (29, 30), this technique physically removes already adhered cells on-demand. Last, we demonstrate an on-demand prototype cell detachment approach with partitioned electrodes that enables biocide-free operation even in media containing chloride ions. Cells are removed from the surface while maintaining high viability in their culture media. This is shown to work for both C. vulgaris microalgae, as well as mammalian cells such as MG-63 cells, a type of cultured human osteosarcoma cells.

RESULTS

Detaching cells using interfacial microbubble generation

To investigate the mechanism of cell detachment, we designed a platform that enables visualization of both electrochemically induced microbubbles and the cells adhering to the surface, shown in Fig. 2A. Freshwater C. vulgaris algae, our model cells, have a diameter of 2 to 10 μm (10). C. vulgaris autofluoresce, allowing us to image them separately from the microbubbles using an inverted microscope equipped with both transmission bright-field and reflective fluorescence imaging. Transmission bright-field microscopy is optimized for rapid imaging of the evolving bubbles, whereas reflective fluorescence microscopy allows for the visualization of the cells’ adherence to and detachment from the surface, unaffected by the overlaying bubbles. For the electrode surface, a transparent gold electrode with film thickness of 10 nm is used (Platypus Technologies). Gold is chosen as a catalyst due to its high stability and resistance to corrosion. For the electrode configuration, we opt for a dual-fingered design, as depicted in Fig. 2B. This design features electrodes with dimensions of 10 mm in width and 1 mm in height, yielding an active area of 10 mm² for both the anode and cathode. A gap of 1 mm between the electrodes is strategically chosen to reduce ohmic losses while ensuring ample separation to prevent mixing between the two streams. A millifluidic channel is created out of polydimethylsiloxane (PDMS) and attached to the electroactive gold surface. The channel has a height of 3 mm, width of 4 mm, and length of 2 cm. The final millifluidic detachment platform is sketched in Fig. 2A.

Fig. 2. Experimental algae detachment results from surface microbubble generation.

(A) Schematic of the experimental setup consisting of a transparent gold electrode, a PDMS millifluidic channel, and an optical microscope. Bright-field imaging is used to image the bubble activity, fluorescence is used to image the algae on the surface. (B) Two-fingered electrode design to minimize ohmic losses and to separate the anodic and cathodic activity. (C) Bright-field (top) and fluorescence (bottom) microscope images of bubbling at three different current densities with attached algae. Scale bar, 200 μm. Measured bubble radius distribution (D) and algae and colloidal coverage (E) as a function of current density. Horizontal line indicates average bubble radius in (D). The dotted line is a guide for the eye in (E).

To understand the role of mechanical detachment and eliminate biocide generation, we implement a chloride-free electrolyte. A 1 M potassium bicarbonate electrolyte solution is chosen, which prevents electrochlorination on the anode and therefore eliminates the generation of chloride-based biocides found when running electrochemical reactions in media containing chloride. Furthermore, the potassium bicarbonate electrolyte acts as a buffer limiting pH fluctuations near the electrodes (31). The resulting electrolyte has a pH of 8.2, within the range of growth conditions for freshwater algae. To verify that the chloride-free potassium bicarbonate electrolyte does not adversely affect the algae during the experiment, we conducted an adhesion test with the algae in this new solution. The adhesion strength of C. vulgaris to our electrode surface is evaluated using a microfluidic device (refer to section S2 for details). The experiment yields an algae adhesion curve as a function of increasing wall shear stress, consistent with prior studies using culture media (26). At a wall shear stress of 9.5 Pa, half of the algae originally adhered to the surface are detached, similar to the adhesion found in algae culture media (26). This outcome demonstrates that using a potassium bicarbonate electrolyte solution does not compromise the algae’s ability to adhere within the time window of the experiments.

To assess the impact of local electrochemical pH shifts on cell adhesion, additional adhesion tests were performed in very alkaline (pH 13) media, mimicking the estimated local pH conditions near the cathode (section S3). In alkaline conditions, adhesion remained comparable to that observed in neutral media, with no substantial difference in adhesion strength measured. Thus, a 50% detachment threshold at 9.5 Pa serves as an appropriate benchmark for subsequent experiments, independent of the expected local alkalinity.

To start the bubbling experiment, C. vulgaris algae are cultured for 5 to 9 days (see Methods for further details). The resulting algae solution is then introduced in the millifluidic channel and allowed to settle for 2 hours. After 2 hours, the algae media is flushed out and replaced with the potassium bicarbonate electrolyte by flowing at 1 ml per minute for 5 min. The channel is now filled with the chloride-free electrolyte, with attached algae cells on the wall. Next, bubbling is induced for 10 s by applying a set current density across the two electrodes using a DC power supply. To remove detached algae, a low electrolyte flowrate of 1 ml per minute is applied to the channel with a syringe pump. This leads to an applied shear stress on the wall of around 3 mPa, well below the required shear stress to remove attached algae.

In Fig. 2C, we show images of the generated bubbles and the algae on the surface at various current densities (see Methods for the full experimental procedure). With increasing current density, we observe an increase in bubbling. The increase in the surface bubble coverage with increasing current density coincides with a slight decrease in average size and in the variance of the bubble radii (Fig. 2D) (32). We find an average bubble radius of around 30 μm for the three higher current densities. In Fig. 2E, we report that increasing current density results in a decreasing algae coverage. At the highest current density tested, the remaining algae coverage is less than 15% of the original coverage. This shows the efficacy of surface-level bubble generation in detaching cells. We note that at intermediate to high current densities, variability in bubble nucleation due to surface microfeatures such as scratches or pits leads to increased SD in detachment performance, as reflected in the larger error bars.

To further confirm that cell detachment is not driven by the biological properties of algae, we repeated the millifluidic bubbling experiments using polystyrene colloids. To this end, we selected colloids with a size and adhesion strength comparable to algae. We ultimately used carboxyl-coated polystyrene beads (carboxyl fluorescent particles, purchased from Spherotech) with a radius (5 to 5.9 μm) closely matching that of the algal cells (~5 μm). Using the microfluidic adhesion platform, we found that these colloids exhibit adhesion strengths comparable to algae in 1 M potassium bicarbonate medium, with 50% detachment occurring at applied shear stresses of 7 to 10 Pa in both systems (fig. S1). When subjected to the same bubbling protocol, surfaces fouled with these beads showed a detachment profile nearly identical to that observed for algae (Fig. 2E and fig. S2). Given that these inert colloids, despite matching the algae’s adhesion characteristics, exhibit indistinguishable detachment behavior, we conclude that the primary mechanism driving detachment is independent of biological activity and instead governed by physical or mechanical interactions.

Shear stress beneath departing bubbles detaches cells

The two proposed physical mechanisms that could explain algae detachment from surface-level bubble generation are the movement of the triple contact line (the line at the foot of the bubble where the gas, liquid, and solid phases meet) before departure and the flows induced by the bubbles after departure (33, 34). In this context, “movement” of the triple contact line refers to the shift in the position of the contact line as the bubble grows or changes shape, which can exert shear forces on nearby cells. The pinned diameter of the triple contact line refers to the diameter of the circular contact patch. In our case, the motion of the contact line as a detachment mechanism is implausible, as no contact line is visible for the experiments performed, as is shown in section S5. This agrees with recent observations made on electrochemically generated bubbles on a platinum electrode (32, 35). The pinned diameter of the contact line is expected to be only a fraction of the size of a single algae cell.

This leaves flows generated by bubble departure as the main culprit in this story. Previous theoretical studies have reported the dynamics and shape of a bubble rising in an unbounded liquid (36). They established a phase diagram as a function of two nondimensional numbers, the Bond and Galileo numbers, $\text{Bo}={\mathrm{\rho }}_{\mathrm{w}}g\frac{R^2}{\mathrm{\sigma }}$ and $G_{\mathrm{a}}={\mathrm{\rho }}_{\mathrm{w}}\frac{\sqrt{\mathit{gR}}R}{{\mathrm{\eta }}_{\mathrm{w}}}$ , with η_w and ρ_w the viscosity and density of water, respectively, g the gravitational acceleration, R the bubble radius, and $\mathrm{\sigma }$ the surface tension of water. $\text{Bo} \text{and} G_{\mathrm{a}}$compare the gravitational force to the surface tension force and the gravitational force to the viscous force, respectively. For our study, the bubble radius ranges from 5 to 100 μm, yielding $\text{Bo}$smaller than 10^–3 and $G_{\mathrm{a}}$ in a range from 0.03 to 3. This means that surface tension forces impose that the bubbles stay spherical (37, 38).

Knowing this, the proposed mechanism of cell detachment is illustrated in Fig. 3A. Namely, spherical bubbles detach and rise from the surface, initiating a water flow beneath them. The generated water flow creates a sufficiently large shear stress on the adhered cells to detach them. To validate this mechanism, we analytically estimate the wall shear stress induced by a departing bubble, which we then compare to the experimentally derived stress necessary to detach algae. The complete derivation of the wall shear calculation can be found in section S6. To summarize, first the dynamics of bubble rise are derived, solving for the velocity of the rising bubble as a function of time from departure to the terminal velocity. Here, the driving force is the gravitational buoyancy force, and the resisting force is Stokesian viscous resistance. This was validated by comparing the predicted terminal velocity, with experimentally measured terminal velocity showing good overlap (fig. S6). Through volume conservation beneath the rising bubble, an average velocity v_w(x,t) of the water flow beneath the bubble is found, which is a function of both time t and lateral distance x from the nucleation site of the bubble. Last, assuming a no-slip boundary condition on the wall and a slip boundary condition on the bubble, the following shear stress on the wall is found

$$\mathrm{\tau }(x,t)=\frac{{\mathrm{\eta }}_{\mathrm{w}}x{v}_b\left(t\right)}{2}\ast \frac{1}{h{(x,t)}^2}$$

Fig. 3. Theoretical model to estimate the shear stress induced by bubble departure.

(A) Schematic of bubble growth and detachment, and algae removal from the surface. (B) Calculated wall shear stress beneath departing bubble as a function of distance from nucleation site at various times for a bubble with 30 μm radius.

Here, v_b(t) is the rise velocity of the bubble, and h(x,t) is the height of the water film beneath the bubble as a function lateral distance x from the nucleation site and of time t.

In Fig. 3B, we plot the calculated shear stress for a 30-μm-radius bubble, the average bubble radius found in our experiments (Fig. 2D), as a function of the distance from the nucleation site at multiple time intervals. We also plot the maximum shear calculated at each position over the entire time interval of bubble rise (black line). Using the microfluidic adhesion experiments in fig. S1, we establish that an applied shear stress of ~9.5 Pa is required to dislodge 50% of the attached algae on gold samples. From our model we find that for a 30-μm-radius bubble, a shear stress of ~9.5 Pa is induced within 7 μm distance from the nucleation site.

Last, we also define a cell detachment radius R_r as the radius from the point of nucleation at which 50% of all algae are removed due to bubble departure, as shown in Fig. 4A. We can analytically estimate this radius by solving for the radial position where a shear stress of 9.5 Pa is reached for different bubbles. We measured and verified cell detachment radius for various bubble sizes. To facilitate single bubble departures, we use the same millifluidic set up as before but decrease the duration of applied voltage from 10 to 1 s and limit the current density to 8 mA/cm². This reduces the probability of two bubbles nucleating sequentially from the same nucleation site or proximal nucleation sites and allows for single nucleation events on our electrode surfaces. In Fig. 4B, microscope images of a single bubble departure event are shown. Here, around the nucleation site, a clear radial decrease in algae coverage is visible after departure. In Fig. 4C, we report the cell detachment radius versus bubble departure radius for the single bubble experiments. On the same curve, the predicted cell detachment radius from the model is also plotted, with a prefactor of 1.25, added to best fit the experimental data. This prefactor could be potentially derived from an exact prediction of the shear stress at the interface solving the Navier-Stokes equation in this three-phase problem, which would go beyond the estimate we established here. Nonetheless, our estimation using mass conservation and Newton’s first principle on the bubble is in remarkable agreement with the experiment. Together, these results validate that shear flow beneath bubbles removes attached cells.

Fig. 4. Comparison of experimental and theoretical results for bubble departure.

(A) Schematic of single bubble departing on fouled surface and leaving a detachment radius of detached algae. Detachment radius R_r is defined as radius at which 50% of the algae are removed postdeparture (B) Fluorescent [before departure (left) and after departure (right)] and bright-field [right before departure (middle)] microscope images of single bubble departing surface covered in algae. Scale bar, 100 μm. (C) Measured and theoretically predicted detachment radius versus bubble departure radius.

Lab-scale prototype of nontoxic cell detachment

We now demonstrate the practicality of our approach for two specific applications by designing a lab-scale prototype of a cell detachment system. We show robust cell detachment for both algae and mammalian cells while maintaining their viability in their respective culture media. For this, the main challenge that needs to be solved to make bubble-driven cell detachment possible is the biocide generation, as in these systems, eliminating chloride from the media is impractical.

Algae photobioreactors

First, in the context of photobioreactors, algae need light to grow through photosynthesis. When the surface of the bioreactor gets fouled due to algae attachment on the walls, light cannot penetrate the bioreactor, severely limiting the growth rate of the algae. Regular cleaning of the inside surface of the bioreactor is therefore necessary. Current techniques necessitate shutting down the reactor for cleaning, increasing operational costs and limiting the viability of photobioreactors (39). To eliminate this cleaning step, we implement our technique to on-demand clean the bioreactor surface through surface-level bubble generation.

Our substrate consists of a 1 by 3 inch glass slide coated with a 2-nm titanium adhesion layer and a 10-nm gold catalyst layer is used. The secondary electrode, the anode, is a 4-cm-long platinum wire. As medium, we use a 7-day-old algae C. vulgaris solution inside its respective medium. The composition of the algae medium can be found in section S7. Unlike our previous electrolyte, the freshwater culture media contains NaCl and other chloride compounds. Hence, if the anode is in direct contact with the medium it will generate toxic chemicals such as chlorine gas and NaClO (bleach). Furthermore, the electrode will degrade rapidly due to the formation of chloride complexes, limiting its practical application (40). To prevent these negative effects, the anode is separated from the primary electrode through a proton exchange membrane (PEM) as indicated in Fig. 5A. The PEM allows only the passage of cations (such as protons) through the membrane, in essence separating the two electrolytic streams. This limits toxic chemicals generated at the anode (the platinum wire) from migrating to the primary electrolyte (the algae medium) and negatively affecting the viability of the biological medium. To prevent chloride complexes from forming at the isolated anode, which would rapidly degrade the electrode, a chloride-free electrolyte consisting of 1 M potassium bicarbonate is chosen for the secondary electrolyte (the anolyte). The PEM also prevents the anolyte and catholyte from mixing.

Fig. 5. Centimetric-scale prototype for robot detaching cells with electrolytic bubble.

(A) Schematic of benchtop-scale algae defouling system with separated anolyte and catholyte streams through a proton exchange membrane (PEM). Linear actuator is used to move the secondary electrode over the surface and minimize ohmic losses. Images of freshwater C. vulgaris (B) and saltwater N. oculata (C) algae after 10 days of growth without electrolysis, with electrolysis with a PEM, and electrolysis without a PEM. Scale bar, 1 cm. Cell density during growth of C. vulgaris (D) and N. oculata (E) algae without electrolysis, with electrolysis with a PEM, and electrolysis without a PEM. (F) Algae coverage before bubbling, and after every sweep. After sweep 3, the medium is replaced with fresh algae-free medium to show effect of detached algae. Scale bar, 1 cm.

To validate the effectiveness of the PEM in preventing biocidal effects, we repeated the electrolysis experiment shown in Fig. 1D using the PEM-separated anode. Electrolysis was conducted in 100 ml of C. vulgaris culture medium at 200 mA for 1 min. The algae were then incubated for 14 days. Control groups included a nonelectrolyzed sample and a sample subjected to electrolysis without the PEM. After bubbling, the solution was placed in a light controlled incubator and the cells are allowed to grow. As shown in Fig. 5B, after 10 days, both the control and PEM-protected samples remained green, indicating healthy algae growth. In contrast, the sample without PEM became transparent, consistent with cell death and similar in appearance to bleach-treated algae solutions. The cell density of this solution was measured every 2 days with a NanoDrop UV-Vis Spectrophotometer (more information in section S9) and is plotted in Fig. 5D.

The experiment was also repeated using Nannochloropsis oculata, a saltwater algae species with NaCl concentrations over 30 times higher than the freshwater medium. The results (Fig. 5, C and E) mirrored those of the freshwater experiment: healthy cell growth in the PEM-protected sample and cell death in the unprotected one. These findings demonstrate the robustness of the PEM-separated anode approach, even in high-salinity environment such as ocean water.

Next, to minimize ohmic losses, which are particularly notable in the freshwater media due to low conductivity, it is essential to reduce the distance between the secondary electrode and the active surface. However, it is also crucial to ensure that the secondary electrode does not obstruct light entering the medium. To achieve this balance, the PEM-separated platinum electrode is mounted on a motorized linear actuator. This setup allows precise movement of the platinum electrode over the surface, thereby minimizing the distance between the two electrodes and effectively reducing ohmic losses. This setup is necessary as the system must remain transparent, precluding the option to make the secondary electrode larger. Last, the catalyst surface is fouled through settling using concentrated 7-day-old C. vulgaris algae solutions. The dense algae medium is allowed to settle for 24 hours. Afterward, the system is flushed twice to remove unadhered algae and left filled with freshwater culture medium to remove nonattached algae. An image of the surface after flushing can be seen in Fig. 5F at t = 0.

To perform the experiment, 15 V (which equates to ~200 mA/cm²) was applied across the two electrodes and the actuator sweeps the surface three times at a velocity of 1 cm/s (movie S1). A relatively high voltage is necessary due to the low conductivity of the freshwater medium and the distance between the two electrodes. A picture of the fouled surface taken after every sweep is shown in Fig. 6B. After the third sweep, the medium is replaced with new culture medium to remove detached algae in the solution. As can be seen in Fig. 5B, adhered cells were successfully removed.

Fig. 6. Detachment of human MG-63 and OVCAR-8 cells via electrochemically generated bubbles.

Gold electrode surface with adhered MG-63 and OVCAR-8 cells before (A and C) and after (B and D) bubbling. Scale bars, 100 μm in (A) and (B) and 200 μm in (C) and (D). (E) Viability of MG-63 human cells after detachment using standard trypsin enzyme approach and surface bubble generation approach with PEM.

Mammalian cell culture

Second, the practicality of our technique for biosensitive applications is further shown in the context of 2D cell culture. Cells are commonly cultured on a 2D surface and later collected using enzymatic approaches such as trypsin to detach them from the substrate, a process largely unchanged for over 100 years (10). However, this enzymatic cell detachment strategy is time intensive and generates large volumes of waste (27). Here, MG-63 human osteosarcoma cancer cells and OVCAR-8 high-grade ovarian serous adenocarcinoma were used as a model cell.

To show the versatility of our technique, MG-63 and OVCAR-8 cells are cultured for 3 days on the same 1 by 3 inch transparent gold electrode. After 3 days (Fig. 6, A and C), they are detached from the surface through bubble generation using the PEM-separated secondary electrode used above. The electrode surface depleted of cells after bubbling is shown in Fig. 6 (B and D). The details of these experiments are described in section S8.

The viability of these cells is measured using a cell counter and is found to not differ from cells detached from the surface using a standard enzymatic based trypsin cell detachment approach as shown in Fig. 6E.

The benefits of bubble detachment of mammalian cells compared to the trypsin protocol are a notable decrease in the amount of time and the quantity of material needed to detach and culture cells. Namely, for a single T25 flask, we estimate a 40 to 50% decrease in plastic usage, up to a 67% decrease in spent liquid (phosphate-buffered saline and culture media), and a decrease in time spent per flask of up to 89% (section S10).

This serves as a proof of concept of using electrochemical bubble generation for detaching and passaging adhered cells. However, additional work is necessary to carefully characterize the adhesion strength of mammalian cells, the effective detachment radius of bubbles, the influence of confluency, and to assess long-term viability through multiple passages.

DISCUSSION

In this study, we demonstrate that bubbles departing from a substrate are capable of detaching adhered cells without affecting their viability. We show with experiments and theory that the flow generated beneath departing bubbles generates sufficient shear stress to remove attached cells. The above strategy relies solely on hydrodynamic flows and hence does not depend on the physicochemical characteristics of the cells, surface, or culture media. Moreover, unlike previous strategies that prevent bioadhesion by avoiding the contact between the surface and the cells or foulant, our strategy allows on-demand, time controlled, removal of the already adhered cells or foulant.

Therefore, we believe that this approach can be implemented for a wide range of situations where cell adhesion needs to be regulated while maintaining high viability. We demonstrate our approach by constructing a lab-scale cell detachment system that allows robust cell removal over a 1 by 3 inch surface while maintaining cell viability. This system is successfully tested on both C. vulgaris microalgae and MG-63 and OVCAR-8 human cells. It is important to note that these experiments were conducted on nonconfluent cell layers, allowing for sufficient electrode exposure to support electrolysis. In fully confluent biofilms, the insulating properties of the biofilm may inhibit bubble generation due to reduced electrolysis.

Future applications of our technology could extend well beyond the confines of photobioreactors, accommodating both seawater and freshwater algae, as well as mammalian cell culturing. By broadening the scope of biofoulants, cells, and organisms, ranging in size from nanoscale proteins to macroscale marine oysters, we can gain insightful data on the capabilities and limitations of our current technologies. For example, for motile algae species, localized oxygen and CO₂ generation at the anode may influence cell behavior, including motility, which could potentially enhance detachment; this interaction between electrochemical gas evolution and biological responses presents an intriguing avenue for future research. These expansions not only enhance our understanding but also pave the way for applications in diverse fields such as sensor technology within the food and biomedical sectors, and in sustainable aquaculture practices. In addition, conducting further research into the impact of surface texture and wetting characteristics on electrochemical surfaces in relation to bubble generation could yield substantial benefits, offering avenues for optimization and application in various industries.

METHODS

Algae culturing

C. vulgaris cells, N. oculata cells, Erdschreiber’s medium, and MB3N medium were purchased from UTEX. All samples of algae were incubated in a humidity-saturated chamber with 5% CO₂ at 25°C with light cycle of 12-hour day and 12-hour night and light intensity of 100 μmol m² s⁻¹. The cultures were constantly shaken on a rotating platform to prevent settling. Settling experiments were conducted with algae cultures of ages 5 to 9 days.

Experimental process bubbling experiments

For the electrode surface, a transparent gold electrode with film thickness of 10 nm on glass is purchased from Platypus Technologies. To create the electrode design seen in Fig. 2B, a 150 W continuous output Nd:yttrium-aluminum-garnet laser scriber is used to create the pattern by removing the 10-nm gold metal layer between the electrodes. After scribing, the samples are cleaned by rinsing with acetone, ethanol, and isopropyl alcohol.

For the bubbling experiments, the millifluidic chip as shown in Fig. 2A is used. To prepare for the experiments, 5- to 9-day-old algae media is introduced into the chip and allowed to settle in the incubator for 2 hours. After 2 hours, the algae media is replaced with the 1 M potassium bicarbonate electrolyte by flowing at 1 ml per min for 5 min. After, the flow rate is adjusted to 1 ml/min to flush detached algae and emulate conditions found in photobioreactors. A set voltage is applied using a DC power supply for 10 s at a time and ramped up from low to high current density. After every step, a fluorescent reflected microscopy image and a bright-field transmitted microscopy image is taken.

An Olympus IX73 inverted microscope is used with a 10× magnification lens. For fluorescence, the beamsplitter has a 475 nm with 40-nm bandpass exciter, 500-nm splitter, and 650-nm longpass receiver lenses.

Adhered algae were counted using the fluorescent images in ImageJ. First, images were transformed into an 8-bit grayscale image. A threshold is applied to the image, allowing only the algae to remain in the image. Next, the image is turned into a binary image and the watershed feature is used to separate algae near to each other. The particle counter feature was then used to count the algae. Algae coverage is then calculated as the percentage of individually counted algae remaining at each current interval as compared to the starting value. For the images of the algae (Figs. 2C and 4B), a green color was added to the binary images using adobe illustrator.

Lab-scale defouling setup

The PEM is an Aquivion membrane purchased from the Fuelcell Store. The PEM was pretreated according to the manufacturer’s guidance in an HCl solution. A SLW-1080 Linear actuator with lead screw purchased from Igus was used to move the electrode. A 7.5-cm platinum wire auxiliary electrode with gold-plated connector was purchased from BASi research products as secondary electrode. A chamber to hold the platinum electrode and PEM was designed and 3D printed using a Formlabs form 2 printer. A DC power supply with a set voltage was used.

Cell culture

MG-63 osteosarcoma cells and OVCAR-8 high-grade ovarian serous adenocarcinoma were cultured in Roswell Park Memorial Institute (RPMI) 1640 Medium (Gibco, 11835030) supplemented with 10% fetal bovine serum (Sigma-Aldrich, F8317), Fungizone (500 ng ml⁻¹; Cytiva, SV30078.01), gentamicin (20 μg ml⁻¹; Thermo Fisher Scientific, 15710064), 100 IU penicillin, and streptomycin (100 μg ml⁻¹; Sigma-Aldrich, P4333-100ML). All cells were maintained at 5% CO₂ and 37°C. Cells were routinely cultured in T-25 flasks and passaged at 70 to 80% confluence with 0.25% trypsin-EDTA (Gibco, 25200056) for use in experiments. The MG-63 and OVCAR-8 cell lines were both used to observe the performance of bubbling on cell detachment. For this, 70 to 80% confluent cells were detached using trypsin, diluted 1:5 with complete RPMI media, and passaged onto the prepared electrode surfaces. These were then grown in the incubator for 3 days. MG-63 cells were used to measure viability after bubbling.

Use of artificial intelligence tools

ChatGPT (version GPT-4o, developed by OpenAI) was used to improve the clarity and language of the manuscript. The authors provided the following prompt to ChatGPT: “Suggest improvements to my writing for clarity and flow, while staying very close to my original style.” This prompt was applied paragraph by paragraph throughout the text. All scientific content, data analysis, and conclusions remain entirely the work of the authors.

Supplementary Materials

The PDF file includes:

Supplementary Text

Figs. S1 to S9

Tables S1 and S2

Legend for movie S1

References

Other Supplementary Material for this manuscript includes the following:

Movie S1