Emergence of debubblers in microfluidics: A critical review

Abstract

Bubbles in microfluidics—even those that appear to be negligibly small—are pervasive and responsible for the failure of many biological and chemical experiments. For instance, they block current conduction, damage cell membranes, and interfere with detection results. To overcome this unavoidable and intractable problem, researchers have developed various methods for capturing and removing bubbles from microfluidics. Such methods are multifarious and their working principles are very different from each other. In this review, bubble-removing methods are divided into two broad categories: active debubblers (that require external auxiliary equipment) and passive debubblers (driven by natural processes). In each category, three main types of methods are discussed along with their advantages and disadvantages. Among the active debubblers, those assisted by lasers, acoustic generators, and negative pressure pumps are discussed. Among the passive debubblers, those driven by buoyancy, the characteristics of gas–liquid interfaces, and the hydrophilic and hydrophobic properties of materials are discussed. Finally, the challenges and prospects of the bubble-removal technologies are reviewed to refer researchers to microfluidics and inspire further investigations in this field.

I. INTRODUCTION

In the past decade, considerable progress and remarkable achievements have been made in microfluidic technologies, and disease diagnosis, point-of-care tests, food safety, in vitro bionic models, and wearable sensors are among the fields that have undergone recent advancements.^1–3 Moreover, microfluidic technologies are becoming increasingly mature and accurate with the rapid development of micromachining and microassembly technologies.^4–6 However, many prominent problems remain to be solved. One such problem of bubbles in microfluidics has recently garnered considerable attention.⁷

During microfluidic experiments, external gas can enter the microchannels of microchips when plugging or removing linked pipes, when switching the reagents, or when the microchannels are not tightened sufficiently.⁸ The gas dissolved in the liquid phase will bubble during the ultra-low-speed operation of the peristaltic pump,⁹ during a temperature change,¹⁰ or when the liquid pressure varies.¹¹ Moreover, owing to the specific microstructure or hydrophobicity of microchannels, such bubbles get trapped and grow.^12,13 Unwanted bubbles in microfluidic experiments are associated with many negative effects. For example, during polymerase chain reactions, bubbles expand when heated and push the samples from the microchamber, thereby causing the failure of experiments.^14,15 In drug-screening experiments, bubbles disturb the diffusion of pharmaceutical molecules, inducing a heterogeneous cell environment in the same culture chamber and thus affecting the experimental results.^16,17 During cell culturing, bubbles introduce toxicity to the cell-culture medium, which is continuously transported to the culture area.¹⁸ Consequently, the cells die from damage to their membranes.^19,20 In organ-on-chip experiments, accumulated bubbles damage the cells and interfere with cell morphology observations in the reaction chamber.^21,22 Most microfluidic devices show right-angle and other special structures, in which bubbles are accumulated, forming the so-called dead zones. In these zones, the microchannel is blocked and liquid flow is disturbed, thereby increasing the pressure and affecting the experimental results.^23,24 Most liquid analytical technologies—such as capillary electrophoresis (CE), liquid chromatography (LC), inductively coupled plasma (ICP)-optical emission spectroscopy, and ICP-mass spectrometry—are adversely influenced by bubbles. In CE, bubbles induce poor separation repeatability and can even cause a current breakdown, particularly during long injection times.^25,26 Bubbles in LC obviously increase the noise and baseline of the detected signal.²⁷ In ICP technologies, bubbles can increase noise intensity and result in broad peaks with sudden signal fluctuations, thus yielding nonreproducible elution profiles of the investigated elements.^28,29 Therefore, the capture and removal of unwanted bubbles are essential in microfluidics. Unwanted bubbles have long puzzled researchers engaged in biological and chemical experiments.³⁰ Conventionally, bubbles have been removed by vacuuming or boiling the liquid in advance, which reduces the solubility of a gas in the liquid and eliminates the generation of bubbles during the experiment.^31,32 Among the various bubble-removal methods, improving the airtightness of the microfluidic device is suggested for reducing the probability of bubble introduction.³³

The aforementioned strategies are mainly preventive and cannot fundamentally solve the bubble problem; moreover, not all liquids are amenable to vacuuming and heating treatments.³⁴ To overcome the limitations of conventional bubble-removal methods, researchers have recently proposed novel bubble-capture or/and bubble-removal methods such as the optical force effect, in which bubbles are drawn toward the target area and irradiated using laser light. Under irradiation, the bubbles are heated, expanded, and ruptured.^35,36 Bubbles can also be manipulated using acoustic waves. Acoustic traveling waves can move bubbles in a predetermined direction, whereas acoustic standing waves can confine the bubbles to a specific region. Therefore, bubbles can be rearranged or captured by exploiting the difference between the dominant positions enforced by standing and traveling waves.^37,38 An air pressure-gradient force can also be used for bubble removal.³⁹ The gas will diffuse from a high-pressure area to a low-pressure area. Bubbles in a liquid will enter a negative-pressure area (generated using a vacuum pump) through a breathable material.

External bubble-removal equipment will increase the volume and cost of the instrument and complicate its operation. To avoid such additional equipment, some researchers have exploited the buoyancy, surface tension, and other characteristics of bubbles in bubble-removal treatments.^40,41 Bubbles are naturally buoyant and will float to the gas–liquid interface, where they can be captured. The surface tension determines whether the bubbles maintain their shape or are broken.⁴² When the surface tension is combined with an external force, bubbles can be pushed into a specific bubble-removal microcavity. Furthermore, bubbles can be trapped using hydrophilic and hydrophobic materials; for example, hydrophobic porous membranes pierced with numerous small holes can expel gas but not liquid because the membrane's hydrophobicity and the liquid's surface tension exert a dual retention effect on the liquid. Such a hydrophobic porous membrane can remove bubbles from the liquid.⁴³

In this review, the recent bubble-capture or/and bubble-removal methods used in microfluidics are introduced and discussed. Depending on whether they require external auxiliary equipment, debubblers are divided into active and passive debubblers. Active debubblers capture and remove air bubbles using an external device (such as a laser, acoustic generator, or negative-pressure pump), whereas passive debubblers exploit the inherent properties of bubbles (such as buoyancy and surface tension) for bubble capture and removal or the hydrophilic or hydrophobic properties of the membrane for bubble filtering. Figure 1 shows a hierarchical diagram of debubbler classification. This review intends to present the importance of capturing or/and removing bubbles in microfluidics and inspire innovation and further research on debubblers in the microfluidic field.

FIG. 1.

Classification of debubblers used in microfluidics.

II. ACTIVE DEBUBBLERS

Active debubblers rely on external auxiliary equipment for bubble capture and removal. Depending on the types of external auxiliary equipment, active debubblers are subdivided into laser, acoustic, and negative-pressure debubblers. Active debubblers control the motion direction of the bubbles by applying lasers or acoustic waves. The rate of bubble trapping or removal can be controlled by adjusting the power of the auxiliary equipment. Bubbles can also be trapped in a collection area by applying negative pressure. Active debubblers are widely used; for example, they can remove blood blisters from cardiopulmonary bypass perfusion systems and bubbles from liquid glass melts. This section focuses on the principles and applications of three active bubble-removal methods.

A. Laser debubblers

Bubbles can be manipulated by the optical force generated using a laser and can be destabilized and ruptured under high-temperature laser irradiation.⁴⁴ As the size and position of a laser can be precisely controlled, laser-based bubble removal is a popular choice for microfluidic experiments.

Under light irradiation, an optical force acts on transparent spherical particles to satisfy the momentum conservation law.⁴⁵ An optical force, which comprises scattering and gradient forces,⁴⁶ manipulates micro- and nano-particles without contacting the particles. The scattering force acts in the direction of the light-wave vector, which is the laser-beam direction in isotropic media such as air, water, and ordinary glass. Therefore, the scattering force pushes the particles in the laser beam direction.⁴⁷ The direction of the gradient force intrinsically depends on the ratio of the refractive indices of the object and surrounding medium.⁴⁸ When this ratio is higher and lower than 1.0, the gradient force pulls the object toward the regions of the highest and lowest laser-beam intensities, respectively.

Murazawa et al.⁴⁹ proposed a debubbler that can capture air bubbles in molten glass under an optical force. When a laser beam focuses on the center of the air bubbles, the bubbles move with the laser spot. However, as a single laser beam cannot simultaneously remove multiple bubbles, the bubble-removal efficiency of this debubbler device is low. Lee et al.³⁶ also proposed a device for capturing bubbles using an optical force [Fig. 2(a)]. In their experiments, the ratio between the refractive indices of the object and surrounding medium is less than 1.0; hence, the optical force generated under laser irradiation pushed the bubbles toward the region of the lowest laser-beam intensity, namely, toward the upper region of the channel. This device can simultaneously remove several hundred microsized air bubbles. Bubbles are continuously removed when the laser is turned on.

FIG. 2.

Laser debubblers. (a) A debubbler that exerts an optical force in a single-layered microfluidic channel (left) and bubble movement in the debubbler without and with laser irradiation (right). Reproduced with permission from Lee et al., Appl. Phys. Lett. 105, 071908 (2014). Copyright 2014 AIP Publishing LLC. (b) A bubble in molten glass under laser irradiation (left) and three-dimensional (3D) finite-difference time-domain simulations of a Gaussian beam focused on bubbles in glass (right). Reproduced with permission from Juodkazis et al., Appl. Phys. A 87, 41–45 (2007). Copyright 2007 Springer Nature. (c) A 3D pulsed laser-triggered cell sorter (top) and a bubble expanding and eventually bursting under laser radiation (bottom). Reproduced with permission from Chen et al., Analyst 138, 7308–7315 (2013). Copyright 2013 Royal Society of Chemistry.

The stability of air bubbles in the liquid phase is related to the Gibbs energy.⁵⁰ At a given temperature, the Gibbs free energy is related to the bubble radius. The radius at which the change in the Gibbs free energy of the bubble is maximized is called the critical radius ( $R_c$). A bubble with a radius greater than $R_c$ will grow, whereas that with a radius smaller than $R_c$ will dissipate. Under thermal equilibrium conditions, laser irradiation increases the temperature of a bubble-containing liquid, causing the dissipation of small bubbles and the expansion and breakage of large bubbles because the surface tension of the liquid is inversely proportional to the temperature. Juodkazis et al.⁵¹ proposed a device that removes microsized bubbles from molten glass using laser irradiation [Fig. 2(b)]. When the laser irradiates the bubbles, the large bubbles are broken, while the others undergo dissipation in the molten liquid. The bubble-removal efficiency of this device is enhanced under scattering and gradient forces, which stretch and compress the bubbles. The device removes bubbles using the high temperature generated using the laser. Moreover, the optical force stretches and compresses the bubbles, accelerating bubble rupture and consequently the bubble-removal efficiency.

In an experiment conducted by Chen et al.,⁵² bubbles in a cell suspension are detected using a photodetector module that triggers a laser and consequently generates a high temperature, owing to which the bubbles rapidly expand and break [Fig. 2(c)]. This module employs a high-speed laser pulse that breaks the bubbles within tens of microseconds.

In summary, bubbles can be manipulated and captured by the optical force generated using a laser. When the liquid temperature is increased to a high value using a laser, the bubbles contained in the liquid can be dissolved or expanded until they burst. However, several shortcomings of laser-based bubble removal must be overcome. First, an excessively high laser intensity or long irradiation time can vaporize the liquid, inducing the formation of bubbles instead of removing them. Second, the high temperature generated using the laser may damage the device.

B. Acoustic debubblers

Bubbles in a liquid can also be manipulated and captured using acoustic standing waves and traveling waves. Acoustic standing waves capture microbubbles at their nodes,⁵³ whereas acoustic traveling waves push the microbubbles along the direction of their propagation.⁵⁴ Acoustic debubblers can be divided into two types with different driving modes: debubblers driven by body acoustic waves (BAWs) and surface acoustic waves (SAWs).

BAWs are sound waves that propagate inside a solid or liquid.⁵⁵ When two waves with the same amplitude and frequency in a medium propagating from opposite directions meet, they form stationary standing waves.⁵⁶ The microbubbles in a liquid can be captured in an acoustic standing-wave field generated using two transducers. A wave propagating outward from the wave source (i.e., a traveling wave) eventually reaches the terminal of the propagation line while continuously transmitting energy.⁵⁷ The traveling wave carries the microbubbles along the propagation direction.⁵⁸ Figure 3(a) shows a BAW-based debubbler proposed by Mino et al.⁵⁹ In the standing-wave mode, the microbubbles are trapped and aggregated around the antinodal lines of the standing waves and their volume expands. As the resonance frequencies of the bubbles are related to their size,⁶⁰ the original sound wave cannot easily regulate the motions of large bubbles. The large-buoyant bubbles flow upward to the liquid level. In the traveling-wave mode, the bubbles are carried away from the excitation transducer by the traveling waves and eventually exit the microchamber. The debubbler is designed to filter the microbubbles and form thrombi from an extracorporeal circulation system, which is a key component of percutaneous cardiopulmonary support (PCPS).

FIG. 3.

Debubblers based on acoustics. (a) Standing-wave mode (left) and traveling-wave mode (right) of a debubbler based on body acoustic waves. Reproduced with permission from Mino et al., Sens. Actuators, A 199, 202–208 (2013). Copyright 2013 Elsevier. (b) Sectional view (upper) and top view (lower) of a debubbler based on surface acoustic waves. Reproduced with permission from Meng et al., Sens. Actuators, B 160, 1599–1605 (2011). Copyright 2011 Elsevier. (c) A debubbler combining electrowetting and acoustic waves. Hyun et al., AIP Adv. 11, 085030 (2021); licensed under a Creative Commons Attribution (CC BY) license. (d) A modular debubbler employing sound waves and a channel composed of a hydrophobic material (left) and bubble enrichment under the action of acoustic waves (right). Reproduced with permission from Peng et al., Sens. Actuators, A 331, 113045 (2021). Copyright 2021 Elsevier.

SAWs are sound waves that propagate along the surface of a solid, confining its energy to the surface.^61,62 Meng et al.⁶³ proposed a debubbler based on SAWs. The SAWs generated using an acoustic generator enter the surface chamber and are reflected at the cavity wall [Fig. 3(b)]. The forward and reflected waves meet, forming standing and traveling waves. When the standing waves dominate, the bubbles are arranged in a curve, whereas the bubbles move around the center of the chamber when the traveling waves dominate. As the linear velocity of the fluid decreases from the outside inward, the bubbles gather toward the center of the chamber under the action of the velocity gradient. This debubbler offers several advantages, such as low power consumption, miniaturization, and rapid and simple fabrication, showing wide application potential in microbubble manipulation in molecular imaging, drug delivery, drug screening, and other biomedical applications.

Hyun et al.⁶⁴ reported a debubbling device combining electrowetting and acoustic waves [Fig. 3(c)]. This device contains a heated surface, electrodes, a dielectric layer, and a hydrophobic layer. Under an applied voltage, the surface wettability of the hydrophobic layer changes and the bubbles attached to the surface are detached. Then, the bubbles are pushed from the chamber under the action of acoustic waves. This device can simultaneously and efficiently remove bubbles with different volumes. This type of debubbler was employed to remove bubbles from a boiling liquid, achieving enhanced heat-transfer performance of the boiling heat-transfer method.

Peng et al.⁶⁵ proposed a modular debubbler that transmits sound waves through a T-shaped tunnel composed of polydimethylsiloxane (PDMS), a hydrophobic material [Fig. 3(d)]. When bubbles enter the channel, small-volume bubbles tend to attach to the hydrophobic interface of the PDMS sidewall, whereas large-volume bubbles are blocked by the obstacles set in the channel. Under water pressure, the blocked bubbles are fragmented into smaller bubbles that attach to the hydrophobic sidewall. Ultrasonic vibrations cause the contraction-expansion motion of the liquid–air interface, generating a steep pressure and streaming velocity gradients near the bubbles. The bubbles attract nearby microbubbles and adhere them to their surface. However, after long-term use, the size of the bubbles exceeds the capacity limit of the device and the bubbles flow out of the microchannel, inducing debubbling failure.

Acoustic waves can achieve noncontact bubble removal. Acoustic standing waves capture the bubbles, whereas acoustic traveling waves control the bubble movement. Bubbles of different sizes can be removed by changing the frequency of the input signals, and the bubble removal efficiency can be controlled by adjusting the power of the input signals. However, in debubblers with such processes, the operating frequency is usually in the ultrasonic frequency band and induces negative effects—such as sonoporation^66,67—that will damage cell membranes and even lead to the death of cells; hence, the application scope of acoustic debubblers is limited, particularly in the biological field.

C. Negative-pressure debubblers

Under a pressure-gradient force, the air is pushed from a high-pressure area to a low-pressure area.⁶⁸ This phenomenon is exploited in debubblers that push bubbles into a negative-pressure area through a breathable membrane. Owing to its excellent biocompatibility and gas permeability, PDMS^69,70 is often used as the raw material of the breathable layer in microfluidic devices.

Skelley et al.³⁹ proposed a multilayer debubbler with a permeable PDMS membrane [Fig. 4(a)]. This device captures bubbles in the flowing liquid within a circular bubble trap. In their debubbler, a vacuum pump is employed to provide a negative pressure. When the vacuum pump is started, bubbles may be drawn through the PDMS membrane into the displacement chamber located in the pneumatic layer. However, the large-volume bubbles are blocked at the trap entrance. Under water pressure, the large-volume bubbles are fractured into microbubbles that flow into the microfluidic channel through both sides of the circular bubble trap. Lochovsky et al.⁷¹ proposed another PDMS debubbler with a cylindrical chamber containing a row of micropillars at the liquid outlet [Fig. 4(b)]. Because the micropillars show minimum clearance, they trap the bubbles while admitting the liquid flow. Under the pressure-gradient force generated using a negative-pressure bump, the bubbles accumulate in the chamber, where they might be drawn into the negative-pressure area through the PDMS membrane. However, the small-volume bubbles tend to flow into the downstream channel through the gaps between the pillars; this undesired behavior is difficult to prevent.

FIG. 4.

Debubblers based on negative pressure pumps. (a) Closed (left) and open (right) states of a multilayer debubbler based on the permeability of a PDMS membrane. Reproduced with permission from Skelley et al., Lab Chip 8, 1733–1737 (2008). Copyright 2008 Royal Society of Chemistry. (b) A PDMS debubbler with closely spaced pillars. Reproduced with permission from Lochovsky et al., Lab Chip 12, 595–601 (2012). Copyright 2012 Royal Society of Chemistry. (c) A modular PDMS debubbler with double layers. Park et al., Membranes 11, 316, 2021; licensed under a Creative Commons Attribution License. (d) A PDMS debubbler with a bevel microstructure. Reproduced with permission from Huang et al., Biomed. Microdevices 22, 76 (2020). Copyright 2020 Springer Nature. (e) A debubbler with a microfluidic network embedded in a PDMS membrane bonded to glass on the bottom and a vacuum chamber on top. Reproduced with permission from Christoforidis et al., Biomed. Microdevices 19, 58 (2017). Copyright 2017 Springer Nature.

Debubblers with multilayer structures require special geometric traps or elaborate negative-pressure methods; consequently, the manufacturing process is cumbersome, expensive, and time-consuming. Attempts to simplify the manufacturing process of debubblers have been made. For example, Park et al.⁷² proposed a double-layered modular debubbler [Fig. 4(c)] in which the disposable PDMS layer and a reusable substrate can be tightly assembled or disassembled using a vacuum pressure device. By exploiting the gas permeability and hydrophobicity of the PDMS layer, this device extracts air bubbles through the PDMS layer under a pressure gradient while retaining liquid in the device. This modular structure achieves low manufacturing costs, simplifies the manufacturing process, and prolongs the service life of the device. Moreover, the device achieves a remarkable bubble-removal efficiency.

Under the action of buoyancy, the bubbles in aqueous solutions rise to the gas–liquid interface and are captured without special geometric traps. Debubblers can also be simplified using bubble floating. Huang et al.⁷³ exploited the bubble-floating property in a debubbler with a bevel structure and a permeable PDMS membrane [Fig. 4(d)]. Under the action of buoyancy, the bevel guides the bubbles upward to the gas-liquid interface of the bubble-capture chamber. Then, the pressure gradient drives the bubbles through the PDMS layer into the negative-pressure area above the capture chamber. By exploiting the negative pressure and buoyancy, this device allows a simplified geometric structure with reduced contact between the liquid and the external environment. Dead zones often emerge in the right-angle and other special structures of microfluidic channels. A vacuum pipe installed near dead zones can remove the bubbles but complicates the device fabrication process. In an experiment conducted by Christoforidis et al.,⁷⁴ an entire fluid-channel network was placed in a negative-pressure environment to remove the geometric limitations of debubblers [Fig. 4(e)] and the need for vacuum pipes.

Negative-pressure debubblers can capture and remove bubbles as well as maintain the continuous flow of a liquid through microfluidic channels. PDMS is generally used as a breathable material. The bubble removal rate can be accelerated by reducing the thickness of the PDMS membrane, increasing the pressure gradient, and increasing the contact area between the bubbles and the membrane. However, the thickness of the PDMS membrane must also meet the mechanical strength requirements.

III. PASSIVE DEBUBBLERS

Unlike active debubblers, passive debubblers can capture or remove bubbles in the absence of external auxiliary equipment such as lasers and ultrasonic transducers. Passive debubblers offer attractive advantages such as a small volume, independent operation, low cost, and easy integration into microfluidic systems. In this section, three types of passive debubblers are discussed: debubblers based on buoyancy, gas-liquid interfaces, and hydrophobic porous films. The principles and applications of each type are also described.

A. Debubblers based on buoyancy

Because the buoyancy force of bubbles is greater than the gravity of that, bubbles float upward and temporarily reside on the liquid surface.^75,76 This characteristic has been exploited for bubble capture in microfluidic devices.

Bubbles can be captured in a single cylindrical chamber in which the height of the outlet tube is higher than that of the inlet tube [Fig. 5(a), left)].⁴⁰ When a gas–liquid mixture flows into the chamber from the inlet tube, the large bubbles are fragmented into small bubbles that gather on the top of the chamber; further, the initially small bubbles float toward the liquid interface. Then, the bubble-free liquid at the bottom of the chamber flows from the outlet tube. During long-duration runs, the bubbles converging at the top of the chamber finally exceed the capacity of the debubbler and the excess bubbles enter the microfluidic channel. To improve the design of this type of debubbler, Zheng et al.⁴⁰ introduced a bypass tube at the top of the chamber and connected the other end of the tube to a filter. The gas is then expelled into the atmosphere [Fig. 5(a), right)]. The bypass tube and filter remove the gathered gas from the chamber, reducing the pollution of the internal liquid from the outside environment and enabling the long-term operation of the device. As described in Sec. II, bubbles in microfluidic devices can be captured by adhering the bubbles to hydrophobic materials such as PDMS.⁷⁷ Recently, Lee et al.³⁴ proposed a debubbler with a cone bubble trap [Fig. 5(b)]. The bubbles in the solution float along the wall toward the top of the cone. The top of the bubble trap is composed of PDMS, which exhibits air permeability. Accordingly, the bubbles are transported to the air chamber and finally expelled into the external environment. This device successfully captures bubbles during long-term operation.

FIG. 5.

Debubblers based on buoyancy. (a) Buoyancy-based bubble traps without (left) and with (right) a bypass channel. Reproduced with permission from Zheng et al., Lab Chip 10, 2906–2910 (2010). Copyright 2010 Royal Society of Chemistry. (b) An integrated debubbler with a cone bubble trap. Reproduced from Lee et al., Biomicrofluidics 15, 014110 (2021), with permission from AIP Publishing LLC. (c) A buoyancy-based 3D-printed bubble trap. Reproduced with permission from Terutsuki et al., Sensors 20, 5779 (2020); licensed under a Creative Commons Attribution License. (d) A bubble-trap module in a 3D cell-culture system. Reproduced with permission from Park et al., J. Micromech. Microeng. 28, 045001 (2018). Copyright 2018 IOP Publishing. (e) Top view (left), bottom view (bottom), and sectional view (right) of a double-layer debubbler with barriers. Reproduced with permission from Sung J et al., Biomed. Microdevices 11, 731–738 (2009). Copyright 2009 Springer Nature.

Figure 5(c) shows the three-dimensional (3D)-printed bubble trap proposed by Terutsuki et al.⁷⁸ The flow channel of this device comprises a curved inlet, flat flow channel, and cylindrical vertical channel at the center. At its inlet side, the tube tip is in slight contact with the inlet hole of the device and is inserted into the outlet hole at its outlet side. When a flow segment containing the bubbles enters the inlet pipe, the bubbles are broken at the inlet hole and the gas is expelled into the external environment. Small bubbles entering the microfluidic channel are affected by buoyancy and float above the cylindrical channel such that the bubble-free liquid flows from the outlet pipe. Park et al.⁷⁹ introduced a similar buoyancy-based bubble-trap module into a microfluidic 3D cell-culture system [Fig. 5(d)]. However, this device is suitable only for short-term use, because it can capture bubbles but cannot remove them.

Bubbles with a size gradient can be sorted through microbarriers set at different intervals in a microfluidic device. Sung et al.⁸⁰ proposed a double-layer debubbler that selectively captures bubbles of specific sizes through successive barriers [Fig. 5(e)]. This debubbler has a top layer and a bottom layer, both composed of PDMS. The top layer contains barriers for capturing the bubbles, while the bottom layer provides a conduit for liquid flow. The intervals of the barriers at the top layer gradually decrease from the inlet to the outlet; hence, the bubbles are captured in progressively small size ranges. However, as the bubbles accumulated in the barriers eventually block the channel, this device is also unsuitable for long-term operation.

Although buoyancy-based debubblers show a simple structure and easy operation, the bubble-removal performance considerably depends on the liquid flow rate through the microfluidic channel. When the flow rate is extremely high, the bubbles are conveyed to the liquid level before they can float and the bubble capture fails. Thus, the bubble-removal efficiency of buoyancy-based debubblers is limited by the fluid flow rate.

B. Debubblers based on gas–liquid interfaces

At the liquid-air interface, the molecules on the liquid surface experience a stronger pulling force from other liquid molecules than from the gas molecules; therefore, the surface area of the bubbles tends to decrease at the interface.^42,81 Owing to the surface tension of the liquid, the spherical shape of the bubbles is maintained. The behavior at the gas–liquid interface can be exploited for bubble removal.

On micrometer and nanometer scales, the viscous forces become much greater than the inertial forces. Therefore, the surface tension of bubbles is greater than the gravity of that in microfluidics.^82,83 When two bubbles approach each other, a thin water film is formed between them. This water film progressively becomes thin and finally breaks, merging the two bubbles. Owing to the surface tension, the water film around a single bubble does not break even when the bubble is compressed or stretched.⁸⁴

Cheng et al.⁴¹ proposed a microchip with an array of pillars composed of a hydrophobic SU-8 photoresist. The SU-8 pillars, together with the micropores at the top layer, form an array of gas-liquid interfaces [Fig. 6(a)], and the bubbles remain intact under surface tension. However, as the bubble volume increases, the hydraulic pressure increases the menisci of the bubbles until the bubbles are broken. The device proposed by Cheng et al. exploits the characteristics of gas–liquid interfaces to capture and remove bubbles without any auxiliary equipment. However, tiny bubbles cannot be captured and instead flow into the downstream channels.

FIG. 6.

Debubblers based on the characteristics of gas–liquid interfaces. (a) Schematic (top) and cross-sectional view (bottom) of a debubbler with paired holes and hydrophobic SU-8 for air-pillar formation. Reproduced from Cheng et al., Appl. Phys. Lett. 95, 214103 (2009), with permission from AIP Publishing. (b) Sectional views of bubbles reaching the entrance of the trapping tube of a debubbler (top) and entering the trapping tube (bottom). Reproduced with permission from Stucki et al., Lab Chip 15, 4393–4397 (2015); licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported License. (c) Schematic (top) and cross-sectional view (bottom) of a debubbler based on hydrophobic and hydrophilic interfaces. Reproduced with permission from Greenwood et al., IEEE Sens. J. 10, 952–959 (2010). Copyright 2010 IEEE.

Stucki et al.⁸⁵ proposed a bubble-removal approach for capturing and removing bubbles using their surface tension and the hydrodynamic forces exerted by the fluid [Fig. 6(b)]. Two hollow tubes (trapping and shield tubes) with different internal diameters are employed in the debubbler. Bubbles are blocked at the entrance of the trapping tube (i.e., the tube with a smaller internal diameter). Owing to the surface tension and hydrodynamic forces, the pushing of small-volume bubbles into the trapping tube is prevented. The bubbles remain at the entrance of the trapping tube until they accumulate, forming large bubbles. When the surface tension inside the trapping tube (F₁) exceeds that around the shield tube (F₂), the trapped bubbles are pushed into the trapping chamber and removed. This device proposed by Stucki et al. removes the bubbles while maintaining the fluid flow. However, bubbles with diameters smaller than the trapping-tube diameter may escape the trap and flow into the microfluidic channel.

The interface can be endowed with hydrophobic and hydrophilic properties by modifying the microchannel.^86,87 When a liquid flows through a hydrophilic/hydrophobic interface, the pressure difference between the liquid and the atmosphere is balanced by the surface tension, which fixes the liquid inside the interfacial boundary. The liquid side wall remains intact unless the pressure at the liquid side exceeds the critical value. Greenwood et al.⁸⁸ constructed a hydrophobic/hydrophilic interface with a hydrophobic self-assembled monolayer [Fig. 6(c)]. The surface tension fixes the liquid walls at the boundary of the hydrophilic/hydrophobic interface and prevents them from entering the hydrophobic interface; moreover, the bubbles are eliminated as the gas migrates to the hydrophobic layers.

The bubble-removal effect of debubblers at gas–liquid interfaces largely depends on the bubble size and the ingenious geometry of the device. Extremely small-volume bubbles are difficult to capture using the aforementioned methods. Moreover, if the gas–liquid interfaces utilized to burst the bubbles in a liquid are set too high to maintain the meniscus shape of the bubbles, the bubble capture can fail.

C. Debubblers based on a hydrophobic porous film

Bubbles can be filtered by using a microporous film made from polytetrafluoroethylene (PTFE).⁸⁹ The hydrophobicity of PTFE⁹⁰ and the surface tension of water prevent the passage of a liquid through the film; alternatively, a gas can escape through the micropores. However, the pressure difference (P_th) between the inside and outside of the film has a threshold related to the surface tension (γ) and contact angle (θ) of the liquid. The threshold is calculated using the following equation:⁹¹

$$P_{\mathrm{th}}={\displaystyle \frac{4\gamma \cos (\pi -\theta )}{d},}$$

where d denotes the diameter of the tiny holes in the PTFE film. When the pressure difference exceeds $P_{\mathrm{th}}$, the liquid will leak from the film.

A PTFE film was used as a bubble eliminator as long as 30 years ago [Fig. 7(a)].⁹² In the previous method, a liquid mixed with bubbles was transferred to a PTFE pipeline using a peristaltic pump and the bubbles were discharged through the pores in the film. Recently, Liu et al.⁹³ employed a PTFE film in a debubbler with a nozzle-film structure [Fig. 7(b)]. The hydrophobic PTFE film is attached to the top of the nozzle, which operates as a valve. When the microporous film is in contact with the liquid, the film deforms under the action of water pressure and the valve opens, that allows the passage of the gas-free liquid beneath the film. When the film is in contact with the bubbles, the water pressure is not maintained and the valve closes, preventing the bubbles from entering the microfluidic channel via expulsion through the micropores in the film. The structure of this device is ingenious but allows the entry of a few tiny bubbles into the microfluidic channel during the open-and-close switching of the valve. As an improvement of this debubbler, Williams et al.⁴³ proposed a rapidly integrated, PTFE film-based debubbler module that discharges bubbles with volumes of nanoliters to microliters [Fig. 7(c)]. When a liquid flows beneath the hydrophobic porous film, the film deflects, exposing a fluidic channel connecting the inlet and outlet. Within this long and narrow fluid channel, the bubbles are in complete contact with the PTFE film and a sufficiently long time can be provided for the discharge of the bubbles from the micropores in the film. Therefore, even the tiniest bubbles are eliminated.

FIG. 7.

Debubblers based on hydrophobic porous films. (a) Inline debubbler coupled with tubing. Reproduced from van Lintel et al., Micromachines 3, 218–224 (2012); licensed under a Creative Commons Attribution License. (b) A PTFE film debubbler with a nozzle-film structure in the open state with a liquid in the debubbler (left) and in the closed state with air bubbles in the debubbler (right). Reproduced with permission from Liu et al., Lab Chip 1, 1688–1693 (2011). Copyright 2011 Royal Society of Chemistry. (c) A rapidly integrated PTFE film-based debubbler module. Reproduced with permission from Williams et al., Micromachines 10, 360 (2019); licensed under a Creative Commons Attribution License. (d) Movement of bubbles from a flat surface to a textured surface to minimize their surface energy. Reproduced with permission from Cheng et al., Microfluid. Nanofluid. 17, 855–862 (2014). Copyright 2014 Springer Nature. (e) A PTFE film-based debubbler with textured V grooves. Reproduced with permission from Cheng et al., Microfluid. Nanofluid. 17, 855–862 (2014). Copyright 2014 Springer Nature.

To minimize their surface energy,⁹⁴ bubbles tend to attach to hydrophobic interfaces.^95,96 At the junction of a hydrophilic/hydrophobic interface, bubbles preferentially move toward the hydrophobic interface [Fig. 7(d)].⁹⁷ Cheng et al.⁹⁷ proposed a hydrophobic debubbler with a textured structure. As shown in Fig. 7(e), V grooves with rectangular concave pits are engraved on smooth horizontal silicon plates and the sidewalls of the grooves are texturized to form a hydrophobic interface. A PTFE film is then bonded to the bottom of the grooves. The bubbles at the junction of the textured grooves and smooth horizontal surfaces tend to move toward the grooves. As the bubbles accumulate, they are enlarged and are finally discharged through the hydrophobic film.

As a breathable and hydrophobic material, the PTFE film enables the removal of bubbles from microfluidics; however, the pressure applied to the film should be suitably controlled. When the pressure is extremely small, the bubbles are not easily expelled from the film pores. Alternatively, the liquid will overflow from the film pores. Therefore, an appropriate fluid pressure on the film is essential for the appropriate operation of the debubbler.

IV. CONCLUSION

In this review, various methods for removing bubbles from microfluidics are presented. Debubblers can be classified as active (requiring external auxiliary equipment) or passive (not requiring external auxiliary equipment). Active debubblers capture and remove bubbles using lasers, acoustic generators, negative-pressure pumps, and other external equipment. The bubble-capture and -removal rates can be controlled by adjusting the power of the input signal. Active debubblers are widely used for removing bubbles from common aqueous solutions and molten glass as well as extracorporeal blood circulation systems. Passive debubblers capture and remove bubbles by exploiting natural phenomena such as buoyancy, the characteristics of gas–liquid interfaces, and the hydrophilic and hydrophobic properties of materials. Passive debubblers require no external auxiliary equipment and create a milder work environment than active debubblers. Passive debubblers offer further advantages such as a simple fabrication process, small size, and independent operation; accordingly, they are easily integrated into a microfluidic system.

The advantages of active debubblers include their usage in wide ranges, high debubbling efficiency, and controlled debubbling speed. However, the external auxiliary equipment increases the expense, bulk, difficulty of integration, and operational complexity of the debubbler. Currently, regular spherical bubbles are usually removed from microfluidic devices using optical forces. The efficacy of the removal of considerably deformed bubbles using the aforementioned methods has not been explored or verified. Moreover, the use of external equipment introduces additional problems; for example, the high temperatures induced because of long-time laser irradiation can melt and damage the device. The work performance of ultrasonic debubbling methods is related to the resonance frequency of the bubbles, which is itself associated with the bubble volume; therefore, such methods can remove bubbles simultaneously only within a certain volume range rather than all bubbles. The application scenarios of ultrasonic debubblers are also limited. For example, when ultrasound is employed for the removal of bubbles from the extracorporeal blood circulation system, it may also destroy red blood cells, causing hemolysis. During ultrasonic diagnosis, the ultrasound emitted by a debubbling device may interfere with the diagnostic result. In negative-pressure debubblers, the liquid in the trap chamber will be extracted through the permeable PDMS film along with the bubbles under a negative pressure, inducing liquid leakage.

In passive bubble-removal methods, the speed of bubble trapping and removal considerably depends on the liquid flow rate, bubble size, and geometry of the microfluidic channels. The efficiency of debubblers based on boundary properties is highly subject to the liquid flow rate in the microchannel. When the flow rate is excessively high, the bubbles have insufficient time to rise to the trap area near the liquid level and the bubble capture fails. Bubble-removal methods based on gas–liquid interfaces cannot easily capture very small bubbles and will fail if the height of the gas-liquid interfaces in the device is excessively high. Alternatively, if the bubble volume is extremely large or the height of the gas-liquid interfaces is set too low, the bubbles may rupture into microbubbles that can flow through the microfluidic device. Finally, the bubble-removal efficiency of debubblers based on hydrophobic films depends on the applied water pressure. When the pressure is too small, the bubbles are not easily discharged from the membrane holes; alternatively, when the pressure is too large, the liquid will overflow from the membrane holes.

As the ideal debubbling goal cannot usually be attained using a single approach, some pioneering researchers have combined several approaches to improve the bubble-removal performance. For instance, acoustic bubble removal has been combined with the electrowetting technology.⁶⁴ In this approach, the substrate surface wettability is changed by applying a voltage to the substrate, allowing the bubbles to escape from the board bonds, and then the released bubbles can be driven to the export of the device by acoustic waves. The combined electrowetting and acoustic waves simultaneously realize the noncontact removal of all bubbles, regardless of their size. The negative-pressure debubbling approach has also been combined with the buoyancy approach.³⁴ Buoyancy confines the bubbles to the trap without requiring a complex geometric structure. The negative-pressure debubbling approach then forces the bubbles into the negative-pressure area through the PDMS film, thereby improving the debubbling efficiency.

V. PROSPECTS

Although the capture and removal of unwanted bubbles from microfluidic systems has significantly progressed, some remaining problems require further investigations and further improvements must be made.

Debubblers with multilayer structures are difficult to manufacture, and the repeatability of bubble-removal performance between microchips is poor. Hence, novel and viable debubbler structures must be developed for simplifying the structure, decreasing the manufacturing cost, and improving the detection reliability and efficiency. For example, a self-assembled bubble-removal chip has been proposed recently,^98,99 which comprises only two layers: a disposable upper layer and a reusable substrate layer. This design decreases the cost and simplifies the device. When the external surroundings of the debubbling are not clean, contact between the solution and the outside air will pollute the sample. The integrated bubble-removal microchip can potentially address the above-mentioned problem because it shows good airtightness⁷³ that can effectively isolate the liquid in the microfluidic device from the outside environment. This type of structure can run stably in different environments.

The components of the microfluidic chips of conventional debubblers are complicated to prepare and time-consuming to assemble. Thus, some advanced micromachining technologies or construction methods for microchips can be introduced to the fabrication process of debubblers. For instance, microchips with ingenious microfluidic networks can be directly constructed using a 3D printing technology,^78,100 which is fast and flexible without requiring component assembly. Constructing microstructures on bubble-removal devices is a difficult process using conventional methods. Femtosecond lasers,^101,102 which deliver an ultrashort pulse width and ultrahigh peak power, can be used to construct complex microstructures on arbitrary substrates. Both femtosecond lasers and 3D printing technology can considerably improve the efficiency of microfluidic device preparation. Paper-based microchips^103–105 can also offer multiple advantages to microfluidics, such as low cost, simple fabrication processes, and mass producibility. Therefore, if debubblers can be maturely constructed using the paper-based method, their progress will jump from the experimental stage to the commercial stage.

Bubbles significantly degrade the detection accuracy and reliability of biological and chemical analysis methods (such as CE and ICP).^25–29 Combining these analytical methods with debubblers will substantially improve their analytical performance. However, the integration of debubbler modules into mature analytical instruments or devices is a challenge. In particular, the modular designs and connection schemes must be universally matched to these instruments or devices while simultaneously meeting the low cost, maintainability, and operability requirements. This aspect has been rarely reported and can be considered as a promising research foray into “no man's land.”

Bubbles are not always unwanted or detrimental to the biological or chemical processes in microfluidics. Recently, they have been utilized in bioengineering and biological applications.^106,107 For example, bubble oscillations can produce destructive microstreaming and shear forces on cell membranes. This technology is an advanced means of lysing cells and obtaining their DNA, RNA, or intracellular proteins.^108,109 Bubbles have also been employed as powerful and ingenious carriers of drugs to lesion zones of the human body in precision medicine and treatment applications.^110,111 When bubbles carrying the drug molecules reach the target area, they can be stimulated and burst using an external ultrasonic wave. Bursting produces asymmetrical microstreaming that temporarily changes the permeability of the cell membranes, allowing the unhindered entry of the released drug molecules into the cell. This technique greatly enhances the drug delivery efficiency. Bubbles have become a hot research topic in microfluidics and have been successfully used in cell mechanics analysis,¹¹² bioassembly,¹¹³ disease diagnosis,¹¹⁴ and other applications.¹¹⁵ One interesting avenue is the extraction of useful information from a gas dissolved in a liquid. For example, the content of ammonia in the breathing gas of patients is usually detected by dissolving gaseous samples in special solutions that remove interfering substances.⁸⁸ However, separating the gaseous samples from the solutions poses certain challenges. As debubblers can separate and collect gases from a liquid, they promise a unique and feasible solution to this problem.

Bubbles are a “double-edged sword” that can both enhance and hinder biological and chemical processes in microfluidics. Researchers in various fields—such as biological engineering, life sciences, analytical chemistry, and micromachining—have pioneered the capture, collection, and/or removal of bubbles in microfluidics. The use of debubblers in microfluidics has guaranteed the successful implementation of biological and chemical processes. The further development of debubblers requires the cooperation of scientists from diverse research fields, including (but not limited to) biological engineering, life sciences, micromachining, and analytical chemistry. Research on debubblers in multiple research fields is expected to reveal richer and more important roles for debubblers in microfluidics, which will accelerate the development of the aforementioned disciplines.

ACKNOWLEDGMENTS

This work was supported by the Natural Science Foundation of the Jiangsu Higher Education Institutions of China (No. 19KJB530013), the Startup Foundation for Introducing Talent of NUIST (2019R16), the Innovation and Entrepreneurship Doctor Program of Jiangsu Province (No. R2020SCB50), and the Jiangsu Government Scholarship for Overseas Studies.

AUTHOR DECLARATIONS

Conflict of Interest

The authors declare no competing interests. No violation of human or animal rights occurred during this investigation.

Author Contributions

Mingpeng Yang: Conceptualization (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal); Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal). Nan Sun: Investigation (equal); Writing – original draft (equal); Writing – review & editing (equal). Yong Luo: Investigation (equal); Writing – original draft (equal); Writing – review & editing (equal). Xiaochen Lai: Investigation (equal); Writing – original draft (equal); Writing – review & editing (equal). Peiru Li: Writing – original draft (equal). Zhenyu Zhang: Writing – original draft (equal).

DATA AVAILABILITY

Data sharing is not applicable to this article as no new data were created or analyzed in this study.