A structure-free digital microfluidic platform for detection of influenza a virus by using magnetic beads and electromagnetic forces

Po-Hsien Lu a, Yu-Dong Ma a, Chien-Yu Fu a and Gwo-Bin Lee *abc
aDepartment of Power Mechanical Engineering, National Tsing Hua University, Hsinchu, 30013, Taiwan. E-mail: gwobin@pme.nthu.edu.tw; Tel: +886 3 5715131 Ext. 33765
bInstitute of NanoEngineering and Microsystems, National Tsing Hua University, Hsinchu, 30013, Taiwan
cInstitute of Biomedical Engineering, National Tsing Hua University, Hsinchu, 30013, Taiwan

Received 14th November 2019 , Accepted 8th January 2020

First published on 9th January 2020

H1N1, a subtype of influenza A virus, has emerged as a global threat in the past decades. Due to its highly infectious nature, an accurate and rapid detection assay is urgently required. Therefore, this study presents a new type of digital microfluidic platform for H1N1 virus detection by utilizing a one-aptamer/two-antibodies assay on magnetic beads. The droplets containing magnetic beads were driven by electromagnetic forces on a structure-free, super-hydrophobic surface to automate the entire assay within 40 min. With different levels of hydrophobic modification, the droplets could be easily controlled and positioned without any assisted microstructure. The tunable electromagnetic forces could be adjusted for three kinds of operating modes for the manipulations of beads and droplets, including movement of droplets containing magnetic beads, mixing of two droplets and beads extraction out of droplets. When compared with previous studies, the manipulations of droplets and magnetic particles in this study are more flexible as they can be easily adjusted by fine-tuning the magnetic flux density. Furthermore, the magnetic beads also served as three-dimensional substrates for the new enzyme-linked immunosorbent assay (ELISA)-like assay. The magnetic beads were conjugated with aptamers, which have high specificity towards H1N1 viruses such that they could be specifically captured and detected. The horseradish peroxidase-conjugated secondary antibody was then used to activate tyramide-tetramethylrhodamine (TTMR) such that fluorescent signals could be amplified. With this approach, the limit of detection was experimentally found to be 0.032 hemagglutination units/reaction, which is sensitive enough for clinical diagnostics. This kind of digital microfluidic platform with the ELISA-like assay could effectively reduce the consumption of samples and reagents such that the volume of all droplets including the H1N1 sample, antibodies, TTMR and wash buffers was only 20 μL. This is the first time that a digital microfluidic platform was demonstrated such that the entire diagnostic process for influenza A H1N1 viruses could be performed by using electromagnetic forces, which could be promising for rapid and accurate diagnosis of influenza.


Influenza, commonly known as “flu”, has been one of the global health threats in the past decades. Influenza viruses could be classified into four groups, namely A, B, C and D.1 Among them, influenza A virus is one of the most hazardous pathogens for human beings. The symptoms of seasonal influenza A are multiple and can range from mild to severe, including cough, headache, sore throat, body aches, high fever, chills and fatigue. For most influenza A patients, the symptoms are not lethal, and they could recover rapidly. However, there are still certain risks for young babies, elderly people and some patients with low immunity.2

In addition to the seasonal influenzas, there are several variants of influenza viruses carried by poultry or pigs.3 Currently, 18 subtypes of hemagglutinin (HA) for influenza A virus and 11 subtypes of neuraminidase (NA) have been identified,4,5 and these kinds of variants have a certain probability of causing serious pandemic diseases with a high mortality rate.6 Due to the diversity of viral subtypes, prevention and prediction of a novel influenza outbreak are extremely challenging. For example, the “Spanish flu” in 1918, which was the first influenza A H1N1 (infA/H1N1) pandemic, caused more than 20 million deaths around the world, and was the most serious global flu in history.7 In 2009, a new swine-origin influenza A virus had emerged in Mexico and the United States, which rapidly developed into another infA/H1N1 pandemic influenza virus and caused over 17[thin space (1/6-em)]000 deaths worldwide.6,8 The subtype of recent influenza A pandemic, infA/H1N1 is considered to possess high potential to evolve into global pandemic9 and cause a large number of security issues and economic impacts. Therefore, it is important to provide a rapid and accurate diagnostic platform such that we could restrain the influenza outbreak at its early stages.

The conventional diagnostic methods for influenza virus are typically classified to viral culture, serological assays, nucleic acid detection (reverse transcription polymerase chain reaction (RT-PCR)), rapid influenza detection tests (RIDTs) and detection of viral antigens. Among them, virus culture is the “gold standard” diagnostic method by physiological observation.10 However, it is highly time-consuming (3–10 days), and requires well-trained technicians to operate for a long period of time in laboratory.11,12 Serological assays rely on viral antibody–antigen interactions with a high sensitivity. However, the reaction time is relative long (3–10 days), and it also requires paired serum specimens and well-trained technicians.13 Alternatively, nucleic acid detection (i.e. RT-PCR) is the most popular approach for diagnosis of influenza viruses. This method has nearly 100% sensitivity, but also requires long operation time (3–24 hours).14 Besides, a precise thermo-cycler and well-trained technicians are required.15 Furthermore, RIDTs have been widely used for influenza A virus diagnosis, which could be performed within a short period of time (about 30 min). However, its sensitivity is relatively low and requires higher viral loads (104–106, equal to 3.2 × 10−2–3.2 × 10−4 HAU of the sample used in this study) to achieve the detection which is still the major challenge in clinical settings.16,17 Similarly, the detection of viral antigens is one of the most common approaches for the influenza virus detection.15 This method has been used for determination of antigenic and genetic characteristics with relatively high sensitivity when compared to RIDTs.15

Moreover, enzyme-linked immunosorbent assay (ELISA) is one of the most popular approaches based on antigen tests in hospitals and laboratories. The conventional ELISA has a relatively short turnaround time in comparison to other influenza virus diagnostic approaches,18 but it still has several disadvantages. For instance, well-trained technicians and relatively large volume of reagents are required. Furthermore, the sensitivity and specificity are not satisfactory.19,20

Recently, several integrated microfluidic platforms using fluorescence immunoassays with specific antibodies conjugated onto magnetic beads for the detection of influenza A virus have been reported.12,21 For instance, an antibodies-conjugated bead-based integrated microfluidic system based on a fluorescent immunoassay (FIA) approach for rapid detection of influenza virus was reported.12 The influenza A viral particles such as H1N1 and H3N2 could be recognized in a short period of time (i.e., 15 min). The limit of detection (LOD) for detection of infA/H1N1 virus in a specimen sample was reported to be 0.031 hemagglutination unit (HAU), which was 1000-fold lower than the conventional serological assays. However, the collection of the magnetic complexes still relied on a permanent magnet, which was attached onto the bottom of the microfluidic chip manually. Similarly, a pneumatically-driven microfluidic system with a sandwich-based aptamer assay was reported.22 The LOD of the dual-aptamer assay was reported to be 0.032 HAU, which is similar to the traditional two-antibody assay. However, the sample pretreatment still relied on manual, off-chip process, and the microfluidic system required bulky equipment such as vacuum pumps or compressors for activating microfluidic devices (such as micropumps, micromixers and microvalves) for liquid transportation, which is a bottleneck for POC applications.

Digital microfluidics is an alternative approach which works on an open surface and involves manipulation of droplets.23 When compared with conventional continuous-flow microfluidics, digital microfluidics has the benefits of lower sample and reagent consumption, independent operation of each droplet, and the capability of discrete droplets acting as independent reaction chambers.23 Several promising digital microfluidics including an electrowetting on dielectric (EWOD) technique24 and a magnetic-based system25 have been demonstrated. The latter one uses magnetic beads to provide functional solid substrates for molecule binding, which is very useful for sample pretreatment.23,26 Furthermore, magnetic digital microfluidics could be classified into three categories including magnetic droplets, magnetic liquid marbles and magnetic substrates.23 For instance, a digital magnetically-actuated droplet manipulation platform with an open hydrophobic surface was reported previously.25 A Teflon-coated surface provided a super-hydrophobic flat surface for digital droplet operation. Droplets with magnetic particles inside could roll on the surface by the attraction from permanent magnets and several manipulations including transportation, particle extraction, droplet merging, and mixing could be successfully performed. However, the magnetic force should be precisely controlled to perform different droplet manipulations such as droplet movement or particle extraction by displacing a stronger or weaker magnet, varying the amount of magnetic particles, altering the moving speed of magnet25 and adjusting the distance between surface and magnet,27 which needs to be improved significantly such that less human intervention and faster response time could be achieved.25

Therefore, a magnetic digital microfluidic platform utilizing ELISA-like assay was developed herein for the detection of infA/H1N1 virus in this work. The magnetic microbeads served as an actuator in the magnetic digital microfluidic system, which could be operated automatically by using a simple magnetic power source. The platform serves for magnetic bead-based droplet system driven with electromagnetic forces, which could be adjusted in real-time. Compared to previous digital microfluidic platforms using permanent magnets, the electromagnetically-driven platform is more flexible and has higher adaptability for the manipulation of droplets and magnetic beads by simply altering the current input. Additionally, the super-hydrophobic modified surface and relatively hydrophilic droplet SETs (surface energy traps) served as virtual reaction chambers for reactions. In comparison to the physical limitations, the limitation of surface affinity is better for droplet transportation. The consumption of samples and reagents could be significantly minimized by using the integrated microfluidic platform (i.e. the volume of each droplet is only 20 μL). Moreover, the power consumption of this magnetic digital microfluidic platform was relatively low, which could be driven at a voltage of 3.23 V provided by a set of batteries. For the ELISA-like assay for the detection of infA/H1N1 virus, H1N1 specific aptamer was used to capture the target virus with high specificity and affinity.22 Furthermore, a tyramide signal amplification (TSA) assay utilizing tyramide-tetramethylrhodamine (TTMR) to amplify the fluorescent signal was used to quantify the magnetic complexes. The LOD of the detection assay was experimentally found to 0.032 HAU, which was comparable to previous works using microfluidic approaches12 and close to the newest RIDTs.16 The entire detection process was only 40 min that demonstrates the great potential for rapid diagnosis of InfA/H1N1.

This is the first time that an electromagnetically-driven digital microfluidic platform for detection of infA/H1N1 virus has been reported.

Materials and methods

Experimental procedure for infA/H1N1 detection

Fig. 1 shows a schematic diagram of the infA/H1N1 virus (97N510H1, stock concentration = 32 HAU, received from the Department of Microbiology and Immunology, National Cheng Kung University, Taiwan) diagnostic process. H1N1 specific aptamer-conjugated magnetic beads (106 beads per μL, 20 μL), surface-blocked magnetic beads without aptamers coating (2 × 108 beads in deionized distilled water (ddH2O), 20 μL), 1st antibodies (GTX127357, infA/H1N1 virus anti-HA antibody, GeneTex Inc., USA) (1[thin space (1/6-em)]:[thin space (1/6-em)]250 in phosphate buffered saline (PBS) with 0.1% bovine serum albumin (BSA), 20 μL), 2nd horseradish peroxidase (HRP)-conjugated antibodies (abs-22, goat anti-rabbit antibody, HRP conjugated, Asia Bioscience Co. Ltd., Taiwan) (1[thin space (1/6-em)]:[thin space (1/6-em)]2000 in PBS with 0.1% BSA, 20 μL), wash buffers (ddH2O for virus isolation and PBS for antibodies conjugation, 20 μL), and TTMR (PerkinElmer Life and Analytical Sciences, Inc., USA, 20 μL) for the TSA assay were pre-loaded into the respective droplet SETs. Note that the concentration of magnetic beads for manipulation of droplet movement was approximately 107 beads per μL, which is more than those used for aptamer-conjugated magnetic beads. Therefore the surface-blocked magnetic beads were used to assist the droplet manipulation. First, the infA/H1N1 virus samples with different concentrations (32 HAU, 3.2 HAU, 0.32 HAU, 0.032 HAU, and 0.0032 HAU, respectively, each for 20 μL) were added into the first SET pre-loaded with aptamer-conjugated magnetic beads, respectively. Next, infA/H1N1 droplet and beads were moved to a mixing area by using a strong electromagnetic force (with a magnetic flux intensity of 350 mT), and mixed by an electromagnetic mixer (110 mT) for 10 min, such that the infA/H1N1 viruses could be captured efficiently by H1N1-specific aptamer-coated magnetic beads. Afterwards, the droplet was transported back to the first SET and the virus-bead complexes were collected and extracted by using a weak electromagnetic force (240 mT) within 10 s. After the virus-bead complexes were moved to the next droplet (which was pre-loaded with wash buffer, ddH2O), the beads and the droplet were moved to the mixing area for washing for 30 s, and sent back to the original SET. After washing, beads were extracted from the droplet and moved to the next droplet containing 1st H1N1 anti-HA antibodies for 10 min mixing. After washing, the beads were extracted from the droplet and moved to the next droplet containing 2nd HRP-conjugated antibody to be linked with 1st antibody for another 10 min mixing. The beads were then washed again and moved to the droplet contained TTMR to carry out the TSA assay to quantify the virus. The TSA mixing was up and down for once, stayed static for 4 min, and then washed for 60 s. Finally, the infA/H1N1 virus was observed under a fluorescence microscope (BX43, Olympus, Japan) and detected with a microplate photometer (ELISA reader, POLARstar Omega, BMG LABTECH, Germany). Note that the TSA assay was used such that the sensitivity of the developed assay could be significantly improved.
image file: c9lc01126a-f1.tif
Fig. 1 Schematic diagram of the infA/H1N1 diagnostic process. Pre-load magnetic beads surface-coated with infA/H1N1-specific aptamers. (a) infA/H1N1 virus loading; (b) virus captured by beads surface-coated with infA/H1N1-specific aptamers; (c) washing out non-specific wastes; (d) 1st antibody adhered on virus-magnetic beads complexes; (e) washing out wastes; (f) HRP-conjugated 2nd antibody adhered on the 1st antibody-virus-magnetic beads complexes; (g) washing out wastes; (h) tyramide signal amplification; (i) washing out wastes; (j) fluorescence detection.

Chip design and fabrication

The microfluidic chip was equipped with multiple “droplet SETs” for pre-loading samples and buffers containing infA/H1N1 samples with aptamer-coated beads, 1st antibody, 2nd antibody, TTMR and wash buffers (ddH2O and PBS) (Fig. 2(a) and (b)). The chip was composed of three layers including a super-hydrophobic layer (with a thickness of 30 μm), a relatively hydrophilic polydimethylsiloxane (PDMS) droplet SET layer (with a thickness of 20 μm) and a glass substrate (with a thickness of 0.7 mm) (Fig. 2(c)). The gradient of hydrophobicity between the super-hydrophobic and the relatively hydrophilic droplet SET enabled the droplet to be trapped without any structure. The dimensions of the chip were measured to be 80 mm × 60 mm × 0.75 mm, and the diameter of droplet SETs was 0.3 mm, indicating that a droplet with a volume of 20 μL could be positioned.
image file: c9lc01126a-f2.tif
Fig. 2 (a) Schematic illustration of the microfluidic chip containing droplet SETs for infA/H1N1 sample, 1st antibody, HRP-conjugated 2nd antibody, TTMR for TSA assay, and washer buffers (ddH2O for virus isolation and PBS for antibodies conjugation); (b) photograph of the microfluidic chip with super-hydrophobic surface and relatively hydrophilic PDMS SETs; (c) exploded view of the multiple layers of the microfluidic chip including PDMS droplet SETs, a super-hydrophobic layer and a glass substrate; (d) schematic diagram of the chip fabrication process. The super-hydrophobic treatment was spin-coated at 300 rpm for 30 s on a glass substrate and dried at room temperature for 15 min. PDMS droplet SETs were stamped by using a PMMA stamp with a specific pattern and dried at room temperature overnight; (e) contact angles of the super-hydrophobic layer (left) and the relatively hydrophilic PDMS droplet SETs surface (right) (n = 3).

The super-hydrophobic layer on the glass substrate was fabricated via a spin-coating process28,29 (Glaco mirror coat “zero”, Soft99 Co., Japan) with a spin-coating speed at 300 rpm for 30 s and then dried at room temperature for 15 min. After the super-hydrophobic spin-coating process, two kinds of PDMS reagents (Sylgad 184A/B, Dow Corning Corp., USA) was pre-mixed at a mass ratio of 10[thin space (1/6-em)]:[thin space (1/6-em)]1, degassed and stamped on the super-hydrophobic layer for the pattern of droplet SETs by using a polymethylmethacrylate (PMMA)-fabricated stamp. After curing at room temperature overnight, the chip was complete (Fig. 2(d)). Compared to previous studies that required microstructures,27,30 this technology based on hydrophobicity gradient to form a structure-free droplet SET surface for droplet manipulation with very simple fabrication processes. The difference of the contact angles between the super-hydrophobic surface (155°) and the hydrophobic PDMS SETs layer (95°) (n = 3) was measured to be approximately 60°, which is satisfactory for effectively distinguishing between these two kinds of operating surfaces (Fig. 2(e)). For a contact angle, it was called hydrophobic when it was above 90 degrees (90–150°), and super-hydrophobic if it was more than 150 degrees (>150°). It is worth mentioning that we called the PDMS-based SETs with a contact angle of 95° “relatively hydrophilic” in this work for emphasizing the contact angle gradient and it was actually hydrophobic in general.

Experimental setup

To establish a magnetic digital microfluidic platform for infA/H1N1 detection, an integrated magnetically-driven platform equipped with tunable electromagnets and a microcontroller (Arduino Uno, Arduino SRL, Italy) was developed (Fig. 3). A structure-free, magnetic digital microfluidic chip was placed at the center of this platform, and electromagnets were placed under the droplet SETs of the chip to attract magnetic beads such that the droplet motion including merging, mixing and beads extraction could be well controlled. The input current of the electromagnets supplied from a power supply (HILA DP-2002, Centenary Materials Co., Ltd., Taiwan) could be adjusted by the Arduino microcontroller, leading to switching between different modes of droplet/bead manipulation. Moreover, the magnitude of the magnetic flux density could be increased by combining a permanent magnet with 240 mT, such that the total magnetic flux density can reach to 240–350 mT (i.e. 0–110 mT from the electromagnet). Thus, the beads extraction and droplet movement could be regulated with an electromagnet. Note that higher electromagnetic forces (>320 mT) enable the droplet movement, merge or mixing; in contrast, weak electromagnetic forces (<320 mT) lead to extraction of beads from the droplets without moving droplets. There was a mixing area with an electromagnet placed above the chip such that the electromagnetic force could agitate the beads in the droplet in order to achieve efficient mixing if necessary.
image file: c9lc01126a-f3.tif
Fig. 3 Schematic diagram of the integrated digital microfluidic platform for infA/H1N1 detection. This platform was composed of a structure-free chip with super-hydrophobic surface and relatively hydrophilic SETs, an electromagnet driving system for operations of droplets and beads, an electromagnetic mixer for liquid mixing, and a control system including a power supply, a laptop and a microcontroller.

Electromagnets were composed of a soft iron core (Centenary Materials Co., Ltd., Taiwan) which is a ferromagnetic material that could concentrate the magnetic flux when the current is supplied, and rapidly demagnetizes when it is turned off. The brass coil (Centenary Materials Co., Taiwan) with a diameter of 0.45 mm was wrapped around the soft iron core to form an electromagnet of 625 turns with a length of 22 mm. The electromagnet could generate a magnetic flux density of 110 mT at an input current and voltage of 1.6 A/3.23 V, respectively, which was measured by using a Tesla meter (WT10A, Weitecidian Co., Ltd., China). Furthermore, for establishing an electromagnetically-driven platform to execute the manipulations of droplets and beads, the magnetic flux density of the electromagnet was too low to work with (<320 mT). Therefore, a permanent magnet with a magnetic flux density of 240 mT and an electromagnet with a magnetic flux density of 110 mT were integrated in this work. The permanent magnet provided a base-line magnetic flux density and the electromagnet served as a regulator to adjust the magnetic flux density for switching between different manipulations. With this approach, a magnetic flux density ranging 240 to 350 mT could be generated.

Furthermore, the heat generated by the electromagnet was a critical issue which has to be addressed properly. To provide an effective electromagnetic force while the heat generation is low, we test different combinations of current and voltage applied on the electromagnet wrap with 0.45 mm brass wire for 625 turns. When an input current and voltage of 1.6 A/3.23 V was applied to the electromagnet for 30 s, the temperature above the chip was measured to be 37.3 °C, and then increased to 41.5 °C after 60 s of actuation (ESI Table S1). Since each working period of the electromagnet was less than 60 s, it is worth mentioning that the heat produced during the actuation of the electromagnet is not high enough to affect the ELISA-like assay for detection of the infA/H1N1 virus. Within 40 min of the detection time, it was observed that the droplet volume was decreased by less than 10% as a result of the heat generated by the electromagnetically-driven platform. Therefore, the developed electromagnetic-driven platform which could generate a magnetic flux density ranging from 240–350 mT was used for the subsequent experiments.

In addition, the mixing mechanism in the droplet-based platform depended on an electromagnet placed on top of the microfluidic chip with a magnetic flux density of 110 mT when an input current/voltage of 1.6 A/3.23 V was applied. The electromagnet could attract the magnetic beads vertically to agitate the liquid such that gentle mixing could be generated. For characterization of the mixing index, two droplets with different colors were merged, and the magnetic beads inside were attracted by the top-side electromagnetic mixer with a driving frequency of 0.5 Hz. Then the images were observed over time and the mixing index was calculated accordingly. Detail information could be found in the following section.

The Arduino microcontroller was further integrated into the microfluidic chip along with the electromagnetically-driven platform with a stepper motor (NEMA17, LDO MOTORS Co., Ltd, China) such that the electromagnets placed underneath could be moved to control the movement of the droplets and beads.

Preparation of aptamer-conjugated magnetic beads

The H1N1-specific aptamer (sequence of the aptamer is as follows: 5′-TTTTT TTTGG CAGGA AGACA AACAG CCAGC GTGAC AGCGA CGCGT AGGGA CCGGC ATCCG CGGGT GGTCT GTGGT GCTGT-3′) was screened by using systematic evolution of ligands by exponential enrichment (SELEX) process.31 The aptamer was demonstrated to have high affinity and specificity toward infA/H1N1 virus with a dissociation constant of 55.14 ± 22.40 nM,31 which is comparable to conventional antibodies. The 5′- end of this aptamer (with a volume of 12 μL and a concentration of 100 μM) was modified with an amine group and conjugated onto carboxylic acid-coated superparamagnetic beads with a volume of 100 μL (1 × 109 beads) (stock concentration = 1 × 1010 beads per mL, beads diameter = 1.05 μm, 65[thin space (1/6-em)]011 Dynabeads® MyOne™ carboxylic acid, Invitrogen, USA)22 and blocked by ethanolamine (2-aminoethanol, monoethanolamine, MEA, Sigma-Aldrich Co., USA) and stored at 4 °C prior to use.

Preparation of virus samples and reagents

For infA/H1N1 detection, the infA/H1N1 virus (97N510H1, stock concentration = 32 HAU, 20 μL) was provided from Department of Microbiology and Immunology, National Cheng Kung University, Taiwan. The reagents including 1st antibodies (GTX127357, influenza A H1N1 virus anti-HA antibody, GeneTex Inc., USA) (1[thin space (1/6-em)]:[thin space (1/6-em)]250 in PBS with 0.1% BSA, 20 μL), 2nd HRP-conjugated antibodies (abs-22, goat anti-rabbit antibody, HRP conjugated, Asia Bioscience Co. Ltd., Taiwan) (1[thin space (1/6-em)]:[thin space (1/6-em)]2000 in PBS with 0.1% BSA, 20 μL), wash buffers (ddH2O for virus isolation and PBS for antibodies conjugation, 20 μL), and TTMR (PerkinElmer Life and Analytical Sciences, Inc., USA, 20 μL) for the TSA assay were prepared.

For specificity test, influenza A H3N2 (InfA/H3N2, A/California/7/2004, 16 HAU) and influenza B (InfB, 94N399IB, 32 HAU) viruses were obtained from the Department of Microbiology and Immunology, National Cheng Kung University, Taiwan. The infA/H3N2 and infB viruses (each 20 μL) replaced the infA/H1N1 virus in the ELISA-like assay when testing the specificity of the developed assay. For spiked sample tests, the spiked samples (throat swab) were obtained from a healthy volunteer with informed consent, and the throat specimen was diluted with 100 μL of PBS. The whole procedure was handled under biosafety level 2 (BSL-2) conditions and the nasopharyngeal swab procedure was followed by the guidance from the National Influenza Research Laboratory in London.32

Results and discussion

Characterization of the magnetic digital microfluidic platform

The electromagnetic-driven platform was composed of two parts. The first part was a permanent magnet, which provided a magnetic flux density of 240 mT. The second part was an electromagnet, as a magnetic regulator, the magnetic flux density generated by the electromagnet could be adjusted by the input current/voltage for distinct droplet/bead manipulations. Two parameters were first optimized for the fabrication of the electromagnets. The first one is the diameter of the brass coils. Experimental results show that the brass wire with a diameter (Φ) of 0.45 mm could decrease the input voltage because of its relatively low resistance when compared to the brass wire with a diameter of 0.23 mm (ESI Table S2). The lower voltage with a stable current had the benefit of the lower power requirement and less heat generation by the electromagnet. The next parameter which was optimized was the turns of the electromagnet. ESI Table S3 shows the relationship between the turns of the electromagnet, input voltage applied and the generated magnetic flux density. From the table, it is clearly observed that a magnetic flux density of 110–115 mT was the operating limit for electromagnets with 490–625 turns. The maximum magnetic flux density was measured to be about 116 mT at an input current/voltage of 1.8A/3.79 V. However, due to the heat generation issue, this study used 3.23 V and 1.6 A to operate the electromagnet with a magnetic flux density of 110 mT. Moreover, within 40 min of the detection time, it was observed that the droplet volume was decreased by less than 10% as a result of the heat generated by the electromagnetically-driven platform. Experimental results showed that for a droplet with the volume of 20 μL containing 2 × 108 beads, it could be moved out of a SET with a diameter of 0.3 mm by a magnetic flux density of 320 mT. It means that the electromagnetically-driven system capable of generating 240–350 mT, which could effectively distinguish the manipulations of droplets and beads (will be discussed in the next section). When compared to the previous magnetic digital microfluidic platforms that only used permanent magnet as the driving force,25 the developed electromagnetically-driven platform in this study had advantages namely, generating higher magnetic flux density, faster response time (0.5 s) and automatic control. Furthermore, it has a great potential for point-of-care applications.

Manipulations of droplets and beads on the super-hydrophobic surface and droplet SETs

In a magnetic digital microfluidic system, three kinds of forces affect the manipulation of droplets, namely, friction force, capillary force and magnetic force.33 The magnetic digital microfluidic platform in this study was established with surface modification treatment. The super-hydrophobic surface with a contact angle of 155° could effectively decrease the resistance of the friction force and the relatively hydrophilic droplet SETs with a contact angle of 95° could effectively define the areas for droplet positioning. As shown in Fig. 4(a), a droplet with a volume of 20 μL could be manipulated with a magnetic flux density of 240 mT on the super-hydrophobic surface and stayed at the circular droplet SET with a diameter of 0.3 mm. For the droplet escape, a magnetic flux density of 350 mT was applied (Fig. 4(b)). Compared with previous studies using physical limitations (i.e. microstructures), this kind of droplet SET mechanism was structure-free, which indicated that the droplets were moveable and no longer permanently fixed.27 Furthermore, in comparison with chemical modification for droplet trapping, the hydrophobicity method is also favorable since it may not adhere biological samples (ESI Fig. S1 and S2). According to a previous study,34 PDMS adsorbed a number of compounds on the surface. In this study, experimental results showed that the magnetic beads may be slightly adhered (i.e. ignorable part of beads) in the PDMS-based SET. However, the issue does not exist on the super-hydrophobic surface.
image file: c9lc01126a-f4.tif
Fig. 4 (a) Manipulation of droplet containing magnetic beads from a super-hydrophobic surface to a droplet SET with a magnetic flux density of 240 mT; (b) manipulation of droplet getting out of a SET with a magnetic flux density of 350 mT; (c) process of beads extraction. Magnetic beads were extracted from the left droplet and entered the right droplet with a magnetic flux density of 240 mT; (d) comparison between natural diffusion (upper) and electromagnetic mixing by an electromagnet with a magnetic flux density of 110 mT (below).

When the magnetic force applied to the droplet was less than the sum of the friction force and the capillary force but higher than the capillary force, it means that the magnetic force was high enough to operate magnetic beads only, indicating that they could be extracted from the droplet while the droplets was constrained within the droplet SETs. Fig. 4(c) shows that 2 × 108 magnetic beads were extracted from a droplet with a volume of 20 μL and entered into another droplet at a magnetic flux density of 240 mT within 10 s (i.e. collecting for 5 s and moving for 5 s). Note that the electromagnetically-driven platform provided a magnetic flux density of 240 mT when the electromagnet was not applied with any power. The magnetic flux density was only generated from the permanent magnet, resulting in no heating issue in this condition.

Characterization of the electromagnetic mixer

The electromagnetic mixer placed on the top of the microfluidic chip was used to attract the magnetic beads vertically to agitate the liquid such that gentle mixing could be generated. The electromagnet could generate a magnetic flux density of 110 mT at an input current/voltage of 1.6A/3.23 V. When the power was turned on, the magnetic beads (107/μL in 20 μL) were attracted to the top of the droplet, and then precipitate later when the power was turned off. With this approach, beads could be mixed efficiently. Fig. 4(d) shows the comparison between natural diffusion and electromagnetic mixing operated at a frequency of 0.5 Hz. Experimental data show that the initial mixing index was only 19.4% and became 82.9% after 60 seconds, which demonstrated an effective mixing process. Compared to conventional magnetic mixing mechanism on a digital microfluidic platform,23 the electromagnetic mixer in this study is more effective since it had three-dimensional mixing process instead of the two-dimensional movement of the beads.26

LOD of the developed assay on magnetic digital microfluidic chip

The entire process described previously was automated on the developed microfluidic system within 40 min. The on-chip fluorescent signals were measured by a microplate photometer for infA/H1N1 virus with concentrations ranging from 32 to 3.2 × 10−3 HAU (Fig. 5(a)). The LOD was experimental found to be 3.2 × 10−2 HAU, which was 1000-fold lower than the conventional serological assays,35 30-fold lower than the commercial RIDTs36,37 and comparable to the newest RIDTs approach16 (104–106 copies = 3.2 × 10−2–3.2 × 10−4 HAU of the sample used in this study). As expected, the fluorescent intensity signals increased along with the concentration of infA/H1N1 virus (Fig. 5(b)), and the R2 value of the calibration curve was found to be 0.9734. The error bars in this figure represented standard deviation of the mean value (n = 3). In the previous studies by using a continuous microfluidic system, the LOD of infA/H1N1 virus was found to be 0.032 HAU.12,22 Therefore, this magnetic digital microfluidic system has great potential as a promising tool for rapid and accurate infA/H1N1 detection.
image file: c9lc01126a-f5.tif
Fig. 5 (a) The sensitivity for the detection of the infA/H1N1 viruses with the concentrations from 32 to 3.2 × 10−3 HAU on the microfluidic chip (n = 3). NC = negative control. Standard deviation from 32 to 3.2 × 10−3 HAU = 0.202, 0.073, 0.091, 0.068, 0.027; (b) the calibration curve of different concentrations of infA/H1N1 viruses from 32 to 3.2 × 10−3 HAU against fluorescent intensity (n = 3). R2 = 0.9734.

Specificity tests

The specificity of the ELISA-like assay was explored by utilizing infA/H3N2 (16 HAU) and infB (32 HAU). Experimental data showed that the fluorescent signal of the infA/H1N1 was 2 times higher than the others, which were the same as the negative control (i.e. full assay process without any kind of virus) (Fig. 6(a)). These results demonstrated that this developed assay is specific to infA/H1N1 detection and may serve as a promising platform. Multiple ELISA-like assays may be feasible if the virus-specific aptamers were used.38
image file: c9lc01126a-f6.tif
Fig. 6 (a) The fluorescent signals of specificity test for detection of the infA/H1N1 viruses measured by microplate photometer (n = 3); NC = negative control; InfB = influenza B virus. Standard deviation from left to right = 8881.989, 3698.578, 5662.915, 5239.809; (b) the fluorescent signals of spiked infA/H1N1 virus with the concentration of 3.2 HAU (n = 3), NC = negative control. Standard deviation = 4989.198 and 2987.373.

Spiked sample tests

Furthermore, the spiked samples (throat swab) were tested. The throat specimen was diluted with 100 μL of PBS. To simulate the positive sample, 2 μL of infA/H1N1 virus with a concentration of 32 HAU was added into 18 μL throat specimen in PBS such that a spiked sample with a final concentration of 3.2 HAU in 20 μL was prepared for the subsequent test. Experimental data show that the fluorescent signal of the positive infA/H1N1 spiked sample was 1.7 times higher than the negative one (Fig. 6(b)). These results indicated that this platform has the great potential to be used in clinical settings.


A rapid, sensitive and accurate infA/H1N1-specific ELISA-like assay was conducted on a magnetic digital microfluidic system. By using the electromagnetically-driven platform, super-hydrophobic surface, relatively hydrophilic droplet SETs and an electromagnetic mixer, the entire assay could be carried out in the droplets with magnetic beads precisely and flexibly. It provided an adjustable magnetic flux density ranging from 240 to 350 mT to attract the magnetic beads for droplet manipulation including droplet transportation (>320 mT) and beads extraction (<320 mT). Additionally, the electromagnetic mixer was designed for liquid agitation such that a gentle mixing in droplets could be achieved. By vertically agitating the magnetic beads, the liquid inside droplets could be gently mixed at a frequency of 0.5 Hz. Moreover, each droplet served as a reaction chamber, and the consumption of samples and reagents could be significantly decreased because of the droplet size, which was only 20 μL. Furthermore, this kind of magnetic digital microfluidic platform could effectively decrease the power consumption such that a compact power supply or a set of batteries with 3.23 V could support the entire system. The LOD of the ELISA-like assay was experimentally found to be 0.032 HAU, which was comparable to the previous studies and close to newest RIDTs. The reaction time of the entire assay on the magnetic digital microfluidic platform was less than 40 min, which demonstrated the great potential for a rapid diagnosis. Therefore, this magnetic digital microfluidic platform may be promising for influenza A H1N1 virus diagnosis.

Conflicts of interest

There are no conflicts to declare.


The authors would like to thank financial support from Ministry of Science and Technology (MOST) of Taiwan (MOST 107-2221-E-013-MY3 and MOST 108-2314-B-007-002). Partial financial supports from the “Higher Education Sprout Project” of Taiwan's Ministry of Education (Grant No. 108Q2713E1) and National Health Research Institutes (NHRI-EX107-10728EI) are also greatly appreciated. The authors also thank Dr. Chih-Peng Chang for providing influenza viruses.


  1. L. Ferguson, A. K. Olivier, S. Genova, W. B. Epperson, D. R. Smith, L. Schneider, K. Barton, K. McCuan, R. J. Webby and X. F. Wan, J. Virol., 2016, 90, 5636–5642 CrossRef CAS PubMed .
  2. H. Nair, W. A. Brooks, M. Katz, A. Roca, J. A. Berkley, S. A. Madhi, J. M. Simmerman, A. Gordon, M. Sato, S. Howie, A. Krishnan, M. Ope, K. A. Lindblade, P. Carosone-Link, M. Lucero, W. Ochieng, L. Kamimoto, E. Dueger, N. Bhat, S. Vong, E. Theodoratou, M. Chittaganpitch, O. Chimah, A. Balmaseda, P. Buchy, E. Harris, V. Evans, M. Katayose, B. Gaur, C. O'Callaghan-Gordo, D. Goswami, W. Arvelo, M. Venter, T. Briese, R. Tokarz, M. A. Widdowson, A. W. Mounts, R. F. Breiman, D. R. Feikin, K. P. Klugman, S. J. Olsen, B. D. Gessner, P. F. Wright, I. Rudan, S. Broor, E. A. F. Simoes and H. Campbell, Lancet, 2011, 378, 1917–1930 CrossRef .
  3. R. G. Webster and E. A. Govorkova, in Antimicrobial Therapeutics Reviews: Infectious Diseases of Current and Emerging Concern, ed. K. Bush, Blackwell Science Publ, Oxford, 2014, vol. 1323, pp. 115–139 Search PubMed .
  4. W. R. Dowdle, J. C. Galphin, M. T. Coleman and G. C. Schild, Bull. W. H. O., 1974, 51, 213–218 CAS .
  5. S. V. Vemula, J. Q. Zhao, J. K. Liu, X. Wang, S. Biswas and I. Hewlett, Viruses, 2016, 8, 15 CrossRef PubMed .
  6. C. Fraser, C. A. Donnelly, S. Cauchemez, W. P. Hanage, M. D. Van Kerkhove, T. D. Hollingsworth, J. Griffin, R. F. Baggaley, H. E. Jenkins, E. J. Lyons, T. Jombart, W. R. Hinsley, N. C. Grassly, F. Balloux, A. C. Ghani, N. M. Ferguson, A. Rambaut, O. G. Pybus, H. Lopez-Gatell, C. M. Alpuche-Aranda, I. B. Chapela, E. P. Zavala, D. M. E. Guevara, F. Checchi, E. Garcia, S. Hugonnet and C. Roth, Science, 2009, 324, 1557–1561 CrossRef CAS PubMed .
  7. E. Ghedin, N. A. Sengamalay, M. Shumway, J. Zaborsky, T. Feldblyum, V. Subbu, D. J. Spiro, J. Sitz, H. Koo, P. Bolotov, D. Dernovoy, T. Tatusova, Y. M. Bao, K. St George, J. Taylor, D. J. Lipman, C. M. Fraser, J. K. Taubenberger and S. L. Salzberg, Nature, 2005, 437, 1162–1166 CrossRef CAS PubMed .
  8. G. J. D. Smith, D. Vijaykrishna, J. Bahl, S. J. Lycett, M. Worobey, O. G. Pybus, S. K. Ma, C. L. Cheung, J. Raghwani, S. Bhatt, J. S. M. Peiris, Y. Guan and A. Rambaut, Nature, 2009, 459, 1122–1125 CrossRef CAS PubMed .
  9. G. Neumann, T. Noda and Y. Kawaoka, Nature, 2009, 459, 931–939 CrossRef CAS PubMed .
  10. V. C. C. Cheng, K. K. W. To, H. Tse, I. F. N. Hung and K. Y. Yuen, Clin. Microbiol. Rev., 2012, 25, 223–263 CrossRef CAS PubMed .
  11. R. W. Y. Chan, M. C. W. Chan, J. M. Nicholls and J. S. M. Peiris, Virus Res., 2013, 178, 133–145 CrossRef CAS PubMed .
  12. L. Y. Hung, T. B. Huang, Y. C. Tsai, C. S. Yeh, H. Y. Lei and G. B. Lee, Biomed. Microdevices, 2013, 15, 539–551 CrossRef CAS PubMed .
  13. E. D. Heegaard, B. L. Petersen, C. J. Heilmann and A. Hornsleth, J. Clin. Microbiol., 2002, 40, 933–936 CrossRef CAS PubMed .
  14. X. Cai, Z. Lei, W.-L. Fu, Z.-Y. Zhu, M.-J. Zou, J. Gao, Y.-Y. Wang, M. Hong, J.-X. Wang and W.-J. Li, Afr. J. Microbiol. Res., 2013, 7, 3919–3925 CAS .
  15. D. E. Dwyer, D. W. Smith, M. G. Catton and I. G. Barr, Med. J. Aust., 2006, 185, S48–S53 Search PubMed .
  16. D. A. Green and K. StGeorge, J. Clin. Microbiol., 2018, 56, e00711–e00718 CAS .
  17. C. H. Wang, C. P. Chang and G. B. Lee, Biosens. Bioelectron., 2016, 86, 247–254 CrossRef CAS PubMed .
  18. M. D. Mavrouli, J. G. Routsias, H. C. Maltezou, N. Spanakis and A. Tsakris, Viral Immunol., 2011, 24, 221–226 CrossRef CAS PubMed .
  19. M. R. Siavashi, H. Taherkhani, K. Rezaei, M. R. R. Deligani and M. Assmar, Iran. Biomed. J., 2005, 9, 91–94 Search PubMed .
  20. H. J. Zheng, J. A. Fuhrman, M. Xu, W. F. Cheng, M. V. R. Reddy and W. F. Piessens, Chin. Med. J., 1990, 103, 709–712 CAS .
  21. S. Carinelli, M. Marti, S. Alegret and M. I. Pividori, New Biotechnol., 2015, 32, 521–532 CrossRef CAS PubMed .
  22. Y. T. Tseng, C. H. Wang, C. P. Chang and G. B. Lee, Biosens. Bioelectron., 2016, 82, 105–111 CrossRef CAS PubMed .
  23. Y. Zhang and N. T. Nguyen, Lab Chip, 2017, 17, 994–1008 RSC .
  24. W. C. Nelson and C. J. Kim, J. Adhes. Sci. Technol., 2012, 26, 1747–1771 CAS .
  25. Z. C. Long, A. M. Shetty, M. J. Solomon and R. G. Larson, Lab Chip, 2009, 9, 1567–1575 RSC .
  26. Y. Zhang and T. H. Wang, Adv. Mater., 2013, 25, 2903–2908 CrossRef CAS PubMed .
  27. Y. Zhang, S. Park, K. Liu, J. Tsuan, S. Yang and T. H. Wang, Lab Chip, 2011, 11, 398–406 RSC .
  28. N. R. Geraldi, L. E. Dodd, B. B. Xu, G. G. Wells, D. Wood, M. I. Newton and G. McHale, Surf. Topogr.: Metrol. Prop., 2017, 5, 034001 CrossRef .
  29. K. R. Langley, E. Q. Li, I. U. Vakarelski and S. T. Thoroddsen, Soft Matter, 2018, 14, 7586–7596 RSC .
  30. C.-H. Chiou, D. Jin Shin, Y. Zhang and T.-H. Wang, Biosens. Bioelectron., 2013, 50, 91–99 CrossRef CAS PubMed .
  31. H.-C. Lai, C.-H. Wang, T.-M. Liou and G.-B. Lee, Lab Chip, 2014, 14, 2002–2013 RSC .
  32. K. G. Nicholson, K. R. Abrams, S. Batham, M. J. Medina, F. C. Warren, M. Barer, A. Bermingham, T. W. Clark, N. Latimer, M. Fraser, N. Perera, K. Rajakumar and M. Zambon, Health Technol. Assess., 2014, 18(36), 1–274 Search PubMed .
  33. D. J. Shin and T. H. Wang, Ann. Biomed. Eng., 2014, 42, 2289–2302 CrossRef PubMed .
  34. M. A. Rad, K. Ibrahim, K. Mohamed and N. Najimudin, Micro Nano Lett., 2013, 8, 305 CrossRef CAS .
  35. A. C. Hurt, R. Alexander, J. Hibbert, N. Deed and I. G. Barr, J. Clin. Virol., 2007, 39, 132–135 CrossRef CAS PubMed .
  36. C. Chartrand, M. M. G. Leeflang, J. Minion, T. Brewer and M. Pai, Ann. Intern. Med., 2012, 156, 500–511 CrossRef PubMed .
  37. Y. Sakai-Tagawa, M. Ozawa, S. Yamada, Y. Uchida, T. Saito, K. Takahashi, N. Sugaya, M. Tashiro and Y. Kawaoka, Microbiol. Immunol., 2014, 58, 600–606 CrossRef CAS PubMed .
  38. V.-T. Nguyen, H. B. Seo, B. C. Kim, S. K. Kim, C.-S. Song and M. B. Gu, Biosens. Bioelectron., 2016, 86, 293–300 CrossRef CAS PubMed .


The preliminary results of the current study were presented at the 22nd International Conference on Miniaturized Systems for Chemistry and Life Sciences (μTAS 2018), held in Kaohsiung, Taiwan (November 11–15, 2018).
Electronic supplementary information (ESI) available. See DOI: 10.1039/c9lc01126a

This journal is © The Royal Society of Chemistry 2020