Micromotor-Based Biosensing Using Directed Transport of Functionalized Beads

Previous micromotor-based biosensing studies used to functionalize the surface of the micromotor with specific molecular probes for binding of target analyte, thus limiting the use of the micromotor for the specific target. In contrast, here, we introduce a novel approach of using a nonfunctionalized micromotor as a generic cargo carrier being able to perform label-free and dynamic loading, transport, and release of functionalized beads. Hence, such an approach enables one to use the same micromotor system for sensing of varying targets via different commercially available functionalized beads, demonstrating the use of micromotors as a practical and versatile means for biosensing. We have also introduced a simplified microfluidic design that can be used for immunosensing or DNA binding tests without necessity for complicated fluid handling (buffer exchange, washing, etc.) steps. We expect this approach to open up new realizations of simplified and generic biosensing platforms.

T he reliable identification of the target analytes using labon-a-chip technology promises an enormous potential in the fields of environmental monitoring, food testing, food/ water safety monitoring, and clinical analysis. 1 In particular, the emerging field of self-propulsion (i.e., "active" particles) combined with microfluidics offers new opportunities for realization of novel micromotor-based biosensing, targeted drug delivery, and cell manipulation/isolation for biomedical application. 2−7 Active particles that convert energy at the particle level from the surrounding environment (e.g., chemical fuel, 8−11 light, 12,13 ultrasound, 14 magnetic field, 15−18 and electric field 19−21 ) into autonomous self-propulsion enables efficient coverage of large areas and volumes while operating under uniform ambient conditions without the need to generate field gradients for driving the particles.
By using different materials and surface coating techniques (e.g., molecularly imprinted polymers, 8,22 biodegradable and biocompatible polymers, 23 and biomimetic material 24 ), the active particles coated with biodegradable materials or functionalized with bioreceptors have been proven to be useful for various tasks such as antibacterial activity, 23 protein detection, 25 and intracellular biosensing of a target miRNA expressed in intact cancer cells. 26 As one notable approach, García et al. 8 demonstrated micromotor-based immunoassay using self-propelled antibody-functionalized micromotors where the different immunoassay steps are obtained via the mobile particle translating between different reservoirs connected using microfluidic channels. This concept eliminates the need to manipulate fluids as common to lab-on-a-chip devices; thus, the washing step is obtained by the motion of the particle itself within the stagnant fluid. However, the coating of the active particle with antibodies makes these micromotors specific and less generic in application.
Here, we demonstrate a generic approach using a nonlabeled micromotor that can selectively load, transport, and release functionalized beads (i.e., treated as cargo) singularly controlled by an external alternating electric field. We use spherical metallodielectric Janus particle (JPs) as the micromotor, 19,21,27,28 where under an applied electric field, the imbalanced polarization at metallic and dielectric hemispheres results in self-propulsive behavior referred to induced-charge electrophoresis (ICEP) 29,30 and self-dielectrophoresis (sDEP). 19 Both the speed and direction of the JP are controlled by the applied electric field frequency, with the JP propelling either with its dielectric (ICEP) or metallic (sDEP) hemisphere forward. 19,28,29 The underlying mechanism of the cargo manipulation used in the current work is dielectrophoresis (DEP), which enables label-free loading and release of a broad range of organic and inorganic cargos. 20 Under a uniform external electric field, the field gradients, necessary for DEP force, are induced at the JP level. By varying the electric field frequency, a cargo particle can shift between attraction (i.e., positive DEP, pDEP) and repulsion (i.e., negative DEP, nDEP) at regions of high electric field intensity according to its geometry and material properties. 31−33 Thus, combining an electrically powered micromotor and DEP-based cargo manipulation enables us to use a uniform electric field to singularly control the selective loading, transport, and release of the functionalized beads in a simple and robust manner. Adding magnetic steering, using a ferromagnetic coating of the JPs with an externally rotating static magnet, enabled us also to precisely navigate the micromotor. 34−36 Here, we present micromotor-based biosensing using simplified microfluidic devices consisting of either a single chamber ( Figure 3) or two chambers with connecting microchannel ( Figure 5). As described in the schematics in Figure 1, the micromotor first loads the functionalized beads as cargos and then transports them toward the sample of the target analyte for some certain incubation time for its "on-thefly" binding with specific molecular probes that are functionalized on the surface of the beads. This is followed by a washing step wherein the beads are transported into a region free of the target analyte for visualization of the binding before release. The micromotor can be reused to load and transport new cargos following release of the previously transported cargos. Without loss of generality, we used biotin-avidin with high affinity as a simplified model for immunosensing or protein and nucleic acid detection. While micromotor-based cargo loading was previously studied, 20,35,36 herein, we investigate for the first time its applicability for biosensing using commercially available functionalized microbeads. This involves consideration of the diffusion of the analyte within the microfluidic chamber and the dependency of the bimolecular binding between the target analyte and surface-immobilized probes on the analyte concentration as well as the retention time within the region where the analyte is introduced.

■ RESULTS AND DISCUSSION
Cargo Loading and Transport. Here, we characterized the cargo (biotin-coated beads) manipulation (i.e., loading, transport, and release). A 10 μm-diameter Ni-Au-coated Janus particle was used as a micromotor within an indium tin oxide (ITO)-sandwiched microchamber wherein the bottom ITOcoated glass slide was covered with thin silicon dioxide layer to suppress adsorption of particles onto the substrate 27 (more details are given in the Supporting Information). In particular, we examined the effect of the conductivity of a phosphatebuffered saline (PBS) solution on the micromotor velocity ( Figure 2b and Figure S1). As shown, the electrokinetic propulsion of the JP becomes ineffective at high solution conductivities (>1 mS/cm; see Figure S1). The chosen solution conductivity at which we performed the study was 0.01% (v/v) PBS at which the JPs move fast enough ( Figure  2b and Figure S1) to shorten the translocation time between the different regions but also enable sufficiently strong biotinavidin binding affinity.
As shown in Figure 2a,b, there are two modes of selfpropulsion, where the JP moves with its dielectric end forward under induced-charge electrophoresis (ICEP) at low frequencies and with its metallic end forward (sDEP) beyond a certain critical frequency. Also, the applied electric field frequency affects the dielectrophoretic response of the cargos, which transitions from pDEP to nDEP response with increasing frequency beyond the crossover frequency (COF) ∼10 kHz ( Figure 2b).
Here, we chose to use the sDEP propulsion mode with nDEP cargo trapping (500 kHz) due to the larger cargo loading capacity (there is less hydrodynamic shear since electroconvection is significantly reduced at such high frequencies) and the relatively large carrier mobility. At lower frequency (5 kHz) with ICEP propulsion and pDEP cargo trapping mode, the cargos were trapped only at the top of the JP (see simulation inset in Figure 2d) due to its inability to penetrate underneath the JP due to size limitation but transported with a nonsmooth motion due to abrupt stops as a result of the loaded cargo ( Figure S1). At the mode of sDEP and nDEP cargo trapping, the micromotor with magnetic steering is directed to capture functionalized beads in a consecutive manner with particles trapped on the equator of its metallic hemisphere, as seen in (Figure 2c, schematic inset in Figure 2d, and Supplementary Video S1). Figure 2d depicts how the increased number of loaded cargo is decreasing the mobility of the micromotor. This can be simply explained due to the increased Stokes drag resulting from the enlarged areal cross section of the loaded carrier due to the trapped cargos. 35,36 The maximal cargo loading capacity the micromotor is able to pick up was ∼7 for the current operating conditions. However, the maximal loading capacity can be changed depending on the JP size and operating conditions ( Figure S2, Figure S3, and Supplementary Video S2).
A Single Microfluidic Chamber Setup. Here, we demonstrate our novel approach using a very simplified microfluidic setup consisting of a single chamber made of a spacer and two drilled inlet holes for introduction of the solution, functionalized beads, micromotor, and sample analyte ( Figure 3). We can qualitatively differentiate between several (a) A Janus particle picks up biotin-functionalized beads and transports them to selected regions. An electric field is used for both JP propulsion and cargo manipulation (negative DEP trapping), while a magnetic field is used for steering of the JP. The metallic coated (Cr/Ni/Au) is colored yellow, while the nonfluorescent bare polystyrene hemisphere is colored dark gray. The red-and graycolored cargos indicate the biotin-functionalized particles with and without avidin-biotin binding, respectively. (b) Biotin-avidin reactions are used as a simplified model of immunosensing and DNA biosensing, wherein the biotin-coated beads bind to avidin molecules that are introduced into the sample solution.

ACS Sensors
pubs.acs.org/acssensors Article zones in between one of the inlets where the target sample analyte is gently introduced and the other inlet which is free of target analyte. Note that the distance between the two holes (∼6.5 mm) is far enough for the target analyte not to reach the other inlet by diffusion during the whole process (∼180 h with an avidin molecule diffusion coefficient of D ≈ 6.5 × 10 −7 cm 2 s −137 ). The micromotors and cargos (i.e., functionalized beads) were first introduced uniformly within the chamber before introducing the target analyte, although they could be as well introduced only into the inlet far from the analyte after filling the chamber with the solution. We have then used the micromotor to trap the cargos to follow a path along the avidin molecule concentration gradients for binding followed by a washing step. The binding of avidin molecules to the biotincoated beads did not seem to change their DEP behavior in a way that necessitates readjustment of the operating parameters. The time of incubation can be then controlled by either increasing the retention time of the micromotor with loaded cargos within region I of the highest avidin concentration or simply picking up the beads within region I after some desired time (see Figure 4). As seen in Figure 3b, a clear binding event occurred between the avidin and the biotin-coated cargos following an incubation time of ∼5 min. This shortened detection time, relative to the long diffusion time of avidin within the chamber, is clearly due to motion of the trapped cargos, via the micromotor, to the region of high analyte concentration. This is further quantified using different avidin concentrations to test the limit of detection of the system, which proves that, for two cycles, concentration as low as 2 μg/ mL can be detected. It is expected that this can be further improved by simply increasing the incubation time ( Figure  4c). As a control, we have used non-biotin-coated microparticles of the same size, which showed no fluorescence signal

ACS Sensors
pubs.acs.org/acssensors Article when compared to the functionalized beads (inset in Figure  4c). Extension to More Complicated Microchannel Geometries. Such a simplified microchamber that does not necessitate any fabrication except drilled holes can be further complicated while still kept simple in terms of fabrication to include additional chambers if one, for example wishes, to perform multiplex (i.e., involving different functionalized beads), sandwich immunoassay, or several washing/buffer exchange/chemical reaction steps. 8,38−41 To prove the applicability of our generic micromotor-based functionalized cargo transport approach also to these potential future realizations of biosensing, we have fabricated a two-microchamber device with connecting microfluidic channel fabricated using direct laser cutting ( Figure 5 and Supplementary Videos S5 and S6). It is clearly demonstrated that the JP can travel very long distances (through the connecting microchannel) while transporting cargo.

■ CONCLUSIONS
We have successfully demonstrated the ability to perform labelfree and dynamic manipulation and transport of biotin-coated beads used as cargo via a JP acting as an active carrier. We have examined the ability to perform cargo loading under various solution conductivities including one (i.e., PBS) that is relevant for the binding of biotin-avidin. It was demonstrated that, using a single microfluidic chamber, simple and robust biosensing could be realized with the advantages of a micromotor-based approach avoiding the need for fluid handling as incubation time and washing steps can be simply controlled by moving the carrier itself. We have demonstrated the ability of transporting such a loaded carrier for very long distances through a microchannel connecting between two microchambers, which may be important for realizations of several buffer exchange/washing steps and/or multiplex sensing. This approach is applicable for a broad range of cargos from ∼100 nm size to several micrometer size. It is also possible to use a mixed population of cargos for multiplex biosensing where their loading/release can be controlled individually by taking advantage of their unique polarizability relative to the medium (i.e., Clausius−Mossotti factor) and the associated different crossover frequencies (COFs) at which the cargo switches between positive-and negative-DEP responses. 36 Although examined for the biotin-avidin reaction, the approach is generic and can be applied for any antibody−

ACS Sensors
pubs.acs.org/acssensors Article antigen or DNA−probe combination. It can also be extended to include more microchambers for several reagents and washing steps as, for example, in a sandwich immunoassay where a reporting antibody is used for the fluorescence signal following the antigen binding to the antibodies immobilized on the surface of the cargos. Although we did not observe any significant effect of the binding on the DEP behavior of the cargos such that it necessitates tuning of the operating conditions, it is possible to easily tune the operating frequency if there is such an effect. The method is of particular use for low concentration of target analytes where the fast transportation of the carrier and trapped cargo toward the analyte region and away from it (i.e., washing step) enables fast detection that otherwise would be long due to the slow diffusion processes. Also, while we have demonstrated this using a single JP guided using magnetic steering so as to precisely control its translocation between the different regions and its incubation time within the region into which the analyte is introduced (region I), an operation mode where several such carriers are simultaneously operating autonomously (i.e., without magnetic steering) so as to enhance sampling of the analyte (i.e., reducing the otherwise long diffusion time) is also possible. Due to the limitation of the electrokinetic propulsion to solution conductivities lower than ∼1 mS/cm, we chose, in the current study, a relatively low conductivity 0.01% (v/v) PBS solution at which the JP moves fast with sufficiently strong biotin-avidin binding affinity. However, since the cargo manipulation via DEP on the surface of the JP is valid also at physiological solutions of higher conductivity, it is only the propulsion mechanism that needs to be improved. One option is to replace the electrokinetic propulsion by other means, e.g., rotating magnetic field. 34,42 Another possible solutions may be to suppress the quenching of the EDL to higher solution conductivity by using highly ion-conductive polyelectrolyte coating. 43 We expect such a novel approach to open new opportunities in realizing simplified micromotor-based biosensing devices.

ACS Sensors
pubs.acs.org/acssensors Article within solutions of higher conductivity, transition of cargo transport from sDEP with nDEP mode to ICEP with pDEP mode, cargo loading and transport using a 27 μm-diameter Janus particle micromotor, and supplementary video legends (PDF) Supplementary Video S1, biotin-coated particles loading and transport by a single JP suspended in 0.01% diluted PBS under an applied electric field of 500 kHz and 20 V pp (corresponds to Figure 2c) (MOV) Supplementary Video S2, loading and transport of avidin-bound biotin cargos using a 27 μm-diameter Janus particle micromotor at a low frequency (20 kHz and 20 V pp ) under ICEP propulsion mode and nDEP trapping mode (corresponds to Figure S2

■ ACKNOWLEDGMENTS
This work was supported by ISF grant 1938/16. Fabrication of the chip was made possible through the financial and technical support of the Russell Berrie Nanotechnology Institute and the Micro-Nano Fabrication Unit. We would like to acknowledge Eric Guyes and Prof. Matt Suss for assisting in laser fabrication of the two-chamber microfluidic device and Dr. Yue Wu for the technical assistance.