Spatiotemporal Control of Electrokinetic Transport in Nanofluidics

May 21, 2015 - Pinpoint Delivery of Molecules by Using Electron Beam Addressing Virtual Cathode Display. Takayuki Hoshino , Moto Yoshioka , Akira ...
1 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/Langmuir

Spatiotemporal Control of Electrokinetic Transport in Nanofluidics Using an Inverted Electron-Beam Lithography System Hiroki Miyazako,*,†,‡ Kunihiko Mabuchi,† and Takayuki Hoshino*,† †

Department of Information Physics and Computing, Graduate School of Information Science and Technology, and ‡Research Fellow of the Japan Society for the Promotion of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan S Supporting Information *

ABSTRACT: Manipulation techniques of biomolecules have been proposed for biochemical analysis which combine electrokinetic dynamics, such as electrophoresis or electroosmotic flow, with optical manipulation to provide high throughput and high spatial degrees of freedom. However, there are still challenging problems in nanoscale manipulation due to the diffraction limit of optics. We propose here a new manipulation technique for spatiotemporal control of chemical transport in nanofluids using an inverted electron-beam (EB) lithography system for liquid samples. By irradiating a 2.5 keV EB to a liquid sample through a 100-nm-thick SiN membrane, negative charges can be generated within the SiN membrane, and these negative charges can induce a highly focused electric field in the liquid sample. We showed that the EB-induced negative charges could induce fluid flow, which was strong enough to manipulate 240 nm nanoparticles in water, and we verified that the main dynamics of this EB-induced fluid flow was electroosmosis caused by changing the zeta potential of the SiN membrane surface. Moreover, we demonstrated manipulation of a single nanoparticle and concentration patterning of nanoparticles by scanning EB. Considering the shortness of the EB wavelength and Debye length in buffer solutions, we expect that our manipulation technique will be applied to nanomanipulation of biomolecules in biochemical analysis and control. have been applied to immuno assay16 and measurement of diffusion coefficient.17 Although these opto-electric manipulation techniques have a lot of advantages as mentioned above, there are still fundamental limitations to spatiotemporal resolution. First, the spatial resolution of laser optics is limited due to the diffraction of the light beams. In the case of opto-electrokinetic tweezers, the minimum size of a virtual electrode, which was generated by light illumination, was 1.52 μm when a digital micromirror device was used,10 and 2.8 μm when a liquid crystal display was used.18 Second, the actual trapping region of the particles was several micrometers or more in length because the driving force of the particles or fluid exists outside the optically excited area. The rapid electrokinetic patterning7,8 achieved the manipulation of 2.0 to 0.1 μm particles, but the cluster size of the particles was about 5.0 to 10.0 μm. Also, the size of the trapping region by opto-electric tweezers19 was also several micrometers or more in length. Third, the temporal resolution of optical scanning is also limited to about the kilohertz order due to the mechanics of the lens. Thus, opto-electrokinetic manipulations are not able to achieve control of fluid or particles at a single nanomolecule level or a microsecond level.

1. INTRODUCTION Manipulation techniques of biomolecules and cells in solutions are basic techniques for biochemical analyses. Especially, electrokinetic manipulations using electroosmosis, such as electrophoresis, dielectrophoresis, or electroosmotic flow (EOF), have been widely studied and applied to lab-on-a-chip devices,1−5 because these manipulations are noninvasive and provide high throughput. Moreover, electrical characteristics such as electrical charges or permittivity are applicable for manipulations of several types of biomolecules. However, the spatial resolution and degrees of freedom are restricted, since these manipulations use several electrodes fixed on an experimental system. In order to achieve high degrees of freedom and resolution, hybrid opto-electrokinetic manipulation techniques6−11 have been rigorously studied in recent years. These techniques use both electrodes and laser optics. Local illumination of light can cause a temperature rise (rapid electrokinetic patterning7,8) or a change in an electric field (opto-electrohydrodynamic assembly9 and opto-electrokinetic tweezers10,11) around the illumination spot, and control of the electric field between the electrodes will enable direct and local control of fluid. By changing the illumination pattern of the light, spatiotemporal control of fluid can be achieved. These hybrid opto-electrokinetic manipulations have achieved spatiotemporal and high-throughput manipulations of bio- and nonbiomaterials,12−15 and these manipulations © 2015 American Chemical Society

Received: March 3, 2015 Revised: May 18, 2015 Published: May 21, 2015 6595

DOI: 10.1021/acs.langmuir.5b00806 Langmuir 2015, 31, 6595−6603

Article

Langmuir

Figure 1. (a) Overview of an inverted EB lithography system for wet samples. Negative charges of scattered electrons generate a virtual electrode on the surface of the SiN membrane (orange region), and it can generate electroosmotic phenomena. (b) Average radial velocity of 240-nm-diameter nanoparticles within 50 μm from the EB (mean ± SD (standard derivation), n = 10). Radial movement of a nanoparticle occurred during irradiation of the EB spot (1.0−2.0 s), and its direction was reversed when the membrane had the PEI coating.

microscope as proposed by Hoshino et al.25,26 (Figure 1a). EB irradiation to a liquid sample through a 100-nm-thick SiN membrane can have mechanical, chemical and electrostatic effects on the liquid sample. We observed the behavior of the suspension around the surface of the SiN membrane with the fluorescent microscope. In this research, we observed fluid flows in deionized water (DI water) during 2.5 keV EB irradiation. Figure 1b plots the mean velocity of fluid flows near the SiN membrane when the EB was irradiated. When the bare SiN membrane was used, the observed flow was toward the EB spot, whereas the observed flow was outward from the EB spot when the polyetheleniminecoated (PEI-coated) SiN membrane was used. The physical model for the generation of these EB-induced fluid flows is as follows. Some of the scattered electrons are charged within the SiN membrane, which is a dielectric. Since the electrons have negative charges, the atoms around the electrons are positively polarized and the atoms at the surface of the SiN membrane are negatively polarized. Therefore, the negative charges are generated at the surface of the SiN membrane, and an electric field can be generated in liquid samples (Figure 2a). When the EB was irradiated to the SiN membrane continuously, the temporal variation of the EB-induced electric field became very small, and the EB-induced electric field could behave as a DC or a low-frequency electric field. Therefore, the EB-induced electric field can cause electrophoresis and EOF. When a spherical particle, which is inherently charged, exists in a liquid sample, the motion equation of the spherical particle is as follows:

One possible approach to overcome these limitations is to use electron beams (EBs). EBs have the advantage of spatiotemporal resolution, because the wavelength of the EB (the de Broglie wavelength) is much shorter than that of visible light, and the mass of an electron is small enough to manipulate it on a microsecond scale. Up to now, many researchers have observed liquid samples using EBs in order to achieve better resolution compared to optical observation.20−23 In addition to these observations, EBs have been applied to manipulation of wet samples using TEM24 or SEM.25,26 Our research group has applied an EB lithography system for actuation of wet samples.25,26 We reported the basic mechanism for mechanical, electrical, and chemical actuations of wet samples by using an inverted EB lithography system,25 and we demonstrated that EB irradiation on a wet sample with a 2.5 keV electron beam through a 100-nm-thick silicon nitride (SiN) membrane can induce chemical deposition on a living cell.26 One of the driving forces of our EB manipulation was the strong and focused electric field induced by negative charges of scattered electrons, so there was a possibility that an inverted EB lithography system for wet samples could induce electroosmosis, which can be applied to the spatiotemporal control of electrokinetic transport of wet samples at a nanometer scale. In this paper, we showed that the change of the electrokinetic field was able to cause EOF and electrophoretic force and that 240-nm-diameter nanoparticles in a liquid suspension could be manipulated with an EB irradiation. In addition, we demonstrated that manipulation of a single nanoparticle and concentration patterning of nanoparticles could be achieved by a simple EB scanning. Finally, we concluded that the proposed technique would be a better spatiotemporal method than the previous techniques and it could be applied for manipulation of nanometer-scale molecules.

mr ̈ = Fdrag (r) + Q E(r) = 6πηa(u(r) − r)̇ + Q E(r)

where η represents the dynamic viscosity coefficient of water; m, a, and Q represent the mass, radius, and electric charge of the particle, respectively; and u(r) and E(r) represent the velocity of the EOF and the electric field induced by the EB at the coordinate r, respectively. The distribution of the EB-induced electric field E(r) is modeled by Poisson equations under the condition that the scattered electrons exist in the SiN membrane (details are shown in Supporting Information S2). The velocity

2. MODELS: EB-INDUCED ELECTROOSMOSIS 2.1. EB-Induced Electrophoresis and Electroosmosis. To make an electrokinetic field for a liquid suspension dynamically, we used opposed coaxial dual optics, which consisted of inverted EB optics and a fluorescent optical 6596

DOI: 10.1021/acs.langmuir.5b00806 Langmuir 2015, 31, 6595−6603

Article

Langmuir

Figure 2. Conceptual diagram of EB-induced virtual electrode and electroosmosis. (a) Distribution of electrical flux lines induced by negative charges of scattered electrons. λD and λE represent the Debye length of wet sample and the maximum of the radial range of the EB, respectively. (b,c,d) Electrical flux lines and direction of electrophoretic force (QE) and drag force of EOF (Fdrag), provided that both the nanoparticles and the SiN membrane are negatively charged.

the EOFs are focused around the EB spot and then fluid instability occurs. Due to these vertical flows, drag force and electrophoretic force are generated in the region where r < λe and z > λd (Figure 2d). It should be noted that the directions of drag force of the EOFs and the electrophoretic force depend on the surface charge of the particles and SiN membrane. 2.2. Simulations. In order to get the spatial distribution of scattered electrons by EB irradiation, we did Monte Carlo simulations using CASINO v 2.4.28 In these simulations, there were three layers, the vacuum layer, 100-nm-thick SiN layer, and the H2O layer. An EB of 2.5 keV (beam radius, 25 nm) was irradiated from the vacuum to the SiN layer, and trajectories of scattered electrons were simulated. Detailed conditions of these simulations are described in Supporting Information S1. Since accelerated electrons had both kinetic and electrostatic energies and the EB kinetic energy was converted into thermal energy through the collisions of atoms, electrostatic and thermal effects induced by the EB were also simulated by the FEM simulator, COMSOL Multiphysics 4.4 (COMSOL Inc.). Electrostatics and species transport were combined to simulate the electric field profile, and heat transport and fluid dynamics were combined to simulate the temperature rise and thermal convection. Since the kinetic energy of electrons was so small (about 10 μW), we assumed that the temperature rise was slight and all the physical parameters used in the FEM simulations were

of the EOF u(r) satisfies the following Navier−Stokes equation with external electrokinetic force:27 η ∂u 1 + (u ·∇)u = − ∇p + ∇2 u + ρe E(r) ∂t ρ ρ

where ρ, p, and ρe represent the density, pressure, and the space charge density, respectively. ρe is determined by the concentration distribution of cations and anions in the liquid sample, and the distribution can be modeled by the Poisson−Boltzmann equation or the Poisson-Nernst-Plank equation.27 Electrophoretic force and drag force of EOF are the main driving forces for manipulation of liquid samples in the proposed method. The direction of electrophoretic force is along that of the electric field, whereas the direction of drag force differs according to the position because EOF, the origin of drag force, is generated at the SiN membrane surface. For symmetry of the surface direction, the coordinate of the position r is hereafter represented in the cylindrical coordinate system (r, θ, z). In the region where r is sufficiently larger than the maximum of the radial range of EB (λe) and z is smaller than the Debye length of the liquid sample (λd) as shown in Figure 2b, the electrical field is almost along the surface of the SiN membrane; thus, the horizontal EOF, which is parallel to the SiN membrane, is generated. In the region where r < λe and z < λd (Figure 2c), vertical flows as well as horizontal flows are generated, because 6597

DOI: 10.1021/acs.langmuir.5b00806 Langmuir 2015, 31, 6595−6603

Article

Langmuir spatially uniform. Detailed conditions of these simulations are described in the Supporting Information S2 and S3.

3. MATERIALS AND METHODS 3.1. Inverted EB Lithography System for Liquid Samples. An inverted EB optics (Mini-EOC, Apco Co.) was used for direct EB irradiation of the liquid sample in an atmospheric environment. The accelerating voltage was set at 2.5 kV. A 100-nm-thick SiN membrane (NT025C, NORCADA) separated the vacuum chamber from the atmospheric environment. A fluorescent microscope (IM-3, Nikon) equipped with a 100× water immersion objective lens (NA = 1.1 and WD = 2.5 mm, Nikon) and an LED white light source (SOLA, Lumencor) was used for observation of nanoparticles in the liquid sample. A fluorescent filter block (B-2A, Nikon) was used for fluorescent observation of nanoparticles. The space between the objective lens and the SiN membrane was filled with DI water in advance before introducing the sample as microdroplets, and the microscope was focused on the surface of the SiN membrane after the sample was introduced. A high sensitivity scientific CMOS camera (pixel size = 6.5 μm × 6.5 μm, NEO sCMOS, Andor) was mounted on the microscope to capture the motion of nanoparticles during the EB irradiation. The images were acquired at 50 frames per second, and the acquisition was started 1 s before the EB irradiation. The start trigger input of the sCMOS and the blanker of the EB were connected to I/O board (BNC2120, National Instruments) to synchronize EB irradiation and image acquisition. The deflector of EB was also connected to the I/O board to scan the EB two-dimensionally. 3.2. Materials. To visualize the EOF, 240 nm yellow fluorescent polystyrene nanoparticles (FP-0252−2, Spherotech) were used. The nanoparticles were diluted in DI water, and 2.0 × 10−2 w/v % (Sample A) and 1.0 × 10−2 w/v % (Sample B) nanoparticle dilutions were prepared. A 1.0 × 10−1 w/v % (Sample C) nanoparticle dilution was also prepared by mixing 10 μL of 1.0 w/v % nanoparticle dilution and 90 μL of 0.1 w/v % Pluronic F-127 (Invitrogen) solution. Two SiN membrane types were used, bare and PEI-coated. The PEI coating process was as follows. First, 10 μL of 0.1 w/v % polyetheleneimine (Sigma-Aldrich) solution was dropped using a micropipette onto a SiN membrane. After leaving this for 1 h at room temperature, the PEI solution was removed by sucking up the solution in a pipet. The SiN was washed with DI water several times and air-dried. 3.3. Observation of Nanoparticles and Luminescence in EB Irradiated Liquid Suspension. A 100 μL aliquot of Sample A was carefully dropped onto the surface of bare SiN membrane and a 100 μL of Sample B was dropped onto the surface of PEI-coated SiN membrane. Spot EB was irradiated onto each sample for 1 s, and the beam current of the EB was set at 4 nA. Fluorescent images of nanoparticles and EBinduced luminescence were acquired by the sCMOS camera. The spot EB irradiation onto Sample A was performed 3 times, and the spot EB irradiation onto Sample B was performed 10 times. 3.4. Two-Dimensional Manipulation of a Single Nanoparticle and Patterning of Concentration. We demonstrated twodimensional manipulation of a single nanoparticle and patterning of concentration by rapid EB scans. To make it easier to observe concentration patterns of nanoparticles, a 10 μL aliquot of Sample C was dropped onto the surface of a bare SiN membrane. The beam current of the EB was set at 13 nA. 3.5. Image Processing. All the images were processed using ImageJ (NIH). To suppress background noise, we averaged images for 1 s before EB manipulation, subtracted the averaged image from each frame of the video, and applied a 3 × 3 median filter to each frame of the video unless otherwise noted. To measure flow velocity during EB irradiation, we summed up the results of each trial, suppressed background noise as mentioned above, and tracked gravity points of nanoparticles by using Move-tr/2D (Library Co.).

Figure 3. Simulation results of scattered primary electrons and the electric field induced by the EB (X−Z plane view). (a) Simulated trajectories of scattered electrons (blue lines) and backscattered electrons (red lines) which were ejected from the vacuum space. (b) z-Axial distribution of scattered electrons (except backscattered electrons). (c) Simulated electrostatic field induced by scattered electrons. The color map represents the voltage and the black lines stand for the electrical flux lines. (d,e) Distribution of x-axial component (d) and z-axial component (e) of volume force by EB-induced electric field. (f) Decomposition of electrostatic field.

simulated trajectories of scattered electrons; their accelerating voltage was set to 2.5 kV. Figure 3 b shows the z-axial distribution of the scattered electrons. Only 0.03% of the scattered electrons were transmitted to the liquid layer. Figure 3c shows the electrostatic potential and electrical flux lines in the liquid layer, which were induced by the external electric field of scattered electrons in the SiN membrane. The electrical flux lines became dense around the EB spot, which means that the electric field was large. Figure 3d,e shows the x-axial and z-axial components of the volume force by the induced electric field, where the decomposition of the electric field was represented as Figure 3f. Each component had a peak at 100−200 nm, which meant that the volume force was mainly applied within 1 μm. The heat transport simulations showed that the 1 s irradiation of the 2.5 keV EB could only raise the temperature several degrees Kelvin, and the flow velocity around the wall was smaller than that of the actual flow (see Supporting Information S3). 4.2. Observation of EB-Induced Fluid Flow and SiN Luminescence in DI Water. In this study, the fluorescent nanoparticles were diluted using DI water; thus, the background intensity of fluorescent images indicated the concentration around focal surface. The brightness within 20 μm from the EB spot decreased monotonically during the EB spot irradiation, which meant that the concentration around the EB spot was decreased. On the other hand, the brightness outside the 20 μm region around the EB spot did not change (Figure 4a).

4. RESULTS 4.1. Simulations of Scattered Electrons and Their Electric Field in Liquid Sample. Figure 3a shows the 6598

DOI: 10.1021/acs.langmuir.5b00806 Langmuir 2015, 31, 6595−6603

Article

Langmuir

Figure 5. Spatiotemporal transition of EB-induced luminescence. (a) Intensity of EB-induced luminescence after start of EB irradiation. As irradiation time passed, the height of the intensity peak decreased. (b) Averaged intensity of EB-induced luminescence within 0.5 μm from the EB spot.

Figure 4. Spatial dependency of EB-induced flow when a bare SiN membrane was used. (a) Relationship between distance from EB spot and intensity change over time. I ̅ means the averaged intensity over the area of each region, and Ii̅ nit means the initial value of I ̅ (the noise suppression was not performed). The intensity within 20 μm decreased during EB irradiation. (b,c) Spatial dependence of the mean radial (b) and mean tangential (c) velocities 40 ms before and 40 ms after the start of the EB irradiation (mean ± SD, n ≥ 10).

over time (Figure 5a). The ratio of mean intensity of the luminescence to the initial intensity monotonically decreased during EB irradiation, and the luminescence was quenched just after the end of EB irradiation (Figure 5b). 4.4. Manipulation of Solutes by Scanning Electron Beam. Figure 6 shows manipulation of a single nanoparticle movement. By scanning the EB triangularly, a single nanoparticle was rotated around a target point (Figure 6a). The coordinates of the gravity point of the tracked single nanoparticle were converted from Cartesian coordinates to polar coordinates where the origin position was set to the target point as shown in Figure 6a, and the rotation angle was defined as the angle between the initial line of the polar coordinates and the line between the origin and the gravity point of the tracked nanoparticle. The variation of rotation angle from the manipulation was as shown in Figure 6b. The nanoparticle was rotated counterclockwise during EB manipulation, whereas it had Brownian motion before and after EB manipulation. The movie of the actual rotational movement of the nanoparticle is shown in Supporting Information Movie3. The scanning EB moved the single nanoparticle in a straight line, where the scanning angle between the direction of the EB scan and the horizontal axis (θobj) was 88° (Figure 6c). The absolute values of velocity and the angle of velocity are shown in Figure 6d. The velocity peaked between 0.54 and 0.60 s, when the difference between the angle of velocity and the scanning angle became

The mean radial velocity of tracked nanoparticles was as shown in Figure 1b. When the bare SiN membrane was used, the nanoparticles moved toward the EB spot, whereas they moved outward from the EB spot when the PEI-coated SiN membrane was used. During EB irradiation, the absolute values of mean velocity decreased monotonically over time, and moved in the opposite direction. The movement of particles returned to Brownian motion about 200 ms after the end of the EB irradiation (Supporting Information Movie1 (n = 3) and Movie2 (n = 10) show the movements of the actual nanoparticles during the EB irradiation when bare and PEI-coated membranes were used, respectively). Figure 4b,c show spatial dependence of the radial and tangential velocity 40 ms before and 40 ms after the start of the EB irradiation. The radial velocity decreased as a function of distance from the EB spot, whereas the tangential velocity was unchanged before and after the EB irradiation. 4.3. Observation of EB-Induced Luminescence in DI Water. The intensity of EB-induced luminescence decreased 6599

DOI: 10.1021/acs.langmuir.5b00806 Langmuir 2015, 31, 6595−6603

Article

Langmuir

Figure 6. Demonstration of rotational and translational manipulations of a single nanoparticle. (a) Trajectory of rotationally manipulated nanoparticle. Yellow circle indicates the target point of the rotational manipulation, and yellow arrows indicate the spatiotemporal pattern of the EB scan. (b) Amount of rotation of the nanoparticle. When the nanoparticle was manipulated as shown in (a), the rotation amount monotonically increased. (c) Trajectory of translationally manipulated nanoparticle. Yellow arrows indicate the spatiotemporal pattern of the EB scan. (d) Error from the target angle (θobj = 88°) and velocity of the manipulated nanoparticle as shown in (c). The error decreased and the velocity was maximized between 0.54 and 0.60 s.

5. DISCUSSION 5.1. Existence of EB-Induced EOF. In this study, we proposed a new control method and its physical model of electrokinetic transport in liquid samples by EB irradiation. The driving forces of manipulation in the proposed method were electrophoretic force and fluid drag force of EOF induced by the EB-induced virtual electrode. Actually, the radial movement of the particles and the reversal of the direction of this movement without the PEI coat on the SiN membrane surface implied the existence of EOF and ruled out the possibility that thermal convection was the main mechanism of bottom flow. Moreover, the intensity decrease of particles around the EB spot implied the coexistence of EOF and electrophoresis. Besides the experimental results, the FEM simulations also supported our findings that the driving force of EOF could exist around the EB spot and the thermal convection could not affect particle transport in the observed region. Since the zeta potential of the nanoparticles was about −4 mV (Manufacturer’s information), the nanoparticles were negatively charged inherently, and were subject to electrophoretic force and the drag force of the electroosmotic flow when an electrostatic field was applied. In spite of the existence of these two forces, the drag force of EOF would be bigger than the electrophoretic force around the surface of the SiN membrane according to the result that the direction of the movement of the nanoparticles was reversed when the bare SiN membrane was coated with PEI as

almost zero. The movie of the actual translational movement of the nanoparticle is shown in Supporting Information Movie4. Figure 7 shows the spatiotemporal patterns of nanoparticle concentration by scanning “μ” and “Λ” patterns. The brightness

Figure 7. Two-dimensional concentration patterning of (a) “μ” and (b) “Λ” shape patterns. (Left) Actual EB scan pattern. (Right) Temporal variation of fluorescence. Degradation of fluorescence was observed in the scanned area.

of the scanned area decreased, and concentration patterns emerged as these letters. Supporting Information Movie5 and Movie6 show the actual patterning of “μ” and “Λ”, respectively. 6600

DOI: 10.1021/acs.langmuir.5b00806 Langmuir 2015, 31, 6595−6603

Article

Langmuir

The Monte Carlo simulation results also implied that the proposed method could generate an EB-induced virtual electrode having a size of about 100 nm. This size was ten times less than that of opto-electric tweezers.10,11 Therefore, the proposed method could be expected to achieve nanoscale manipulations of biomolecules, which would not be achievable by opto-electrical manipulations. Although the proposed method achieved the size of the virtual electrode as simulated, the experimental results showed that the nanoparticles within 30 μm from the EB spot moved in the radial direction during the irradiation. The spatial resolution of the proposed technique, however, was tens of micrometers, and we could not achieve better spatial resolution than in previous studies. One of the reasons for this was the Debye length. In our paper, we used DI water as solvent in order to suppress the effect of electric shielding of ions. The Debye length of DI water is about 1 μm,27 so the range of the electrokinetic effect was large. Future work should include experiments with changes of the ion concentrations in the liquid suspension to shorten the Debye length and achieve better spatial resolution. In our experiments, the EB was irradiated continuously, so the effect of dielectrophoresis was considered to be small. However, the FEM simulation results implied that the profile of the EBinduced electric field was nonuniform. Therefore, the dielectrophoretic force as well as electrophoretic force should be considered when the EB is manipulated at high frequency such as rapid and repeated EB scanning or blanking. Our future research will include dielectric phenomena of electron-beam manipulation and its effects on liquid samples. Our findings indicated that the effect of electroosmosis should be considered in SEM observations for wet samples.20−23 The accelerating voltages of these studies were higher than that in our study (2.5 keV), so the electric field generated by the scattered electrons would have induced EOF the same as in our experiments, and the fluid motion would affect the spatial resolution of images of wet samples. The SEM observation for solid materials was reported to be affected by the collisions of scattered electrons and their electrostatic force.32 These facts implied that the effect of not only kinetic and electrostatic energies of the EB, but also the EOF induced by the EB, should be considered for SEM observations of wet samples to achieve better resolution. 5.2. Time Response of EOF. The monotonic decrease of the average radial velocity and EB-induced luminescence during EB spot irradiation for 1 s implied that the strength of the EB-induced electric field, the driving force of electrophoresis and EOF, would decrease. We thought that one of the reasons for this decrease was deceleration of electrons by negative charges of the SiN membrane. The negative charges of the SiN membrane induced an electric field not only in the liquid sample, but also in the vacuum space, and the electrons were subjected to repulsion electrostatic force and decelerated by the EB-induced electric field, which would result in the deceleration of electrons. At the same time, the range of the EB also decreased, and thus the effect of the EB-induced electric field on the liquid sample decreased. This deceleration of electrons by charges of sample is known as “retarding” and it has been used to increase resolution of SEM images.33 The return to Brownian motion 200 ms after the end of EB irradiation implied that the surface charge was released and there would be no external electric field due to this charge. This meant that the proposed method could easily switch the EB-induced electric field on and off at intervals of several hundreds of

shown in Figure 1b. The PEI coating could affect the zeta potential of the surface of the SiN membrane, which determined the direction of EOF, and could not affect the zeta potential of the nanoparticles. Therefore, the drag force of EOF would contribute to the reversal of nanoparticle movement when the PEI was coated on the SiN membrane, which indicated the drag force of EOF would be bigger than electrophoretic force. The electrophoretic force was small in the experiments, because the zeta potential of the nanoparticles was small (about −4 mV) in the experiments and the electrophoretic velocity of the nanoparticles were decelerated by the electric force applied to the excess cations (H+) within the electric double layer around the nanoparticles. If the nanoparticles are increasingly charged, the electrophoretic effects due to the inherent charges and polarization charges due to the EB-induced electric field will become more significant. Our future work will include the details of inherent and polarization charges of highly charged nanoparticles during EB irradiation and their effects on the manipulation. The direction of observed movement of particles was consistent with previous studies. These studies showed that the zeta potential of solid SiN in pH 7.0 aqueous solution was about −40.0 mV,29,30 so the cations (H+) in DI water were thought to be attracted to the bare SiN membrane when it was used, which meant that the spatial charge density ρe was positive near the surface of the bare SiN membrane. On the other hand, PEI is a cationic polymer,31 so the anions (OH−) in DI water were thought to be attracted to the PEI-coated SiN membrane when it was used, which meant that ρe was negative near the surface of the coated SiN membrane. The radial electric field Er was negative, because the direction of the EB-induced electric field E was toward the EB spot. Thus, the radial component of the driving force of EOF ρeEr was negative when the bare SiN membrane was used, and positive when the PEI-coated SiN membrane was used, which was consistent with the directions of actual movement of particles shown in Figure 1b. Since the EB-induced electric field was highly focused according to the FEM simulations, there was a possibility that cathodic reactions such as electrolysis of DI water occurred, and the anionic chemical species generated by the cathodic reactions might affect the movement of the nanoparticles during the EB irradiation. However, the effect of electroosmotic flow due to the electric field of electrons would be larger than that of the generation of anionic chemical species even if cathodic reactions had occurred during the EB irradiation, because our experiments showed that the direction of particle movement was reversed when PEI was coated on the SiN membrane, and the PEI coating had nothing to do with cathodic reactions. Therefore, our results provided strong evidence supporting the physical model of the proposed method, which was based on EB-induced electroosmosis, and the possibility of manipulation of liquid samples by electrophoretic force and drag force of EOF. From the viewpoint of the manipulation mechanism, our method was quite different from that of hybrid manipulation methods using optics and electrodes,6−11 because our method could directly induce a focused electric field and easily control it. This direct control of an electric field implied that our manipulation method did not need any special coating or layer such as an indium tin oxide layer or photoconductive layer on an electrode or membrane, nor did it need a complicated control system for the optics and electrodes. Thus, our manipulation method would achieve both simplicity of the experimental system and highthroughput of manipulation. 6601

DOI: 10.1021/acs.langmuir.5b00806 Langmuir 2015, 31, 6595−6603

Article

Langmuir

buffer, and the design for control of electrokinetic transport by EB scanning.

milliseconds. The return to Brownian motion after the end of EB irradiation, however, could have several causes. One possible cause would be the discharge of the injected electrons from the SiN membrane, such as secondary electron emission into a vacuum or leakage of current to the silicon nitride membrane and the liquid sample. This discharge of electrons would induce the change of the electric field. Another cause would be the sudden change of the electric field generated by induced-charge electroosmosis.34 Future work should include study of the physical model for the discharge of electrons and its temporal effects on manipulation. 5.3. Two-Dimensional Control of Electrokinetic Transport. The proposed method achieved translational manipulation of a single particle as well as two-dimensional patterning of particle concentrations by using EB-induced electroosmosis. This meant the possibility exists for spatiotemporal manipulation of biomaterials the same as in opto-electrical manipulation methods.6−11 Considering that the mass of electrons is very small and thus the time resolution of the EB scan can be achieved on a microsecond scale, we expect that simultaneous generation of several virtual electrodes in the two-dimensional space by rapid EB scanning would be possible, and that the proposed method would provide high throughput and more degrees of freedom than the opto-electrical manipulation methods.6−11 The control rule of EB-induced electrodes, however, would be more complicated than that of the opto-electrical manipulation methods,6−11 because the EB spot generates both vertical and horizontal flows around the EB spot, and thus the trapping point cannot be generated by a simple EB spot irradiation, whereas the trapping point can be generated by the opto-electrical manipulation methods.6−11 However, it would be possible to generate various electric field lines and stream lines of EOF by spatiotemporal arrangement of EB-induced virtual electrodes by rapid EB scanning in order to generate trapping points or get the desired control of particles. Therefore, future work should include the design of an electric field by EB scanning to generate desired particle concentration patterns or to control particle movement.

■

ASSOCIATED CONTENT

* Supporting Information S

Detailed description of simulation conditions and 6 movies of the electron-beam manipulation. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.5b00806.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was supported by a research grant from the Precise Measurement Technology Promotion Foundation and JSPS KAKENHI Grant Number 26282162, 15H05511, 15J09841.

■

REFERENCES

(1) Meighan, M. M.; Staton, S. J. R.; Hayes, M. A. Bioanalytical Separations Using Electric Field Gradient Techniques. Electrophoresis 2009, 30, 852−865. (2) Voldman, J. Electrical Forces for Microscale Cell Manipulation. Annu. Rev. Biomed. Eng. 2006, 8, 425−454. (3) Wang, X.; Cheng, C.; Wang, S.; Liu, S. Electroosmotic Pumps and Their Applications in Microfluidic Systems. Microfluid. Nanofluidics 2009, 6, 1−34. (4) Stone, H. A.; Stroock, A. D.; Ajdari, A. Engineering Flows in Small Devices. Annu. Rev. Fluid Mech. 2004, 36, 381−411. (5) Henares, T. G.; Funano, S.; Sueyoshi, K.; Endo, T.; Hisamoto, H. Advancements in Capillary-Assembled Microchip (CAs-CHIP) Development for Multiple Analyte Sensing and Microchip Electrophoresis. Anal. Sci. 2014, 30, 7−15. (6) Kumar, A.; Williams, S. J.; Chuang, H. S.; Green, N. G.; Wereley, S. T. Hybrid Opto-Electric Manipulation in Microfluidics-Opportunities and Challenges. Lab Chip 2011, 11, 2135−2148. (7) Williams, S. J.; Kumar, A.; Wereley, S. T. Electrokinetic Patterning of Colloidal Particles with Optical Landscapes. Lab Chip 2008, 8, 1879− 1882. (8) Kumar, A.; Kwon, J. S.; Williams, S. J.; Green, N. G.; Yip, N. K.; Wereley, S. T. Optically Modulated Electrokinetic Manipulation and Concentration of Colloidal Particles near an Electrode Surface. Langmuir 2010, 26, 5262−5272. (9) Hayward, R.; Saville, D.; Aksay, I. Electrophoretic Assembly of Colloidal Crystals with Optically Tunable Micropatterns. Nature 2000, 404. (10) Chiou, P. Y.; Ohta, A. T.; Wu, M. C. Massively Parallel Manipulation of Single Cells and Microparticles Using Optical Images. Nature 2005, 436, 370−372. (11) Hwang, H.; Park, J. K. Optoelectrofluidic Platforms for Chemistry and Biology. Lab Chip 2011, 11, 33−47. (12) Choi, W.; Nam, S. W.; Hwang, H.; Park, S.; Park, J. K. Programmable Manipulation of Motile Cells in Optoelectronic Tweezers Using a Grayscale Image. Appl. Phys. Lett. 2008, 93, 143901. (13) Daw, R.; Finkelstein, J. Opto-Electrokinetic Manipulation for High-Performance on-Chip Bioassays. Lab Chip 2006, 12, 4955−4959. (14) Hwang, H.; Park, J. K. Rapid and Selective Concentration of Microparticles in an Optoelectrofluidic Platform. Lab Chip 2009, 9, 199−206. (15) Hwang, H.; Park, J. K. Dynamic Light-Activated Control of Local Chemical Concentration in a Fluid. Anal. Chem. 2009, 81, 5865−5870.

6. CONCLUSION In this paper, we proposed a new control method of electrokinetic transport in liquid samples by using an inverted EB lithography system. We first presented a physical model of the proposed method, which was based on the electroosmosis induced by the negative charges of EBs in an SiN membrane. Second, we showed that the 2.5 keV EB could generate EOF in DI water, and that two characteristic flows could be generated, vertical flow around the EB spot and horizontal flow which was parallel to the surface of SiN membrane. These results were consistent with the presented physical model. We also showed the time-dependent behavior of the EB-induced EOF. Finally we demonstrated applications of the proposed method to rotational and translational manipulations of a 240-nm-diameter polyethylene nanoparticle in two-dimensional space, and twodimensional patterning of concentration of 240-nm-diameter nanoparticles. We expect the proposed method to enable both highthroughput and nanoscale manipulation of biomolecules to be achieved because of the smallness of the EB-induced virtual electrode, controllability of EOF, and high response of the induced electric field. To realize this, future work should include studies on the relationships of ion concentrations of aqueous solution and the effect of electrostatic shielding in a 6602

DOI: 10.1021/acs.langmuir.5b00806 Langmuir 2015, 31, 6595−6603

Article

Langmuir (16) Hwang, H.; Chon, H.; Choo, J.; Park, J. K. Optoelectrofluidic Sandwich Immunoassays for Detection of Human Tumor Marker Using Surface-Enhanced Raman Scattering. Anal. Chem. 2010, 82, 7603−7610. (17) Hwang, H.; Park, J. K. Measurement of Molecular Diffusion Based on Optoelectrofluidic Fluorescence Microscopy. Anal. Chem. 2009, 81, 9163−9167. (18) Hwang, H.; Choi, Y. J.; Choi, W.; Kim, S. H.; Jang, J.; Park, J. K. Interactive Manipulation of Blood Cells Using a Lens-Integrated Liquid Crystal Display Based Optoelectronic Tweezers System. Electrophoresis 2008, 29, 1203−1212. (19) Jamshidi, A.; Neale, S. L.; Yu, K.; Pauzauskie, P. J.; Schuck, P. J.; Valley, J. K.; Hsu, H.-Y.; Ohta, A. T.; Wu, M. C. NanoPen: Dynamic, Low-Power, and Light-Actuated Patterning of Nanoparticles. Nano Lett. 2009, 9, 2921−2925. (20) Nawa, Y.; Inami, W.; Chiba, A.; Ono, A.; Miyakawa, A.; Kawata, Y.; Lin, S.; Terakawa, S. Dynamic and High-Resolution Live Cell Imaging by Direct Electron Beam Excitation. Opt. Express 2012, 20, 5629−5635. (21) Sato, C.; Manaka, S.; Nakane, D.; Nishiyama, H.; Suga, M.; Nishizaka, T.; Miyata, M.; Maruyama, Y. Rapid Imaging of Mycoplasma in Solution Using Atmospheric Scanning Electron Microscopy (ASEM). Biochem. Biophys. Res. Commun. 2012, 417, 1213−1218. (22) Sugi, H.; Minoda, H.; Inayoshi, Y.; Yumoto, F.; Miyakawa, T.; Miyauchi, Y.; Tanokura, M.; Akimoto, T.; Kobayashi, T.; Chaen, S.; et al. Direct Demonstration of the Cross-Bridge Recovery Stroke in Muscle Thick Filaments in Aqueous Solution by Using the Hydration Chamber. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 17396−17401. (23) Ogura, T. Direct Observation of Unstained Wet Biological Samples by Scanning-Electron Generation X-Ray Microscopy. Biochem. Biophys. Res. Commun. 2010, 391, 198−202. (24) Mirsaidov, U.; Mokkapati, V. R. S. S.; Bhattacharya, D.; Andersen, H.; Bosman, M.; Ö zyilmaz, B.; Matsudaira, P. Scrolling Graphene into Nanofluidic Channels. Lab Chip 2013, 13, 2874−2878. (25) Hoshino, T.; Morishima, K. Electron-Beam Direct Processing on Living Cell Membrane. Appl. Phys. Lett. 2011, 99, 174102. (26) Hoshino, T.; Mabuchi, K. Closed-Looped in Situ Nano Processing on a Culturing Cell Using an Inverted Electron Beam Lithography System. Biochem. Biophys. Res. Commun. 2013, 432, 345− 349. (27) Kirby, B. J. Micro- and Nanoscale Fluid Mechanics: Transport in Microfluidic Devices; Cambridge University Press, 2010. (28) Drouin, D.; Couture, A.; Joly, D.; Tastet, X. CASINO V2. 42A Fast and Easy-to-use Modeling Tool for Scanning Electron Microscopy and Microanalysis Users. Scanning 2007, 29, 92−101. (29) Jaffrezic-Renault, N.; De, A.; Clechet, P.; Maaref, A. Study of the Silicon Nitride/aqueous Electrolyte Interface on Colloidal Aqueous Suspensions and on Electrolyte/insulator/semiconductor Structures. Colloids Surf. 1989, 36, 59−68. (30) Bousse, L.; Mostarshed, S. The Zeta Potential of Silicon Nitride Thin Films. J. Electroanal. Chem. Interfacial Electrochem. 1991, 302, 269− 274. (31) Suh, J.; Paik, H.; Hwang, B. K. Ionization of Poly(ethylenimine) and Poly(allylamine) at Various pH′s. Bioorg. Chem. 1994, 22, 318−327. (32) Egerton, R. F.; Li, P.; Malac, M. Radiation Damage in the TEM and SEM. Micron 2004, 35, 399−409. (33) Phifer, D.; Tuma, L.; Vystavel, T.; Wandrol, P.; Young, R. J. Improving SEM Imaging Performance Using Beam Deceleration. Microsc. Today 2009, 17, 40−49. (34) Squires, T. M.; Bazant, M. Z. Induced-Charge Electro-Osmosis. J. Fluid Mech. 2004, 509, 217−252.

6603

DOI: 10.1021/acs.langmuir.5b00806 Langmuir 2015, 31, 6595−6603