Microfluidics and Nanofluidics

, Volume 9, Issue 4, pp 755–763

Dielectrophoretically assembled particles: feasibility for optofluidic systems


    • Centre for Intelligent Systems ResearchDeakin University
  • Chen Zhang
    • School of Electrical and Computer EngineeringRMIT University
  • Jos L. Campbell
    • School of Applied SciencesRMIT University
  • Aminuddin A. Kayani
    • School of Electrical and Computer EngineeringRMIT University
  • Saeid Nahavandi
    • Centre for Intelligent Systems ResearchDeakin University
  • Arnan Mitchell
    • School of Electrical and Computer EngineeringRMIT University
  • Kourosh Kalantar-zadeh
    • School of Electrical and Computer EngineeringRMIT University
    • Chemical Engineering DepartmentMassachusetts Institute of Technology (MIT)
Research Paper

DOI: 10.1007/s10404-010-0590-7

Cite this article as:
Khoshmanesh, K., Zhang, C., Campbell, J.L. et al. Microfluid Nanofluid (2010) 9: 755. doi:10.1007/s10404-010-0590-7


This work presents the dielectrophoretic manipulation of sub-micron particles suspended in water and the investigation of their optical responses using a microfluidic system. The particles are made of silica and have different diameters of 600, 450, and 250 nm. Experiments show a very interesting feature of the curved microelectrodes, in which the particles are pushed toward or away from the microchannel centerline depending on their levitation heights, which is further analyzed by numerical simulations. In doing so, applying an AC signal of 12 Vp–p and 5 MHz across the microelectrodes along with a flow rate of 1 μl/min within the microchannel leads to the formation of a tunable band of particles along the centerline. Experiments show that the 250 nm particles guide the longitudinal light along the microchannel due to their small scattering. This arrangement is employed to study the feasibility of developing an optofluidic system, which can be potentially used for the formation of particles-core/liquid-cladding optical waveguides.


DielectrophoresisManipulationMicrofluidicsOptofluidicsSilica particles

1 Introduction

Micro and nano particles can be manipulated in microfluidic systems using dielectrophoretic (DEP) forces. A DEP force is generated when a neutral particle is suspended in a non-uniform electric field. This electric field induces electrical charges within the particle to establish a dipole. If a non-uniform electric field is applied, the ends of the dipole experience unequal Columbic forces, which result in a total non-zero, imposed force. If the particle is less polarizable than the suspending medium, it is repelled from the regions of higher electric field and the motion is called negative dielectrophoresis, while the opposite case is referred to as positive dielectrophoresis.

Already, DEP systems are used in applications such as separation, filtering, and permanent localization of micro/nano particles (Cummins et al. 2006; Lao et al. 2006, Ofir, et al. 2008; Webb and Li 2005; Zhang et al. 2009). However, the possibility of forcing particles into forming suspended three-dimensional (3D) objects using dielectrophoresis is yet to be investigated. The advantages of such systems stems from the fact that while suspended particles in liquid can be brought into intimate contact with each other, they can still remain unbounded. Such suspended particles can then be employed to form suspended devices with exceptional functionalities due to the unique properties of the micro/nano materials.

Using DEP forces, suspended particles can be pushed into distinct regions and later be scattered from those regions by applying electric fields. Hence, DEP forces can be employed for the ‘forced-assembly’ of particles, positioning them to establish 3D optical devices within a liquid. The accessibility of a variety of micro, sub-micron, and nanoscale particles can allow us to form such optofluidic systems with various optical scattering and guiding properties. As a result, particles of different dimensions can be used as the building blocks of devices for targeted applications. For instance, it is possible to make suspended optical waveguides that scatter certain wavelengths while allowing others to pass through, using the nanoparticles with dimensions comparable to the wavelengths. Furthermore, a mixture of various materials with distinctly different optical and electrical properties can be mixed and manipulated to create new optical materials suspended in microfluidic systems.

In this article, we describe the DEP patterning of silica sub-micron particles using an array of curved microelectrodes. The magnitude and frequency of the AC signal, which is applied across the microelectrodes, as well as the medium flow rate, are selected such that the microelectrodes act as a funnel, smoothly forming a narrow band of particles along the centerline of the microchannel. The width of the patterned particles remains constant and can be tuned along the microchannel due to the curved configuration of the microelectrodes.

The silica particles have a higher refractive index than the surrounding flow and can confine the coming light. Therefore, our system is a platform to create a particles-core/liquid-cladding optical waveguide, in which the light is confined inside the high refractive index patterned particles (the core) by the low refractive index liquid (the cladding). Optical waveguides with solid-core/liquid-cladding (Schmidt et al. 2007; Yang and Erickson 2008), liquid-core/solid-cladding (Wang and Fang 2005) and liquid-core/liquid-cladding (Bernini et al. 2008; de Matos et al. 2007; Wolfe, et al. 2004) configurations have already been reported. It has also been shown that the characteristics of these optical systems can only be tuned by changing the flow rate and the composition of the liquids, while the solid component remains unchanged (Bernini et al. 2008; Conroy et al. 2005; de Matos et al. 2007; Wolfe et al. 2004, 2005). However, in our waveguides the optical characteristics can be controlled on demand to a much higher degree by changing the composition and size of sub-micron particles as well as the applied signal and frequency.

2 Theory

For a homogenous spherical particle, the time-averaged DEP force is given as (Jones 2003):
$$ \bar{F}_{\text{DEP}} = 2\pi r^{3} \varepsilon_{0} \varepsilon_{\text{m}} \text{Re} [f_{\text{CM}} (\omega )]\nabla E_{\text{rms}}^{2} , $$
where r is the radius of the particle, ε0 is the permittivity of the vacuum, εm is the relative permittivity of the suspending medium, Erms is the root-mean-square (rms) value of the electric field, and Re[fCM] is the real part of the Clausius–Mossotti (CM) factor, calculated as:
$$ f_{\text{CM}} (\omega ) = {\frac{{\varepsilon_{\text{p}}^{*} - \varepsilon_{\text{m}}^{*} }}{{\varepsilon_{\text{p}}^{*} + 2\varepsilon_{\text{m}}^{*} }}}, $$
where εp* and εm* are the complex permittivities of the particle and medium, each defined as:
$$ \varepsilon^{*} = \varepsilon_{0}\,\varepsilon - {\frac{i\sigma }{\omega }}, $$
where \( i = \sqrt { - 1} , \) ε is the relative permittivity, σ is the electric conductivity, and ω is the angular frequency of the applied electric field.The DI water had a conductivity and relative permittivity of 2 × 10−4 S/m and 78, respectively. The silica particles, used in the experiment were fabricated using the Stöber method (Stober et al. 1968). The particles had an initial weight ratio of 5% and were diluted in DI water in a 1:20 volumetric ratio. The overall conductivity of particles is described as σbulk + 2Ks/r, in which σbulk is the bulk conductivity that is negligible for silica particles, and Ks is the total surface conductance composed of the contributions of the Stern layer and diffuse layer formed around the particles (Morgan and Green 2003), while the relative permittivity of particles was taken 3.5. In our experiments, the crossover frequencies of 3.2, 1.45, and 1.35 MHz were obtained, respectively for 250, 450, and 600 nm particles, above which the particles demonstrated negative DEP response. Using this data, we calculated Ks = 1.1 ± 0.15 nS (Hughes et al. 1999), and sketched the Re[fCM] spectra, as given in Appendix 1. At the frequency of 5 MHz, which was applied during our experiments, the Re[fCM] values of −0.26, −0.39, and −0.42, were obtained, respectively for 250, 450, and 600 nm particles.
The sedimentation force applied on the particle is the interaction of the weight and buoyancy given as below:
$$ F_{\text{sedimentation}} = \frac{4}{3}\pi r^{3} (\rho_{\text{p}} - \rho_{\text{m}} )g $$
where ρp and ρm are the densities of the particle and medium, and g is the gravitational acceleration. The densities of silica particles and DI water are ~2,000 and 1,000 kg/m3, respectively.

3 Experimental

Figure 1 presents the plan view of the proposed DEP system. The system consists of a microchannel integrated into a DEP platform. The microchannel is fabricated from poly(dimethylesiloxane) (PDMS) with a width of 1,000 μm and a height of 80 μm. The DEP platform is comprised of a glass slide that supports five pairs of microelectrodes on its surface. To fabricate the microelectrodes, thin films of chrome/gold were deposited on the glass substrate with a thickness of 500 Å/1500 Å. The pattern of the microelectrodes and the microchannel were formed using photolithography techniques (Kalantar-zadeh and Fry 2007). The microelectrodes have a curved configuration with a width of 50 μm, a minimum gap of 40 μm and a distance of 1,000 μm between the consequential pairs. The electric potential is applied through the pads of 6 × 2.25 mm. Due to their curved shape, the microelectrodes create a strong, spatially varying electric field within the microchannel, which is crucial in creating a DEP force that increases smoothly along the microelectrodes, avoiding abrupt movement of particles (Khoshmanesh et al. 2009, 2010).
Fig. 1

The layout of the DEP platform, which is comprised of a PDMS block housing the integrated microchannel fabricated on a glass substrate

4 Results and discussions

The principle of the system performance is illustrated in Fig. 2. The particles that are moving far from the centerline are not affected by the DEP force and keep moving under the drag force (path-1). Conversely, the particles that are moving close to the centerline are affected by the DEP force. Due to their negative DEP behavior, these particles are repelled from the regions of high electric field that are produced at the microelectrode tips. The microelectrodes act as a funnel due to their curved configuration, decelerating the particles and patterning them parallel to their edges (paths-2). Approaching the tips, the DEP forces become stronger and the particles exhibit two distinct responses as described below.
Fig. 2

Under appropriate conditions, the microelectrodes act as a funnel, concentrating the particles and patterning them as a narrow band along the centerline of the microchannel. The light passes through the narrow band of particles

The experiments showed that applying a combination of a high electric potential across the microelectrodes and a low flow rate within the microchannel pushed the particles toward the sidewalls, producing a particle-free region along the centerline (Khoshmanesh et al. 2009). By contrast, the combination of a low potential and a high flow rate funneled the particles, producing a particle-rich region along the centerline.

To comprehend this behavior, we assessed the induced DEP force using the simulation method described in (Khoshmanesh, et al. 2009). In our simulations, the size of particles was set to 450 nm, while the voltage and frequency were set to 12 Vp–p and 5 MHz, respectively. At this frequency, the particles demonstrated negative DEP response, as shown in Appendix 1. The simulation results revealed a very interesting feature of the curved microelectrodes, in which the direction of the DEP-y force strongly depends on the location of particles along the width and height of the microchannel. This response was analyzed at different locations of the microchannel, as shown in Fig. 3A. In doing so, Fig. 3B1–B4 show the variation of DEP-y force at the entry region of the microelectrode, and at different heights of 10, 20, 30, and 40 μm, in which the location of the microelectrode is addressed by dashed lines. The graphs showed that if the particles were located at the inner edge of electrodes and levitated at lower heights (z < 40 μm), they were pushed toward the centerline under the DEP-y force. Alternatively, if they were located at the outer edge of electrodes and levitated at higher heights (z > 40 μm), they were pushed toward the sidewalls. Since the microelectrodes had a curved configuration and the silica particles were denser than water, all particles were initially located at the inner edge of microelectrodes and levitated at lower heights. Applying a low potential allowed the particles to remain at lower heights. Under these conditions, all particles were pushed toward the centerline and funnelled between the microelectrodes. A similar trend was seen at the middle parts of the microelectrode, as shown in Fig. 3C1–C4. However, nearby the tip, the positive DEP-y forces were only observed at z < 27 μm heights, as shown in Fig. 3D1–D4. Applying a high flow rate through the microchannel imposed a strong drag force on the particles, which avoided the further lateral motion of particles at the tips.
Fig. 3

The variations of DEP-y force on 450 nm particles at different locations of the DEP system. The particles are pushed toward or away from the centerline, according to their levitation heights. A: location of planes, B1B4: variations of DEP-y force at x = 3,000 μm, C1C4: variations of DEP-y force at x = 3,500 μm, D1D4: variations of DEP-y force at x = 4,000 μm

In our simulations, we have applied the single-particle model of DEP force throughout the microchannel including the centerline, where we expect a high concentration of particles. Although this model is reasonably consistent with the experimental observations (Doh and Cho 2005; Kang, et al. 2008), the modified models of DEP force can be applied in calculations, in which the particle-particle interactions are taken into account (Khusid and Acrivos 1995; Kumar et al. 2007).

The sedimentation force applied on particles is 8.2 × 10−17, 4.8 × 10−16, and 2.2 × 10−15 N, respectively for 230, 450, and 600 nm particles, as obtained from Eq. 4. The levitation height of the particles is obtained after the DEP-z and sedimentation forces reach a balance along the z-axis. Applying a low potential across the microelectrodes decreased the DEP-z force and levitated the particles at lower heights. For example, applying an AC signal of 12 Vp–p and 5 MHz most particles were levitated at z = 15–35 μm nearby the tips, as observed by an inverted microscope (Nikon, TE2000-U) while increasing the signal magnitude to 20 Vp–p pushed the particles upward and levitated most of them at z = 35–50 μm. The experimental results are consistent with the predicted DEP-z forces obtained from simulation (not shown). The effects of other forces induced by Joule heating effect, fluid heating by light, and optical radiation pressure are ignored due to their negligible influence, as given in Appendix 2.

In our experiments, the combination of a low potential and a high flow rate (12 Vp–p and 1 μl/min) was employed to concentrate the silica particles along the centerline for the purpose of producing a silica particle-rich region (path-3) in DI water medium. These silica particles have a higher refractive index (nd = 1.475 at 20°C) (Khlebtsov et al. 2008) than DI water (nd = 1.333 at 20°C). Therefore, this configuration can establish a particles-core/liquid-cladding optical waveguide, as shown in Fig. 2. The refractive index contrast (nd core − nd cladding) of our system is 0.142, which is higher than that of 0.11 obtained with aqueous solution of CaCl2 and DI water (Wolfe et al. 2004), and 0.017 obtained by ethylene glycol and ethylene glycol–water mixture (Wolfe et al. 2005) used to develop diffusion controlled optical waveguides. We also apply silica particle of different diameters to investigate the influence of light scattering.

Figure 4 shows the response of the system when a broadband light is applied: vertically at the top side of the microchannel (Fig. 4a–c), and horizontally at the inlet side of the microchannel (Fig. 4d–f), when particles of different diameters were injected into the system.
Fig. 4

ac The 600, 450, and 250 nm silica particles are patterned along the centerline under the DEP field. df The 600 and 450 nm particles scattered the horizontally applied light due to their large size and looked as a bright band while the 250 nm particles guided the light more effectively due to their small size and appeared as a dark band

Figure 4a–c shows the response of 600, 450, and 250 nm silica particles to the DEP field while a 1 mW light was applied perpendicular to the substrate. The magnitude and frequency of the applied signal were 12 Vp–p and 5 MHz, respectively while the flow rate of the suspension was 1 μl/min (equivalent to an average velocity of 0.2 mm/s) using a syringe pump (Harvard Apparatus, PHD 2000). These conditions were applied to levitate the particles at z = 15–30 μm nearby the tips, as described before. All particles exhibited negative DEP behavior and were repelled from the microelectrodes. The particles formed a funnel before the first microelectrode tips and subsequently a narrow band along the centerline afterwards. The width of the band, Δ could be tuned by varying the applied voltage and flow rate. For example, in the case of 450 nm particles, the maximum Δ was obtained at 12 Vp–p and 1 μl/min as ~20 μm. The Δ was very sensitive against voltage variations and decreased to ~10 μm by varying the voltage between 10–14 Vp–p; however it was more resilient against flow rate variations and only decreased to ~15 μm by varying the flow rate between 0.8–1.5 μl/min. Interestingly, the width of the band remained largely constant along the centerline within those ranges. This is another unique feature of curved microelectrodes, in which the magnitude of DEP force increases smoothly over the length of electrodes and abrupt motions of particles are avoided. In contrast, using oblique microelectrodes the width of the band could not remain constant along the centerline due to the abrupt increase of DEP forces over the microelectrode tips. Increasing the applied signal to 20 Vp–p pushed the particles toward the sidewalls, which was not of interest in this work for forming an optical waveguide. Further increasing of flow rate to 2 μl/min destabilized the funneling of particles and also caused the width of the narrow band to change between the consequent pairs, as shown in Appendix 3.

In order to further assess the feasibility of developing a particles-core/liquid-cladding optical waveguide, we conducted a series of experiments by placing a broadband light source horizontally at the inlet side of the microchannel, while the vertical light was turned off. This provided a longitudinal beam of light along the microchannel, as shown in Fig. 2. Results are shown in Fig. 4d–f for different dimensions of particles.

It was seen that the 600 nm particles were sufficiently large to scatter the horizontal light, and hence the narrow band of particles along the centerline was observed brighter than the surrounding medium, as shown in Fig. 4d. A significant degree of scattering was also observed near the sidewalls where the image looked brighter. A similar response was also seen for 450 nm particles, as shown in Fig. 4e. However, the response of 250 nm particles to the horizontal light was quite different. The 250 nm particles guided the light, due to their small diameters and their intimate contacts, along the centerline. As a result, the light passed through this narrow band with much less extinction and the band was observed darker than the surrounding medium, as seen in Fig. 3F.

Mie scattering theory (Bohren and Huffman 1983; Conroy et al. 2005) can be employed to show the dependence of light scattering intensity versus silica particle diameters at different wavelengths, (as detailed in Appendix 4) as shown in Fig. 5. The scattering is stronger at lower wavelengths and is proportional to the diameters of particles. For example at λlight = 632 nm (red light), the scattering efficiencies of light were 0.043, 0.185, and 0.349 for 250, 450, and 600 nm particles, respectively, meaning that the 600 nm particles scattered the red light 8.1 times stronger than 250 nm ones.
Fig. 5

The variations of the scattering efficiency versus the diameter of silica particles in different wavelengths of the light source

Figure 6 shows the schematic of light scattering and guiding for different sized particles using cross sectional and top views. When the light was applied normal to the microchannel (Fig. 4a–c) the light was scattered in all cases, and consequently the centerline was always dark. To comprehend this response, it is useful to divide the microchannel into five regions according to the density of particles (Fig. 6a, b—cross sectional view): a particle-rich band along the centerline, two particle-depleted bands on both sides of the centerline and two particle-poor bands along the sidewalls. Light was scattered by the 450 and 600 nm diameter particles in the particle-rich (centerline) and particle-poor (sidewalls) regions. As a result, images were darker in these regions (Fig. 4a, b). Interestingly, even in the case of the small particles of 250 nm, the light within the particle-rich region was also scattered (Fig. 4c). Mie theory calculations showed that the light scattering for 250 nm is much smaller than the 450 and 650 nm, while the observed scattering intensity was comparable. We ascribe this strong scattering to the ‘separation of particles’ rather than the diameters of the particles for the 250 nm case. Significant particle separation close to the top and bottom surfaces of the microchannel occurred, since most particles were levitated at z = 15–35 μm under the interaction of DEP-z and sedimentation forces, reducing the density of particles at those regions. In this case, the scattering of light occurred due to the large distance between the silica particles, which is comparable to the wavelength of the impinged light. The separation of the particles is also responsible for scattering in the particle-poor and particle-depleted regions (which is denser than of 450 and 600 nm) for 250 nm particles.
Fig. 6

The schematic of the light scattering by small and large size particles. a, b cross sectional; c, d: top views

Alternatively, when the light was applied along the microchannel (Fig. 4d–f), the large 450 and 600 nm particles scattered the light within the particle-rich and particle-depleted regions while the small particles of 250 nm guided the light through the particle-rich (centerline) region. This region was enclosed by liquid streams on both sides that had a lower refractive index (Fig. 6c, d—top view). Under these conditions, the small particles confined the light to form a multimode waveguide of less than 20 μm width and 40 μm height, while the large particles scattered the light.

5 Conclusions

In summary, the optical properties of a DEP-activated system were demonstrated using silica submicron particles. The particles were focused along the microchannel centerline by energizing the microelectrodes at 12 Vp–p and 5 MHz, and applying a flow rate of the 1 μl/min. The system was applied to analyze the possibility of establishing a particles-core/liquid-cladding optical waveguide when 250 nm silica particles were used. Additional optics (optical fibers, lens and camera) are necessary at the outlet to analyze the output light profile. The performance of the system can be optimized continuously by providing feedback to change in real-time the magnitude and frequency of the AC signal, as well as the flow rate of the medium. We believe that DEP-activated optical waveguides have a great potential to be integrated into lab-on-a-chip systems to facilitate the optical excitation and detection of target samples.

Copyright information

© Springer-Verlag 2010