Janus particle microshuttle: 1D directional self-propulsion modulated by AC electrical field

A catalytic Janus particle is capable of gaining energy from the surrounding fuel solution to drive itself to move continuously, which has an important impact in different fields, especially the field of micro-systems. However, the randomness of self-propulsion at the microscale restricts its use in practice. Achieving a directed self-propelled movement would greatly promote the application of the Janus particle. We proved experimentally that an AC electric field was an effective way to suppress Brownian motion and control the direction of self-propelled movement. The self-propulsion and dielectrophoretic response of a 2 mu m Janus particle were observed and the related basic data were collected. Interdigital electrodes, 20 mu m in width, were energized in pulsed style to modulate the self-propulsion, which resulted in a shuttle-style motion in which a single Janus particle moved to and fro inside the strip electrode. The change of direction depends on its unique position: the catalyst side is always pointed outward and the orientation angle relative to the electrode is about 60 degrees. Numerical simulation also proved that this position is reasonable. The present study could be beneficial with regard to self-propulsion and AC electrokinetics of the Janus particle. (C) 2014 Author(s).


I. INTRODUCTION
The Janus particle, proposed by de Gennes in 1991, has a non-homogeneous structure on its two hemispheres, which enables some novel physical and chemical properties to be induced. 1 At present, the Janus particle has been fabricated by different methods including microfluidics, self-assembly, sputtering, and so on. 2,3 Among all applications, a very interesting example is self-propulsion ( Fig. 1(a)), where one hemisphere of the Janus particle covered with a thin layer of catalyst (e.g. platinum (Pt)) serves to decompose the fuel reactants (e.g. 2H 2 O 2 → 2H 2 O + O 2 ) surrounding the particle. 4 Because the catalytic reaction generates asymmetry of the molecule number at the two sides of the Janus particle, the concentration gradient of molecules will force the Janus particle to move autonomously, 5 with the direction opposite to the catalyst side. Considering this compact structure and the fact that an additional battery is not required, self-propulsion of the Janus particle has attracted widespread interest in fields such as MEMS, self-assembly, and sensors, so far. 2,3 For microsized Janus particles, the self-propulsion exhibits very complex behaviours and the related mechanism has been studied thoroughly. Commonly, it was believed that the movement of Janus microparticles is due to the competition between two different motions: random Brownian motion and self-propellant motion. 6 This results in inconvenience when one wants to control the direction of motion of Janus microparticles, such as in targeted drug delivery. Hence, artificial intervention has to be introduced for the self-propulsion of Janus particles. The key to generating an expected directional movement is to overcome the random rotation from Brownian motion. In 2012, Baraban et al. 7 published their results on how to realize directed self-propulsion of Janus particles in a magnetic manner. An additional magnetic cap was coated between the catalytic metal layer and SiO 2 particles, and an external magnetic field was used to direct the movement of the Janus particles. The strength of the magnetic field was very weak, so it served only to suppress the random rotation rather than driving the particles to move. Long range 1D directional motion was realized under a fixed magnetic field, and the flexible 2D planar motion of Janus particles was carried out by further adjusting the applied magnetic field in time. In principle, more complicated functions including transport, manipulation, and separation of Janus particles could be achieved too. The magnetic field method was also found in the case of using non-spherical Janus particles or other micro-devices. 8 In addition to the magnetic field, a light-driven method and a DC electrical field were also available to direct the self-propulsion of Janus particles. [9][10][11] In this paper, we present a novel design utilizing an AC electrical field to control the orientation of the Janus particles and further to realize a controllable shuttle-style self-propulsion ( Fig. 1(b)). In Section II A, the experimental procedure is introduced briefly. In Section II B, we show the stochastic characteristics using the effective diffusion coefficient D eff and the probability distribution of displacement rotation angleγ . Next, in Section II C, we examine the dielectrophoresis (DEP) of the whole Janus particles under a high frequency AC electric field and determine the orientation angle of the Janus particle as a two-hemisphere structure. Based on the characteristics from Section II B and the manipulation manner described in Section II C, we put forward a strategy in Section II C in which by switching the AC electric field "on" and "off" with a certain frequency we can eliminate the negative effect of random motion and direct the Janus particles to move. Finally, the conclusions are drawn in Section III and information about the fabrication and materials is provided in the Appendix.

A. Experiment
The experiment was conducted on a MicroPIV platform (LNM, Institute of Mechanics, CAS). The trajectories of particles immersed into H 2 O 2 solution were observed by video microscopy with an image field of view of 512 × 512 pixels. Then, the test solution was prepared by mixing the Janus particle aqueous solution and the hydrogen peroxide solution to give a certain H 2 O 2 concentration. After the preparation of the solution, a 70 μL droplet was put on the glass slides. A CCD camera (Andor 897) was used to record videos. Image series consisting of 600-1000 frames were captured in one position located about 2 μm above the glass substrate. The particle positions (x, y) can be tracked by the software video spot tracker (V07.02). Then, a high frequency AC electric field generated by indium tin oxide (ITO) interdigital electrodes (IDEs) was applied. ITO material provides high transparency and good electrical conductivity, enabling visualization of the movement of particles by an inverted microscope. The electrode pad accesses the signal generator to load the pulsed signal. Similar observations were carried out to record the motion of Janus particles affected by the DEP force.

B. Stochastic characteristics of self-propulsion of Janus particle
A Pt-SiO 2 type Janus microsphere is a common experimental model used to carry out the study of self-propulsion phenomena and was adopted in most of the works in the literature. 4 Its nonuniform surface structure consists of one dielectric silica hemisphere and one Pt-coated hemisphere. When Pt-SiO 2 microspheres were dispensed into hydrogen peroxide solution, a catalytic reaction happened on the Pt surface, where the molecular number ratio of reactant and product is 2:3. Obviously, the overall number of molecules in solution is increased and the distribution becomes uneven at the same time. At the Pt side, there are more product molecules and the concentration is higher than at the opposite SiO 2 side. According to the viewpoint of thermal dynamics, Brownian motion is the result of random collision of small molecules with the particle surface. The more molecules collide, the higher the resultant random force is. Therefore, the Janus particle will be pushed away from the Pt-coated hemisphere.
Brownian motion originates from random collisions of surrounded liquid molecules. This process is equivalent to applying a random Brownian force on a particle. The expression of Brownian force is given by, 12 Here, k B is the Boltzmann constant, T is the thermodynamic temperature, μ is dynamic viscosity, R p is the radius of the particle, t is the time interval, and ξ is a random number. Under the action of Brownian force, the particle will give rise to random translation motion. Furthermore, considering that Brownian force does not always pass through the centre of the sphere, it also results in a Brownian force moment that makes the particle rotate randomly at the same time. The rotation angular velocity depends on the torque θ and viscous friction factor f r of fluid and is decided by The relationship between the rotation angle θ and the time interval t is obtained by From Eq. (3), it can be clearly seen that the rotation angle of the particle is proportional to the time interval. For a movement with a very small rotation angle θ , it can be deemed as a straight motion.
In the present case, the characteristic θ is set to ± π 2 and the resultant characteristic time τ R is about 10 s. Below this τ R , the particle is able to maintain its direction and move forward.
We carried out the observation of the Janus particles (2 μm in diameter) and collected basic data in different concentrations of H 2 O 2 solutions as shown in Fig. 2. The effective diffusion coefficient D eff and the displacement rotational angle γ as a function of the time interval of observation t are used to describe the self-propulsion quantitatively. It can be seen that D eff increases with t at the relatively short times, and then approaches the constant D eff when t exceeds a certain threshold τ R . The time threshold indicated that Janus particles change their motion from self-propulsion to Brownian-like motion. This transition can also be regarded as a change from straight to random motion. For different H 2 O 2 solutions, the new constant D eff is increased and the threshold τ R is decreased as the concentration of H 2 O 2 increases. The typical behaviours of self-propulsion have been studied in detail recently by Howse et al. 13 and also using our measurements. 14 In Fig. 2(b), we show the probability of the γ angle at different time intervals in 5% H 2 O 2 concentration. Obviously, in a short time period, small-angle rotations have large probability; for example, the particles' rotation is mainly concentrated between -50 • and 50 • when the time interval is 1 s. This means that the particles possess good directionality. The directionality decreases when the probability of small-angle rotations is slightly lower as the time interval increases. The randomness of the particles rotation appears until the time interval is 12 s. At this moment, the angle of rotation has the same probability. So, the directionality of particles motion is available for our experiment for a short time interval. For example, for the short time t = 2.5s, where the particles have good directional motion, the effective diffusion coefficient is measured to be approximately 12 μm 2 /s in 5% H 2 O 2 , and the mean squared displacement L 2 is about 120 μm based on the well-known equation This means that the Janus particles have a large probability of undergoing almost straight propulsion with a typical length of 11 μm. This typical length is the basis for our design of microshuttle electrodes.
C. Controllability of Janus particles

Dielectrophoretic response of Pt-SiO 2 microsphere
According to bipolar theory, particles suspended in fluid media will be polarized under an external electric field and polarization charges of equal size and opposite sign are generated. Even for a neutral particle, the non-homogenous electrical field may induce the particle to move because the force at different poles cannot be counteracted, which is called DEP. For a homogenous sphere, the expression of the DEP force is given by, 15 where r is the radius of the particle, ε 0 is the permittivity of a vacuum, ε f is the permittivity of the fluid medium, ∇ E 2 rms is the gradient of the RMS electric field, and C M * is the Clausius-Mossotti factor. Re[C M * ] is the real part of C M * , where ε p is the permittivity of the particle, σ p is the conductivity of the particle, σ f is the conductivity of the fluid medium, ω is the angular frequency of the AC electrical field, and j = √ −1. Depending on the sign of C M * , the sign of the DEP force could be positive or negative, in turn, to generate an attractive or repulsive force.
Currently, extensive works on the DEP response of micro-or nano-particles, cells, and nanotubes have been done, [16][17][18]  Here, white particles are the control particles of uncoated SiO 2 particles, whose DEP response is nDEP.
SiO 2 particles) because there is a big difference in dielectric constant between metal and dielectric materials. Here, as reported by the literature, 19,20 a simple linear superposition model was used to consider the two-hemisphere structure. The integrate C M * is calculated as the sum of the Pt-SiO 2shell and the bare SiO 2 sphere, The response curve of C M * versus applied AC frequency is given in Fig. 3(a). From this figure, it can be seen that C M * of ordinary inorganic particles gradually decreases with increases in frequency and that the DEP force changes from pDEP to nDEP. The crossover frequency is about a few hundred kilohertz. The response of pure metal particles has the opposite trend. Different from two kinds of homogenous particles, the DEP properties of Pt-SiO 2 -type Janus particles combine the properties of SiO 2 dielectric spherical particles and a Pt-SiO 2 -shell. 19 We observed the DEP response of Pt-SiO 2 Janus particles with a frequency range from 1 KHz to 10 MHz. The crossover frequency from nDEP to pDEP is about 10∼20 KHz, which agrees with the experimental observation. A typical result in Fig. 3(b) shows that Janus particles are pDEP when the frequency is 10 MHz and the electrical potential is 2V pp . This is in agreement with the model above. It can be seen that Janus particles are attracted to two edges of the electrodes and keep a steady state. Especially, the side of the Pt-coated hemisphere faces away from the electrodes and the silica side points to the electrodes inside. Controlling the Janus particle's orientation is a key factor in our design and its mechanism will be explained by a numerical model below.

Orientation angle of Pt-SiO 2 microsphere
We calculated the electric field distribution and the energy of the system for different particle positions in order to understand why individual Janus particles are oriented in a unique direction as we observed in the experiment. The total electric energy E e of the system can be obtained by integrating the local energy density, E es , over the subdomain volume of the particle.
where D is the electric flux density, ε 0 is the dielectric permittivity of a vacuum (8.854 × 10 −12 ), and E is the electric field intensity. For our system, which is not in a vacuum, the electric flux density D is given by D = ε 0 ε r E, where ε r is the relative permittivity of the particle. We performed a 2D simulation with Comsol Multiphysics 4.1. As stated above, the Janus particle is attracted to the high field intensity area in the present electric field gradient so that the minimum potential energy should be reached. The voltage is kept constant and only the orientation angle of the Janus particle is changed. We then calculated the total stored electric energy (since it was a 2D simulation and we integrated over the area rather than the volume). The total electric energy of the system was calculated from -180 • to 180 • by rotating the particle about its Pt/SiO 2 interface axis in 30 • increments. The potential energy is maximal at -120 • and minimal at -60 • . The latter corresponds well to the experimentally observed orientation of the particles (Fig. 4). It can also be straightforwardly judged from the fact that blue colour (cold colour represents low electrical potential) occupies almost the whole volume inside the particle.

Time scale of modulation signal
The typical time scale τ DE P of the self-propellant particle on IDEs is given by L/U f , where L is the characteristic length dimension and U f is the velocity of self-propulsion. For typical values of L (which is of the order of 10 μm as given by the half width of the electrode) and U f (which is of the order of ∼10 μm/s as given purely by the self-propellant velocity) in our experimental, the resulting time scale is about 1 s. It should be noted that the rotation time scale τ R from Eq. (3) is used to depict the transition from self-propulsion to Brownian-like motion. The value of this characteristic time is about 10 s for our experiment (2-μm Janus particle and 5% H 2 O 2 solution). Below the rotation time scale of t < τ R , the particle will move in a straight line in a direction from the Pt side to the SiO 2 side, and above the rotation time scale t > τ R the particle will move randomly and can no longer be steered in only one direction. When the DEP force is modulated at the time scale ranging from τ DE P < τ < τ R , we have the possibility to control the motion in time before the Janus particle begins to change its direction. The strength of the DEP force will be kept at a very low level so that it is just sufficient to change the orientation angle; in this paper, the voltage of 2V PP is chosen.

Prototype verification of directed self-propulsion
To demonstrate the ability to direct the movement of the Janus particle in 1D mode, we performed prototype experiments where we introduced 2-μm Janus particles suspended in a 5% H 2 O 2 solution and then applied pulsed DEP. The electrode parameters are W I DEs = W S P AC E = 20 μm, W I DEs is the electrode width, and W S P AC E is the space between two adjacent electrodes. In the present design, we pulse the DEP force by turning it on and off with a 50% duty-cycle square wave with a frequency of 0.2 Hz. By tuning the time period, we can direct the movement of the Janus particles.
The results in Fig. 5 depict a time-series of images under pulsed operating conditions. In the first "on" period, it is shown that one Janus particle is attracted to the left edge of the electrode and the Pt-side of Janus particle faces the left (Fig. 5(1)). During this period, the DEP force at the edge of the IDEs is maximal and strong enough to overcome the Brownian motion, and the Janus particle is trapped steadily. Once the electrical field is powered off, no DEP force is applied to the Janus particle and self-propellant force is dominant. At the beginning of the first "off" period, the Janus particle moves from the left edge to the right edge inside the electrode. The path of the particle is nearly a straight line (Fig. 5(2)). At the end of the first "off" period, the Janus particle has moved to the vicinity of the right edge (Fig. 5(3)). In the second "on" period, the particle is immediately attracted to the right edge of the electrode, and at the same time the Janus particle instantly rotates by 180 • and keeps its Pt-side opposite to the electrode again ( Fig. 5(4)). In the following periods ( Fig. 5(5,6,7)), the Janus particle recovers its self-propelled motion and comes back to the left edge of the electrode from the right edge. In the last image ( Fig. 5(8)), the Janus particle returns to the initial state and the whole shuttle-style cycle is completed. The above procedure can be repeated many times only when pulsed DEP is applied and a fuel solution is supplied. The result has successfully verified the prototype design. Based on the present work, more complicated operations with directional self-propellant motion could be carried out in the future.

III. CONCLUSION
We have introduced a DEP-based modulation method that makes use of time-pulsing to generate 1D self-propulsion for Janus particles. Although it was found that the randomness of self-propulsion is significant due to the appearance of Brownian-like motion, at short time scales, directional movement is still approached. With the help of an AC electric field generated by IDEs, one can control the orientation of Janus particles when they experience positive DEP so that they are attracted to the edge of the IDEs. The time scale τ for a pulsing AC electrical field is decided by the nature of self-propulsion and the geometric dimension of the electrodes. By modulating the time precisely, Janus particles have enough time to arrive at the opposite side of the electrode and maintain approximately straight movement during this period. By repeatedly turning the electric field "on" and "off" many times, the shuttle-style motion of Janus particles is generated. We also developed a numerical model to explain the unique orientation of Janus particles and a simple analytical model to calculate the DEP response under an AC electric field, and the model predictions were consistent with the experimental observations.