INTRODUCTION

Over the past decades, there has been interest in the use of monodisperse systems, especially those consisting of micro- and nanosized particles, in various technological processes. There are known directions for using monodisperse powders in studies of the petrochemical industry, in studying the evolution of space objects, and in studies of dusty plasma [1–5]. The identity of the particle size parameters of such mixtures in some cases can significantly improve the technology or create a completely new process or substance. Particle size uniformity is important when measuring the kinematic parameters of a medium, for example, in PIV digital tracer imaging methods, where particles (usually submicron or micron in size) with negligible mass and volumetric concentration are introduced into the medium under study. Also of particular interest are monodisperse aerosols (due to the strong dependence of the properties of dispersed systems on the droplet size) and solid particles [6–8]. Since particles almost always have some charge under natural conditions, there is a growing interest in new devices that can separate large numbers of charged particles by size at atmospheric pressure in the air.

The electrodynamic Paul trap, which ensures the confinement of several thousand charged particles, is successfully used in the study of Coulomb structures [9, 10], the propagation of solitary density waves in them [11–13], and the occurrence of instabilities. Traps can also find practical application in cleaning various surfaces or solar cell panels [14, 15], for purifying gases from particles of given sizes, as well as in scientific and laboratory research to determine physical [16, 17] and thermodynamic [18, 19] parameters of captured individual particles and structures. The retention of particles in such a trap largely depends on the parameters of the trap, the frequency and amplitude of the voltage on the electrodes, and the ratio of charge to particle mass [20–22].

Objective—To study the separation of a polydisperse mixture of solid particles by size using an electrodynamic linear quadrupole Paul trap at atmospheric pressure in air. The variable parameter is the amplitude of the alternating voltage on the electrodes of the trap.

1. EXPERIMENT

The experimental design and linear quadrupole trap were described in detail by us earlier [21, 23]. A linear quadrupole trap is four cylindrical electrodes located at the vertices of a square with a side length of 23 mm, the diagonal pairs of which are supplied with an alternating voltage of opposite phase with a frequency of 50 Hz. The diameter of horizontally located dynamic electrodes is 4 mm and length is 300 mm. The confinement of charged particles in a trap is described by the system of Mathieu equations [20]. The main parameter that determines the retention of a particle by a trap is the ratio of the particle’s charge to its mass q/m. There are stable solutions for certain combinations of amplitude and frequency of voltage applied to the electrodes at a given ratio. When the voltage decreases, the electric field of the trap will not be able to retain particles with the lowest values of q/m, which will fall out of the trap under the influence of gravity. The lowest q/m will be for larger particles, since the charge on a particle is proportional to the radius or surface area (radius squared), depending on the charging method; and the mass is proportional to the radius cubed [24].

The experiment used polydisperse aluminum oxide powder Al2O3. Figure 1 shows the experimental particle size distribution in the original oxide powder, obtained from a sample of 300 particles.

Fig. 1.
figure 1

Distribution of initial particles by size for a sample of 300 particles.

Full size image

The maximum amplitude of alternating voltage on the electrodes was experimentally established to be 11 kV at which most of the charged particles of the polydisperse mixture are retained inside the trap. To impart a charge to the particles sufficient to be captured and retained inside the trap [24], the method of inductive charging on a metal plate with a potential of 7 kV was used. The potential value was chosen experimentally, since at it the particles were detached from the charging plate and captured by the trap. After the system of particles captured by the trap has reached a stationary stable state, a glass slide is introduced under the lower electrodes of the trap. The amplitude of the alternating voltage at the electrodes is reduced by 2 kV. After the system returned to a new stable state, the slide with the crumbled particles was removed, and a new slide was placed in its place. Then, the voltage was lowered again and the cycles were repeated until all particles left the trap.

2. RESULTS AND DISCUSSION

Figure 2 shows the size distributions of particles falling out of the trap with each decrease in voltage. The red line in the graphs indicates the approximation curve of the normal distribution density function. The initial particle size distribution in the mixture ranges from 5 to 80 µm. The absence of a fraction of sizes 5–10 µm in the distribution of the initial mixture (see Fig. 1) is explained by the low content of this fraction and possible adhesion to larger particles. It should also be noted that with the induction method of charging particles, preliminary separation occurs at the stage of separation of particles from the plate and their capture by a trap. Even with a small content of small-sized fraction in the initial mixture, the trap will capture more small particles, since the retention area for them is wider. If there are adhered particles, they will disperse when charged or trapped due to electrostatic repulsion. After the first stage of decreasing the voltage from 8 to 6 kV, particles with sizes in the range of 30–45 μm fell out of the trap according to the approximation curve. The size of 300 particles was measured at this stage. As expected, the heaviest particles measuring 50–80 µm also left the trap. At other stages of voltage change, the number of particles of such large sizes is extremely small. With the next voltage decrease from 6 to 4 kV, the distribution of the bulk of the precipitated particles (260 particles measured) lies in the range of 25–35 µm. When decreasing from 4 to 2 kV (152 particles measured), the range narrowed to 20–30 μm, and at the final stage the lightest fraction (184 particles measured), predominantly with sizes of 5–10 μm, left the trap. The presence of outliers in particle sizes as the size range narrows at each step of the experiment can be explained by the Coulomb interaction between particles and structure instabilities. The experiment with filling and subsequent complete cleaning of the trap was carried out 5 times; the data for the most uniform separation are presented.

Fig. 2.
figure 2

Size distribution of particles falling out of the trap when the voltage decreases: (a) from 8 to 6 kV. Sample of 300 particles; (b) from 6 to 4 kV. Sample of 260 particles; (c) from 4 to 2 kV. Sample of 152 particles; (d) from 2 to 0 kV. Sample is 184 particles.

Full size image

CONCLUSIONS

The separation of a polydisperse mixture of inductively charged aluminum oxide microparticles in air using a quadrupole linear Paul trap by changing the voltage on the linear electrodes was experimentally demonstrated. When the voltage decreases from a maximum of 8 kV, heavier particles fall out of the trap in a stepwise manner with a step of 2 kV at each stage, and the sizes of the precipitated particles decrease by an average of 10–15 microns. A wider spread of sizes after the first stage may be a consequence of an abrupt decrease in voltage, during which a stronger disturbance of the particle cloud occurs in the filled trap. At the next stages, the cloud size is smaller, so the particles are more stable and more clearly structured in size throughout the volume of the trap. For finer filtration, use the primary filtered fraction and reduce the voltage in smaller steps.