Vacuum System Setup

The specific vacuum system, illustrated in Fig. 3.10, contains the essential elements typically required to obtain high vacuum. In the designed vacuum system, there are three pumping devices, including two rotary vane vacuum pumps and the turbo molecular pump. Other components of the vacuum system, such as valves and gauges, aid the actions of these pumps. Besides, the charge-couple device (CCD) camera was adopted to capture the glow images. Accordingly, the electrical source was supplied by pulse DC controller. The details functions of the components are described below.

Fig. 3.10. The schematic diagram of the specific vacuum system.

a. Rotary Vane Pump:

The rotary vane pump's hollow body has a rotating cylinder mounted off-axis. In the rotor, two diametrically opposed radially directed vanes are spring-loaded to force contact with the pump body. Since the rotor is positioned off-axis, its motion causes the volume between the vanes and the body to vary during each half turn. The gas inlet port is so positioned that the volume behind the last vane to pass increases, allowing gas to expand into it until the next vane passes. As the volume exposed to the inlet increases, the volume trapped between the vane and the exhaust port decreases. In a single stage pump, the exhaust port has a valve connected to the atmosphere. In a two stage pump, the first stage's exhaust connects directly to the second stage's inlet. Gas exits the pump by bubbling up through the pump's oil reservoir. In either version, the gas is compressed against the final exhaust valve until it exceeds atmospheric pressure, at which point it forces open the exhaust valve and escapes. All the sliding surfaces, the bearings, and gaskets are lubricated and sealed

against gas leaks by the pump's oil. Considering the exhaust pressure illustrates an important but frequently overlooked factor of vane pump operation. With an operating temperature of, typically, 80°C and a gas pressure greater than atmospheric at the exhaust, chemical reaction between the gas and the fluid can increase when pumping corrosive gases and even explode if a hydrocarbon fluid is used when pumping 100%

oxygen.

In our designed vacuum system, two rotary vane pumps were used to be fore pumps. First one is a fore pump to vacuum chamber and the other is a fore pump to turbo molecular pump. When the chamber was exhausted below the critical pressure, the turbo molecular pump was turned on to obtain the expectative vacuum degree.

b. Turbo Molecular Pump:

A turbo pump has a stack of rotors resembling a jet engine. Each of rotors which has multiple and angled blades drive at very high speed with tangential direction. Gas molecules, hit by the underside of the angled blades, move with momentum in the direction of the higher pressure exhaust. Turbo pumps come in two basic designs. In the SNECMA design (named after a French jet-engine manufacturer) all gas enters through the main flange at the visible single-end of the pump; the Pfeiffer (named after the original turbo pump's manufacturer) double-ended design, has two rotors set-mounted on a common axle. Gas enters the pump through a right angle port between the two rotor sets and exits into an exhaust manifold connecting the two ends of the pump. For normal commercially available pumps, pumping speeds range from approximately 20 L/s to 3,000 L/s (although some turbo pumps manufactured in Russia are quoted with pumping speeds of 20,000 L/s). All gases are pumped at roughly the same rate. One pump, for example, listed with 450 L/s for nitrogen, has a pumping speed for hydrogen of 310 L/s. Turbo pumps reach full operating speed

within a few minutes of switch-on, making a separate roughing line unnecessary since the accelerating turbo can rough the chamber. The trick is to match chamber volume and the effective pumping speed, and to time the turbo pump's start so its rotational speed is high enough to prevent backstreaming when the chamber reaches 10^-1 Torr (that is, when an oil-sealed mechanical pump's relative backstreaming rate starts its rapid rise). With proper venting, a turbo pump can be entirely halted in under one minute. This slight delay before the chamber reaches atmospheric pressure and can be opened usually works for most applications. The benefit is clear. If the pump does not run while venting the chamber, we have no need for a high vacuum valve between pump and chamber. Further, in a correctly designed and operated system, the turbo pump does not allow oil vapor to backstream so that the system does not require an LN2 trap between chamber and pump. A turbo pump's high rotational speeds (some small diameter units operate at 60,000 rpm) put serious strain on the shaft bearings.

Most manufacturers now offer light-weight ceramic ball bearings with grease lubrication since this combination is both light, lowering the momentum of the bearing, and has a low bearing-lubricant vapor pressure at the pump's working temperature. More recently, manufacturers have begun offering magnetically levitated bearings which are either permanent magnets for the small sized pumps or a combination of permanent and dynamic magnetic fields, supporting the shaft without contact. Magnetically levitated pumps (such as those manufactured by Shimadzu) include non-contact bearings, resulting in a turbo pump that has true zero oil-vapor backstreaming with bearings that never wear down.

Using turbo molecular pump to obtain the expectative vacuum degree can avoid the oil contamination. The exhaust was connected to the second rotary vane pump, which produces a pressure low enough for the turbo molecular pump to work efficiently. In general, the turbo molecular pump can be a very versatile pump. It can

generate many degrees of vacuum from intermediate vacuum (~10^-4 Torr) up to ultra-high vacuum levels (~10^-10 Torr).

c. Gauge:

In order to check the vacuum degree of the chamber, two kinds of gauges were used to measure the different range of the vacuum degree.. The Baratron gauge is general purpose pressure transducers designed to provide accurate and repeatable pressure measurement in the range from 1 K Torr to as low as 0.05 Torr Full Scale (FS). However, the higher vacuum degree can be detected by ionization gauge which is used to measure the residual pressure of vacuum in the high vacuum and ultra-high vacuum ranges. Most of the ionization gauge cannot measure the higher pressures of the medium vacuum range, where the Pirani gauges, thermocouple gauges and convention gauges are use instead.

Fig. 3.11. The schematic of ionization gauge.

The ionization gauge consists of three distinct parts, shown in Fig. 3.11, including the filament, the grid and the collector. The filament is used for the

production of electrons by thermal emission. A + charge on the grid attracts the electrons away from the filament; they circulate around the grid passing through the fine structure many times until eventually they collide with the grid. Gas molecules inside the grid may collide with circulating electrons. The collision can result in the gas molecule being ionized. The collector inside the grid is - charged and attracts these + charged ions. Likewise they are repelled away from the + grid at the same time. The number of ions collected by the collector is directly proportional to the number of molecules inside the vacuum system. By this method, measuring the collected ion current gives a direct reading of the pressure. The above paragraph is a simplification of what happens. The design of the gauge head effects how efficiently electrons are produced, how long they survive, and how likely they are to collide with a molecule. Combining these factors together gives the gauge a sensitivity. As a general rule, the higher the sensitivity, the more efficient the operation of the gauge.

ve

ve

ve v_e

d. Pulse DC Controller:

The fabricated device was excited by pulse DC controller, including a DC-pulse power controller (SPIK 2000A, SHENCHANG ELECTRIC CO., LTD) and a direct current (DC) power supply. The power supply provides the voltage to DC-pulse power controller and forms the output of bipolar voltage waveform. The bipolar voltage waveform, having the duty ratio with 20%, was applied with different frequencies between the interdigitated Chromium (Cr) electrodes which has a 20 µm gap by the pulse DC controller. Fig. 3.12 shows the photograph of the pulse DC controller set using in this thesis.

Fig. 3.12. Photograph of pulse DC controller and DC power supply.

e. Charge-Couple Device Camera:

The images of the glow dicharge from the micro-plasma device were captured by using a color charge-coupled device (CCD) camera (Watec, WAT-202D) with zoom lenes. The signal of the color charge-coupled device (CCD) cam era captured was connected to a personal computer by a capture card (UPMOST, MTV Video Capture).

Chapter 4

Experimental Results and Discussion

4.1 Introduction

The characterizations of the interdigitated electrode micro-plasma device will be presented in this chapter. The structure of interdigitated electrode micro-plasma device was fabricated agreement with the designed structure by photolithographic techniques. The interdigitated electrode micro-plasma device which has a simple fabrication process offers the advantages, such as low ignition voltage, low cost and stable glow discharge. In the following sections, the electrical properties on neon discharge have been examined by bipolar voltage waveform with different frequencies of 10 kHz, 14.3 kHz and 25 kHz. The relationship between the ignition voltage and dielectric layer thickness is also investigated. In addition, the electrical properties on neon discharge have been examined by bipolar voltage waveform with different argon gas concentration.

4.2 The Features of the Interdigitated Electrode Micro-Plasma Device

The thick photoresist (XP SU-8 3050) is design specifically for ultra-thick, high-aspect-ratio microelectrical mechanical systems (MEMs). In addition, the decomposition temperature which is about 380 °C for SU-8 can resist the plasma.

Therefore, the thick photoresist can be served as a barrier rib to take the place of the photo-definable glass. Since the thick photoresit was spin-coated on the substrate, the

(a) Cross-section

(b) Cross-section with a tilt angle

incompact issue of the photo-definable glass can be solved. Besides, the geometric pattern was obtained accurately with the interdigitated electrode by aligner mask. The photo-definable glass was attached with the screen electrodes in order to realize the proposed device easily.

The structure of the interdigitated electrode micro-plasma device was measured by the scanning electron microscope (Hitachi S-4800). The electrode thickness, dielectric thickness, barrier rib height, aperture size of geometric pattern and cross-section of fabricated device can be accurately measured. Fig. 4.1 shows the cross-section of the square-shape interdigitated electrode micro-plasma device. The barrier rib has an aspect ratio (height over width) of 3 where height (60 µm) is high enough to avoid the cross-talk.

Fig. 4.1. The SEM image of the square-shape interdigitated electrode micro-plasma device.

Several geometric patterns are also developed by the designed mask. By the lithographic techniques, the geometric pattern can obtain at the right position, including two electrodes. It need not make screen electrodes and the larger array can be realized efficiently. The features of the interdigitated electrode micro-plasma device, shown in Fig. 4.2, were measured and different patterned shapes are also illustrated in Fig. 4.3.

(a) (b)

(c) (d)

Fig. 4.2. The features of the hexagon-shaped interdigitated electrode micro-plasma device. (a) cross-section with a tilt angle (b) two electrodes confined to the patterned shape (c) cross-section of the dielectric layer and electrode (d) surface morphology of HfO2.

(a) (b)

(c) (d)

Fig. 4.3. The different patterns of the interdigitated electrode micro-plasma device. (a) circle-shaped (b) diamond-shaped (c) square-shaped and (d) star-shaped.

4.3 The Electrical Properties in Neon Gas

The photograph of the fabricated device, shown in Fig. 4.4, was investigated in a vacuum chamber. Before measuring the fabricated device, the electrode pads should be connected with electric wires by sliver paste. The torr seal was used to cover the electric wires with the sliver paste.

The electrical properties of the device are investigated in the specially designed vacuum system. The vacuum chamber was pumped down to 2x10^-6 Torr by a turbo molecular pump and back-filled with the research vacuum grade of neon gas between

300 and 800 Torr. The bipolar voltage waveform with different frequencies is applied between the Cr electrodes.

Fig. 4.4. The photograph of the fabricated device with the electric wire.

4.3.1 Relationship between voltage and pressure with different bipolar voltage waveform frequency

The voltage margin, the ignition voltage (Vf) and the minimum sustain voltage (Vs), for hexagon-shaped interdigitated electrode micro-plasma device operating in neon gas at pressure between 300 and 800 Torr is presented in Fig. 4.5. The interdigitated electrode micro-plasma device is driven by a bipolar voltage waveform which has the frequencies of 10, 14.3 and 25 kHz. It should be noted that the ignition voltage reduces by increasing the operating pressure. The higher operating pressure causes the decreased mean free path of the neon gas (greater densities of gas molecules). The mean free path is the average distance traveled by a gas atom between collisions with another. In addition, the probability of collision is related to the collision frequency which is the number of gas atom collisions per unit time. It means that increasing the pressure causes a higher probability of collision. Therefore, the ignition voltage is reduced by increasing the operating pressure.

300 400 500 600 700 800

Bipolar voltage waveform frequency of 10 kHz

300 400 500 600 700 800

Bipolar voltage waveform frequency of 14.3 kHz

300 400 500 600 700 800

Bipolar voltage waveform frequency of 25 kHz

Fig. 4.5. The margin of the firing voltage and the minimum sustain voltage for the interdigitated electrode micro-plasma device with different bipolar voltage waveform frequency of (a) 10, (b) 14.3 and (c) 25 kHz in neon gas at pressure from 300 to 800 Torr.

Accordingly, with increasing the frequency of bipolar voltage waveform, the ignition voltage is getting lower at the same pressure. It is due to the impedance related to the frequency.

1 ZC

j Cω

= (1)

where ω, angular frequency, is equal to 2 fπ . From Eq. (1), higher frequency causes the impedance to decrease; however, the ignition voltage reduces by increasing the frequency of the bipolar voltage waveform, shown in Fig. 4.6.

300 400 500 600 700 800

Fig. 4.6. The ignition voltage for the interdigitated electrode micro-plasma device with different bipolar voltage waveform frequencies in neon gas at pressure 300 to 800 Torr.

4.3.2 Relationship between ignition voltage and pressure with different dielectric layer thickness

The characteristics of ignition voltage and pressure in neo gas with different dielectric layer thickness are shown in Fig. 4.7. The ignition voltage also decreases with reducing dielectric layer thickness, meaning the dielectric layer capacitance increases. This is due to the division of the externally applied voltage over the gas gap and dielectric layer. The applied sustaining voltage Vs is deduced from voltage across the gas gap Vg and the voltage across the dielectric layer Vd by

s g

V =V +V_d (2) with

( )

d d T

V V S J dt

= +C

∫

where JT is the instantaneous total current density and C the equivalent capacitance of the dielectric layer, and S the discharge surface [33][34]. Here, we assume the gas gap has the same condition, meaning the Vg is considered as constant. According to the capacitance formula, C ^A

d

=ε , reducing the dielectric layer thickness causes the

capacitance to increase. The voltage across the dielectric layer, Vd, reduces relatively at the same dielectric layer. The equivalent circuit model to estimate the capacitance of the micro-plasma device is shown in Fig. 4.8.

300 400 500 600 700 800

Bipolar voltage waveform frequency of 10 kHz

Fig. 4.7. The firing voltage for the interdigitated electrode micro-plasma device with different dielectric layer thickness in neon gas at pressure from 300 to 800 Torr.

Fig. 4.8. The equivalent circuit model to estimate the capacitance of the interdigitated electrode micro-plasma device.

4.4 The Electrical Properties in Neon-Argon Gas Mixtures

The ignition voltage of the diamond-shaped interdigitated electrode micro-plasma device is also investigated in pure neon gas at pressure from 300 to 800 Torr. The ignition voltage, where the fabricated device has 300 nm dielectric layer thickness, reduces as a function of neon pressure. In addition, the characteristic of the pure argon gas is also measured at the same pressure range. Fig. 4.9 shows the comparison of ignition voltage with pure neon and pure argon gas. The ignition voltage of pure argon gas reaches a lower minimum value and becomes larger after 500 Torr. The main reason for the fact that the relationship of ignition voltage versus gas pressure becomes a curve is due to the Paschen’s Law. The law essentially states that, at higher pressures (above a few Torr) the breakdown characteristics of a gap are a function (generally not linear) of the product of the gas pressure and the gap length,

usually written as V= f( pd ), where p is the pressure and d is the gap distance. The reduction of the dimensions of micro-plasma device afforded by MEMS can be operated at atmospheric pressure. However, the ignition voltage of neon reduces by increasing the operating pressure because the values of pd are small enough. In addition, after the minimum value of the critical pd, the ignition voltage of the argon gas increases.

300 400 500 600 700 800

180 200 220 240 260 280

Ignition Voltage (V)

Pressure (Torr)

100% Ne 100% Ar

Bipolar voltage waveform frequency of 10 kHz

Fig. 4.9. The ignition voltage for the interdigitated electrode micro-plasma device in pure neon and argon gas at pressure from 300 to 800 Torr.

The mixtures, containing 1% and 5% of the argon gas, all behave in a similar manner to the pure neon gas. The ignition voltage of the neon-argon mixtures, shown in Fig.4.10, has a lower value than the pure neon gas under similar conditions. Also, the voltage margin decreases by the increase of argon gas concentration.

300 400 500 600 700 800

Fig. 4.10. The characteristics of firing voltage and different concentrations of argon gas for the interdigitated micro-plasma device with bipolar voltage

waveform frequency of 10 kHz at pressure from 300 to 800 Torr.

However, the mixtures are used to provide an improvement in efficiency of ionization due to the Penning effect and a reduction in dissociative recombination losses. It could have an advantage over single pure neon gas in two ways. Firstly, the energy required to produce a give number of ions should be reduced since energy consumed in the production of metastable states is still available for ionization.

Secondly, the probability of ion loss by dissociative recombination is lower, since the partial pressure of minority gas is so low that the probability of molecular ion formation, required for dissociative recombination, is very small.

The metastable state of neon is larger than the ionization potential of argon.

Besides, the lifetime of a neon atom in this state is long enough to make many thousands of collisions at moderate argon gas concentrations. When the excited neon atoms collide with argon atoms, the energy exchange takes place and the ionization of argon atoms occur. A relatively small amount of argon in the neon will ensure that the excited neon atoms will collide with an argon atom in a very high probability. In such a collision, the excited state of the neon atom will return to the ground state and loss its energy to ionize the argon atom.

4.5 The Glow Images of the Interdigitated Electrode Micro-Plasma Device

The glow of the 5x5 hexagon-shaped arrays, shown in Fig. 4.11, is uniform and operating in 300 Torr of neon gas. The driving voltage and frequency are 250 V and 10kHz, respectively. After operating the micro-plasma device, however, the barrier rib is still well-found.

10 kHz Hexagon-shaped

Fig. 4.11. Photograph of the 5x5 arrays of the interdigitated micro-plasma device with a hexagonal structure operating in 300 Torr of neon gas.

300 Torr Diamond-shaped

Ne 100% Ne-1% Ar 300 Torr

(a) (b)

300 Torr Ne-5% Ar

(c)

Fig. 4.12. Photograph of the 5x5 arrays of the micro-plasma device with a diamond structure operating in 300 Torr of (a) neon gas, (b)neon(99%)-argon(1%), and (c) neon(99%)-argon(5%) mixtures.

In addition, the glow of the 5x5 diamond-shaped arrays, shown in Fig. 4.12(a), is

In addition, the glow of the 5x5 diamond-shaped arrays, shown in Fig. 4.12(a), is