FACULTY OF ENGINEERING AND SUSTAINABLE DEVELOPMENT
.
Design and Development of Gigahertz Range VCO Based
on Intrinsically Tunable Film Bulk Acoustic Resonator
Danial Tayari
June 2012
Master’s Thesis in Electronics
Master’s Program in Electronics/Telecommunications
Examiner: Prof. Daniel Rönnow
i
PREFACE
The Master‟s thesis is the outcome of my research at Terahertz and millimeter wave laboratory of Chalmers University of Technology, Sweden, for the Master‟s program in Electronics/ Telecommunication Engineering at University of Gävle, Sweden.
Professor Spartak Gevorgian at Chalmers University of Technology supervised me during the thesis. The thesis is examined by Professor Daniel Rönnow at University of Gävle.
ii
Abstract
The purpose of this thesis is to des ign and fabricate Gigahertz range voltage controlled oscillator based on intrinsically tunable film bulk acoustic resonator.
Modified Butterworth Van Dyke (MBVD) model was studied and implemented to simulate FBAR behavior. Advanced designed system (ADS) was used as the simulation tool.
Oscillator theory is studied and an oscillator based on non-tunable FBAR at 2GHz is simulated which shows -132 dBc/Hz phase noise @ 100 kHz offset frequency.
iii
Acknowledgments
I would like to thank Professor Spartak Ge vorgian for considering me as a member of his research group.H is constant supervision during the thesis period guided me through the right pass to achieve my goal.
Dr.John Berge who helped me in the fabrication process of the VCO,without his help the fabrication wouldn‟t have been done in the proposed time.
My special thanks to Professor Daniel Rönnow for accepting to be the examiner of my work. To my friends at Gävle and Göteborg who always helped and supported me during my stay in Sweden.
iv
Notations and Abbreviations
Notations
C Capacitance
Cm Motional capacitance
foff , ωoff Offset (angular) frequency
g Coplanar wave guide gap width
kt2 (effective) piezoelectric coupling coefficient
Phase-noise relative to carrier at offset frequency
Lm Motional inductance
PDC DC power
Qp Parallel resonance quality factor
Qs Series resonance quality factor
Rm Motional resistance
s Coplanar wave guide strip width
Zo Characteristics impedance
𝜀eff Effective permittivity
α Attenuation constant
v
Abbreviations
AC Alternating Current
AIN Aluminum Nitride
BAW Bulk Acoustic Resonator
BJT Bipolar Junction Transistor BSTO Barium Strontium Titanate
CPW Coplanar Waveguide
DC Direct Current
FBAR Film Bulk Acoustic Resonator
FOM Figure of Merit
GSG Ground Signal Ground
HFO Hafnium Oxide
IF Intermediate Frequency
LO Local Oscillator
MBVD Modified Butterworth Van-Dyke
PLD Pulsed Laser Deposition
PN Phase Noise
RF Radio Frequency
SAW Surface Acoustic Wave
VCO Voltage Controlled Oscillator
vi
Contents
1 Introduction ... 1
1.1 Introduction and Motivation ... 1
2 Film bulk acoustic resonators... 3
2.1 Characteristics ... 6
2.1.1 Quality factor ... 6
2.1.2 Effective Coupling coefficient... 6
2.2 Modeling ... 7 2.3 Tunable FBAR ... 8 3 Oscillator theory ...11 3.1 Types of oscillators ...11 3.1.1 Relaxation oscillators ...11 3.1.2 Harmonic oscillators...12 3.2 Oscillation criteria...12 3.2.1 Reflection oscillator ...14 3.2.2 Transistor oscillator ...15
3.3 Oscillator Phase noise ...16
3.4 Oscillator figure of merit ...17
4 FBAR Oscillators ...18
4.1 Fixed frequency FBAR Oscillator ...19
4.2 Voltage controlled oscillator (VCO) Based on Non-tunable FBARs. ...22
5 Tunable FBAR VCO ...25
5.1 Design...25
5.1.1 ADS momentum design ...26
5.1.2 Substrate definition ...26
5.1.3 Coplanar waveguide ...27
5.1.4 Grounding capacitors...29
5.1.5 Tunable FBAR resonator...31
5.1.6 Decoupling capacitor ...32
5.1.7 Meandered circuit ...33
5.1.8 Co-Simulation ...34
6 Device Fabrication and Measurement ...37
vii 6.1.1 PLD overview ...38 6.1.2 Laser-target interaction. ...39 6.1.3 The plume...39 6.1.4 Pulsed laser ...39 6.2 Measurements ...42
6.2.1 Test resonator measurement...42
6.2.2 Oscillator measurements ...44
7 Conclusion and future work ...45
8 References ...46
Appendix1: Tunable FBAR Resonator MBVD model Extracted Parameters ...49
Appendix2: Transmission line parameter extraction ...50
1
1 Introduction
The motivation for the work presented in the thesis and thesis organization are provided in this chapter.
1.1 Introduction and Motivation
Oscillators are devices which create a periodic AC signal at a defined frequency. All RF and microwave devices containing transmitters or receivers use oscillators. Usually oscillators are deployed in receivers and transmitters with mixers in which the signal can be up or down converted in frequency. In contrast to fixed frequency oscillators , - the voltage control oscillators (VCOs) can produce a range of frequencies enabling the radio device using them to operate on many different frequencies.
Communication systems frequency range can be determined by the frequency range of the oscillator. So this frequency range is desirable t o be large enough for covering the operating communication band.
In today‟s communication systems, staying within the allocated frequency band and not disturbing other users of adjacent frequencies is very important. Therefore the frequency stability of the devices should be very high. In case of oscillators the frequency stability is defined by its phase noise and it is a critical issue since it determines the frequency stability of the complete radio transceiver.
There are number of parameters affecting the oscillator phase noise. One of the most important one is the reactive components quality factors. In LC oscillators reactive components such as inductors and capacitors are employed in the resonator part .The LC resonator defines the oscillation frequency and the quality factor of the resonator has significant impact on oscillator phase noise. So the concern is to have resonators with higher quality factor. LC resonator quality factor is generally very low (around 20) making it a challenge for designers to design low phase noise oscillators based on LC resonators. However designing broad frequency range oscillators is easier when using these low quality factor components.
To have high quality resonators, surface acoustic wave (SAW) and bulk acoustic wave (BAW) devices can be used. The film bulk acoustic resonator (FBAR) is a type of recently developed BAW devices with very high quality factor.
2
These types of FBARs make it possible for the oscillator to have relatively high tun ing range compared to traditional AlN FBAR oscillators and are suitable in wireless communication systems and sensor applications, e.g. bio sensing.
The contents of this thesis are organized as follows.
3
2 Film bulk acoustic resonators
Film bulk acoustic resonator (FBAR) consists of a piezoelectric thin film which is sandwiched by a pair of electrodes. Depending on the applied frequency of the electrical signal to the FBAR; its piezoelectric material may expand or contract resulting in generation of the acoustic waves. When the thickness of the piezoelectric layer is the same as an integer of half acoustic wavelength the resonance occurs which means the resonance frequencies are determined by thickness and are independent from lateral dimensions.
At some frequency the impedance of FBAR reaches its minimum magnitude, which means the generated acoustic wave travels in the most efficient way through the physical material. This frequency is referred to the series resonant frequency fs . On the other hand when FBAR
impedance magnitude reaches its maximum there is no response from the piezoelectric meaning that no acoustic wave transfers energy through the FBAR. This happens at parallel resonance frequency or anti resonance frequency fp [1].
Fig. 2.1 FBAR structure under bias voltage
4
Fig. 2.2 illustrates the relation between the FBAR Impedance and its resonance frequencies
In comparison to LC resonators FBAR offers very high quality factor. In recent years FBARs using different material with very high quality factors of more than 2000 at a few gigahertzes have been introduced and FBAR duplexer filters for mobile phones now are commercially in use. Integrated LC resonators with varactor typically have Q factor of around 20.
In order to have high quality factor of the FBAR, the resonator must be acoustically isolated from the substrate. Depending on the Type of isolation FBARs are categorized as solidly mounted or membrane mounted, Fig. 2.3.
The first type uses an acoustic reflector, a Bragg reflector, consisting of λ/4 layers with high and low acoustic impedance alternatively. The second type based on an air cavity formed below the bottom electrode.
Series resonance
Parallel resonance
5
(a) (b) Fig 2.3(a) solidly mounted and (b)membrane mounted FBAR
Dimensions of FBAR are relative to acoustic wavelength. Acoustic propagation speed in solid materials is about 103_104 m/s. For e lectromagnetic waves however it is of order 107_108 m/s. So acoustic wavelength is four to five order of the magnitude lower than electrical wavelength, consequently FBARs are much smaller than electromagnetic resonators based on transmission line segments, for example. In addition acoustic loss is fairly low for piezoelectric materials at gigahertz range making them useful for high quality factor resonators at those frequencies. For example AlN_FBARs with quality factor of 280 demonstrated at 20 GHz at [2]. Some reported FBARs are given in Table 2.1.
T able 2.1: some recently reported FBARS
6
2.1 Characteristics
The important characteristics of FBAR relevant for Oscillators are the Q-factor (quality factor) and the resonance frequencies. Regarding these, the coupling coefficient is important in that the quality factors and impedance response depend on it.
2.1.1 Quality factor
For a resonator the quality factor is defined as
(2.1)
Where the maximum Energy stored in the resonator and is the energy dissipated in the lossy sections in one resonance period.
In simple resonators Q is commonly obtained from 3_dB bandwidth of the impedance. In FBARs however, the Q can be determined by the equation given in [10].
| | (2.2)
2.1.2 Effective Coupling coefficient
The coupling coefficient shows the percentage of the energy converted from mechanical to electrical and vice versa. [11]
7
2.2 Modeling
One of the most commonly used models for FBAR is the modified Butterworth-Van Dyke (MBVD) model. [12] Which is shown in Fig. 2.4
Rm Cm Lm
Rs
R0 C0
Fig. 2.4 MBVD model
In this model C0 represents the parallel plate capacitance, Cm, Lm, and Rm show the acoustic
resonance, Rs represents the ohmic loss of the electrodes while R0 defines the dielectric loss.
Typically the MBVD model is extracted from the measurements and used as a model in circuit simulation.
The measured reflection coefficient of an FBAR resonator and its equivalent MBVD model is plotted in Fig. 2.5. The parameters of MBVD model should be tuned in a way that simulated and measured plots fit each other excellently.
8
2.3 Tunable FBAR
Traditional FBARs are not tunable, however in case of intrinsically tunable FBAR a DC voltage is used to tune the resonance frequency of FBAR [13]. A tunable FBAR can be obtained from the DC field dependency of dielectric constant, acoustic velocity and electromechanical coupling coefficient
As an example of tunable FBAR, presented in [14] is shown in Fig. 2.6(a). For this FBAR, BSTO as the ferroelectric material and HfO2 and SiO2 as the layers of the Bragg reflector are
used.
As the bias voltage increases the resonance loop grows. For comparison Fig. 2.6(b) represent a plot of traditional FBAR. Although the resonance frequencies are not the same, we can see the difference between these two. For a given resonance frequency and capacitance, the resonance loop size is governed by the effective coupling coefficient and acoustic quality factor. This shows that FBAR Based on AlN gives much higher quality factor than the BSTO tunable one.
Table 2.2 gives the resonance frequencies of the resonator for different bias voltages. The resonator is then connected to the rest of the circuit.
(a) (b)
9
„T able 2.2: Extracted resonance frequencies from resonator MBVD model
Bias voltage(V) fs(GHz) fp(GHz) 5 5.729 5.752 10 5.658 5.743 15 5.612 5.722 20 5.58 5.733 25 5.553 5.712
Extracted parameters of the MBVD model are given in Figure 2.7.
(a)
Motional Inductance vs. DC Bias
10
(b)
(c)
Fig. 2.7 Extracted MBVD model parameters vs. bias voltage (a) motional inductance Lm (b) motional resistance Rm (c) motional
and static capacitance Cm and C0
Motional Resistance vs. DC Bias
1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 0 5 10 15 20 25 30 Voltage(V) M o ti o n a l R e si st a n c e ( O h m s)
Motional & Static capacitance vs. DC Bias
11
3 Oscillator theory
An oscillator is a circuit which uses DC power to generate periodic AC signal at its output. In an oscillator there is no need for any input signal except for the DC power supply. This chapter discusses briefly the two main types of oscillators, criteria to have the oscillation and the important merits to evaluate the oscillator performance. More explanations about working principle of the oscillators can be found in [15].
3.1 Types of oscillators
Oscillators are divided into two main groups, relaxation oscillators and harmonic Oscillators.
3.1.1 Relaxation oscillators
These oscillators switch repetitively between two states e.g. charging or discharging a capacitor or inductor. The amplitude of the charging current and the time constant determines the frequency of oscillation.
A simple relaxation oscillator is shown in Fig. 3.1
12
3.1.2 Harmonic oscillators
A harmonic oscillator consists of a resonator and an active device. The active device cancels the losses in the resonator, resulting constant oscillation amplitude at the frequency defined by the resonator.
In the oscillators the resonator is made of an inductor and a capacitor. The resonance frequency of a circuit is dependent on the values of the inductor and capacitor and defined as
√ Where is the value of the inductor and is the value of the capacitor.
For microwave frequencies, harmonic oscillators are preferred due to better phase noise performance. The oscillators in this work are of harmonic type.
3.2 Oscillation criteria
Harmonic oscillators use positive feedback. If a transistor is sufficiently biased it can provide enough feedback from the output for oscillation. This type of design is known as reflection oscillation which is explained later in this chapter. A schematic of a feedback oscillator is shown in Fig. 3.2
An amplifier is shown by block α while β represents feedback block connecting output to the input. A small input signal is used to understand the oscillation.
α
β
𝑉𝑜𝑢𝑡
𝛿
𝑣13
The output signal is having a transfer function given in (3.2)
(3.2)
Where α is the forward gain and a function of the amplitude of the signal, while dependence of the feedback β, is typically on the signal frequency .The open-loop gain is identified as
(3.3)
Analyzing small-signal closed loop transfer function can be done to see if the circuit has necessary condition for oscillation. It is important that the function has a pair of poles in the right–hand plane (RHP), or
Number of poles in the RHP + Number of poles in LHP > 0
Identifying the closed-loop transfer function and its poles is usually a difficult task. Instead small signal open-loop gain can be analyzed. To do this in simulation one can break open the circuit appropriately for open loop analyses.
The Nyquist criterion represents an oscillation criterion which is based on the open-loop gain analyses. According to this criterion when the small signal loop gain encircles the point 1+j 0 in the clock wise direction with increasing frequency, the closed-loop system is unstable.
An example of the Nyquist plot is shown in Fig. 3.3
14
The oscillation is guaranteed by Nyquist criterion, but since this criterion is based on the small-signal loop gains, no information regarding steady state frequency and amplitude oscillation can be obtained from it.
To have stable oscillation, loop gain must decrease and eventually be equal to 1+j 0.This is known as Barkhusen criterion [16]
In practice after the oscillation has been started the magnitude of the loop gain, due to the device non-linearity will be reduced to 1 for stable amplitude.
Having a zero phase in the loop gain means all the signals are summed together, producing a sum that is greater than any of the single signals. If they were in opposite phase for example, they would have cancelled out, resulting in no oscillation.
3.2.1 Reflection oscillator
Reflection oscillators are common topology for microwave oscillators since the necessary feedback to make the circuit unstable can be provided by parasitic elements of the amplifier. Fig. 3.4 shows RF-circuit for one port negative-resistance oscillator.
𝛤𝑖𝑛 𝑅𝑖𝑛− 𝑍0 + 𝑗𝑋𝑖𝑛 𝑅𝑖𝑛+ 𝑍0 + 𝑗𝑋𝑖𝑛 𝛤𝐿 𝑅𝐿− 𝑍0 + 𝑗𝑋𝐿 𝑅𝐿+ 𝑍0 + 𝑗𝑋𝐿 𝑋𝑖𝑛 𝑅𝑖𝑛 𝑋𝐿 𝑅𝐿
15
Where Zin = Rin+ jXin is the input impedance of the active device and ZL = RL+ jXL is the
impedance of the passive load.
For the oscillation to occur the following conditions must be satisfied: +
+
For a passive load , indicating . So the negative resistance refers to an energy
source. When the magnitude of the signal which causes the negative resistance and the loss of the passive component (resonator) are balanced, steady state oscillation happens.
3.2.2 Transistor oscillator
In this type of oscillator the negative resistance is provided by terminating a potentially unstable transistor. The circuit model is shown in Fig. 3.5.
Negative resistance Transistor Terminating Network Load Network
16
3.3 Oscillator Phase noise
Phase noise is an important parameter for performance of an oscillator and is a way to qualif y frequency stability. Short term random fluctuations in the output signal are referred as phase noise. For a given signal as
0 +
both the amplitude and phase are constant but in realistic oscillators either external or internal noise-sources to the oscillator cause both quantities to have fluctuations.
Output spectrum of a realistic oscillator is given in Fig. 3.6
Amplitude variation is limited and due to circuit non-linearities for steady state oscillation has less impact on the oscillator performance. Phase variation on the other hand may be random and discrete- making it the main contributor in oscillation-noise.
In communication systems it is important for devices to stay in their defined operating frequency band, thus phase noise plays an important role since the frequency stability of the total system is
Random phase variations
Discrete spurious signals
frequency f0
Amplitude
17
determined by the frequency stability of the oscillator. For instance a local oscilla tor may cause channel interference due to its high phase noise when used in a down converter as shown in Fig. 3.7.
Power
Frequency
Oscillator phase noise is defined as the ratio of the power of a side band to the power of the carrier frequency at an offset frequency from the carrier. Typically the side band is
normalized by the unit bandwidth. ( )
( )
0 ⁄ 3.4 Oscillator figure of merit
Figure of merit (FOM) is used to rank oscillators. The most common FOM is given by (3.8) − +
−
(3.8)
Where is the phase noise in dBc/Hz , is the offset frequency , 0 is the oscillation
frequency and PDC is the transistor DC power consumption in milliwatt.
Interference wanted and adjacent channels
fIF fLO
18
4 FBAR Oscillators
The small size and high quality factor of FBAR resonators make them to be of interest in microwave oscillators.
Table 4.1 shows some publications of Oscillators based on FBAR.
T able 4.1: Summary of FBAR.based Oscillators
year Ref FBAR1 Resonator
integration T echnology f0 [GHz] Pout [dBm] PDC [mW] PN [dBc/Hz] FOM [dB] 1984 [17] ZnO,mm Mounted on pcb 0.259 -24 -66@100 kHz 2001 [18] ZnO,mm Mounted on pcb AlGaAS HBT 2.2 -108@100kHz 2003 [19] AlN,mm Mounted on pcb bipolar 1.985 10 -112@ 10 kHz 197 2005 [20] ZnO,mm Mounted on pcb 1.1 -13.6 -115@ 10 kHz 2008 [21] sm Mounted on pcb 130 nm CMOS 2.2 6 -136@1 MHz 195 2008 [22] ALN,sm Mounted on pcb 65 nm CMOS 22 0.062 -124@100kHz2 222 2008 [23] sm Mounted on pcb 65 nm CMOS 22 0.92 -128@100kHz2 214 2009 [24] ALN,sm Mounted on pcb BiCMOS 2.11 21.6 -136@100 kHz 209 2011 [25] ALN,mm Mounted on pcb 0.18 µm CMOS 2 0.022 -121@100kHz 222.9
1mm=membrane mounted, sm=solidly mounted 2
19
4.1 Fixed frequency FBAR Oscillator
The topology used in this wor k to design a fixed frequency FBAR oscillator is shown in Fig. 4.1. A commercial BJT transistor (Infineon BFP-420) on common-base topology is used and the circuit is designed as negative resistance Oscillator so that the transistor parasitic provides necessary feedback required.
ZR Za
Fig.4.1 common base topology
For the circuit to oscillate the oscillation condition ZR+Za=0 must be satisfied. This can be
achieved by designing a proper output matching network.The resonator is placed at the emitter port via a matching network. The resonator implemented in the circuit is of AlN solidly mounted type presented in [26].
The equivalent MBVD model of the resonator is given in Fig 4.2, while Table 4.2 summarizes corresponding parameters, - the series and parallel resonance frequencies, as well as the Q factors.
Rm Cm Lm
Rs
R0 C0
Fig. 4.2 2 GHz FBAR MBVD model
T able 4.2: 2 GHz FBAR extracted parameters [26]
20
As it is discussed in chapter 2 FBAR shows much lower impedance at the series resonance than of the parallel one making it easier to design the matching network for compensating the losses of the resonator. Regarding this the Oscillator is designed functioning at the series resonance of the resonator. The circuit is based on 0.6 mm Alumina substrate and microstrip transmission lines are used as matching networks.
Simulation is done using Agilent ADS software. Large signal model provided by the manufacturer is used to characterize the transistor, which is biased at Vcc=4 V and Ic=20 mA The
output matching is designed to create an impedance which cancels out the resonator losses so there is no need for extra matching network at the resonator side.
Quarterwave transmission lines and capacitors are used as RF chokes, Harmonic balance is used for large signal analysis of the steady state oscillation and finally the procedure is completed by fine tuning the circuit. Fig. 4.3 shows the ADS schematic of the circuit.
Fig. 4.3 ADS schematic of the designed 2 GHz fixed frequency FBAR Oscillator
Fig. 4.4(a) shows the output spectrum of the circuit. It can be seen that the circuit oscillates at 2.044 GHz which agrees with the resonator series resonance frequency.
The Oscillator phase noise is given in Fig. 4.4(b). At 100 kHz offset from the fundamental, the oscillator has a phase noise of -132 dBc/Hz which is better than the one reported in[22].
The output waveform is given Fig. 4.3(c) in two periods.
21
Fig. 4.3 2 GHZ FBAR Oscillator (a) output spectrum (b)phase noise plot (c) output voltage in time domain
(a) (b)
22
4.2 Voltage controlled oscillator (VCO) Based on Non-tunable FBARs.
In LC oscillators frequency tuning can usually be achieved by directly connecting a control voltage at the varactor part of the resonator. In AlN FBAR oscillators however this method doesn‟t give frequency tuning range higher than 1 MHz. Some of the reported VCOs based on non-tunable FBAR are listed in Table 4.3.
T able 4.3: comparisons of non-tunable FBAR VCOs
Reference f0(GHz) Power Tuning
range[MHz] Best PN@1MHz[dBc/Hz] FOM[dB] Technology [20] 1.1 - 0.2 -123 - Discrete [19] 2 3.3V/35 mA 2.5 -150 -195 Bipolar [27] 2.1 2.4V/24. 3mA 37 -144 -193 0.25µm BiCMOS+ aboveIC [28] 2 67µw 10 -149 -220 0.13µmCM OS
Fig. 4.4 shows a colpitts based PCB oscillator, with the FBAR wire bonded to the circuit. This topology is used in [20] and by applying a DC control voltage a very small frequency tuning range is achieved. Vtune Vdd FBAR
23
The VCO in [19] is based on common-collector topology. The FBAR is wire bonded to the Oscillator and a varactor is coupled to the FBAR to reach the tunability of 2.5 MHz at 2 GHz.
Fig. 4.5 FBAR VCO with tuning varactor [19]
The circuit diagrams in [27] and [28] are given Fig. 4.6 and Fig. 4.7.
24
The parasitic, non-tunable, reactive elements of the circuit always reduce the frequency tunability. By comparing the works that have been done until now it can be understood that to have more tuning range of the FBAR the circuit topology gets more complicated and yet the tuning range is quite low compared to LC VCOs.
Due to limitations in fabrication process, in this work the purpose was to design a high tuning range FBAR VCO keeping the circuit topology as simple as possible by using only one transistor. Chapter 5 discusses the design of a voltage control oscillator based on intrinsically tunable FBARs.
25
5 Tunable FBAR VCO
As it is mentioned in section [2.3] tuning of an FBAR is possible by having ferroelectric material like BSTO as the piezoelectric of the FBAR resonator. Oscillators based on tunable FBARs are not studied previously and this work presents VCOs using tunable FBARS for the first time.
5.1 Design
In this section an integrated VCO based on Tunable FBAR is presented. The design and simulation are done in Agilent ADS software and the final mask is prepared for fabrication process.
A single –transistor topology is chosen in order to reduce the fabrication complexity. The tunable FBAR is located on the emitter of the transistor as shown in Fig. 5.1. The oscillator frequency is defined by the series resonance frequency of the device which is tunable according to the applied bias voltage VR through the inductor L.
Fig. 5.1 T unable FBAR Oscillator circuit Schematics
Decoupling capacitor C1 is used to isolate the resonator from the circuit in DC. An open stub
matching network at the output and inductive stub at the resonator ports make the compensation for the resonator loss. Stubs S1 and S2 are quarter waves and AC shorted by capacitors C2 and
C3The output is taken from the collector by a 50 load. Coplanar wave guide which.is used for
the stubs and matching network and the transistor is the same as in section [4.1] (base, two emitters and collector).One of the emitters connects to the FBAR via S3 and the other is used for
dc-bias through stub S1.
26
5.1.1 ADS momentum design
To do electromagnetic simulation of the circuit, the design is done in ADS momentum.
Each part of the circuit is separately simulated and finally C o-simulation is done to analyze the total circuit including DC analyses. The following section describes the design in momentum.
5.1.2 Substrate definition
The substrate used in this work was presented in [14] to achieve a tunable FBAR. Fig. 5.2 shows the substrate layers with the corresponding thicknesses. The circuit is designed using 0.5µm thick gold layer on top of the BSTO layer. For the resonator and circuit fabrication the metal layers are patterned according the design as explained in sections [5.1.5-5.1.8].
27
5.1.3 Coplanar waveguide
The conventional coplanar waveguide (CPW) consists of conductors on top of a dielectric surface [29]. The two ground planes are separated from the center strip by the gap as shown in Fig. 5.3.
W
The thickness and permittivity of the substrate, the dimensions of the center strip and the gap width determine the characteristic impedance (Z0), the attenuation constant and the effective
dielectric constant (𝜀eff) of the CPW. [29].
Using CPW simplifies the fabrication, eliminates the need for via holes and reduces radiation loss [30]. In CPW characteristic impedance is determined not only by the strip but also the slot width, making possible to reduce the size without limit. The only drawback is higher losses[31]. To make low Z0 in conventional CPW, a very wide center strip conductor and a very narrow slot
width can be fabricated. This however shows high current density at the slot edges which increases conductor losses; moreover a wide strip conductor can potentially couple power from the dominant CPW mode to unwanted spurious propagation modes. Therefore it is not recommended to have conventional CPW lines with Z0 less than 30 [29].
Fig. 5.3 shows the CPW used in this circuit design in ADS momentum.
28
Fig. 5.3 CPW in ADS momentum
In this design, GSG ports are defined. Ports1 and 2 are defined as signal ports, ports 3 and 5 are ground reference ports associated with port 1 and ports 4 and 6 are ground reference associated with port 2.Table 5.1 gives the line parameters extracted from the software by keeping t he center strip width constant.
T able 5.1: Extracted CPW parameters from ADS momentum at 5.421 GHz
29
It can be seen that the strip and gap widths of 100 and 48 µm respectively, determined the CPW characteristic impedance of the 50 𝜴 which was later used in the circuit design.
Using this data in the general transmission line model, the Oscillator circuit was designed based on the measured data of the tunable FBAR resonator introduced in [14].
The ADS schematic of the circuit is shown in Fig. 5.4.
Fig. 5.4 ADS schematic of designed tunable FBAR VCO circuit
Due to the relatively low Q factor of tunable FBAR resonator it was decided to integrate the circuit to reduce the noise caused by parasitic effects as much as possible.
5.1.4 Grounding capacitors
In order to AC ground the quaterwave stubs in the design, the platinum layer forms bottom electrode of the capacitor. As shown in Fig. 5.5.
Since there are conductive holes in BSTO a 100 nm silicon dioxide layer is considered over the BSTO layer to isolate the top and bottom electrode in DC.
30
Fig. 5.6 Reflection coefficient of the AC shorted quarterwave length stub
The reflection coefficient of the grounded quarterwave length stubs is shown in Fig. 5.6.The stubs represent high impedance at the desired oscillation frequency range, suitable for isolating RF signal from the DC bias voltage
Fig. 5.5 large size AC grounding capacitor
Pt+Au
SiO2
Au
31
5.1.5 Tunable FBAR resonator
The resonator was modeled and designed according to the measured data obtained in [14]. The resonator is connected to the emitter by an integrated inductor made from the Al top electrode. The active area of the resonator is 700µm2 Fig. 5.7 gives illustration of the resonator in Momentum.
Fig. 5.7 T unable FBAR resonator in ADS momentum
The geometry of the top-electrode in the active area is designed in way to suppress spurious lateral acoustic resonances. The DC probe will be applied to the gold patch on the top of aluminum inductor.To get the resonance frequency of the resonator Co-simulation is done by adding the motional parameters from MBVD model as shown in Fig. 5.8.
Resonator active area Resonator DC Bias location MBVD model motional parameters Spiral inductor made from Top Al electrode
Fig. 5.8 T unable FBAR Co-simulation in ADS
Au
Al
32
5.1.6 Decoupling capacitor
To isolate the DC bias of the resonator from other part of the circuit a series decoupling capacitor near the emitter leg is used, as shown in Fig. 5.9. Here again the gold layer and bottom electrode are isolated using silicon dioxide layer.
Fig. 5.9 Decoupling capacitor (C1 and C2 form a series capacitor letting through the RF and isolating DC of the resonator).
Capacitors C1 and C2 are formed between the platinum and gold layers with BSTO and SiO2 as
The dielectric material. The series capacitor is large enough not to load the resonator to affect the RF signal and the discontinuity at the gold layer prevents the DC bias signal disturbing the rest parts of the circuit. The decoupling capacitor is illustrated in Fig. 5.10 which shows the total designed circuit in momentum
33
5.1.7 Meandered circuit
Due to limitation in the fabrication and dimension of the sample (1 10mm) it was decided to make the circuit as compact as possible so the transmission lines were meander in order to have two circuits in one sample. The circuits with meandered lines are represented in Fig. 5.11
34
5.1.8 Co-Simulation
Some modification such as placing the resonators in the middle of the mask and adding some test resonators needed to be done for the final mask to get the optimum layout for fabrication. To see the oscillator performance Co-simulation is done which is shown in Fig. 5.12
35
(c)
Co-simulation results for different bias voltages are given in Table 5.2.
T able 5.2: Oscillator Co-simulation results
Resonator bias voltage(V) Oscillation frequency(GHz) PN@100kHz(dBc/Hz)
5 5.655 -101 10 5.596 -106 15 5.561 -106.3 20 5.545 -106.3 25 5.526 -106.4 § Output Spectrum
36
The Co-simulation results show that the Oscillation frequency tunability range of 129 MHz. The oscillation frequency is slightly lower than the resonator series resonance frequency which is expected for the loaded resonator. The phase noise is -106 dBc/Hz @ 100kHz frequency offset from the carrier. The final mask for fabrication is given in figure 5.14
Fig. 5.14 final oscillator mask with the frame and alignment
37
6 Device Fabrication and Measurement
Fabrication of the device is done at Chalmers C leanroom. A 20 nm thick layer of TiO2 for better
adhesion between SiO2 and Pt using magnetron sputtering . The device layer is sputtered on the
1 1cm sample containing the Bragg reflector. Bottom electrode pattern has been mapped from the mask to the platinum bottom electrode by photolithography as shown in Fig. 6.1
38
6.1 BSTO film growth by the PLD
The growth of the BSTO film has been performed using Pulsed Laser Deposition .(PLD) and explained in following section.
6.1.1 PLD overview
PLD concept is basically simple and is shown in Fig. 6.2.
Fig. 6.2 Schematic of PLD device [32]
A short pulsed laser beam is focused onto a target (BSTO in this work). Plasma is formed immediately on the target surface due to the pulse energy. The plasma then reaches the substrate which had been mounted on a heater and heated to the defined temperature and causes the target material to be deposited on the substrate.
39
6.1.2 Laser-target interaction.
When the laser beam strikethrough the target surface, the material ablates out with same stoichiometry as in the target. The vapor pressure, absorption of the material and pulse laser wavelength determines the amount of ablated material [33].
6.1.3 The plume
The target forms a plume after ablation. This high energy plume tends to move towards the substrate and presents forward peaking phenomenon [34].
Oxygen is often introduced into the chamber to keep constant the stoichiometry of the oxide material and reduce the kinetic energy [33].
The pressure of the background gas and the distance between the substrate and target define the shape of the plume.
6.1.4 Pulsed laser
The laser energy significantly affects the film quality. Higher energy increases the vapor pressure and consequently the kinetic energy which could result in defects in the surface of the deposited film due to re-sputtering.
Substrate temperature dramatically effects the deposited film quality in a way that higher temperature results in better quality.
Table 6.1 summarizes the process parameters used by PLD system in BSTO film deposition.
T able 6.1: PLD system Setting
parameters comments Laser source KrF Laser wavelength 248 nm Energy density 1.5 J.cm-2 Target BSTO Oxygen pressure 20 Pa Repetition rate 10 Hz
Substrate Si/Hfo2/Sio2/…./Tio2/Pt
Substrate temperature 6200 -640o C
40 Fig. 6.3 shows the sample after BSTO deposition
Fig. 6.3. 10 10 mm sample after BSTO deposition. The rain blow color shows the thickness difference of BSTO surface
Isolating SiO2, 100 nm aluminum top electrode and finally 500 nm gold layers were sputtered
and patterned on the sample. A step by step fabrication process is presented in Appendix 3.The AFM picture of the BSTO film in the middle of the sample is shown in Fig. 6.4.
Fig. 6.4.AFM picture of the BST O film (a) 2D (b) 3D view of the BST O film surface
41
Fig. 6.5 fabricated sample after (a) Sio2 deposition & lift -off (b) Gold deposition & image reversal resist removal
(c) Al deposition & lift -off
(a) (b)
42
6.2 Measurements
6.2.1 Test resonator measurement
Test resonator one port measurement was done using Agilent PNA N5230A and probe station with 150- µm GSG mounted microprobes. Figures 6.7, 6.8 and 6.9 show the fabricated test resonator and measurement results.
Fig. 6.8 T est resonator reflection coefficient @ different DC bias voltages Fig. 6.7 fabricated and measured T est resonator
43
The test resonator measured results in comparison to the previously measured data which was used in the oscillator circuit design show a shift of around 200 MHz downwards in the resonance frequencies. It was also observed that maximum bias voltage for the resonator before the breakdown is 20V.These effects are due to the integrated inductor and fabrication process technology which differs slightly from the previously used resonator.
Fig. 6.9 measured series and parallel resonance frequencies Test resonator series and parallel resonace frequencies
44
6.2.2 Oscillator measurements
To measure the oscillator, BJT transistors were mounted on the circuit by silver epoxy as shown in Fig. 6.10.
Fig. 6.10 Final oscillator circuits including transistors
Prior to RF measurement of the oscillator output spectrum and the phase noise, a DC measurement was done to verify the transistor is consuming the expected power as considered in the simulation.
DC measurement showed that unfortunately the grounding capacitors were shorted to ground in DC due to the pin holes in SiO2 layer. That could be because of FHR device which heats the
45
7 Conclusion and future work
BSTO intrinsically tunable FBAR and passive components have been monolithically integrated on high resistivity silicon substrate. To demonstrate the optional of the technology, 5.5 GHz voltage controlled Oscillator have been designed and fabricated on the substrate.
The simulated results showed high tunability with low phase noise compared to LC tank oscillators.
Measured Test resonators represent tenability of 114MHz @ 5.5 GHz. DC measurement of the Oscillator revealed short circuit in the integrated RF grounding capacitors due to pin holes in the SiO2 and BSTO layer.
Future work will contain another fabrication round to prevent the pin holes by using E-beam evaporation technology for SiO2 deposition. Depending on the Oscillator performance, the
fabrication process can be further optimized.
46
8 References
[1] R. Lanz and P. Muralt, “Bandpass filters for 8 GHz using solidly mounted bulk acoustic wave resonators.,” IEEE transactions on ultrasonics, ferroelectrics, and frequency control, vol. 52, Jun. 2005, pp. 936-46.
[2] Y. Yoshino, “Piezoelectric thin films and their applications for electronics,” J. Appl. Phys., vol. 105, no. 6, pp. 061623-7, Mar. 2009.
[3] M. Ylilammi, J. Ellä, M.Partanen, and J. Kaitila, ”Thin Film Bulk Acoustic Wave Filter,” in IEEE Transactions an Ultrasonics, Ferroelectrics, and Frequency Control, vol. 49, no. 4,Apr.2002,pp.535-539.
[4] S.-H. Lee, J.-H. Kim, G.D. Mansfeld, K.H. Yoon, and J.-K. Lee, “Influence of Electrodes and Bragg Reflector on the Quality of Thin Film bulk Acoustic wave Resonators, ”in IEEE International Frequency Control Symposium and PDA Exhibition , 2002, pp. 45-49.
[5] B.P. Otis and J.M: Rabaey, “A 300-µW 1.9-GHz CMOS Oscillator Utilizing Micromachined Resonators,” in IEEE J. Solid-State Circuits, vol. 38,no. 7, July 2003,pp. 1271-1274.
[6] T. Yokoyama, T. Nishihara, S. Taniguchi, M. Iwaki, Y. Satoh, M. Ueda, and T. Miyashita,”New Electrode Material for low-loss and High-Q FBAR Filters,” in Proc. IEEE 2004 Ultrasonics symposium, vol. 1, pp. 429-432.
[7] B. Ha, I. Song, Y. Park, D. kim, W. kim, K. Nam, and J.J. Pak, ”Novel 1-chip FBAR Filter for Wireless Handsets,” in Int. Conf. On Solid-State Sensors, Actuators, and Microsystems Dig. Tech. Papers, 2005, vol. 2, pp. 2069-2073.
[8] H. Zhang, J. Kim, W. Pang, H. Yu, and E.S:Kim, ” 5GHz lLow-phase-noise Oscillator Based on FBAR with Low TCF,” Int. Conf. on Solid-State Sensonrs, Actuators, and Microsystems Dig. Tech. papers, 2005, vol. 1, pp. 1100-1101.
[9] E. Lee, L. Mai, and G. Yoon, “Development of High-Quality FBAR Devices for Wireless Applications Employing Two Step Annealing Treatments” in IEEE Microwave and Wireless Components Letters, vol. 21, no. 11, November 2011.
[10] M. Norling, J. Berge, and S. Gevorgian “Paraeter extraction for tunable TFBARs based on BaxSr1-xTio3” in IEEE MTT-S Int. Microwave Symp. Dig., 2009, pp. 101-104.
47
[12] J. D. Larson III, P. D. Bradley, S. Wartenburg, and R.C Ruby, “Modified Butterworth-Van Dyke circuit for FBAR resonators and automated measurement system,” in Proc. IEEE Ultrason. Symp., 2000, pp. 863-868.
[13] S. Gevorgian, a Vorobiev, and T. Lewin, “dc field and temperature dependent acoustic resonances in parallel-plate capacitors based on SrTiO[sub 3] and Ba[sub 0.25]Sr[sub 0.75]TiO[sub 3] films: Experiment and modeling,” Journal of Applied Physics, vol. 99, 2006, p. 124112.
[14] J. Berge and S. Gevorgian, “ Tunable bulk acoustic wave resonators based on Ba0.25Sr0.75TiO3 thin films and a HfO2/SiO2 Bragg reflector,” in IEEE Transactions on
Ultrasonics, Ferroelectrics and Frequency Control, vol. 58 ,no. 12, 2011, pp. 2768-2771, [15] Nave. R. (2010). Relaxation Oscillator Concept. Available: http://hyperphysics.phy-astr.gsu.edu/hbase/electronic/relaxo.html. Last accessed 16th April 2012.
[16] A. R. Hambley, Electronics, 2nd ed., Prentice Hall, 2000, pp.636-640.
[17] S.G Burns and R.S:Ketcham, “Fundamental-mode Pierce oscillators utilizing bulk-acoustic-wave resonators in the 250-300MHz range,”IEEE Trans.Microw.Theory Tech., vol.MTT-32,no.12,pp.16668-1671,Dec.1984.
[18]Y.S.Park,S.Pinkett,J.S.Kenney,and W.D:Hunt,”A 2.4 GHz VCO with an integrated acoustic solidly mounted resonator,” in Proc.IEEE Ultrason. Symp.,2001,pp.839-842.
[19] A.Khanna,E.Gane,and T.Chong, “A 2 GHz voltage tunable FBAR oscillator,” in IEEE MTT-S Int. Microwave Synp. Dig.,2003,pp. 717-720.
[20] J.J Kim, H. ZHANG, W. Pang, H. Yu, and E.S: Kim, “low phase-noise FBAR-based voltage controlled oscillator without varactor,” in Proc. IEEE solid-state Sensors, Actuators and microsys.,2005,pp.1063-1066.
[21] P. Vincent, J. B. david, I. Burciu, J. Prouvee, C. Billard, C. Fuchs, G. Parat, E.Defoucaud, and A. Reinhardt, “ A 1 V 220 MHz-tuning-range 2.2 GHz VCO using a BAW resonator,” in IEEE Int. Solid-State Circiuts Conf. Dig. Papers, 2008, pp. 478-479.
[22] S. Dossou, N. Abele, E. Cesar, P. Ancey, J.F. Carpentier, P. Vincent, and J. M. Fournier, “ 60 µW SMR BAW oscillator designed in 65 nm CMOS technology,” in IEEE Int. Symp. Circiuts Syst. Dig. Papers, 2008, pp. 1456-1459.
48
[24] H. El-Aabbaoui, J.-B. David, E. D. Foucauld, and P. Vincent, “Ultra low phase noise 2.1 GHz Colpitts oscillators using BAW resonator.” in IEEE MTT-S Int. Microwave Symp. Dig., 2009, pp. 1285-1288.
[25] A. Nelson, J. Hu, J. Kaitila, R. Ruby, B. Otis, “A 22_W, 2.0GHz FBAR Oscillator.” in IEEE Radio Frequency Integrated Circuits Symp (RFIC)., 2011
[26]. M. Norling, J. Enlund, S. Gevorgian, and I. Katardjiev, “ A 2 GHz oscillator based on a solidly mounted thin film bulk acoustic wave resonator, ” in IEEE MTT-S Int. Microwave Symp. Dig., 2006, pp. 1813-1816.
[27] Kim B. Ostman, and al., “ Novel VCO Architecture Using Series Above- IC FBAR and Parallel LC Resonance”,in IEEE J. of Solid-State circuits, vol. 41, no. 10, October 2006
[28] J. Shi, B. Otis, “ A Sub-100μW 2GHz Differential Colpitts CMOS/FBAR VCO,” in IEEE Custom Integrated Circuits Conf (CICC)., 2011
[29] Rainee N. Simons, Coplanar Waveguide Circuits, Components, and Systems. John Wiley & Sons, Inc. 2001
[30] D. A. Blackwell, D. E. Dawson, and D. C. Buck, „„X-Band MMIC Switch with 70 dB Isolation and 0. 5 dB Insertion Loss,‟‟ 1995 IEEE Microwave Millimeter-Wave Monolithic Circuits Symp. Dig., pp. 97—100, Orlando, FL, May 15—16, 1995.
[31]A. Borgioli, Y. Liu, A. S. Nagra, and R. A. York, „„Low-Loss Distributed MEMS Phase Shifter,‟‟ IEEE Microwave Guided Wave Lett., Vol. 10, No. 1, pp. 7—9,
January 2000.
[32] L. W. Martin, Y.H. Chu, and R. Ramesh. “Advances in the growth and characterization of magnetic, ferroelectric, and multiferroic oxide thin films” .Mater. Sci. Eng., R, 68(4-6):89-133, May 2010
[33] R. Easen. “pulsed laser deposition of thin films. Applications-led growth of functional materials”. Wiley, 2007.
49
Appendix1: Tunable FBAR Resonator MBVD model Extracted
Parameters
Fig.1T unable FBAR resonator MBVD model extracted parameters (a) series and parallel resonance frequencies (b) series and parallel Q factors (c) effective coupling coefficient (d) impedance magnitude
50
Appendix2: Transmission line parameter extraction
Conversion S-parameter to ABCD
Conversion of S to ABCD-parameter is given in [1]
ABCD network for Transmission line
,Z0
Z
ois a characteristic impedance of the transmission line.
is the length of the line. Note that+
Complex propagation constant= attenuation constant NP/m
β
= wave propagation constant51 For a Lossless line
=
0When the transmission line is lossless this reduces to
For TEM wave propagation the effective permittivity and Loss tangent can be obtained from [1]
𝜀
√
Where
Where is the attenuation constant due to dielectric in NP/m.
52
Fig. 1 extracted parameters for a 1.4 mm length CPW with g=48µm and S=100 µm on the multilayer substrate (a) characteristic impedance (b)effect permittivity (c)attenuation constant
Reference
[1] David M. Pozar, Microwave engineering Third edition. John Wiley & Sons, Inc., 2005
(a) (b)
53
Appendix3: Fabrication Steps and recipe
Note: the thicknesses are not to the same scale
Process Recipe Schematic
1.New sample including Bragg reflector
Parameters:Hfo2 260nm/
SiO2 284nm(3 pairs)
Size:10 10 mm Thickness:501 m
2.Cleaning Tool: Wet bench
Parameters: Acetone, Ultrasonic bath for 3 min @ 100% power
Intention: To clean the photo resist used for protecting the wafer during cutting
3.TiO2 Deposition Tool: Sputter –
NORDIKO 2000
Parameters: Ti and O2 for 8
min.
Intention: To deposit Tio2 on
the sample for better adhesion between Sio2 and Pt.
4.Platinum Deposition Tool: Sputter – NORDIKO 2000 Parameters:
Pt for 2 min @60 w
Intention: To deposit Pt on the sample as the bottom
54 pt 100 nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Silicon
5.Cleaning Tool: Wet bench
Parameters: Acetone, Ultrasonic bath for 3 min @ 100% power
Intention: To clean surface from impurities before applying the photo resist
6.Resist applying and spinning
Tool: Hot plates and resist spinner
Parameters:
photo resist S1813 @4000rpm for 30 sec.
Hot Plates@ 900 C for 1 min. Intention: To apply resist evenly on the sample for edge removal
7.Photo resist pattering-I (edge removal)
Tool: Mask aligner – KS MJB3-UV 400 Parameters:
Soft contact, exposure time 1 min.
Intention: To remove resist edges
8.Developing Tool: Wet bench
Parameters:
Developer MF-319 for 1.5 min.
55 pt 100 nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Silicon pt 100 nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Silicon pt 100 nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Silicon
9.Photo resist patterning -II Tool: Mask aligner – KS MJB3-UV 400 Parameters: soft contact, exposure time 15 sec. Intention: To expose photo resist according to the bottom electrode mask pattern
10.Developing Tool: Wet bench
Parameters:
Developer MF-319 for 15 sec.
Intention: to develop the exposed photo resist
11.Etching Tool: Ion Beam Milling
Oxford Chamber.
Parameters: Argon gas flow for 20 min.
Intention: To pattern the Pt bottom electrode layer
12.Resist removal Tool: Wet bench, Ultra sonic bath
Parameter: Microposit
56 pt 100 nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Silicon BSTO 234n m pt 100 nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Silicon BSTO 234n m pt 100 nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Silicon
13.Oxygen plasma strip Tool: Plasma Therm Batch Top
Parameters:O2 plasma for 1
min @ 250 W.
Intention: to remove organic residue from photo resists.
14.BSTO deposition Tool: Pulsed Laser Deposition-(PLD)
Parameters: Target BSTO Temperature 620-640o C- 3100 laser pulses in 5 min. Intention: To deposit BSTO film
15.Cleaning Tool: Wet bench
Parameters: Acetone
Intention: to clean the sample surface after BSTO deposition Ready to be taken to the main cleanroom
16.LOR Lift off- Resist Tool: Wet bench –Hot plates –Resist spinner
57 17.Coat and prebake imaging
resist
Tool: Wet bench –Hot plates –Resist spinner
Parameters: Photo resist S1813@4000 rpm for 30 sec. Hot plates @1100 C for 2 min. Intention: Coat and prebake resist making it ready for patterning.
18.Expose imaging resist Tool: Mask aligner – KS MJB3-UV 400
Parameters: soft contact exposure time 10 sec. Intention: To expose photo resist according to the SiO2
mask pattern
19.Develop the resist and LOR
Tool: Wet bench
Parameters: Developer MF-319 for 2 min
Intention: To develop the resist and LOR for SiO2
58 BSTO 234nm pt 100 nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Sio2 284nm Hfo2 260nm Silicon
20.Sio2 layer sputtering Tool: FHR MS 150x4-L
Sputter Deposition system. Parameters:O2 and Si
combination for 429 sec. Intention: To deposit SiO2
layer on the sample.
21.SiO2 Lift-off Tool: Wet bench-Ultra sonic
bath
Parameters: Microsit Remover 1165@750C for 5 min. Ultra Sonic bath @%20 for 1 min.
Intention: To lift off SiO2 and
have the pattern of it.
22.Resist applying for image reversal work(Gold layer)
Tool: Wet bench–Resist spinner
Parameters: Photo resist S1813@4000 rpm for 30 sec.
59
s oluble 23.Exposure using inverted
mask
Tool: Mask aligner – KS MJB3-UV 400
Parameters: exposure time 6 sec
Intention: to expose the sample for Gold deposition and patterning(the gold layer finally remains at exposed area)
24.Reversal bake Tool: Hot plates
Parameters: 1250C for 2 min. Intention: To make the exposed area inert while the unexposed area remains photo active.
25.Flood exposure without mask
Tool: Mask aligner – KS MJB3-UV 400
Parameters: flood exposure for 60 sec.
60
26.Developing Tool: Wet bench
Parameters: Developer AZ 351 B for 1 min
Intention: To develop the resist according to the mask pattern
27.Gold Deposition Tool: AVAC E-beam evaporator.
Parameters: 8.8 Å/ sec to reach 0.5 µm gold thickness. Intention: To deposit the Gold layer on the sample.
28.Lift-off Tool: Wet bench
Parameters: Acetone @750 C for 5 min.
61 29.Deposition of Aluminum
top electrode (same procedure as the Sio2 Layer)
Tool: FHR MS 150x4-L Parameter: Al deposition for 51 sec.