This dissertation has two research topics, which is organized as follows:
In chapter 1, we will introduce this dissertation at beginning and describes the motivation of this paper.
In chapter 2, we will present the on-chip antenna structure. And before that, we will review some basic theories such as half-wave dipole, image theory, and monopole. Then we begin to develop our first topic, or dual-band millimeter-wave on-chip monopole antenna.
Next, in chapter 3, we will demonstrate a novel structure for the end-fire pattern antenna. Firstly, we will introduce some antenna type whose radiation pattern is end-fire. And then, we will show the results of this design.
Then, in chapter 4, the design of the high F/B ratio quasi-Yagi antenna has been presented. Here, we will review the theory of the Yagi-Uda antenna first. And then, we will demonstrate our study of the high F/B ratio quasi-Yagi antenna. After introduction of the high F/B ratio quasi-Yagi antenna, we will present the enhance type high F/B ratio quasi-Yagi antenna.
The last, chapter 5, we will give the summary and the conclusion of all and the future study.
Chapter 2
A 24/60GHz Dual-Band Millimeter-Wave On-Chip Monopole Antenna
In this chapter, a 24 / 60GHz dual-band millimeter-wave on-chip antenna is presented here. We design the feeding network by using the coplanar waveguide (CPW) structure. For the dual-band design, there are two major current paths to radiate. To avoid the harmonic frequency band of the low frequency-band, we add two strips to couple the harmonic frequency of the low frequency-band. Besides, the two strips let the higher frequency-band performance better. At this work, the simulator is based 3-D full-wave EM solver, Ansoft HFSS. Here, we will display the simulation result of this design.
2.1 Basic Theory
2.1.1 Theory of the Coplanar Waveguide (CPW) Structure
In recent years, the CPW structure, or the GSG for the single input or GSGSG for the differential input, is very widely used in the feed structure of the MMIC or RFIC system because of its several advantages.
A coplanar waveguide (CPW) fabricated on a dielectric substrate was first demonstrated by C. P. Wen [20] in 1969. Since that time, tremendous progress has been made in CPW based microwave integrated circuits (MICs) as well as monolithic microwave integrated circuits (MMICs) [21]-[24]. A conventional CPW on a dielectric substrate consists of a center strip conductor with semi-infinite ground planes on either side shown in Fig. 2.1.
The CPW offers several advantages over conventional microstrip line:
The first is that it simplifies fabrication, which no any metal on the back side and all the metal are on the same plane. That is to say it eliminates the need for wraparound and via holes [25] and [26].
Second, a ground plane exists between any two adjacent lines so cross talk effects between adjacent lines are very week [25] and the CPW has low dispersion and hence offers the potential to construct wide band circuits and components.
Third, the characteristic impedance is determined by the ratio of the signal path metal width s, or 2a in the figure, and the distance of the two ground plane s+2w, or 2b in the figure, shown in Fig. 2.2, so size reduction is possible without limit, the only penalty being higher losses [27].
And the forth, it reduces radiation loss when delivering the signal by CPW structure. So it is very appropriate to being the feed network of the antenna.
According these advantages of the CPW structure, hence we use the CPW structure to being the feed structure of the on-chip antenna in this paper.
The characteristic impedance is determined by the width of the signal line and the width of the gap on either side, which the approximate formula is shown below [26].
Where εre is the effective dielectric constant of the CPW structure. The approximate formula of the effective dielectric constant can be written that
1 (tanh(1.785log( ) 1.75) (0.04 0.7 0.01(1 0.1 )(0.25 )))
If the thick of the substrate is approximately infinite, the effective dielectric constant can be written as
(a)
(b)
Fig. 2.1 The conventional coplanar waveguide (CPW) structure (a) 3D structure and (b) cross-section view
Fig. 2.2 The design parameters of the conventional CPW structure[28]
Metal：Ground
Metal：Ground Metal：Signal
Substrate
2.1.2 Theory of Half-Wave Dipole Antenna[29]-[31]
In dipole antenna, the very widely used antenna is the half-wave dipole antenna whose structure is shown in Fig. 2.3(a). It is a linear current whose amplitude varies as one-half of a sine wave with the maximum at the center of the half-wave dipole antenna and the current distribution is shown in Fig. 2.3(a). And then the radiation pattern is shown in Fig. 2.3(b). The current distribution is placed along the z-axis and for the half-sine wave current on the half-wave dipole, the current distribution can be written as
( ) sin[ ( )],
4 4
I z =Im β λ− z z ≤ λ (2.7)
Where β=2π/λ, λ is the wavelength of the operating frequency of the antenna.
I(z)
Im
z
λ/2
(a) (b)
Fig. 2.3 The half-wave dipole
(a) Current distribution I(z) and (b) Far-field radiation pattern F(θ)
This current will have the maximum value I at the center (m z=0) and will be zero at the ends (
z λ4
= ± ). According to the current distribution, the far-field radiation pattern can be calculated as
cos( cos )
The definition of the field pattern function, F(θ)=g(θ)f(θ), then the complete (normalized) far-field pattern of the half-wave dipole antenna is
cos[( 2) cos ]
( )
-F πsin θ half wave dipole
θ = θ (2.11)
And then we define the antenna directivity, D, which defines as that:
max rad 4 D U
P π
= (2.12)
In theory, the input power can be radiated totally by antenna. In practical, the antenna has the loss, however, the radiated power will less than the input power.
gain, G, which are defined as
r rad in
P
η = P (2.13)
Note that
0≤ηr ≤ 1 (2.14)
And the antenna gain is
G= × ηr D (2.15)
Since gain is a power ratio it can be calculated in decibels as follows
10log
GdB = G (2.16)
10log
DdB = D (2.17)
Gain relative to a half-wave dipole carries the units of dBd. And the unit dBi is often used instead of dB to emphasize that an isotropic antenna is the reference. The relation between the dBi and the dBd is:
2.15
dBi dBd= + (2.18)
According to the antenna parameters mentioned above, we can know that the radiation efficiency is the higher the better. That is to say the input power can radiate
by antenna almost and then the antenna gain will be higher.
2.1.3 Theory of the Image Theory[29]-[31]
Consider an ideal dipole near a perfect ground plane and oriented perpendicular to the ground plane shown in Fig. 2.4. The uniqueness of the solution to a differential equation (wave equation) plus its boundary conditions introduces an equivalent system that is different below the ground plane (GG’). However, it satisfies the same boundary conditions on the ground plane (GG’) and has the same sources above the plane. Using this equivalent model, the solution will be different for the initial problem which below the plane. However, we can find the same solution for the problem above the ground plane and satisfies the boundary conditions. As a result, the image for this case is equidistant below the image plane and the same direction.
Fig. 2.4 Ideal dipole above and perpendicular to a perfectly conducting ground plane
(a) Physical model and (b) Equivalent model using image theory
An ideal dipole oriented parallel to a perfect ground plane has an image that again is equidistant below the image plane. However, the direction is oppositely as shown in Fig. 2.5.
Fig. 2.5 Ideal dipole above and parallel to a perfectly ground plane (a) Physical model and (b) Equivalent model using image theory
The image of a current element oriented in any direction with respect to a perfect ground plane can be calculated by decomposing the element into perpendicular and parallel components, shaping the images of the components, and constructing the image from these image components as shown in Fig. 2.6.
Fig. 2.6 Ideal dipole above and obliquely oriented relative to a perfectly ground plane
(a) Physical model and (b) Equivalent model using image theory
2.1.4 Theory of the Monopole[29]-[31]
A monopole is a dipole that has been divided in half at its center feed point and fed against a ground plane shown in Fig. 2.7. According to the image theory, if the current distribution over the monopole antenna is equal to the dipole then the electric
field of the monopole and the dipole will be the same. However, the image current of the monopole is generated by the ground metal. Hence, the length of the monopole is one-quarter wavelength, which is half of the dipole.
Fig. 2.7 Monopole antenna over perfect ground plane with their image (dashed)
The current and charges on a monopole are the same as on the upper half of its dipole counterpart, but the terminal voltage is only half that of the dipole. The input impedance for a monopole is half of its dipole counterpart, or
1 ,
Where ZA,mono is the input impedance of the monopole and ZA,dipole is for dipole.
Therefore, the radiation resistance of the monopole can be written as:
12
Where Rr,mono is the radiation resistance of the monopole and Rr,dipole is for dipole.
as a dipole. However, a monopole fed against a perfect ground plane radiates one-half the total power of a similar dipole in free space. Because of the reason, it is leading to a doubling of the directivity:
1
, 2 ,
4 4
mono 2 dipole
A mono A dipole
D = π = π = D
Ω Ω (2.21)
Where Dmono is the directivity and Ddipole is for dipole.
2.2 Design of the 24 / 60 GHz Dual-Band Millimeter-Wave On-Chip Monopole Antenna
The structure of the proposed dual-band millimeter-wave on-chip CMOS antenna is shown in Fig. 2.8 and the layout photo of the on-chip antenna is shown in Fig. 2.9. The used process of this on-chip antenna is TSMC 0.13 CMOS process. The CPW line structure is adopted for feeding network of this antenna [1] and [2]. The structure consists of three parts: longer path monopole antenna, shorter path monopole antenna, and two stubs at the middle.
The longer current path which is near a loop and about quarter wavelength of the lower operating frequency operates at low frequency band. On the contrary, the shorter current path which is on the right side of this antenna and approximated quarter wavelength of the higher operating frequency works at high frequency band.
Here, the wavelength is the effective wavelength of the lower or higher frequency calculated by the effective dielectric constant of the substrate [32] and [33].
Besides, it is important that the distance of the CPW gap, g = 48 μm, and the distance, d = 40 μm, between the radiator and the ground metal of the CPW structure are the key factors of the matching feeding and the return loss [34].
The two strips in the middle of the antenna structure are designed by Ls = 335 μm, Ws = 30 μm, and the gap between the two stubs, s = 20 μm. The feeding line is designed by W = 80 μm. The on-chip antenna size is about 0.76 × 1.045 mm2.
g g
d
Ls
Ws
W s
d
GND
GND
z SUB
Y
x
(a) (b)
Fig. 2.8 Configuration of the dual-band millimeter-wave on-chip antenna:
(a) The Geometry and (b) The 3D structure
Fig. 2.9 The layout photo of the on-chip antenna
2.3 Simulation Results
The current distribution of the on-chip antenna is shown in Fig. 2.10 for operating at 24GHz and 60GHz respectively. When operating at lower band (24GHz), the main current distributes over the longer path of the on-chip antenna shown in Fig.
2.10(a). And for higher band (60GHz), the main current distributes over the shorter path of the on-chip antenna shown in Fig. 2.10(b)
Fig. 2.11 shows the simulated input return loss of the antenna. At 24GHz band, the minimum of input return loss is about 15dB and the bandwidth is about 180MHz.
For the higher operating frequency band, 60GHz, the minimum of input return loss is about 38dB and the bandwidth is also about 700MHz.
For the radiation pattern at the lower operating frequency band shown in Fig.
2.12, the XY-plane approximates an omni-directional pattern and the gain is about -9dB. At this band, we can expect the pattern result by the current distribution that this antenna acts as a CPW monopole antenna. That is to say, the pattern of this antenna seems like the monopole antenna.
For higher operating frequency band, the radiation pattern shown in Fig. 2.13 has higher directivity so that the gain is about 1dB. We can see in Fig. 2.10(b) that there have two major current paths. So the pattern of this band has higher directivity. The simulated result summary is shown in Table 2.1
Table 2.1 simulated result summary
Simulated results 24 GHz band 60 GHz band
Bandwidth 180 MHz ; 0.75% 700 MHz; 1.67%
Max. gain in XY-plane -9 dB 1 dB
2
(a)
(b)
Fig. 2.10 The current distribution: (a) at 24 GHz and (b) at 60 GHz
(a)
(b) Fig. 2.11 The input return loss
(a) near the 24 GHz band and (b) near the 60 GHz band
0
0
(a) The XY-plane, (b) The XZ-plane and (c) The YZ-plane
2.4 Conclusion
A 24 / 60 GHz dual-band millimeter-wave CPW-fed on-chip antenna is presented not only for a 24 GHz ISM-band application but also for 60 GHz WPAN CMOS on-chip antenna application. This on-chip CMOS antenna is fabricated with a 0.13-μm standard CMOS process. The whole on-chip antenna size is about 0.76 × 1.045 mm2. The bandwidth of the lower band is about 0.75% and about 1.67% for higher band. The radiation pattern of the lower band approximates an omni-directional at the XY-plane and the gain is about -9 dB. For the higher band, the radiation pattern has higher directivity so the maximum gain is about 1 dB.
Chapter 3
A Novel Structure for the End-Fire Pattern Antenna
In recent years, there several antenna configurations which can provide the end-fire pattern have been proposed such as Yagi-Uda antenna [11]-[19], traveling-wave long wire antenna [35]-[36] and antenna array systems [37]-[38].
However, traditional end-fire radiated antenna is middle gain antenna and is not suitable for a radar system. In addition, most traditional end-fire antenna arrays are along z or x axes direction, which will has excellent directivity but the cross aperture figure will be limited in some special application. Therefore, it is very necessary to develop a new structure which is located along the end-fire direction, or y-axis.
In this chapter, we demonstrated a new configuration of the end-fire radiated antenna based on the leaky-wave antenna structure which has the frequency scanning characteristics that will radiate to near end-fire direction at high frequencies [39]-[42].
The proposed structure is composed of the conventional open end leaky-wave antenna with reducing ground plane and the adding short-circuited stubs and the connecting metal squares. According to the measured results, the impedance bandwidth is about 2.1GHz from 3GHz to 5.1GHz of 10-dB. The antenna peak gain we obtained is about 5.4dBi in the end-fire direction. Furthermore, the F/B ratio of this design is better than 30dB and the measurement results show that the F/B ratio is increased while the operating frequency is increased.
3.1 Some end-fire antenna structures
There are some type antennas which can obtain the end-fire radiation pattern
shown in Fig. 3.2[44] and the antenna array systems shown in Fig. 3.3[45]. All of these antennas, the main operating mode is based on the array theory. Like the antenna array, the Yagi-Uda antenna structure is similar to the array but it only excides one element. For the traveling-wave long wire antenna, the current distribution of the antenna is also like the array systems.
According to the mention above, we know that the main design idea is the antenna array theory if we want design end-fire pattern antenna. In the array theory, there are many factors will affect the radiation pattern, like the distance between the each element or the excided power of each element. That is to say, we can change these factors to design the end-fire antenna.
Fig. 3.1 Yagi-Uda antenna[43]
Fig. 3.2 Traveling-wave long wire antenna[44]
Fig. 3.3 Antenna array system[45]
3.2 The Antenna Design
Figure 3.4 shows the structure of the proposed antenna which has three parts: a) conventional open end leaky-wave antenna structure with reducing ground plane, b) adding the stubs which are connected to the ground plane and c) adding the connecting metal squares which connect the antenna body and the short-circuited stubs. These configurations are fabricated on the substrate, or FR-4, with a dielectric constant of εr = 4.4 and a thickness of h = 1.6mm. This research adopts an asymmetrical feed line to excite the proposed antenna which is similar to leaky-wave antenna. Detailed dimensions are listed in Table 3.1. Figure 3.4(a) shows the structure of the convention open end leaky-wave antenna of which the length L is 50mm, the width W is 15mm and the ground plane on the back side of the substrate is 45mm × 10mm.
In Fig. 3.4(b), it shows the design which we add the short-circuited stubs in order to improve the antenna gain and the impedance bandwidth. The stubs can be divided into three parts. The first stub connects to the ground plane and the distance, g1, between this stub and the feed line is 3mm. The second stub is a vertical one which is connecting to the first stub and with the gap, g2 = 1mm, between this stub and the antenna body. The third stub is parallel to the bottom side of the antenna body with the gap g3 = 0.5mm and it connects to the second one. Since these stubs are short-circuited by the first stub and the gap distance between the antenna body and the third stub is very slight, it will cause that most of the current will distribute on the edges of the gap. For all the stubs, the width Ws of them are equal to 1mm and the length Ls1 of the first stub is 14mm, the second one, or Ls2, is 11.5mm and the third one, or Ls3, is 52mm.
Figure 3.4(c) shows the configuration in which we add the connecting metal
the connecting metal squares is 0.5mm × 0.5mm. Here, the position and the size of the metal squares which are connecting the antenna body and the short-circuited stubs act extremely critical roles of this design. They not only match the impedance bandwidth but also affect the antenna gain and F/B ratio since the connecting metal squares will cause that the main current will be directed to the connecting metal squares between the short-circuited stubs and antenna body. That is to say, comparing configuration (b) and configuration (c), the current distribution changes from that most of the current flows on the edges of the gap to that most of current distributes at the connecting metal squares points.
Table 3.1 Dimension of the proposed antenna structure
L 50 mm Ls1 14 mm g2 1 mm
W 15 mm Ls2 11.5 mm g3 0.5 mm
Wg 10 mm Ls3 52 mm L1 8 mm
Ws 1 mm g1 3 mm
80mm
Fig. 3.4. Configuration of the proposed antenna,
(a) conventional open end leaky-wave antenna with reducing ground plane.
(b) Adding the stubs which are connected to the ground plane.
(c) Adding the connecting metal squares which connect the antenna body and the short-circuited stubs
3.3 Simulation and Measurement Results
In our study, for the original configuration which is similar to the conventional open end leaky wave antenna but the ground plane is reduced as shown in Fig. 3.4(a), we can see the current distribution in Fig. 3.5(a) which demonstrates that most of the current distributes at the bottom side of the antenna body and the current flows along the y-axis. Meanwhile, the length of antenna body is about a wavelength for the operating frequencies so the wave will travel on antenna body but there still has slight
In our study, for the original configuration which is similar to the conventional open end leaky wave antenna but the ground plane is reduced as shown in Fig. 3.4(a), we can see the current distribution in Fig. 3.5(a) which demonstrates that most of the current distributes at the bottom side of the antenna body and the current flows along the y-axis. Meanwhile, the length of antenna body is about a wavelength for the operating frequencies so the wave will travel on antenna body but there still has slight