Embed Size (px)
Citation preview
UNIVERSIDAD POLITÉCNICA DE
MADRID
ESCUELA TÉCNICA SUPERIOR DE INGENIEROS
DE TELECOMUNICACIÓN
DOCTORAL THESIS
ULTRA HIGH FREQUENCY THIN
FILM SAW DEVICES
Author: Juan Gabriel Rodríguez Madrid
Ingeniero Electrónico
Thesis Supervisors: Prof. Fernando Calle Gómez
Catedrático de Universidad
Prof. Gonzalo Fuentes Iriarte
Profesor Contratado Doctor I3
2013
ii
iii
Doctoral Thesis: Ultra High Frequency Thin Film SAW Devices.
Author: Juan Gabriel Rodríguez Madrid
Thesis Supervisors: Prof. Fernando Calle Gómez
Prof. Gonzalo Fuentes Iriarte
El tribunal nombrado por el Mgfco. y Excmo. Sr. Rector de la Universidad Politécnica
de Madrid, el día … de …………… de 2013 para juzgar la Tesis arriba indicada,
compuesto por los siguientes doctores:
Dr. ……………………………………………………………….(Presidente)
Dr. ……………………………………………………………………(Vocal)
Dr. ……………………………………………………………………(Vocal)
Dr. ……………………………………………………………………(Vocal)
Dr. ………………………………………………………………..(Secretario)
Realizado el acta de lectura y defensa de la Tesis el día …. de ………………… de 2013
en Madrid, acuerda otorgarle la calificación de ……………………………
El Presidente:
El Secretario:
Los Vocales:
iv
v
A mi madre,
por apoyarme
en todo momento.
vi
vii
RESUMEN
Durante los últimos años el flujo de datos en la transmisión que tiene lugar en
los sistemas de comunicación ha aumentado considerablemente de forma que día a día
se requieren más aplicaciones trabajando en un rango de frecuencias muy alto (3-30
GHz). Muchos de estos sistemas de comunicación incluyen dispositivos de onda
acústica superficial (SAW) y por tanto se hace necesario el aumento de frecuencia a la
que éstos trabajan. Pero este incremento de frecuencia de los dispositivos SAW no sólo
es utilizado en los sistemas de comunicación, varios tipos de sensores, por ejemplo,
aumentan su sensibilidad cuando la frecuencia a la que trabajan también lo hace.
Tradicionalmente los dispositivos SAW se han fabricado sobre cuarzo, LiNbO3
y LiTaO3 principalmente. Sin embargo la principal limitación de estos materiales es su
velocidad SAW. Además, debido a la alta temperatura a la que se depositan no pueden
ser integrados en la tecnología de fabricación CMOS. El uso de la tecnología de capa
delgada, en la que un material piezoeléctrico es depositado sobre un substrato, se está
utilizando en las últimas décadas para incrementar la velocidad SAW de la estructura y
poder obtener dispositivos trabajando en el rango de frecuencias requerido en la
actualidad. Por otra parte, esta tecnología podría ser integrada en el proceso de
fabricación CMOS.
Durante esta tesis nos hemos centrado en la fabricación de dispositivos SAW
trabajando a muy alta frecuencia. Para ello, utilizando la tecnología de capa delgada,
hemos utilizado la estructura nitruro de aluminio (AlN) sobre diamante que permite
conseguir velocidades SAW del sustrato que no se pueden alcanzar con otros
materiales. El depósito de AlN se realizó mediante sputtering reactivo. Durante esta
tesis se han realizado diferentes experimentos para optimizar dicho depósito de forma
que se han obtenido los parámetros óptimos para los cuales se pueden obtener capas de
AlN de alta calidad sobre cualquier tipo de sustrato. Además todo el proceso se realizó
a baja temperatura para que el procesado de estos dispositivos pueda ser compatible con
la tecnología CMOS.
viii
Una vez optimizada la estructura AlN/diamante, mediante litografía por haz de
electrones se fabricaron resonadores SAW de tamaño nanométrico que sumado a la alta
velocidad resultante de la combinación AlN/diamante nos ha permitido obtener
dispositivos trabajando en el rango de 10-28 GHz con un alto factor de calidad y
rechazo fuera de la banda. Estás frecuencias y prestaciones no han sido alcanzadas por
el momento en resonadores de este tipo.
Por otra parte, se han utilizado estos dispositivos para fabricar sensores de
presión de alta sensibilidad. Estos dispositivos son afectados altamente por los cambios
de temperatura. Se realizó también un exhaustivo estudio de cómo se comportan en
temperatura estos resonadores, entre -250ºC y 250ºC (rango de temperaturas no
estudiado hasta el momento) diferenciándose dos regiones una a muy baja temperatura
en la que el dispositivo muestra un coeficiente de retraso en frecuencia (TCF)
relativamente bajo y otra a partir de los -100ºC en la que el TCF es similar al observado
en la bibliografía.
Por tanto, durante esta tesis se ha optimizado el depósito de AlN sobre diamante
para que sea compatible con la tecnología CMOS y permita el procesado de dispositivos
trabajando a muy alta frecuencia con altas prestaciones para comunicaciones y sensores.
ix
ABSTRACT
The increasing volume of information in data transmission systems results in a
growing demand of applications working in the super-high-frequency band (3–30 GHz).
Most of these systems work with surface acoustic wave (SAW) devices and thus there is
a necessity of increasing their resonance frequency. Moreover, sensor application
includes this kind of devices. The sensitivity of them is proportional with its frequency.
Traditionally, quartz, LiNbO3 and LiTaO3 have been used in the fabrication of
SAW devices. These materials suffer from a variety of limitations and in particular they
have low SAW velocity as well as being incompatible with the CMOS technology. In
order to overcome these problems, thin film technology, where a piezoelectric material
is deposited on top of a substrate, has been used during the last decades. The
piezoelectric/substrate structure allows to reach the frequencies required nowadays and
could be compatible with the mass electronic production CMOS technology.
This thesis work focuses on the fabrication of SAW devices working in the
super-high-frequency range. Thin film technology has been used in order to get it,
especially aluminum nitride (AlN) deposited by reactive sputtering on diamond has
been used to increase the SAW velocity. Different experiments were carried out to
optimize the parameters for the deposit of high quality AlN on any kind of substrates. In
addition, the system was optimized under low temperature and thus this process is
CMOS compatible.
Once the AlN/diamond was optimized, thanks to the used e-beam lithography,
nanometric SAW resonators were fabricated. The combination of the structure and the
size of the devices allow the fabrication of devices working in the range of 10-28 GHz
with a high quality factor and out of band rejection. These high performances and
frequencies have not been reached so far for this kind of devices.
Moreover, these devices have been used as high sensitivity pressure sensors.
They are affected by temperature changes and thus a wide temperature range (-250ºC to
250ºC) study was done. From this study two regions were observed. At very low
x
temperature, the temperature coefficient of frequency (TCF) is low. From -100ºC
upwards the TCF is similar to the one appearing in the literature.
Therefore, during this thesis work, the sputtering of AlN on diamond substrates
was optimized for the CMOS compatible fabrication of high frequency and high
performance SAW devices for communication and sensor application.
xi
CONTENTS
1. INTRODUCTION 13
1.1 SAW devices 13
1.2 Thin film technology 18
1.3 Thesis outline 19
2. THEORY OF ACOUSTIC WAVES 21
2.1 Basics of acoustic waves 21
2.2 Piezoelectricity 23
2.3 Surface acoustic waves in layered structures 25
2.4 Materials for SAW devices 28
2.4.1 Material properties for SAW devices 28
2.4.2 Some properties of AlN 30
2.4.3 Diamond as substrate 32
3. METHODOLOGY 33 3.1 Synthesis of AlN by sputtering 33
3.2 Material characterization techniques 38
3.2.1 X-ray diffraction (XRD) 39
3.2.2 Photoluminiscence (PL) 41
3.2.3 Atomic force microscopy (AFM) 41
3.2.4 Scanning electron microscopy (SEM) 42
3.2.5 Transmission electron microscopy (TEM) 43
3.3 SAW device fabrication 45
4. AlN FILMS: RESULTS AND DISCUSSION 48 4.1 Substrate 49
4.2 Target-substrate distance 54
4.3 Gas flow ratio (Ar/N2) 55
4.4 Discharge power 57
4.5 Process pressure 57
4.6 Film 58
4.7 Summary and conclusions 67
5. AlN THIN FILM SAW RESONATORS 68 5.1 Influence of the layer thickness on device performance 68
5.2 Extraction of coupling-of-mode parameters 74
5.3 SAW devices on “free-standing” diamond 81
5.4 Summary and conclusions 85
xii
6. SAW SENSORS 86
6.1 Pressure sensors 86
6.2 Temperature study 94
6.3 Summary and conclusions 98
7. GENERAL CONCLUSIONS AND FUTURE WORK 100
Acknowledgments 102
References 105
13
1. INTRODUCTION
This chapter describes the evolution of the surface acoustic wave (SAW) not
only for telecommunication systems but also for sensor applications. The importance of
thin film technology for high frequency SAW devices fabrication and CMOS
compatibility is also introduced in this chapter.
1.1 SAW devices
Surface acoustic waves (SAW) is a general term used for mechanical waves
which concentrate all or most of their energy on the surface region of a solid. Their
behavior is in contrast to that of bulk waves, where the energy propagates throughout
the solid. SAW are often referred to as Rayleigh waves because they were first
explained by lord Rayleigh in 1885 [Rayleigh.1885]. These are two dimensional waves
confined to the surface of a solid, down to a depth of approximately two wavelengths.
Since this first study, surface acoustic waves have been extensively investigated in the
seismology field. In 1927, so-called Sezawa waves were analyzed [Sezawa.1927].
These waves were later shown to be higher order Rayleigh modes.
The most popular method to excite acoustic waves in solids is by means of the
piezoelectric effect. In the field of signal processing, SAW devices were not exploited
until the discovery of the interdigital transducer (IDT) in 1965 by White and Voltmer
[White.1965]. From then, the basics of SAW device technology were settled down
during the period 1970-85, mainly design techniques and characterization of SAW
propagation on quartz and LiNbO3 [Morgan.1998]. Afterwards, SAW devices were
developed in different applications such as mobile phones and other wireless
communication systems [Campbell.1998]. Nowadays, SAW devices are widely used as
band-pass filters and resonators in communication systems to filter a transmission or
reception frequency (which basically means all devices requiring wireless transmission).
The fast growth of communication media, such as mobile and satellite services, together
with new fields of applications in sensing and radiofrequency (RF) identification (RF-
ID tags), has induced a growing demand of high-performance, low loss SAW devices.
The telecom industry alone consumes annually nearly four billion SAW-based radio
14
frequency (RF) bandpass filters, mainly for mobile phones and base-stations
[Weigel.2002]. Other applications for RF filters are GPS, navigation system and
relatively new systems for wireless data transfer (Bluetooth, WLAN). Table 1.1 shows a
summary of the main applications of SAW devices [Ballantine.1997, Schmidt.2000,
Pedros.2007, Epcos.2013, Rfsaw.2013]. The frequency of operation for the above
applications lies typically in the range of 400 MHz to 6 GHz, where conventional LC
filters suffer from unacceptable losses. This problem could be solved by
electromagnetic filters based on ceramic materials. In addition, the latter exhibits a good
power handling capability. On the other hand they are very large and therefore
impractical for applications where a high level of miniaturization is required. Among
other advantages, SAW devices are a trade-off between high frequency and
miniaturization.
Table 1.1. Principle applications of SAW devices.
SYSTEMS AND
INFRASTRUCTURES
MOBILE
COMMUNICATION
MULTIMEDIA
APPLICATIONS
SENSORS
AND
ACTUATORS
REMOTE
CONTROL
AUTOMO
TIVE
SECTOR
WLL and WiMAX WCDMA systems
(UMTS)
Digital audio
and video
Chemical
and
biochemical
sensors
Tag RFID Remote
key
WLAN/Bluetooth Mobil telephones
(GMS and
UMTS)
Digital
television
(DTV)
Microfluidic
actuators
Remote
switches
Tire
pressure
sensors
GPS Beepers
Wireless
audio
Radar
15
Figure 1.1. Telecommunication devices and electromagnetic spectrum
The electromagnetic spectrum and some communication technologies operating
at different frequencies are shown in figure 1.1 [Goleniewski.2006]. The terminology
that the International Telecommunication Union (ITU) applies to the various bands;
Extremely low (ELF) (3-30 Hz), super low (SLF) (30-300 Hz), ultra low (ULF) (300-
3000 Hz), very low (VLF) (3-30 kHz), low (LF) (30-300 kHz), medium (MF) (300
kHz-3 MHz), high (HF) (3-30 MHz), very high (VHF) (30-300 MHz), ultrahigh (UHF)
(300 MHz-3 GHz), superhigh (SHF) (3-30 GHz), extremely high (EHF) (30-300 GHz),
and tremendously high frequencies (THF) (300-3000 GHz) are all various forms of
radio bands. The principal applications for SAW devices are located along these
frequency bands.
On the other hand, for the last few decades, SAW devices have been used not
only for communication systems, but also in sensor applications [Ballantine.1997]. In
fact, virtually all acoustic wave devices are sensors in which wave propagation
characteristics are highly sensitive to minimum perturbations along the propagation path
and/or at its boundaries. Hence, these devices can be designed to exhibit extreme
sensitivities towards small variations in the surrounding medium. Changes in the
characteristics of the acoustic wave propagation path result in modifications of the wave
velocity and/or amplitude, which in turn result in a change of the output signal, i.e. a
frequency/phase shift or wave attenuation. Thus, SAW devices have been used in a
16
wide range of physical, chemical and biochemical sensors applications such as for gas
[Ricco.1985; Mauder.1995; Changbao.2011; Jinyi.2013] , temperature and pressure
sensors [Schimetta.2000; Kang.2012; Binder.2013] By bending stretching and
compressing the SAW substrate, sensors of different magnitudes such as torque, force,
displacement, vibration, acceleration, etc. have been fabricated [Seifert.1994;
Pohl.1998; Stoney.2012; Zukowski.2012]. In addition, SAW sensors can compete in
special applications and harsh environments, where the sensors have to withstand severe
conditions related to heat, radiation, or electromagnetic fields, or they are required to
have a long life time free of maintenance.
SAW sensor systems have used delay lines or resonators based on the following
principles: one-port SAW devices, directly affected by the measurand or property that
want to be measured, and two-port SAW devices, which are electrically loaded by a
conventional sensor and, therefore, indirectly affected by the measurand. Wireless
measurement systems with passive SAW sensors offer new perspectives for remote
monitoring and control of moving parts. As an example, the interest in tire pressure
monitoring systems (TPMS) for the automotive industry has increased in recent years.
First generation of TPMS was based on MEMS pressure sensors and battery-powered
active transmitters positioned in every wheel [Tjiu.2004]. The use of SAW sensors
offers the most promising solution for battery-less second generation TPMS
[Dixon.2006; Cook.2009].
SAW devices were traditionally fabricated on bulk piezoelectric substrates such
as LiNbO3, KNbO3, BaTiO3, etc. These are ferroelectric materials and thus exhibit a
piezoelectric effect. The substrate is generally a single crystalline material, thus the
main problem is the expensive and time consuming manufacturing process.
Furthermore, other disadvantages such as high cost and very difficult integration into
CMOS technology due to its high manufacturing temperature have pushed the
introduction of thin-film technologies. Moreover, with increased spectrum crowding,
high bandwidth requirements, miniaturization, and low cost requirements, this
technology is in the way to replace conventional SAW RF-filters in mobile
17
communication applications above 3 GHz due principally to both, cost and
performance.
Therefore, the recent trend is to improve low loss high frequency SAW devices
for current high speed communication systems. Since f = v/λ, there are two procedures
to increase the frequency. One way to achieve this aim is to reduce the wavelength of
the SAW, λ. This implies a reduction of the IDT period (i.e., the electrode width), which
however is limited by the lithographic resolution in the fabrication of the IDT. The
second approach for high frequency devices is to increase the phase velocity of the
SAW, v. In this context, new materials with higher SAW velocity are being studied, of
which diamond, with high phase velocity of over 10.000 m/s, has attracted a great deal
of attention. However, diamond substrates do not exhibit the property of
piezoelectricity. Thus, a piezoelectric layer is required for the excitation and detection
of the electroacoustic wave. The necessity of a film technology is clear in order to
develop this kind of structures.
Thin piezoelectric films deposited on a substrate have been recently developed
in order to overcome the disadvantages of the traditional piezoelectric substrates
explained before. In this approach, by properly choosing a material/substrate
combination, the electro-acoustic (and other) properties of the devices can be tuned. For
example, if the substrate SAW velocity is higher than the thin film piezoelectric
material, the frequency of operation of SAW devices increases for a certain lithographic
resolution [Assouar.2007]. The quality factor and thermal stability can also be improved
in a similar way [Wu.2001]. Furthermore, another important advantage of using thin
films is that the process can be easily integrated with planar technology. This enables
the possibility of fabricating thin film SAW/BAW components on chip as another step
in the silicon-based IC-processing line. The compatibility with CMOS integration
(when both materials, thin piezoelectric film and substrate are synthesized at
temperatures below 400ºC) of these devices will bring substantial benefits such as the
lower device size and fabrication costs, reduced losses and power consumption,
increased sensitivity, etc.
18
1.2 Thin film technology
Thin films are material layers, with thickness ranging from a few atomic layers
(nm) to a few micrometers, that are deposited on a substrate. That allows combining the
bulk properties of the substrate material with the surface properties of the deposited thin
film. As a demand for high technology products increases, more advanced materials are
needed. Thin film technology is an efficient and most often less expensive way to solve
the increased demands. This technology has been developed for more than 40 years in
several ways [Kline.1983; Lakin.2005]. The potential of this technology for
electroacoustic devices was demonstrated around 20 years ago [Lakin.1992;
Ruby.1994] but the first mass-produced component entered the RF market was an
FBAR-based 1900 MHz antenna duplexer for mobile phones, in 2001 [Ruby.2001;
Ruby.2004].
SAW devices based on thin piezoelectric materials deposited on a high SAW
velocity substrate, are the best candidates to replace present devices. The low insertion
loss, large bandwidth, temperature stability, and capability of handling powers up to 1
W, are the main stringent characteristics required by competitive filters nowadays.
Moreover, resonators with high coupling coefficient and quality factors are required, not
forgetting the low fabrication cost which could be affordable thanks to the thin film
technology.
Historically, zinc oxide (ZnO) films on silicon by reactive sputter deposition was
one of the first material used in thin film acoustic wave devices [Hickernell.1996]. Thin
aluminum nitride (AlN) films were also studied due to their higher acoustic wave
velocity, lower electrical leakage and relatively low temperature coefficient
[Okano.1993; Ishihara.1998; Takagaki.2002]. At the beginning, the main concern was
to grow high quality AlN with sufficient control of impurities during the deposition
(oxygen and water), which required the use of ultra-high vacuum systems. Advances in
deposition techniques such as physical vapor deposition (PVD) systems, allowed highly
textured piezoelectric AlN to be grown on a variety of substrate materials at low
temperatures (<500ºC) using reactive magnetron sputtering [Shiosaki.1980;
Iriarte.2002; Clement.2003; Rodríguez-Madrid.2012].
19
Nowadays, AlN, ZnO and lead zirconium titanate (PZT) are the most common
materials used in thin film technology for electroacoustic devices [Chandran.2012;
Pedros.2011; Rodríguez-Madrid.2012]. Among these materials, the AlN synthesis for
high frequency SAW and BAW application has been extensive researched and
developed and it appears to be the best compromise between performance and
manufacturability [Aigner.2003]. Other important advantage of the AlN thin film
technology is its process compatibility with CMOS integration. During last years,
microcantilevers for energy harvesting, bulk acoustic resonators (FBARs) and MEMS
actuators based on this technology have been fabricated through a fully CMOS
compatible process [Hara.2005; Defosseux.2012; Tran.2013]. Thus, high frequency
SAW resonators based on thin film AlN technology are very attractive in order to
replace present devices for future applications.
1.3 Thesis outline
The principal aim of this thesis research is to develop high performance and high
frequency SAW devices at low temperature for CMOS integration. The first step
consists of improving the quality of AlN layers deposited by reactive sputtering. A
number of factors are under consideration during the reactive sputtering process,
including among others the target-to-substrate distance, the gas flow, the process
pressure, and the target power, which have been studied in order to achieve the optimum
quality of c-axis oriented AlN, direction where the crystal orientation offers maximum
piezoelectric effect.
Silicon (Si), magnesium oxide (MgO), alumina (Al2O3) and silicon carbide (SiC)
are promising substrates for SAW devices, although we have mainly focused in
synthetic diamond. Its excellent properties, such as high SAW velocity and the
possibility of working under harsh environments, make it the best candidate for the
processing of SAW devices working in the high frequency range (GHz).
Before device manufacture, AlN thin films have been characterized by atomic
force microscopy (AFM), scanning electron microscopy (SEM), transmission electron
microscopy (TEM) and X-ray diffraction. High quality films were used to fabricate
SAW devices using e-beam lithography. The combination of AlN/diamond
20
heterostructures together with high resolution lithography allows us to fabricate devices
working above the 10 GHz range, not reached for this type of devices so far. Finally,
these ultra-high frequency devices have been implemented in applications such as
pressure sensors.
The thesis is organized as follows. The Introduction chapter briefly describes the
evolution of the SAW applications and the importance of thin film technology for high
frequency SAW devices fabrication and CMOS compatibility and for sensor
applications. Chapter 2 presents a summary of the main theoretical models used for the
design, description, parameter extraction and characterization of thin film resonators.
During chapter 3, the methodology followed to fabricate SAW devices is described;
process details regarding the deposition of piezoelectric AlN thin films, its
characterization techniques and the technological steps to process SAW devices are
presented here. Related to material synthesis, Chapter 4 presents all the results obtained
during the optimization of the process to deposit piezoelectric c-axis oriented AlN on
different substrates, focusing on diamond.
The results obtained after the characterization of super high frequency SAW
devices, are presented in chapter 5. In chapter 6, “free-standing” structures are
fabricated to work as high sensitivity pressure sensors, due to the high frequency range
of the devices. A temperature study of them is also shown in this chapter. Finally, in
chapter 7, the principal conclusions of this thesis work as well as the future work are
presented.
21
2. THEORY OF ACOUSTIC WAVES
The purpose of this chapter is to give a brief introduction on the electroacoustic
wave theory, focusing on the devices studied during this thesis work. The
piezoelectricity and the materials properties used for the fabrication of the high
frequency devices are studied through this chapter. Finally, acoustic wave propagation
in layered structures is described.
2.1 Basics of acoustic waves
Elastic waves or acoustic waves in solids are composed of mechanical
deformations, strain (S), of a material and the associated internal forces, stresses (T)
[Nye.1957]. Deformations in a solid are displacements of its constituent particles from
their steady state position. A plane wave generates displacements, and thus
deformations, that vary harmonically in the direction of wave propagation.
For linear plane wave theory the following assumptions are made; first, for a
small deformation of a body, the stress and strain relation is linear, which is known as
the Hooke’s law, 𝑇 = c𝑆 (c is the stiffness contstant). Second, the contours of constant
displacements for a plane wave in isotropic solids (those whose properties are the same
in all directions) are planes perpendicular to the propagation direction. For the case of
anisotropic solids, those whose properties vary with different crystallographic
orientations, the propagation of acoustic waves generally shows a strong dependence on
the propagation direction. The Christoffel equation, that describes the unbounded
acoustic wave propagation in an infinite solid, results in three general solutions or three
mutually orthogonal plane waves [Rosenbaum.1988], one quasi-longitudinal and two
quasi-shear, each associated with its acoustic phase velocity and particle displacement
vector (polarization). The first wave exhibits the highest phase velocity, and it is
polarized mainly in the direction of the propagation direction (longitudinal wave),
whereas the latest solutions, also known as transverse waves, are slower and polarized
mainly perpendicularly to the propagation direction. They are called “quasi”-
longitudinal or “quasi”-transverse because of the slight deviation (parallel or normal) of
the polarization vector from that in the isotropic case. Nevertheless, certain
22
crystallographic directions or symmetries yield polarizations that are exactly parallel or
perpendicular to the wave propagation, and they are therefore denoted as pure waves.
Stresses and strains representation in solids require tensor notation. The elastic
stiffness of a solid is fully described by a four rank tensor with a maximum of 81
elements. However, due to the symmetry properties of crystals, a reduced notation can
be used resulting in a 6 by 6 matrix, which for the most general solid consists of 21
independent elements. The number of independent elements reflects the degree of
crystal symmetry; as the symmetry increases the number of independent constants
decreases. For isotropic solids there exists only two independent constants representing
one pure longitudinal and one pure shear mode (the two shear waves are identical in
isotropic media and are therefore degenerate). The phase velocities, independent of the
propagation direction, are given by:
𝑣𝑙𝑜𝑛𝑔 ,𝑠𝑒𝑎𝑟 = 𝑐𝑙𝑜𝑛𝑔 𝑐𝑠𝑒𝑎𝑟
𝜌 (1)
where 𝜌 is the mass density of the solid and 𝑐𝑙𝑜𝑛𝑔 and 𝑐𝑠𝑒𝑎𝑟 are the longitudinal and
shear stiffness constants respectively.
Wave propagation in unbounded media can always be represented by these three
mutually independent non-dispersive plane waves. Electroacoustic devices, however,
are inherently associated with boundaries and wave reflections, which can, in general,
breed other types of modes that are composed of coupled plane waves where the
propagation is strongly dependent of the waveguide. In some cases they are dispersive,
i.e. the wave velocity is frequency dependent.
23
2.2 Piezoelectricity
Piezoelectricity is the ability of some materials (typically crystals and certain
ceramics) to generate an electrical potential in response to an applied mechanical stress.
This is the most common method to generate and detect acoustic waves.
Piezoelectric materials provide a direct transduction mechanism to convert
signals from mechanical to electrical domains and vice versa. The reversible and linear
piezoelectric effect manifests as the production of a charge (voltage) upon application of
stress, which is known as direct effect. On the other hand, the production of mechanical
stress upon application of an electric field is called the converse piezoelectric effect. To
produce these effects, the structure of a crystal lacks a center of inversion symmetry;
thus, the application of strain changes the distribution of charge on the atoms and bonds
comprising the crystal in such a manner that a net electrical polarization of the crystal
results. That is, for an unstrained crystal the arrangement of ions (charges) is in such a
way that the net charge dipole is zero. When mechanical deformation (compressive or
shear) happens, ions of opposite signs are displaced relative to each other, producing a
non-zero dipole and thus an electrical polarization in the crystal, as shown in figure 2.1.
Crystals exhibiting this direct piezoelectric effect always exhibit the converse effect as
well. The former effect is used for electrical detection of acoustic waves while the latter
is employed for electrical generation of acoustic waves.
Figure 2.1 Schematic illustration of the piezoelectric effect. Compressive or shear stress leads to dipole
formations.
24
The elementary piezoelectric effects are given by the piezoelectric constitutive
relations:
𝐷𝑖 = 𝑑𝑖𝑗σ𝑗 + 𝜀𝑖𝑖𝑇𝐸𝑖 𝐷𝑖 = 𝑒𝑖𝑗 S𝑗 + 𝜀𝑖𝑖
𝑆𝐸𝑖 (2)
𝑆𝑗 = 𝑠𝑖𝑗𝐸σ𝑗 + 𝑑𝑖𝑗𝐸𝑖 𝑇𝑗 = c𝑖j
Eσ𝑗 − 𝑒𝑖𝑗𝐸𝑖 , (3)
where Di is the electrical displacement, σj is the mechanical stress, Sj is the mechanical
strain, Ei is the electric field, cij is the elastic stiffness constant, sij is the elastic
compliance coefficient and εii is the permittivity. The piezoelectric coefficients, dij and
eij , are third rank tensors which in reduced tensor notation correspond to a 3x6 matrix
[Giacovazzo.2002]. Furthermore, the indices (i=1…3) define normal electric field or
displacement orientation, (j =1…3) define normal mechanical stresses or strains, and (j
=4…6) represent shear strains or stresses. Other common way to show the piezoelectric
constitutive relations is:
𝑇𝑖𝑗 = 𝑐𝑖𝑗𝑘𝑙𝐸 𝑆𝑘𝑙 − 𝑒𝑘𝑖𝑗 𝐸𝑘 (4)
𝐷𝑖 = 𝑒𝑖𝑗 𝑆𝑗𝑘 + 𝜀𝑖𝑗𝑆𝐸𝑗 (5)
In the direct effect using equation (2), a mechanical stress σj or strain Sj causes a
net electrical displacement, Di, on i faces of the material, the magnitude of which
depends on dij and eij, respectively. Similarly, the converse effect expressed by equation
(3) relates the induced normal and shear stress or strain to the applied electric field via
the piezoelectric coefficient and the elastic stiffness tensors [Nye.1957].
Transducers using piezoelectric materials can be configured either as actuators
(the design of the device is optimized for generating strain or stress using the converse
piezoelectric effect) or sensors (the design of the device is optimized for the generation
of an electric signal, in response to mechanical input). As a result, large piezoelectric
coefficients dij are desired in actuator applications, whereas sensor applications take
advantage of large piezoelectric eij coefficients [Gerfers.2010].
25
2.3 Surface acoustic waves in layered structures
A structure that converts electric field energy into mechanical wave energy and
vice versa is called an electroacoustic transducer. Interdigital transducer (IDT), first
developed by White and Voltmer in 1965 [White.1965], is one of the most common
transducers used to generate and detect SAW. This kind of transducer consists of a
comb-like metal structure (strip lines) defined on the surface of a piezoelectric material
(see figure 2.2). When an alternating electric field is applied between the electrodes of
an IDT, and as a consequence of the piezoelectric effect, surface acoustic waves will
travel away from the transducer in both directions. At the frequency for which the
acoustic wavelength is equal to the period of the IDT, the electric field between adjacent
electrodes will reach the highest level and all the electrodes will contribute to the
generation of the acoustic wave in a cooperative manner. Hence, at that particular
frequency, often referred to as the center or the Bragg frequency, electromagnetic waves
applied are converted into an acoustic signal with maximum efficiency.
Figure 2.2 Surface acoustic waves generated by an interdigital transducer (IDT) on a piezoelectric
substrate. The coordinate system used for equation definition is also shown.
The equations that govern the propagation of acoustic waves in piezoelectric
media are deduced after the quasistatic approximation assumption and applying the
mechanical equations of motion [Nye.1957; Rosenbaum.1988]. These are called the
wave equations or equations of motion of the system,
𝜌𝜕2𝑢𝑗
𝜕𝑡2− 𝑐𝑖𝑗𝑘𝑙
𝜕2𝑢𝑘
𝜕𝑥𝑖𝜕𝑥𝑙− 𝑒𝑘𝑖𝑗
𝜕2Φ
𝜕𝑥𝑖𝜕𝑥𝑘= 0 (6)
𝑒𝑖𝑘𝑙𝜕2𝑢𝑘
𝜕𝑥𝑖𝜕𝑥𝑙− 𝜀𝑖𝑘
𝜕2Φ
𝜕𝑥𝑖𝜕𝑥𝑘= 0 (7)
26
where Φ is potential, 𝑢𝑗 is elastic displacement. Following the method described in
[Campbell.1968], the general solution to the wave equation (6) and (7) is assumed to be
a linear combination of partial waves given by:
𝑢𝑗 = 𝛼𝑗𝑒𝑖𝑘𝑏 𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡), 𝑗 = 1,2,3 (8)
Φ = 𝛼4𝑒𝑖𝑘𝑏 𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡) (9)
where the constants 𝛼 give the relative amplitudes of the displacement components of
each partial wave, and the decay constant 𝑏 describes the variation with depth of the
amplitude and phase of the partial wave measured on a plane constant phase, i.e. a plane
perpendicular to the XZ-plane (figure 2.2). Substituting the partial waves (8) and (9)
into the wave equation (6) and (7), an eighth-order algebraic equation in the decay
constant 𝑏 is obtained [Iriarte.2003a].
The particle displacement and potential in the equations for the partial waves, (8)
and (9), are given as eight linear combinations of plane waves. These linear
combinations constitute the general solution of the system:
𝑢𝑗 = 𝐶𝑚𝛼𝑗𝑚𝑒𝑖𝑘𝑏𝑚 𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡)8
𝑚=1 𝑗 = 1,2,3 (10)
Φ = 𝐶𝑚𝛼4𝑚8
𝑚=1 𝑒𝑖𝑘𝑏𝑚 𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡) (11)
where 𝐶𝑚 are weighting factors that can be find out by determining as many boundary
conditions as necessary, that means, as many boundary conditions as weighting factors
Cm appear in the equation system.
So far, discussions concern bulk materials. For layered structures, another set of
boundary conditions must be satisfied at the interface between adjacent layers. The
assumption of the semi-infinite substrate is only applicable for the lowest layer, so that
eight unknown constants exist for each layer except for the lowest one. This is so
because at the interface between two layers of different materials, the wave is likely to
be reflected back towards the surface, hence giving rise to positive values of the decay
constants 𝑏. Once the equations describing the system are defined, boundary conditions
have to be applied to solve it. These are the follows:
27
The particle displacements and the traction components of stress (T13, T23 and
T33) must be continuous across the interface, because of the intimate nature of the
contact assumed between the two materials. Since the surface is supposed to be
mechanically stress free, the three traction components of stress must vanish thereon.
The electrical boundary conditions to the problem are provided by the continuity of the
potential and of the normal component of electric displacement across both the interface
and the free surface. In the case of a layered structure, the conditions of each layer
boundary are also required for 𝑢, 𝑇, 𝐷 and the potential for being continuous. The
particular case of AlN/diamond structure is deeply studied in [Iriarte.2003b;
Iriarte.2003c]. For this structure, the general solution for the displacement and the
potential satisfying (6) and (7) is given by:
𝑢𝑗𝐷𝑖𝑎 = 𝛼𝑗
𝐷𝑖𝑎𝑒𝑖𝑘𝑏𝐷𝑖𝑎 𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡) (12)
ΦDia = 𝛼4𝐷𝑖𝑎𝑒𝑖𝑘𝑏𝐷𝑖𝑎 𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡) (13)
𝑢𝑗𝐴𝑙𝑁 = 𝛼𝑗
𝐴𝑙𝑁𝑒𝑖𝑘𝑏𝐴𝑙𝑁 𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡) (14)
ΦAlN = 𝛼4𝐴𝑙𝑁𝑒𝑖𝑘𝑏𝐴𝑙𝑵𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡) (15)
where 𝛼𝑗𝐷𝑖𝑎 and 𝛼𝑗
𝐴𝑙𝑁 (j=1,2,3) are constants. Substituting (12)-(13) and (14)-(15)
equations into (6) and (7), the eight-order equation is obtained for 𝑏𝐷𝑖𝑎 and 𝑏𝐴𝑙𝑁
respectively and eight roots are obtained. Moreover, assuming that the diamond layer is
a bulk material, four out of the eight b roots satisfy the condition that the SAW wave
does not exist at negative infinity for –x3 (see figure 2.2). Therefore, the displacement
and the potential are expressed as the addition of the linear combinations of the eight
waves for AlN layer and four waves for diamond substrate, which are given as:
𝑢𝑗𝐴𝑙𝑁 = 𝐶𝑚
𝐴𝑙𝑁𝛼𝑗𝐴𝑙𝑁(𝑚)
𝑒𝑖𝑘𝑏𝐴𝑙𝑁(𝑚 )
𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡)8𝑚=1 (16)
ΦAlN = 𝐶𝑚𝐴𝑙𝑁𝛼4
𝐴𝑙𝑁(𝑚)8𝑚=1 𝑒𝑖𝑘𝑏𝐴𝑙𝑁
(𝑚 )𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡) (17)
𝑢𝑗𝐷𝑖𝑎 = 𝐶𝑚
𝐷𝑖𝑎𝛼𝑗𝐷𝑖𝑎 (𝑚)
𝑒𝑖𝑘𝑏𝐷𝑖𝑎(𝑚 )
𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡)4𝑚=1 (18)
ΦDia = 𝐶𝑚𝐷𝑖𝑎𝛼4
𝐷𝑖𝑎 (𝑚)4𝑚=1 𝑒𝑖𝑘𝑏𝐷𝑖𝑎
(𝑚 )𝑥3𝑒𝑖𝑘(𝑥𝑙−𝑣𝑡) (19)
28
Most of the SAW energy concentrates within one wavelength near the surface of
the substrate. Thus, the piezoelectric substrate is assumed to be semi-infinite when its
thickness is at least 3 times the acoustic wavelength. In a layered system as the one
studied during this thesis work, the SAW velocity value is between the SAW velocity of
the piezoelectric thin film and the SAW velocity of the substrate. Therefore, the SAW
velocity is dispersive for this kind of structures, and its value depends on the kH ratio,
where H is the thickness of the piezoelectric film and 𝑘 = 2𝜋𝜆 is the wave vector.
When vsubstrate > vpiezothinfilm, the SAW velocity of the structure is higher for smaller kH;
otherwise, for vsubstrate < vpiezothinfilm, the SAW velocity of the structure is higher for
higher values of kH.
2.4 Material for SAW devices
2.4.1 Material properties for SAW devices
The main properties of the piezoelectric material in SAW devices which
determine their performance are the phase velocity, the electromechanical coupling, and
the temperature coefficient of delay.
- Phase velocity(v): SAW propagate along the material with a phase velocity
given by:
𝑣 =𝑤
𝑘, (20)
where 𝑤 is the angular velocity and k is the wave vector. The resonance
frequency (𝑓0) of SAW devices is determined by the material phase velocity
and the SAW wavelength, λ,
𝑓0 = 𝑣
𝜆 . (21)
The wavelength of a SAW device is determined by its IDT periodicity
(=λ).
29
- The electromechanical coupling coefficient (K2) indicates the degree to
which a piezoelectric material converts electrical energy into mechanical
energy, or vice versa. The relative increase of the stiffness constant for a
given propagation direction is dependent on the piezoelectric stress, the
stiffness and the dielectric constants associated to that direction, according to
𝐾2 =𝑒2
𝜀𝑆𝑐𝐸 , (22)
where 𝑒, 𝑐 and are the piezoelectric coefficient, the elastic coefficient, and
electrical permittivity, respectively, and the superscripts 𝐸 and 𝑆 denote that
they have been measured at constant electric field and constant strain,
respectively. 𝐾2, called the piezoelectric coupling constant, is related to the
electromechanical coupling coefficient, kt, by
𝑘𝑡2 =
𝐾2
1+𝐾2 . (23)
In the literature the former formalism is typically used for laterally excited
waves (LTE), whereas the latter is used for thickness excited (TE) waves
[Rosenbaum.1988; Iriarte.2003a]. It should be noted that for 𝐾2 ≪ 1
𝑘𝑡2 ≈ 𝐾2. (24)
In an experimental way, 𝐾2can be obtained from [Ingebrigtsen.1969]:
𝐾2 ≅ 2𝑣𝑓−𝑣𝑚
𝑣𝑓, (25)
where 𝑣𝑓 and 𝑣𝑚 are propagation velocity of the wave with “free” and
metalized surface.
- Temperature coefficient (of) frequency (TCF) or delay (TCD): This
coefficient measures the temperature effect in the SAW device performance,
and indicates the SAW device stability under low or high temperature
conditions. It can be calculated as follows:
30
𝑇𝐶𝐹 = −𝑇𝐶𝐷 =1
𝑓0
𝑑𝑓
𝑑𝑇=
1
𝑣
𝑑𝑣
𝑑𝑇−
1
𝐿
𝑑𝐿
𝑑𝑇=
1
𝑣
𝑑𝑣
𝑑𝑇− 𝛼, (26)
where 𝐿 is the IDT distance between input and output IDTs, and 𝑇 is the
temperature. The first term of the right side of the equation defines the
variation of the phase velocity with temperature, and the second part is the
expansion coefficient of the material 𝛼.
2.4.2 Some properties of AlN
Even though AlN is used in many other microelectronic applications such as
passivation layers, dielectric films, solar cell coatings, and short wave-length emitters, it
is one of the most promising piezoelectric materials for high frequency SAW devices.
Furthermore, AlN is an ideal piezoelectric layer for diamond-based SAW applications,
due to the good thermal stability, the excellent mechanical and chemical properties, and
the high acoustic wave velocity present in the heterostructure.
Figure 2.3. Wurtzite crystal lattice (the larger, yellow spheres are N ions, whereas the smaller gray
spheres are Al ions).
Stoichiometric AlN is a very stable compound with a strong covalent bonding.
Crystalline AlN occurs in two polytypes: the hexagonal wurtzite (2H) and the cubic
zincblende (3C). They differ in several ways, e.g. the energy bandgap, lower for the
zincblende polytype, and the piezoelectric effect, maximum in the wurtzite structure.
Figure 2.3 shows schematically the AlN wurtzite crystal. This AlN has a direct band
gap of 6.2 eV [Yim.1973], a resistivity around 1016
V.cm, a refractive index of 2.15, a
density of 3.2 g/cm3, a melting point over 2200°C, a high thermal conductivity of 2
31
W/K/cm [Strite.1993], as well as high phase velocities of 5750 m/s [Okano.1993] and
11350 m/s [Dubois.1999] for SAW and BAW, respectively. Moreover, its
electromechanical coupling factor is K2= 0.25%, and the temperature coefficient of
delay, TCD=19 ppm/ºC [Takagaki.2002; Bu.2003]. All these properties turn AlN into a
very interesting candidate for high-power, high-temperature and high frequency
microelectronics [Iriarte.2003b]. In Table 2.1, the phase velocity, K2
and TCD of SAWs
on different piezoelectric material are shown.
Table 2.1. Phase velocity v; electromechanical coupling coefficient K2; and temperature coefficient
delay, TCD, of SAWs in different piezoelectric material. Data are taken from [Campbell.1998;
Emanetoglu.1999; Morgan.2000; Bu.2003; Nishimura.2005; Pedros.2007].
Material Orientation/cut v (m/s) K2 (%) TCD
(ppm/ºC)
SiO2 47.5º Y-X(ST) 3158 0.11 ~0
LiNbO3 Y-Z
128º Y-X
3488
3992
4.5
5.3
94
75
LiTaO3 77.1º Y-Z
112º X-Y
3254
3300
0.72
0.70
35
18
GaAs (001) 2867 0.059 35
GaN (0001) 3693 0.13 28.3
AlN (0001) 5790 0.25 19
ZnO (11-20) 2700 1.1 59-42
4H-SiC (0001)/(1120) 6832 0.011
For applications involving AlN layers, it is important to control the crystallographic
orientation since the piezoelectric properties of highly c-axis oriented films are
approximately the same as those of single crystalline films. It is well known that Al has
a strong chemical affinity to oxygen to the formation of Al2O3. The formation of Al2O3
in the AlN films will retard the single grain columnar growth, and lead to a poor
orientation as well as a poor surface topography. Preferred orientation of the
polycrystalline grains in the AlN films is a critical factor affecting its piezoelectric
properties. Therefore, poor c-axis orientation can lead to reduced piezoelectric
performance.
Crystalline aluminum nitride can be formed in several ways such as chemical
vapor deposition (CVD), physical vapor deposition (PVD), reactive evaporation,
32
molecular beam epitaxy (MBE), etc. The method used for the film deposition in this
thesis is reactive sputtering, with the main objective of achieving high quality
polycrystalline AlN thin films at low temperature.
2.4.3 Diamond as a substrate
Because of exceptional mechanical and chemical properties, diamond has a great
potential to be used as a substrate for the development of high-performance SAW
devices compatible with harsh environments [Werner.2001]. The main disadvantage of
the conventional diamond thin films is the relatively high substrate temperatures (600-
800°C) required in the synthesis process that difficult the diamond-CMOS integration.
Ultra-nano-crystalline diamond (UNCD) is a novel material that can be grown at lower
temperatures, such as 400°C, retaining its main properties comparable to that of single
crystal diamond [George.2009].
CVD diamond is an appropriate material to work in harsh environments
[Werner.2001], and its SAW velocity is approximately 12000 m/s [Assouar.2007], the
largest of all materials. Thanks to AlN high SAW velocity, AlN/diamond structure has
not a high velocity dispersion. Moreover, the temperature coefficient of delay (TCD) of
the AlN/diamond system is between -21 and -9 ppm/ºC [Assouar.2007]. These
properties make CVD diamond combined with AlN an attractive bilayer system for the
manufacture of high frequency SAW sensors. In this thesis work, the SAW devices are
composed by an AlN layer synthesized over micro and nanonocrystalline diamond
substrates, with aluminum paths acting as interdigital transducers (IDT).
33
3. METHODOLOGY
This chapter describes process details regarding the deposition of piezoelectric
AlN thin films, its characterization techniques and SAW devices fabrication. Reactive
sputtering technique, which is used during this thesis work to deposit AlN, is described
here. The methods used for the crystallographic and structural characterization of this
piezoelectric film are also introduced in this chapter. Finally, the methodology to
process one port SAW resonators is described.
3.1 Synthesis of AlN by sputtering
The effect of cathodic sputtering was first described in 1852 by W.R. Grove
[Grove.1852]. Nowadays, the deposition of thin films by sputtering is typical of many
production processes for communications (e.g. SAW and BAW devices), optics (e.g. x-
ray lenses, reflectors), optoelectronics (solar cells, photodiodes, LCD), electronics (e.g.
microchips), surface protection (tools, machine parts) and medical devices (e.g.
implants, anti-rejection coatings) [Wessling.2006; Chan.2012; Rodríguez-
Madrid.2012b; Wang.2013].
Physical sputtering is a process that takes place under high vacuum conditions.
Atoms (or cluster of atoms) are mechanically ejected from a surface (target) as a result
of the bombardment of the latter by heavy energetic particles (ions), generated from a
partially ionized gas (plasma) to finally deposit onto a substrate, see figure 3.1. Most
plasmas are formed by electron impact ionization of the gas in a controlled electrical gas
discharge; when argon is the gas, the ionization is described by:
𝑒− + 𝐴𝑟 → 𝐴𝑟+ + 2𝑒− (27)
Moreover, the flux of particles onto a surface at a given pressure 𝑝 is given by:
Γ =𝑝
2𝜋𝑚𝑘𝑇 [𝑚−2𝑠−1] , (28)
where 𝑚 is the mass of the element, 𝑘 is Boltzmann constant, and 𝑇 is the temperature
in Kelvin.
34
Figure 3.1 Reactive sputtering process.
Typically, in a DC sputtering system the cathode is the target while the grounded
chamber walls act as the anode. When a negative potential is applied to the cathode, the
electric field created between the electrodes accelerates all charged particles,
stochastically presented in the gas in small concentrations; when an electron ionizes a
neutral molecule according to equation (27), an electron/ion par is created. This process
continues in such a way that the plasma is produced. The positive ions are accelerated
by the electric field towards the cathode and attain enough energy to cause both the
physical sputtering and equally importantly the emission of secondary electrons from
the cathode surface. Those electrons continue the ionization process and sustain the gas
discharge. At the beginning, the electric field is to a first order approximation constant
in the space between the two electrodes, but as the plasma is excited and equilibrium
conditions are established, the potential distribution changes dramatically. This
distribution is visualized in Figure 3.2. A surface in contact with the plasma will build
up a negative potential with respect to the plasma. This potential will repel the low
energy electrons such that at equilibrium the ion current is equal to the electron current.
Thus, the plasma always acquires a positive potential with respect to the chamber walls.
This potential is normally referred to as the plasma potential Vp.
35
Figure 3.2. Potential distribution of a DC sputtering system.
A magnetron is normally used in reactive sputtering systems in order to increase
the efficiency of the electron ionization, which means an increase in the plasma density.
The magnetron creates a strong magnetic field close to the target surface. This magnetic
field confines the fast electrons to the “dark space” (see fig. 3.2) by forcing them to
travel along spiral trajectories due to the Lorentz force:
𝐹 = 𝑞(𝑣 × 𝐵) (29)
where 𝑞 is the elementary charge, 𝑣 is the velocity of the particle, and 𝐵 is the magnetic
field. As the energetic electrons are close to the target surface, the ionization efficiency
increases, which in turn results in higher density plasmas. Due to the higher efficient
ionization, the discharge can be sustained at much lower pressures.
In order to allow the sputtering of insulating materials, an alternating power is
often applied to electrodes in a diode discharge. Furthermore, to increase the efficiency
of gas discharges, the so-called pulsed DC discharge is applied. In this case the polarity
of the target is periodically switched from negative to positive for a short time interval.
This is particularly useful during the sputtering of metallic targets in a reactive
atmosphere. The positive pulse is used to discharge the surface of the compounded
target and thus particle generation, which is known as poisoning.
Sputtering of pure metal targets in a reactive gas atmosphere, i.e. reactive
sputtering, is used to form a compound film on the substrate, as seen in figure 3.1.
Considering AlN sputter deposition, the aluminum atoms are sputtered from the Al
target while the nitrogen atoms are introduced into the chamber. Typically a mixture of
36
Ar and N2 is used, because the Ar ions lead to a higher sputtering rate owing to their
larger mass. Reactive sputtering is characterized by its high deposition rate, good
control of the film texture and stoichiometry, very good smoothness, etc. Furthermore,
sputtering deposition is a planar process and thus it could be part of the IC fabrication
[Vossen.1991].
Sputtering depends on some process parameters that offer an enormous
flexibility in order to achieve the required film properties. Nevertheless, this also brings
an extreme complexity of the full set of parameters to achieve a desirable and
reproducible result. The orientation of the deposited sputtered film depends on many
parameters related with the adatom mobility [Engelmark.2002]. In addition, the
optimization of the whole set of parameters is generally system dependent [Naik.1999]
and it requires a deep understanding of the process.
The main parameters for the control of adatom mobility are the substrate
temperature, process pressure, discharge power, and gases flow ratio [Cheng.2003;
Clement.2003; Iriarte.2011].
Substrate temperature plays an important role in the sputtering process and it
can be considered the most relevant parameter for promoting the growth. However, low
temperature means compatibility with CMOS standard procedures, higher throughput,
and thus lower manufacturing costs. In this work, the c-axis oriented AlN is obtained by
optimizing process pressure, discharge power and gas flow ratio at room temperature.
The crystal orientation of the film growth depends also on the kinetic energy that
particles have when they arrive to the surface of the growing film. Process pressure is
the most important parameter in a sputtering system related to this kinetic energy. The
value of the mean free path of a gas molecule at constant temperature is inversely
proportional to the pressure. At low sputtering pressures, the mean free path of the
species increases, a low scattering of particles is produced, and thus the energetic
particles in the plasma can easily transfer their energy to the adatoms of the surface to
promote the oriented growth. However, a bombardment of the growing film with other
ions, such as the fast neutrals, is also observed at low pressures. The fast neutrals are
ions reflected and neutralized upon impact with the target, and they may have sufficient
energy to cause re-sputtering, damage the growing film and reduce the deposition rate
[Jonsson.1999]. In general, the c-axis orientation in the AlN sputtering process is
37
favoured by the increase of kinetic energy transfer at the film surface, and thus by a
decrease of the sputtering pressure.
Concerning the discharge power, it is directly related with the deposition rate.
At higher discharge powers, higher deposition rates are achieved. The latter affects the
growing film in several ways. High deposition rates reduce the impurity incorporation
into the film from background gases. On the other hand, ion/atom flux ratio at the
substrate decreases during ion assisted deposition, resulting in tensile stress and low
textured films. Moreover, for extremely high deposition rates, structural defects and
non-stoichiometric films are obtained. Conversely, low deposition rates results in poor
film quality because of the gas impurities incorporation.
The addition of an inert component to the gas mixture in a reactive sputtering
process is normally used because it offers some advantages. In the particular case of an
AlN sputtering system, Ar is selected as the inert gas in order to increase the sputtering
yield and hence the deposition rate. Moreover, the Ar/N2 flow ratio can be modified in
order to control both, the film stoichiometry and structure. For an increase of the
nitrogen concentration while the sputtering pressure is kept constant, the crystallization
of AlN thin film is improved. Further, the gases flow ratio can be modified in order to
obtain compressive or tensile stressed films [Iriarte.2003d].
The sputtering system process parameters are not the only ones that affect the
crystal quality of the sputtered films. In a piezoelectric/substrate, the thickness of the
piezoelectric thin film is one of the key parameters for the subsequence device
fabrication. The film material quality worsens when the thickness decreases and hence
the device performance also depends on that thickness.
In all cases, the AlN thin film synthesis during this thesis has been done using a
homemade reactive magnetron sputtering deposition system installed in the ISOM
laboratories, which is shown in figure 3.3. This system consists of two chambers, the
preload and the sputtering chamber. The main chamber (left) contains a substrate heater
to control and measure the temperature during the sputtering process. The substrate
heater can handle samples up to 7 cm in diameter and can raise the sample temperature
up to 950ºC. The process gases, nitrogen (N2) and argon (Ar), are introduced by 100 and
200 sccm mass flow controllers, respectively. The Al target, water-cooled on the
backside, is powered by an ENI RPG50 asymmetric bipolar-pulsed-DC generator. The
38
applied frequency has been 250 kHz while the duration of the positive pulse (pulse
width) has been 496 ns (13% of the total signal period), with an amplitude of +40V. The
fact that the system has been dedicated exclusively to the deposition of AlN has
facilitated the synthesis of high purity and highly c-axis oriented material.
Figure 3.3: Home built "reactive sputtering" AlN system.
3.2 Material characterization techniques
There are different analytical techniques that are used in this work to determine
the structural properties and physical features of the AlN thin film. The principal
characterization aspects are its crystallographic orientation, that determines film quality;
and its surface roughness, which determines the crystal orientation; as well as the film
thickness (an important parameter for the orientation of the AlN and to improve the
device performance).
The crystallographic characterization has been principally examined by XRD
using a Phillips X-Pert Pro MRD diffractometer, figure 3.4a, and surface roughness has
been analyzed by AFM (Digital Instruments MMAFM-2), figure 3.4b, and SEM
(Crestec CABL-9500C), figure 3.4c. The thickness of the sputtered AlN has been
characterized by a contact profilometer (KLA-Tencor alpha step iq surface profiler) and
different steps during the process, as surface cleaning before each deposition, has been
check with a Nomarski microscope (Leica Leitz DMRX), figure 3.4d. All these systems
can be found in the ISOM facilities.
39
TEM and PL studies were possible by different collaborations. D. Araujo, M.P.
Villar and R. Garcia from “Departamento de Ciencia de los Materiales e Ingeniería
Metalúrgica” (University of Cádiz) prepared and measured several samples by TEM.
On the other hand, PL was performed by B. Neuschl and K. Thonke from the
Semiconductor Physics group in the Institute of Quantum Matter (University of Ulm).
a) b)
c) d)
Figure 3.4. Equipments used for the AlN characterization installed in ISOM facilities; a) Phillips X-Pert
Pro MRD diffractometer; b) AFM, Digital Instruments MMAFM-2; c) SEM, Crestec CABL-9500C and
d) Nomarski microscope, Leica Leitz DMRX.
3.2.1 X-ray Diffraction (XRD)
The determination of the preferred orientation of the crystallites in a
polycrystalline aggregate is referred to as texture analysis. The term texture is used as a
broad synonym for preferred crystallographic orientation in a polycrystalline material,
normally a single phase. A texture analysis is done by measuring the diffraction
intensity of a given reflection at a large number of different angular orientations of the
40
sample. The intensity of a given reflection (h, k, l) is proportional to the number of h, k,
and l planes in reflecting condition (Bragg‟s law). Hence, the texture analysis gives the
probability of finding a given (h, k, l) plane normal as a function of the specimen
orientation.
The texture analysis is provided by the scan θ/2θ and rocking measurements. The
scan θ/2θ is the most suitable to show the different crystallographic orientations of, in
this case, AlN (see Figure 3.5). In this measurement the incident angle is equal to the
scattered one. The X-ray source and the x-ray detector are coupled during the scan. The
peaks obtained in the XRD diffractogram provide the information needed to know the
sample orientation [Scintag.1999].
Figure 3.5. Different crystallographic orientations for hexagonal AlN.
The degree of texture for a given crystallographic plane is extracted from the
presence of the different peaks and their full width at half maximum (FWHM) values in
the measured rocking curve. This X-ray diffraction analysis is one of the most widely
used techniques in crystallography. In a rocking curve measurement, the sample is
rotated (rocked) through an angular range around the orientation peak obtained in the
scan θ/2θ, bringing the plane in and out of the Bragg condition. The width of the
measured peak contains information of the amount by which the measured plane is off
the surface normal, usually referred to the „„degree of orientation‟‟ of the analyzed
material. This value shows an inverse relationship with the sample quality, and is used
41
in this work to determine the crystal quality of the AlN samples [Iriarte.2010]. The θ/2θ
diffraction pattern of a highly oriented AlN shows the (002) peak at 2θ ≈ 36º, and a
minimum FWHM of the (002) rocking curve centered at θ ≈ 18º.
3.2.2 Photoluminescence (PL)
Photoluminescence is the spontaneous emission of light from a material under
optical excitation. The excitation energy and intensity are chosen to characterize a
variety of material parameters. PL spectroscopy is a contactless, nondestructive method
that provides optical characterization, and it is a selective and extremely sensitive probe
of discrete electronic states. In PL, light is directed onto a sample; the material absorbs
photons (electromagnetic radiation) and the excess energy can be dissipated by the
sample through the emission of light. The energy of this photoluminescence relates to
the difference in energy levels between the two electron states involved in the transition
between the excited state and the equilibrium state. The quantity of the emitted light is
related to the relative contribution of the radiative process. The intensity and spectral
content of this photoluminescence is a direct measure of various important material
properties, i.e. impurity levels and defects, material quality, band gap, etc.
3.2.3 Atomic force microscopy (AFM)
Atomic force microscopy is a very powerful technique in materials science, since
morphological information on the nanometer scale can be obtained relatively fast, due
to the fact that this type of microscope does not requires operation under vacuum.
During this thesis work, surface morphology has been analyzed by this technique,
which relies on Van der Waals forces between a microscope tip in close interaction with
the sample. By knowing these forces we can obtain different kinds of data such as the
surface roughness.
AFM provides a 3D profile of the surface with nanoscale resolution, by
measuring forces between a sharp probe (<10 nm) and surface at very short distance
(0.2-10 nm probe-sample separation). The probe is supported on a flexible piezoelectric
42
cantilever. The AFM tip gently touches the surface and records the small force between
the probe and the surface.
The system is composed by a scanner which emits a laser beam which is
reflected by a photodetector when the cantilever is moved. This optical system converts
the beam movement into a voltage signal that can be measured. The principal operation
modes are the static mode, when the tip is in contact with the surface, also known as
contact mode; and tapping mode, when the cantilever oscillates around its resonance
frequency, and the tip is intermittently in contact with the surface.
In the contact mode, the force between the tip and the surface is kept constant
during scanning by maintaining a constant deflection. In tapping mode, the cantilever
oscillates at a frequency slightly above its resonant frequency, the Van der Waals forces
acts to decrease the amplitude of the oscillation as the tip gets closer to the sample. This
decrease combined with the feedback loop system maintains constant the height of the
cantilever above the sample. An AFM image is therefore produced by imaging the force
of the intermittent contacts of the tip with the sample surface. Thus samples or/and tip
don‟t suffer degradation after taking numerous scans with tapping mode compare to
what is sometimes observed in contact mode. This makes tapping mode preferable for
measuring soft sample, e.g. organic thin film. The materials analyzed by this technique
during this thesis work are considered rigid samples, thus contact mode has been used.
3.2.4 Scanning electron microscopy (SEM)
Scanning electron microscopy produces images of a sample by scanning over its
surface with a focused beam of high energy electrons. The electrons that interact with
the sample produce secondary electrons, back-scattered electrons and characteristic X-
rays that reveal information about the sample including external morphology (texture),
chemical composition, crystalline structure and orientation of materials making up the
sample. This kind of microscope can obtain images of the sample from several microns
down to a few nanometers. This technique has been used to analyze topographically the
AlN films and also to image the finished SAW devices to verify the different fabrication
steps.
43
3.2.5 Transmission electron microscopy (TEM)
Transmission electron microscopy uses a focused beam of high-energy electrons
that is transmitted through an ultra thin specimen, interacting with the specimen as it
passes through. The transmitted beam is then projected onto a screen. The signal
obtained from the interaction of the electrons transmitted through the specimen reveal
information about the sample including external morphology (texture), chemical
composition, and crystalline structure and orientation of materials making up the
sample. The specimens for TEM analysis must be specially prepared to thicknesses
which allow electrons to transmit through the sample. Preparation of TEM specimens is
specific to the material under analysis and the desired information to obtain form the
specimen.
The most common mode of operation for TEM is the bright field imaging mode.
In this mode the contrast is formed directly by occlusion and absorption of electrons in
the sample. Due to the fact that the wavelength of electrons is much smaller than that of
visible light, the optimal resolution attainable for TEM images is several orders of
magnitude better than that of an optical microscope. Besides, diffraction pattern can be
generated by adjusting the magnetic lenses such that the back focal plane of the lens
rather than the imaging plane is placed on the imaging apparatus. This produces an
image that consists of a pattern of dots in for single crystal samples or a series of rings
in the case of polycrystalline or amorphous solid material. This image provides
information about the space group symmetries in the crystal and the crystal orientation.
This technique has been used to image the c-axis oriented AlN in order to verify the
grain columnar growth and orientation and the AlN/diamond interface.
44
3.3 SAW device fabrication
One port SAW resonators are based on the use of SAW reflection gratings to
form resonant structures. Figure 3.6 illustrates the geometric structure of a one-port
SAW resonator. The devices fabricated in this work have the same center frequency for
the reflection gratings and IDT, while the IDT metallization ratio is 0.5. In the one-port
SAW resonators, surface waves emitted from both sides of the excited IDT are
constructively reflected at the center frequency by SAW reflection gratings, which give
rise to SAW standing waves within the IDT.
Figure 3.6. One-port SAW resonator with open-circuit reflection-grating electrode strips.
Once the piezoelectric AlN film was deposited on the diamond substrate, the
process to develop the SAW one port resonators was the following:
- The AlN/diamond was cleaned using a standard RCA procedure
- The photoresist was spinnned on the AlN surface.
- The IDT and reflectors were patterned with e-beam lithography and then
developed.
- A lift-off was done after the Ni/Al metallization.
- The IDT/AlN/diamond structure is cleaned again with a standard RCA
procedure.
- The photoresist is spun on the IDT-reflectors metalized on top of the AlN
surface.
- A lift-off of the PADs was done after the photolithography process and the
Ti /Au metallization.
45
These process steps, including the previous diamond and AlN deposit, are done
under low temperature conditions and could be compatible with CMOS technology,
figure 3.7.
Figure 3.7. Simplified SAW device fabrication process integrated in the CMOS technology.
SAW resonators fabricated in this thesis work has a period of either 400 nm or
800 nm, i.e. 100 nm or 200 nm finger width, respectively. E-beam lithography
technique is then necessary in order to develop these devices. This kind of lithography
offers a better line resolution and cost effective solution for iteration design-test
structure than the conventional optical lithography. However, depending on the
substrate conductivity, this kind of lithography has some disadvantages related to
charge accumulation and dose effects that must be carefully considered. In that way,
different substrates of different conductivity character were used to fabricate IDT with
line widths from 300 nm to 10 nm [Iriarte.2012].
46
The e-beam lithography system used for the development of these nano devices
was the CRESTEC CABL-9500C, installed in the ISOM clean room. This system has a
50 KeV column and resolves a minimum line width of 10 nm with a minimum beam
diameter of 2 nm. The lowest voltage is 5 kV, and the beam current can be varied from
5 pA to 100 nA. The resist, ZEP520 (diluted 1:2), was spinned at 5000 rpm during 1
minute, which resulted in a resist thickness of approximately 150 nm. Then, the resist
was hard baked for 2 minutes at 190ºC. An organic anti-static layer (Espacer 300Z,
Showa Denko) was used to avoid the charge accumulation due to the insulating
character of the AlN/diamond substrate. This layer strongly influences the minimum
resolution when working directly on insulating substrates as demonstrate in
[Iriarte.2012].
ZEP520 resist has a sensitivity of 220 μC/cm2. According to this, the calculated
dose for the field size used (600 μm x 600 μm), voltage and current of e-beam used as
well as for the amount of dots per field used (60000) should be 0.44 μs. The thickness
of the metal of the IDT is limited by the resist thickness. Different metallization were
carried out in order to optimize the lift-off process, the IDT metallization for the final
fabricated devices was Ni/Al (10 nm/40 nm). Nickel (Ni) was used to enhance the
adhesion of the Al layer to the AlN/diamond substrate. This metallization was done by
evaporation in a Varian VT 118, available within the ISOM facilities. After evaporation,
the samples were introduced on a 1-methyl-2-pyrrolidone bath at 85ºC for 10 min and
additional 10 seconds in ultrasonic bath in order to lift-off the resist and the metal on
top of it. The IDT remaining on the substrate were analyzed by SEM and AFM after the
lift-off process, see figure 3.8.
Figure 3.8. (a) SEM and (b) AFM images of a = 800 nm IDT on AlN/diamond.
47
The SAW one port resonators fabricated during this thesis work (following the
process previously explained) consist of IDT (and reflectors) with electrodes widths of
100 and 200 nm fabricated on AlN/MCD and AlN/NCD. In order to optimize the design
of these resonators, a matrix composed of different devices was processed, figure 3.9.
The number of periods were varied from 100 to 200 and the aperture from 6.2 μm to 13
μm while the metallization ratio was 0.5 for all designs and the nominal period was kept
constant at 400 nm or 800 nm i.e. finger width and pitch of 100 nm and 200 nm,
respectively. The design of the final SAW resonator used in this thesis work for both
periods consists of 100 finger pairs for both the IDT and the reflectors and a finger
length of 6.2 μm.
After the fabrication, the high frequency characterization of the present SAW
devices has been accomplished by measuring the S parameters with network analyzer
8720ES (Agilent, USA) connected to a standard probe station such as the Microtech
RF-1 (Cascade, USA). The response obtained thus permits the direct determination of
the center frequency (f0) and the insertion loss.
Figure 3.9. Matrix devices used to optimize the design.
48
4. AlN FILMS: RESULTS AND DISCUSSION
The results obtained during the optimization of the sputtering system in order to
obtain c-axis oriented AlN on different substrates are presented here. Different
substrates have been used throughout this optimization; magnesium oxide, MgO was
used because of its high stability under pressure and high temperature conditions
[Belonoshko.2010]. Alumina, Al2O3, and spinel, MgAl2O4, were used owing to their
high SAW velocity and excellent properties such as good thermal and chemical
stability, high hardness and abrasive resistance [Shih.2010, Didenko.2000]. These
substrates have a similar lattice constant to AlN. Nevertheless, diamond has been the
principal substrate material used for the subsequent SAW device fabrication. Nano and
microcrystalline diamond wafers were provided by O.A. Williams and W. Müller-
Sebert from the “Institut für Angewandte Festköperphysik, Fraunhofer Gesellschaft”
(Germany). The other substrates were commercial samples.
The aim of this chapter is to show the optimization of the deposition conditions
to grow highly c-axis oriented AlN thin films at room temperature on arbitrary
substrates for the subsequent SAW device fabrication. This chapter includes some
results reported in references [Iriarte.2010; Rodríguez-Madrid.2012].
Substrate cleaning is the first step to achieve a good film. All the substrates used
here were cleaned before each deposition using a standard procedure consisting of
rinsing in pirrolidone for 5 minutes at 80ºC, acetone for 5 min at 60ºC, and methanol in
an ultrasonic bath. After this, the sample was loaded into the chamber, and the
sputtering process started when the base pressure was lower than 5 × 10−8 mbar. In
fact, in our growth system base pressures below 10-7
mbar are required in order to
prevent the incorporation of oxygen into the growing AlN thin film. The process gases,
nitrogen (N2) and argon (Ar), were N70 quality or better, and the Al target was 99.999%
pure.
Normally, the process prior to each deposition was the following. The Al target
and the chamber were conditioned for 5 min in pure Ar plasma with a flow of 30 sccm.
An additional 5 min in reactive atmosphere plasma conditioning with the deposition
gases flow were done afterwards. At the beginning of all sputtering deposits the
temperature was 25ºC. During the process, this temperature was increased until 90ºC
49
because of the plasma heating. If not otherwise specified, the substrate was Si (100) and
the optimized parameters used on each experiment series in the following sections were
as shown in Table 4.1 [Iriarte.2010; Rodríguez-Madrid.2012].
Table 4.1. Processing parameters for sputtering AlN on diamond substrates [Rodríguez-Madrid. 2012].
Parameter Value
Power [W] 700
Base pressure [mbar] <5x10-8
Gas composition Ar/N2 [sccm] 3/9
Process pressure [mTorr] 3
Substrate temperature [ºC] 25
Target substrate distance [mm] 45
Target diameter [mm] 101.6
Target thickness [mm] 6.35
Sputtering time [min] 15
The most critical parameters normally considered to play a major role in the
degree of c-axis orientation of AlN thin films deposited by reactive sputtering, such as
target-substrate distance, flow gas ratio, discharge power and process pressure were
independently varied. The results obtained on each parameter study are presented in the
following sections.
4.1 Substrate
The good lattice match between the substrate and the layer is an important
crystal growth promoting factor. AlN has a large lattice mismatch with most of the
substrates; one of the exceptions is SiC whose lattice mismatch with AlN is 1%
[Chung.2009]. On the contrary, the lattice mismatch between AlN (0002) and Si (100)
is 42.7% [Zhang.2005]. In addition to high defect density and residual stress, a poor c-
axis orientation may also arise from the large difference in lattice parameters.
The lattice mismatch is not the main substrate characteristic influencing the
growth of AlN thin films by sputtering. In fact, substrate roughness plays an important
role. Therefore, prior to each deposition, substrate roughness of the different materials
used to deposit AlN have been analysed by AFM. Figure 4.1 shows an example of
these measurements for spinel substrate. A 3D and height AFM micrographs (3 μm
50
size) of a MgAl2O4 is shown. The surface roughness obtained was of 3.26 nm rms and
the data scale was 20 nm for the 3D image.
Figure 4.1. 3D and height AFM micrographs of MgAl2O4. The roughness is of 3.26 nm rms, and data
scale for the 3D image is 20 nm.
Figure 4.2 shows the rocking curve FWHM values of c-axis oriented AlN thin
films deposited on substrates of different nature such as Si, glass, Al2O3, MgO,
MgAl2O4 and micro and nanocrystalline diamond, showing variable surface roughness
as well as different crystallographic properties. It can be observed that highly c-axis
oriented AlN thin films can be grown on all these substrates as long as the substrate
surface roughness is smooth enough, roughly below 4 nm root mean square (rms),
independently of the lattice mismatch between substrate and deposited layer.
51
0 5 100
10
20
30
FW
HM
00
2-r
oc
kin
g c
urv
e (
º)
Roughness (nm RMS)
Silicon
Glass
Al2O3
MgO
MgAl2O4
MCD
NCD1
NCD2
Figure 4.2 Rocking curve FWHM values of c-axis oriented AlN thin films deposited on substrates of
different material and surfaces roughness.
During this thesis work principally two kinds of diamond samples have been
used MCD and NCD. MCD samples are defined as films for which the grain size is in
the range of several micrometers to slightly less than a micrometer. NCD films are
defined here as films for which the grain sizes are typically measured in several tens to
hundreds of nanometers. The principal disadvantage of these synthetic diamonds
substrates is its high surface roughness. Several methods to overcome this problem
[Malshe.1999] have been investigated. Mechanical polishing of the diamond surface
has been used for MCD in order to solve it [Fujii.1997; Nakahata.2003]. However, this
procedure is expensive, time-consuming and not yet optimized for NCD. Alternative
methods consist of using the smooth unpolished nucleation side of freestanding
diamond layers [Mortet.2002; Elmazria.2003; Mortet.2003]. This solution requires the
whole Si substrate to be sacrificially etched. At least 100 µm thick films are required for
structural stability, and thus this method is expensive, as the growth of a free-standing
diamond film is a long duration process, approximately 1 µm/hour. In this work we
propose an alternative “free-standing” diamond structure, see Figure 4.3; an alumina
substrate was used to get a more robust structure so the thickness requirement was
alleviated; in this case, 20 µm thick diamond films were employed. Throughout this
work we refer to this NCD glued on top of an Alumina substrate as free-standing NCD.
52
Figure 4.3. Schematic process followed to obtain the free-standing diamond to use the unpolished nucleation side
to deposit AlN.
Figure 4.4 shows the AFM analysis of both MCD and NCD substrates. The
average results for polished MCD substrates were 3.7 nm rms. In the figure, the
difference between surface roughness for the as grown NCD, 262.4 nm rms, and for the
nucleation side of the free-standing NCD, 2.1 nm rms, can also be observed. For the
lowest roughness values, polished MCD and free-standing NCD, the AlN thin film
could be deposited with a good orientation on the c-axis.
Figure 4.4. Height AFM micrographs (20 µm size) of polished microcrystalline diamond. The roughness
is 3.7 nm rms, top; rough nanocrystalline diamond, the roughness is 262.4 nm rms, middle; and
nucleation side of free-standing nanocrystalline diamond, the roughness is 2.1 nm rms), bottom.
53
For polycrystalline substrates materials, the surface roughness is not the only
aspect that has influence on the degree of c-axis AlN orientation that can be obtained.
Grain size plays an important role on that influence for polycrystalline substrates.
Figure 4.5 visualizes the impact of the substrate grain size on the growth of AlN thin
films as measured by analyzing the rocking curve FWHM values of the (0002)-AlN
peak deposited on three diamond substrates: a MCD substrate with a surface roughness
of 3.70 nm rms, and three NCD substrates with surface roughness of 2.10 nm rms, 5.85
nm rms and 10.70 nm rms, respectively. Clearly, the AlN thin film grows with a much
better degree of c-axis orientation when the substrate has a microcrystalline grain size,
even if its surface roughness is higher than the one for the smoothest NCD. A similar
FWHM of the 0002-AlN rocking curve was obtained for both rough nanocrystalline
diamonds. This grain size of the substrate seems to be a better “buffer” for the growth of
polycrystalline AlN thin films with grain sizes also in the micrometer range. As smaller
is the grain size of polycrystalline substrate, much smoother surfaces are required in
order to obtain c-axis oriented AlN thin films. The graph shows also how increasing the
discharge power enhance the growth in the (0002)-direction for both micro and
nanocrystalline substrates.
400 500 600 7000
5
10
15
20
25
30
35
FW
HM
00
2-r
oc
kin
g c
urv
e (
º)
Power (W)
NCD 5.85 nm rms
NCD 10.77 nm rms
NCD 2.10 nm rms
MCD 3.70 nm rms
Figure 4.5. Influence of the substrate grain size on the c-axis orientation of AlN thin films on different
diamond substrates with different grain sized, microcrystalline (MCD 3.70 rms) and nanocrystalline
(NCD 2.10 nm rms, NCD 5.85 nm rms and NCD 10.70 nm rms).
54
4.2 Target-substrate distance
The effect of the distance between target and substrate on the AlN growth was
also analyzed. Figure 4.6 shows the rocking curve FWHM values of the (0002)-peak of
AlN thin films deposited on Si using two different target-substrate distances, namely 30,
45 and 70 mm, varying the discharge power from 400 to 700 W. For a low discharge
power, a slight improvement is obtained by increasing the target-substrate distance to 70
mm. However, better results are obtained at shorter distances when using higher target
powers. These results, as well as the amount of data available in the literature showing
highly textured AlN thin films at a multitude of target-substrate distance values
[Chen.2005; You.2007; Kar.2008], lead us to the conclusion that this parameter by itself
does not play a significant role in the optimization of the quality of the AlN thin films,
i.e., highly oriented thin films can be obtained in a wide range of target-substrate
distances going from 30 mm to 70 mm and likely above. If not otherwise specified, the
target-substrate distance uses from now on experiments will be kept at 45 mm.
400 500 600 700
2
3
4
5
FW
HM
00
2-r
ockin
g c
urv
e (
º)
Power (W)
30 mm
45 mm
70 mm
Figure 4.6. Rocking curve values of AlN deposited on Si by reactive sputtering at three different target-
substrate distances (30, 45 and 70 mm) for different discharge powers (400-700 W).
55
4.3 Gas flow ratio (Ar/N2)
Higher nitrogen concentrations likely correspond to AlN films showing higher
crystallinity and refractive index, because of less nitrogen deficiency and improved
stoichiometry. So, nitrogen partial flow rate could determine the preferred orientation of
deposited films. Different films were grown under excess of Ar and N2 varying the
discharge power. Figure 4.7 shows how an N2 excess in the gas flow enhances the
FWHM of the 0002-AlN rocking curve, and thus the c-axis growth orientation.
For a deep understanding of how this parameter affects the growth of AlN,
different experiments were done under the optimized parameters (see Table 4.1) on
silicon wafers varying the Ar/N2 ratio. As shown in figure 4.8, the synthesis of AlN
films in the (0002) direction is possible for the whole range of ratios studied here.
Moreover, the amount of Ar enhances a faster processing and thus the deposition rate
increases with higher ratios. According to these results, an Ar/N2 ratio of 2/3 is
recommended; however, there is a trade-off between high deposition rates and film
quality, as structural defects and nonstoichiometric films are likely to be obtained with
extremely high deposition rates. For that reason the optimized Ar/N2 ratio is 1/3 for
most of the analyzed substrates and the other process parameters.
300 400 500 600 700
5
10
15
FW
HM
00
2-r
ockin
g c
urv
e (
º)
Power (W)
ArExcess
N2Excess
Figure 4.7. Rocking curve FWHM values of highly oriented AlN films deposited under different gas
flow regimes [Iriarte.2010] .
56
0/3 1/3 2/3 3/3
1,0
1,2
1,4
1,6
1,8
2,0
50
60
70
FWHM
FW
HM
00
2-r
ockin
g c
urv
e (
º)
Ar/N2 ratio
Rate
Ra
te (
nm
/min
)
Figure 4.8. Rocking curve FWHM values and deposition rate of sputtered AlN films as a function of the
Ar/N2 ratio [Iriarte.2010].
4.4 Discharge power
Theoretically, an increase of the discharge power gives rise to an increase of the
crystal quality and higher deposition rates. The effect of the discharge power on the
orientation of the growing films is shown in figure 4.9. The applied power was in the
range 350-900 W. For values below 300 W the AlN thin films deposited do not exhibit
any crystalline orientation, even if the process pressure is low. The process is likely to
be outside the desired energy window to promote crystal growth.
300 400 500 600 700 800 900
2
3
4
5
0
20
40
60
80
FW
HM
002-r
ockin
g c
urv
e (
º)
Power (W)
FWHM
Rate
(nm
/min
)
Rate
Figure 4.9. Rocking curve FWHM values and deposition rate of highly oriented AlN films as a function
of the target power [Iriarte.2010].
57
The higher deposition rates were obtained for higher powers, as expected. There
is the same trade-off as the observed for the gas flow ratio. An adequate power value is
700 W, that assures a highly c-axis orientation of the thin AlN taken in account the
other process parameters.
4.5 Process pressure
Together with the discharge power, the other most influencing parameter on the
degree of c-axis orientation of the AlN thin films is the process pressure. An increase of
the pressure affects negatively the preferential growth in the (0002) direction, although
it is possible to obtain highly oriented films even at pressures as high as 8 mTorr, as
shown in figure 4.10. Nevertheless, it can be stated that a low pressure promotes the
growth of the AlN thin film in the c-axis orientation; although a pressure below 2 mTorr
may cause instability in the plasma and introduce a high amount of stress at the
film/substrate interface [Dubois.2001; Rajamani.2001; Iriarte.2003d].
2 4 6 8
1,6
1,8
2,0
2,2
2,4
0
20
40
60
80
FW
HM
002-r
ockin
g c
urv
e (
º)
Pressure (mTorr)
FWHM R
ate
(nm
/min
)
Rate
Figure 4.10. Rocking curve FWHM values and deposition rate of highly oriented AlN films as a function
of the process pressure.
To optimize the two most influential parameters on the degree of the c-axis
orientation of the AlN thin films, a number of films grown at different discharge power
and pressures were done. Figure 4.11 shows the FWHM of the AlN (0002) rocking
58
curve for three pressures varying the discharge power from 300 W up to 700 W. The
results show a trend in which higher power promotes the growth of the AlN thin films
in the (0002) direction. The optimized value of pressure is 3 mTorr. The sputtering
chamber had to be opened to be repaired. The different values of the FWHM of the
rocking curve under pressure of 2 mTorr in figure 4.10 and the one obtained for the
same pressure at 700 W in figure 4.11 are due to a change in the sputtering chamber
conditions because of this opening.
300 400 500 600 700
3
6
9
12
15
FW
HM
002-r
ockin
g c
urv
e (
º)
Power (W)
2mTorr
2.5mTorr
3mTorr
Figure 4.11. Rocking curve FWHM of (0002)-AlN peak values of thin films deposited under different
discharge powers (300-700 W) and process pressure (2, 2.5 and 3 mTorr) [Iriarte.2010].
4.6 Film thickness
The AlN layer thickness influences the FWHM value of the (0002)-peak in the
rocking curve. As shown in figure 4.12, the degree of c-axis orientation depends on the
film thickness up to a value of approximately 1 μm. Once this value is reached, there is
no influence on the layer thickness on the FWHM value obtained in our X-ray
diffraction analysis. A similar effect has also been observed in [Saravanan.2005]. It
should be noticed that this conclusion is valid for three substrates belonging to the three
different crystallographic structures: monocrystalline (silicon), polycrystalline
(synthetic diamond) and amorphous (glass).
59
A deep study of the AlN/diamond structure has been done because of the
importance of this structure for the objectives of this thesis regarding SAW devices. In
the AlN/diamond structure the thickness of the piezoelectric AlN must be optimized
[Benetti.2005], not only in terms of its crystal quality, but also to maximize the
penetration of the acoustic waves into the fast propagation substrate. After the
optimization of the sputtering parameters in order to obtain c-axis AlN films, AlN films
from approximately 150 nm to 1.2 μm on MCD and free-standing NCD substrates have
been deposited under the conditions shown in Table 4.1.
0 1000 2000 3000
1,5
3,0
4,5
6,0
FW
HM
00
2-r
ockin
g c
urv
e (
º)
Film thickness (nm)
Glass
Silicon
Diamond
Figure 4.12. Rocking curve FWHM of (0002)-AlN peak values of AlN layers as a function of the AlN
film thickness, for three substrates with different crystalline structures [Iriarte.2010].
Figure 4.13 shows the XRD spectra in the theta/2theta scan mode of AlN films
with different thickness deposited on polished MCD. It can be seen that the AlN films
are highly oriented on the c-axis, (0002) direction, perpendicular to the surface, and that
no other diffraction peaks are observed. The intensity of that peak is reduced and its full
width at half maximum (FWHM) is increased when reducing the deposited film
thickness.
60
35 40 45 50
2
3
4
5
6
Inte
ns
ity
(a
.u.)
(º)
1200 nm
600 nm
300 nm
150 nm
AlN (0002)
Diamond (111)
Figure 4.13. Influence of the AlN layer thickness deposited on MCD on the /2 XRD spectra.
In order to estimate the crystal quality of these AlN films, the rocking curve on
the (0002) direction was analyzed, see figure 4.14. The FWHM corresponding to the
(0002)-peak rocking curve is reduced for thicker AlN layers down to 2º for layers
thicker than 1 µm. These results are comparable with those found in the literature
[Iriarte.2003b; Chou.2006; Iriarte.2010]. Therefore, the crystal quality for the sputtered
AlN films on diamond improves as their thickness increases, up to a saturation value.
The same experiments were done for free-standing diamond, and similar results
were obtained. In figure 4.15 it can be observed that the intensity of the peak is reduced
when reducing the deposited film thickness, as it happens for the AlN deposited on
MCD. The AlN layers grow along the (0002) direction at 36.16º but a peak at 35.3º is
also observed. The incorporation of carbon in the interlayer could explain this peak as
shown in [Okamoto.2001]. For all these X-ray diffraction studies the sample was only
mechanically aligned because of the polycrystalline diamond substrate; this is the
reason why the amplitude of the peaks in theta/2theta graphs (fig. 4.13 and 4.15) is not
related with the AlN thickness and the displacement of the peaks on the rocking curves
(fig. 4.14).
Figure 4.16 shows the influence of the AlN layer thickness on the FWHM of the
0002-roking curve for the films deposited on free-standing NCD and MCD. The width
61
and thus the quality film is comparable for both substrates, though a bit larger for NCD
than for MCD.
5 10 15 20 25
3
4
5
6
1200 nm
600 nm
300 nm
150 nm
In
ten
sit
y (
a.u
.)
(º)
Figure 4.14. Influence of the AlN layer thickness deposited on MCD on the 0002-rocking curves.
35 40 45 50
3
4
AlN (101)
Inte
ns
ity
(c
ou
nts
)
2(º)
1200 nm
600 nm
300 nm
150 nm
AlN (0002)
Diamond (111)
Figure 4.15. Influence of the AlN layer thickness deposited on free-standing NCD on the /2 XRD
spectra [Rodríguez-Madrid.2013].
62
0 300 600 900 12002
3
4
5
6
7
8
9
10
FW
HM
00
02
-ro
ck
ing
cu
rve
(º)
AlN thickness (nm)
AlN on free standing NCD
AlN on polished MCD
Figure 4.16. Influence of the AlN layer thickness on the FWHM 0002-rocking curve for both substrates,
free-standing NCD and MCD.
TEM analysis has been done to confirm these results and to understand them
from a structural point of view. Figure 4.17 shows the structure and the columnar
growth of the AlN on the diamond substrate observed by TEM.
Figure 4.17. Structure and columnar growth of the AlN film deposited on MCD.
Figure 4.18 shows the three regions of the AlN layer grown directly on the
diamond substrate recorded in bright field (BF) conditions. The respective insets
correspond to the electron diffraction of the regions indicated by the dashed circles, for
which the micrographs were recorded using the selective area aperture. The aperture of
location 1, 2 and 3 is located at the bottom, the center and the top of the AlN layer,
respectively. The corresponding selected area electron diffraction (SAED) patterns
63
show that the spots gradually change from circular shape to a well defined spot. This
indicates that the grain disorientation is reduced with the layer thickness. Near the
interface, the grains have an estimated misorientation of around 20º, while in the center
of the layer this value is reduced to 6º, and at the top of the layer grains are almost
perfectly oriented.
Figure 4.18. Micrographs of the AlN layer recorded in BF conditions. Columnar grains above the mid-
layer appear, while in the first steps the equiaxial growth dominates. The orientation of the grains is
evidenced by the SAED recorded at the dashed circle locations. The elongated spots (location 1 and 2)
result from the addition of small spots each one corresponding to a grain [Rodríguez-Madrid.2012].
A more accurate analysis of the SAED of location 3, figure 4.19, shows two
grain families with different orientations that correspond to two superimposed patterns:
(i) one spot system corresponding to the [2110] pole (i.e. crystal oriented along this
direction), (ii) and the other parallel to the [0110] direction, as shown in figure 4.18 by
the rectangles in the inset of location 3. The orientation of these two patterns
corresponding to the two columnar grain families has the same c-orientation, i.e. the
grains are rotated around this c-axis.
64
Figure 4.19 SAED recorded at the top of the AlN layer (location 3). Two grain families with different
orientations are detected: (a) with the [2110] direction parallel to the e-beam direction (i.e. crystal
oriented on the [2110] pole), (b) with the [0110] direction parallel to the e-beam direction (i.e. crystal
oriented on the [0110] pole).
These results agree with those obtained by X-ray diffraction, figure 4.20. For
thinner AlN layers, a larger FWHM of the 0002-rocking curve is observed, in other
words, the grain misorientation increases.
In addition to the XRD and TEM analysis, a photoluminescence study of the
samples grown on MCD substrates with different thickness was done. The spectra are
shown in Figure 4.21. The spectrum at the very bottom was taken directly from the
MCD substrate sample, without AlN. At least two broad emission bands between 250
and 400 nm are observed, probably related to certain defects in diamond, since the
emission is hampered by the AlN on top of the other layers.
65
Figure 4.20. Comparison between FWHM of the AlN (0002) rocking curve and the misorientation
analyzed by TEM.
The total intensity for all the AlN samples is quite weak. For the thinnest AlN
layer (150 nm), the spectrum exhibits an emission band at about 300 nm. This emission
is related to carbon, likely incorporated at the beginning of growth of the AlN from the
diamond substrate [Collazo.2012]. For the thicker AlN layers this emission is not
observed anymore, also suggesting that carbon incorporation takes place in the
AlN/diamond interlayer, during the first steps of the AlN sputtering. these results are in
agreement with X-ray diffraction study observed before. The 35.3º peak shown in the
θ/2 θ graph (figure 4.15) correspond to the carbon incorporation.
With the exception of this carbon-related band, all the 150 nm-, 300 nm- and
600 nm-thick layers of AlN do not show any other emission band, most likely because
of defects in the polycrystal which cause nonradiative processes. For the thickest layer,
a broad band around 440 nm (2.78 eV) is observed, perhaps due to aluminum vacancies
which cause band-to-impurity transitions, in agreement with [Sedhain.2009].
66
Figure 4.21. Low temperature photoluminescence for different thickness of AlN films sputtered on MCD
substrate.
4.7 Summary and conclusions
The sputtering parameters that affect the growth of AlN are related to the
substrate, the growth conditions, and/or the film thickness.
Concerning to the substrate characteristics, a good lattice match between the
substrate and the deposit layer is not strictly necessary to achieve highly c-axis oriented
polycrystalline AlN thin films on substrates of different nature (different material as
well as different crystallographic properties). Otherwise, the substrate surface roughness
and grain size are important parameters in order to obtain the AlN in the (0002)
direction. For the analyzed substrates, the surface roughness should be kept below 4 nm
rms to guarantee the desired orientation of the film.
Regarding the system parameters involved during the sputtering process, target
to substrate distance is not a critical parameter. AlN oriented on the c-axis can be
obtained for different distances. Gas flow ratio is neither an important process
parameter when the stoichiometry is guaranteed, (0002) oriented AlN can be obtained
for different Ar/N2 ratios. The optimized value for this parameter is 3/9 sccm for Ar/N2
respectively. On the other hand, process pressure and discharge power are crucial
67
parameters promoting the c-axis orientation of the layers. The optimized values are 3
mTorr and 700 W, respectively.
The principal aim of this thesis work is to process high frequency SAW devices,
in that sense, different thin film thickness are necessary. AlN films from 150 nm to
1200 nm were deposited in order to study the influence of the thickness on the
piezoelectric film orientation. Using the optimized sputtering parameters, highly c-axis
AlN were obtained for all the thicknesses on different substrates. The FWHM of the
002-rocking curve increases when the thickness decreases.
A special study of the thickness influence was done for MCD and NCD
substrates, which are chosen for the high frequency devices. The deposition of AlN by
sputtering was optimized for different film thicknesses. As well as for other substrates
analyzed in this work, when the thickness increases, the orientation on the (0002)
direction improves, as proved by X-ray diffraction and TEM analysis. Grown on
diamond substrates at room temperature, a very high degree of c-axis orientation of the
AlN films is observed, even for film thickness as narrow as 150 nm.
68
5. AlN THIN FILM SAW RESONATORS
This chapter presents the results obtained of SAW resonators processed on
different diamond substrates, namely, AlN/MCD and AlN/free standing NCD. The
combination of both nanolithography and high speed substrates, allows us to fabricate
SAW devices working at very high frequencies [Benetti.2005; Hashimoto.2013] in fact,
at frequencies not reached so far for the fundamental harmonics. In this chapter the
thickness influence on the device performance is also studied. This chapter includes
results already published in the references [Iriarte.2012; Rodríguez-Madrid.2012b;
Rodríguez-Madrid.2013].
5.1 Influence of the layer thickness on device performance
In the design of SAW devices on AlN/diamond systems, the thickness of the
piezoelectric layer is a key parameter. Its influence on the SAW device response
(frequency, quality factor, etc.) is presented in this section.
In AlN/diamond system, the electromechanical coupling coefficient (K2) and the
propagation velocity v of the acoustic waves depend on the film-thickness-to-
wavelength ratio (H/λ). At small H/λ, the SAW extends far into the underlying solid,
and v approaches the large velocity value of the bare substrate, though K2 is reduced due
to the non-piezoelectricity of diamond. Conversely, at large H/λ, the SAW is mostly
concentrated in the film region, and thus K2 is large but v is limited by the smaller
velocity value of the AlN overlayer. In addition to the intrinsic losses imposed by the
dispersion of the material system, the crystal quality and orientation of sputtered AlN
thin films, and thus their piezoelectricity, is limited by the film thickness [Naik.1999;
Iriarte.2010]. Therefore, the design of high-frequency and low-loss AlN/diamond based
SAW one-port resonators requires careful selection of the film thickness, which should
be a trade-off among the different aspects mentioned so far.
In order to study the influence of the piezoelectric film thickness on the SAW
response, one-port SAW resonators with a period of 800 nm were manufactured using
the AlN films on MCD studied in section 4.6, with thicknesses of 150 nm, 300 nm, 600
nm, and 1200 nm (kH=2πH/λ= 1.18, 2.36, 4.71, and 9.42, respectively). Thus, for an
69
optimal design structure (IDT/piezoelectric-film/diamond substrate), the kH providing a
mode with the best combination of v and K2 is selected. This choice implies a
compromise between the frequency and the intensity of the targeted resonances, i.e., a
larger frequency can be achieved at the cost of larger losses or vice versa due to the K2
behavior. Once kH is chosen, and since f= v/λ, the smallest λ that can be reliably
patterned is selected. This requires also a film thickness large enough to ensure a good
crystallographic quality. Since thickness is a limiting factor, as shown in chapter 4, a
compromise should be made again between achievable frequency and intensity (loss) of
the targeted resonances.
Rayleigh and confined (Sezawa) modes are presented in similar structures as the
one used in this thesis due to the higher v of the substrate in comparison with the v of
the piezoelectric film [Takagaki.2004; Pedros.2005]. Figure 5.1 shows the reflection
spectra S11 of the fabricated devices.
For the SAW resonators on the 150, 300, and 600 nm-thick AlN films, the main
resonance corresponds to first order Sezawa mode (labeled as S1), observed at 14.0, 12.6
and 10.9 GHz respectively. For the SAW device on the 1200 nm-thick film S1 appears at
9.9 GHz, whereas the main resonance at 11.5 GHz corresponds to the second Sezawa
mode (S2). The frequencies achieved in these devices are higher than the largest value of
5.4 GHz obtained for this material system so far [Benetti.2005; Fujii.2012] In addition,
the quality factor Q and the out-of-band rejection for the S1 mode in the 300 nm-thick
film are 12560 and 36.5 dB, respectively, values among the largest reported for a SAW
resonator [Benetti.2005; Hashimoto.2013].
The identification of the acoustic modes reported above is based on the
numerical calculation of the velocity dispersion using a method based on a Green‟s
function formalism [Every.2002], which was developed to calculate surface Brillouin
spectra. In addition, the theoretical values of K2 were determined using the transfer
matrix approach [Fahmy.1973].
70
Figure 5.1. Reflection coefficient (S11) spectra for identical one-port SAW resonators with = 800 nm on
(a) 1200 nm-, (b) 600 nm-, (c) 300 nm-, and (d) 150 nm-thick AlN films deposited on MCD. R and Si
denote the Rayleigh and Sezawa modes, respectively [Rodríguez-Madrid.2012b].
The material parameters used, shown in table 5.1, are extracted from the
literature [Tsubouchi.1981; Benetti.2005].
Table 5.1. Elastic (c), piezoelectric (e), dielectric (ε) and density (ρ) material constant for AlN
(hexagonal) and poly-diamond, used to calculate dispersion curves and K2 [Tsubouchi.1981;
Benetti.2005].
c11,c22 c33 c12
(GPa)
c13
(GPa)
c44 ,c55 c66 ε11,ε22
(pF/m)
ε33 e15, e24 e33
(C/m2)
e31, e32 ρ(kg/m3)
AlN 345 395 125 120 118 110 80 95 -0.480 1.550 -0.580 3260
diamond 1153.12 86.44 533.34 55 - 0 - 3512
71
Figure 5.2 shows the calculated dispersion of the SAWs and pseudo-SAWs
propagating in the AlN(0002)/diamond(111) heterostructure. The spectrum of SAWs
comprises the Rayleigh (R) and the Sezawa (Si) modes that propagate with velocities
below that of the transverse bulk wave in the substrate [Pedros.2005]. The greyscale
represents the magnitude of the shear vertical component of the waves. In this scale,
brighter areas indicate larger values. The experimental data, obtained from the
resonance frequencies in figure 5.1, are indicated by white solid stars at each kH (the
experimental data obtained for devices fabricated on free-standing NCD substrates and
devices with a period of 400 nm are also indicated by white solid squares and circles
respectively).
Figure 5.2. Dispersion of the velocity of the Rayleigh (R), Sezawa (Si), and pseudo-bulk (PBi) modes on
AlN/diamond. The greyscale traces correspond to the calculated modes, whereas the experimental data
for devices with 800 nm period fabricated on AlN/MCD are represented by white solid stars, for devices
with 800 nm period fabricated on AlN/“free-standing” NCD are represented by white solid squares and
for devices with 400 nm period fabricated on AlN/MCD are represented by white solid circles.
The velocity of the modes decreases from kH=0 (diamond) to kH=∞ (AlN),
indicating that the thickness of the piezoelectric film needs to be reduced in order to
obtain higher frequencies. However, the dispersion of K2 is not a monothonic function
72
of kH, as shown elsewhere [Iriarte.2003b; Benetti.2005]. Figure 5.3 shows the
calculated K2 for the different kH studied in this section. The best K
2 value for the S1
mode (1.21 %) was obtained for the 300 nm-thick AlN film (kH=2.36), in agreement
with previous calculations [Iriarte.2003b; Benetti.2005].
From the point of view of the film quality, the degree of c-axis orientation of the
AlN film (and thus its piezoelectricity) decreases as its thickness is reduced, as seen in
figure 4.16. According to these and previous [Tonisch.2006] results on sputtered AlN,
the FWHM of the (0002)-rocking curve must typically be smaller than 4º in order to
yield a strong piezoelectric response. Therefore, the compromise between selecting the
right thickness to increase the frequency, and the requirement to maintain the
crystallographic quality of the piezoelectric film, is evident with the experimental data
seen in this section. For the case studied here, AlN/MCD, the 300 nm-thick AlN film
(kH=2.36) corresponds to the best case, where a very high frequency is achieved while
keeping a maximum K2 for the S1 mode and a FWHM of the AlN layer around 4º, both
conditions ensuring low loss.
Figure 5.3. Calculated electromechanical coupling coefficient (K2) for the Rayleigh (R) and Sezawa (Si)
modes as a function of the normalized piezoelectric film thickness on the AlN/diamond structure (kH).
The wavelength is 800 nm in all the cases [Rodríguez-Madrid.2012b].
In order to increase the frequency of these devices, the IDT period λ was
reduced. Thus, using 100 nm instead of 200 nm for the finger width, the wavelength
73
was shortened from λ = 800 nm to λ = 400 nm. These one-port SAW resonators were
manufactured using the same AlN/MCD structures studied before with a thickness of
150 nm, 300 nm, 600 nm, and 1200 nm (kH=2πH/λ= 2.36, 4.71, 9.42, and 18.84,
respectively).
Figure 5.4 shows the reflection coefficient of these devices whose main
resonance at frequencies of 17.53 GHz, 19.37 GHz, 20.13 GHz and 22.78 GHz
correspond to first Sezawa mode for 1200 nm and 600 nm 300 nm and 150 nm-thick
AlN layer respectively. Other important resonances are observed, see figure 5.4. The
influence of the diamond substrate on the device response can be clearly observed for
this case. The wave penetrates less on the substrate because of the IDT period is half the
period of the devices analyzed so far. The devices fabricated on the 1200 nm- and 600
nm-thick AlN/MCD structures show similar frequencies due to this fact.
-12
-8
-4
-12
-8
-4
-12
-8
-4
16 20 24 28-12
-8
-4
S4
S3
S2S
1
c)
b)
a)
d)
600 nm
R
S2
S1R
S3S
2
S1
S3
S1
S2
1200 nm
R
300 nm
S1
1 (
dB
)
Frequency (GHz)
150 nm
Figure 5.4. Reflection coefficient (S11) spectra for identical (= 400 nm) one-port SAW resonators on (a)
1200 nm, (b) 600 nm, (c) 300 nm, and (d) 150 nm-thick AlN films on MCD. R and Si denote the
Rayleigh and Sezawa modes, respectively.
74
5.2 Extraction of coupling-of-mode parameters
In this section coupling-of-mode COM parameters of an IDT/AlN/diamond
structure were extracted by finite element method (FEM) and by data fitting.
The requirements of low-loss and high frequency are more and more stringent
for SAW devices, as discussed in chapters 1 and 2. Precise extraction of coupling-of-
mode (COM) parameters plays a key role in the design for high performance surface
acoustic wave (SAW) filters. A COM model allows to calculate the admittance of the
device, while providing an accurate description of multiple reflections on the counter
propagating waves. Compared with other models for the design and analysis of SAW
devices, the COM model provides an efficient and highly flexible approach for
modeling various types of electrodes, and therefore, this method has been frequently
used in the design of low-loss filters [Chen.1985; Hashimoto.2000; Plessky.2000].
Moreover, the finite element method (FEM) is commonly employed to evaluate
the reflectivity phenomena affected by the thickness and other material properties of the
electrode as well as its metallization ratio. This method has also been applied to
calculate the admittance responses, the effects of electrodes on phase velocity, coupling
coefficient, and reflectivity in similar structures as the one evaluated in this thesis work
[Iriarte.2003c; Brizoual.2007; Ro.2010].
In collaboration with Professor Ruyen Ro from the Department of Electrical
Engineering, I-Shou University in Taiwan, the FEM is used to extract COM parameters
for an interdigital transducer (IDT)/AlN/MCD structure. These parameters are
compared with the ones obtained with the analytic solution and the extracted by a fitting
procedure.
Figure 5.5 shows the frequency response obtained for the device processed on
600 nm thick AlN layer deposited on MCD, with an IDT wavelength of λ = 800 nm.
Three resonant peaks can be observed in this response. The first peak at a frequency of
around 11 GHz corresponds to the first Sezawa SAW mode, travelling at v around 8809
m/s, which is higher than the sound velocity for the AlN due to wave propagation
through the substrate. The other peaks, at 13 GHz and 14 GHz, correspond to the
second and third order Sezawa SAW modes.
75
The FEM was employed in order to evaluate COM parameters such as phase
velocity, coupling coefficient and reflectivity. For each propagating mode, a resonance
frequency can be observed using the FEM when ignoring the effects of the electrodes.
Also, it is assumed the IDT interface to be an electrically free or shorted surface
[Ro.2010].
10 12 14 16 18
-20
-15
-10
-5
0
S3
S1
S1
1 (
dB
)
Frequency (GHz)
S2
Figure 5.5. The reflection coefficient S11 for a SAW resonator fabricated on AlN/MCD.
Figure 5.6 shows the idealized lumped-element equivalent circuit for a one port
SAW resonator in the vicinity of resonance. This consists of a series inductance-
capacitance-resistance (LCR) branch shunted by a capacitor. The shunt capacitor CT
represents the IDT capacitance, while the elements Lr, Cr and Rr relate to equivalent
motional parameters for the series-resonance condition. The capacitance ratio CT / Cr is
an important one in one-port SAW resonators design (for more details, see
[Campbell.1998]). The IDT static capacitance CT is given by:
𝐶𝑇 = 𝑁𝑝𝐶𝑠 = 𝑁𝑝𝐶0𝑊 , (30)
where 𝑁𝑝 is the number of IDT finger-pairs, 𝐶𝑠is capacitance/finger-pair, 𝐶0 is
capacitance/finger pair/unit length, and 𝑊 is the acoustic aperture. Besides, the
equivalent series resistance Rr is approximated by:
𝑅𝑟 ≈ 1
𝐺𝑎 (𝑓0)
1− 𝜌
1+ 𝜌 , (31)
76
where 𝐺𝑎 𝑓0 = 8𝐾2𝑓0𝐶𝑠𝑁𝑝2 is the conductance at the IDT center frequency, and 𝜌 is a
dimensionless reflection coefficient related to the ratio of reflected-to-incident surface
waves entering the reflection grating.
The equivalent series inductance Lr may be approximated by:
𝐿𝑟 ≈𝑑𝑒
𝜆0
1
(4𝑓0𝐺𝑎 𝑓0 ) , (32)
where 𝑑𝑒 is the effective cavity length, and 𝜆0 = 𝑣𝑓0
. The equivalent series
capacitance 𝐶𝑟 is obtained as:
𝐶𝑟 =1
4𝜋2𝑓02𝐿𝑟
. (33)
Figure 5.6. LCR lumped-equivalent circuit in vicinity of resonance.
Data fitting has been used when the material constants are not exactly known,
which is the case of polycrystalline diamond [Iriarte.2003c]. Figure 5.7a shows the real
part of the conductance of the SAW resonance peaks correspond to all observed SAW
modes. The measured (blue line) curve was fitted according to the COM model until the
simulated curve (red line) reproduces pretty well the main and others resonant peaks.
Apart from the main resonances peak, two minor responses were observed in the
vicinity of mode 1, marked as A and B in figure 5.7b. The response A could be caused
by the effects of the gratings, and the response B by the spurious response due to the
multiple reflections of the IDT. These phenomena can be also observed on the fitted
curve, A‟ and B‟.
The procedure explained before was followed for the susceptance (figure 5.8)
and the reflection coefficient S11 (figure 5.9).The two minor responses were also
observed in figure 5.8b and figure 5.9b.
77
Figure 5.7. a) Fitted and measured conductance (G) vs. frequency for IDT/AlN/diamond, and b)
expanded view of G around mode 1.
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
10 11 12 13 14 15 16 17 18 19
Frequency (GHz)
G
measured
fitted
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
10.6 10.65 10.7 10.75 10.8 10.85 10.9
Frequency (GHz)
G
measured
fitted
AA'B
B'
a)
b)
78
Figure5.8.a) Fitted and measured susceptance (B) vs. frequency for IDT/AlN/diamond and b) expanded
view of B around mode 1.
-0.01
-0.005
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
10 11 12 13 14 15 16 17 18 19
Frequency (GHz)
Bmeasured
fitted
-0.01
-0.005
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
10.6 10.65 10.7 10.75 10.8 10.85 10.9
Frequency (GHz)
B
measured
fitted
a)
b)
79
Figure 5.9. a) Fitted and measured S11for IDT/AlN/diamond and b) expanded view of S11 around mode
1.
The values of the COM parameters extracted from the fitting procedure for all
the observed modes are shown in Table 5.2. The list of COM parameters extracted from
this device is the following:
- vf is theoretical phase velocity.
- veff is measured phase velocity:
𝑣𝑒𝑓𝑓 = 𝑓0𝜆0 (34)
where 𝑓0 is the center frequency estimated in S11, and 𝜆0, is the IDT period.
- k12 is the reflective per electrode obtained by:
-25
-20
-15
-10
-5
0
10 11 12 13 14 15 16 17 18 19
Frequency (GHz)
S1
1 (
dB
)
measured
fitted
-25
-20
-15
-10
-5
0
10.6 10.65 10.7 10.75 10.8 10.85 10.9
Frequency (GHz)
S1
1 (
dB
)
measured
fitted
b)
a)
80
𝑘12 = 𝜋(𝑓𝑎𝑠𝑦𝑚 − 𝑓𝑠𝑦𝑚 )
(𝑓𝑎𝑠𝑦𝑚 + 𝑓𝑠𝑦𝑚 ) (35)
where 𝑓𝑎𝑠𝑦𝑚 and 𝑓𝑠𝑦𝑚 are the eigenfrequencies of asymmetric and symmetric
modes, respectively.
- K2 is the coupling coefficient calculated as:
𝐾2 =𝜋
4𝑁
𝐺0
𝐵0, 𝑓 = 𝑓0 (36)
where N is the number of finger pairs, and 𝐺0 and 𝐵0 are radiation conductance
and susceptance, respectively.
- R is the electrical equivalent of mechanical impedance gives by:
𝑅 = 2𝜋
𝑤0𝐶𝐾2 (37)
where 𝐶0 = 𝐶2 is the electrode capacitance per section.
Table 5.2. Extracted COM Parameters for all the observed modes.
Mode
(f, GHz)
First
(10.777)
Second
(12.91)
Third
(14.72)
Fourth
(15.07)
veff(m/s) 8635 10349 11795 12100
k12 -0.0047 -0.0031 -0.0005 -0.0001
K2 0.0143 0.0033 0.001 0.0007
The values of the COM parameters extracted from the fitting procedure, FEM
analysis and by analytic solution for the first mode are shown in Table 5.3. In the used
FEM model the resistance of electrode was assumed zero, which means the conductivity
is approaching infinity. The large value of R means that the value of reflection
coefficient S11 is not close to 0 dB, starts from -1.4 dB at 10 GHz to -2.6 at 19 GHz
(figure 5.5). The veff is higher if we compare the extracted data with the FEM one and
K2, k12 and C are smaller. These differences are due to the finite value of the resistance
electrode.
81
Table 5.3. COM parameters extracted from fitting, FEM and analytic solution for the first mode (f: GHz).
Parameter Analytic solution FEM Extracted
vf (m/s) 8809 - -
veff (m/s) - 8580 8635
K2 0.011 0.0171 0.0143
k12 - -0.022 -0.0047
C (F/m) - 1.247×10-10
1.165×10-10
R (Ω) - - 0.278f+12.125
Loss - - 654.88 f
5.3 SAW devices on “free-standing” diamond
A very important structure used during this thesis work was the AlN/”free-
standing” diamond, as they were used to fabricate the SAW pressure sensors addressed
in next chapter. Also, resonator devices were developed on these free-standing
structures.
Figure 5.10 shows the reflection coefficient of a SAW resonators of λ = 800 nm
fabricated on 1200 nm, 600 nm, 300 nm and 150 nm-thick AlN sputtered on the free-
standing NCD (see figure 4.3). The film-thickness-to-wavelength ratio (H/) are kH =
2πH/λ = 9.42, 4.71, 2.36 and 1.18, respectively. As it was shown above, the
electromechanical coupling coefficient (K2) and the propagation velocity of the acoustic
waves depend on H/. This effect can also be observed on Figure 5.10 for the device
fabricated on free-standing NCD. The v obtained for the first Sezawa mode, S1,
increases from 7816 m/s for the 1200 nm-thick AlN, to 8425 m/s for the 600 nm-thick
AlN, 9995 m/s for the 300 nm-thick AlN and 10751 m/s for the 150 nm-thick AlN.
Moreover, the out-of-band rejection and quality factor Q also improve when the AlN
thickness is reduced down to 300 nm. On the other hand, for the device fabricated on
150 nm-thick AlN, the response is degraded due to the fact that the crystal quality also
worsens for this narrow film thickness range, a behavior common to all the observed
modes.
Figure 5.11 shows the difference observed in all devices response processed on
polished MCD and “free-standing NCD”. The observed shift in frequency is due to
82
small differences in the AlN thickness. The influence of the crystal quality in the device
response is clearly observed in these graphs. As seen in figure 4.16, there is a difference
between the FWHM of the 0002-AlN-rocking curve for the films deposited on MCD
and NCD. This difference is significantly higher when the thickness decreases below
600 nm. The effect of this behavior can be observed also in the device response. For the
AlN layer of 600 nm, the FWHM of the 0002-rocking curve was practically the same
value for both substrates, see figure 4.16, and similar device responses, quality factor
and the out of band rejection (figure 5.12), are observed. For the higher value of the
layer thickness, in addition to the SAW is more concentrate in the AlN layer and thus
the frequency is lower, the crystal quality worsens for the layer deposited on the free-
standing NCD. This is translated in a worse out of band rejection and Q of the device
response respect to the same layer deposited on MCD (figure 5.12). The latter was also
observed for the device response fabricated on the 300 nm- and 150 nm-thick AlN
layers.
8 10 12 14
-12
-9
-6
-3
0
-12
-9
-6
-3
0
-12
-9
-6
-3
0
-12
-9
-6
-3
0
1200 nm
S4
S4
S3
S3
S2
S2
S1
S1
S1
Frequency (GHz)
S1
1 (
dB
)
150 nm
S1
600 nm
300 nm
Figure 5.10. Reflection coefficient (S11) spectra for identical (= 800 nm) one-port SAW resonators on
(a) 1200 nm-, (b) 600 nm-, (c) 300 nm-, and (d) 150 nm-thick AlN films on “free-standing” NCD. Si
denotes the Sezawa modes.
83
As concluded before, the choice of the piezoelectric film thickness follows from
a compromise between the frequency and the intensity of the targeted resonances. For
the structure presented in this section, the optimal AlN film thickness is 300 nm. The
principal resonance for the SAW one port resonator fabricated on this film is 12.5 GHz,
corresponding to the first Sezawa mode, and it shows a Q factor of 643 and an out of
band rejection of 11 dB. The Q and out of band rejection are much better for the device
processed on AlN/MCD due to the better quality of the AlN film. Nevertheless, the
results obtained for the devices processed on AlN/free-standing NCD are the highest
obtained so far on nanocrystalline diamond, and the achieved frequency is the highest
reached among other free-standing NCD structures [Mortet.2002; Elmazria.2003;
Bénédic.2004]. Besides, the frequency results obtained during this thesis work for both
substrates MCD and NCD are much higher than the published so far as shown in figure
4.13. In addition, these results confirm that the AlN/free-standing NCD structure is an
alternative to the more expensive polished MCD for the processing of high performance
and high frequency SAW devices.
10 12 14-20
-15
-10
-5
0
10 12 14
-10
-5
0
10 12 14
-15
-10
-5
0
10 12 14
-40
-30
-20
-10
0
S4
S4 S3S3
S2
S2
S2
S1
S1
S1
Frequency (GHz)
S1
1 (
dB
)
Frequency (GHz)
S1
1 (
dB
)
Frequency (GHz)
S1
1200 nm free standing NCD
MCD
Frequency (GHz)
600 nm
150 nm300 nm
Figure 5.11. Comparison of the reflection coefficient (S11) spectra for identical (= 800 nm) one-port
SAW resonators on (a) 1200 nm-, (b) 600 nm-, (c) 300 nm-, and (d) 150 nm-thick AlN films on MCD
and “free-standing” NCD.
84
250 500 750 1000 1250
0
500
1000
1500
12000
13000
Q
Thickness (nm)
MCD
NCD
a)
250 500 750 1000 1250
0
5
10
15
20
25
30
35
40
Ou
t o
f B
an
d R
eje
cti
on
(d
B)
Thickness (nm)
MCD
NCD
b)
Figure 5.12. Comparison of quality factor Q (a) and Out of band rejection (b) of the one port SAW
resonators fabricated on different thicknesses of AlN on MCD and “free-standing” NCD.
2002 2003 2004 2005 2012 2013
0
1
2
3
4
5
10
15
20
2002 2003 2004 2005 2012 2013
0
1
2
3
4
5
10
15
20
2002 2003 2004 2005 2012 2013
0
1
2
3
4
5
10
15
20
2002 2003 2004 2005 2012 2013
0
1
2
3
4
5
10
15
20
2002 2003 2004 2005 2012 2013
0
1
2
3
4
5
10
15
20 AlN/free-standing diamond
[Fu
jii.2
01
2]
[Ha
sh
imo
to.2
01
3]
[Ro
drí
gu
ez-M
ad
rid
.20
12
]
[Be
ne
tti.2
00
5]
[Bé
né
dic
.20
04
]
[Elm
azri
a.2
00
3]
[Mo
rte
t.2
00
2]
Fre
qu
en
cy
(G
Hz)
Publication Year
AlN/diamond
[Fu
jii.2
01
2]
[Ha
sh
imo
to.2
01
3]
[Ro
drí
gu
ez-M
ad
rid
.20
12
]
[Be
ne
tti.2
00
5]
[Bé
né
dic
.20
04
]
[Elm
azri
a.2
00
3]
[Mo
rte
t.2
00
2]
Fre
qu
en
cy
(G
Hz)
Publication Year
[Fu
jii.2
01
2]
[Ha
sh
imo
to.2
01
3]
[Ro
drí
gu
ez-M
ad
rid
.20
12
]
[Be
ne
tti.2
00
5]
[Bé
né
dic
.20
04
]
[Elm
azri
a.2
00
3]
[Mo
rte
t.2
00
2]
Fre
qu
en
cy
(G
Hz)
Publication Year
ScAlN/6H-SiC
[Fu
jii.2
01
2]
[Ha
sh
imo
to.2
01
3]
[Ro
drí
gu
ez-M
ad
rid
.20
12
]
[Be
ne
tti.2
00
5]
[Bé
né
dic
.20
04
]
[Elm
azri
a.2
00
3]
[Mo
rte
t.2
00
2]
Fre
qu
en
cy
(G
Hz)
Publication Year
[Fu
jii.2
01
2]
[Ha
sh
imo
to.2
01
3]
[Ro
drí
gu
ez-M
ad
rid
.20
12
]
[Be
ne
tti.2
00
5]
[Bé
né
dic
.20
04
]
[Elm
azri
a.2
00
3]
[Mo
rte
t.2
00
2]
Fre
qu
en
cy
(G
Hz)
Publication Year
Figure 5.13. Comparison of frequency results for SAW devices published so far.
85
5.4 Summary and conclusions
In this chapter AlN layers deposited on MCD and “free-standing” NCD were
used to develop one-port SAW resonators with = 800 nm, which work in the 10-14
GHz range with good out-of-band rejection. By reducing the IDT period to = 400 nm,
higher frequencies were obtained (16-28 GHz). These frequencies are the highest
obtained for this material system.
Moreover, we have analyzed the influence of the film thickness on the response
of super-high-frequency SAW devices fabricated on AlN/diamond systems. From this
study on the influence of the layer thickness on the device performance, we can
conclude that there is compromise to optimize very-high-frequency and low-loss SAW
devices. The film thickness should be kept well below the SAW wavelength (kH=1.2-
2.4), but it has to be large enough to ensure a rocking curve FWHM value of the
AlN(0002) reflection of at least 4º as measured by X-ray diffraction. In particular, for a
λ of 800 nm, the optimum AlN layer thickness is 300 nm. This structure provides a S1
mode operating at 12.6 GHz with a Q factor of 12560 and an out-of-band rejection of
36.5 dB. The frequency of this mode has been shifted to 14.0 GHz using a 150 nm-thick
AlN film, though the reduced crystal quality and K2 yield lower Q factor and out-of-
band rejection values. COM parameters were obtained by fitting FEM and analytic
solution and minor responses from gratings and IDT reflections were observed.
The same thickness study with same devices was done for AlN//”free-standing”
NCD and with SAW resonators of λ = 400 nm for AlN/MCD. The trade-off between
selecting the right thickness while maintaining the crystallographic quality of the
piezoelectric film was also observed for both configurations. Moreover, AlN/”free-
standing” NCD structure can be proposed as alternative of the AlN/MCD for the
processing of SAW one port resonator due to similar results were obtained for both
structures.
86
6. SAW SENSORS
In this chapter, the AlN/free standing diamond structures have been used for the
processing of SAW resonators for pressure sensing applications. The devices fabricated
allow detecting changes in pressure with a sensibility of 0.33 MHz/bar.
On the other hand, the temperature stability of those SAW devices has been
studied. A low temperature coefficient of frequency (TFC) was observed over a wide
temperature range, from -250 to 250 ºC. This resonator structure is promising for future
communication and sensor applications in harsh environments for which temperature
may vary significantly, such as the space or industrial fields. In the case of pressure
sensors, the study of the evolution of the sensitivity with the temperature is required for
eventual recalibrations at different temperature conditions. This chapter includes the
results described in [Rodríguez-Madrid.2013].
6.1 Pressure sensors
During the two last decades, MEMS sensors played a major role in pressure
measurements [Eaton.1997], although other kind of sensors have been developed.
Capacitive pressure sensors using a diaphragm and a pressure cavity to create a variable
capacitance; piezoelectric pressure sensors that exploit the piezoelectric effect to
measure the strain caused by pressure; piezoresistive pressure sensors, for which the
resistance of the piezoresistive material can be altered by the pressure applied on it; and
SAW pressure sensors, that measure the strain caused by pressure, are some of the types
of sensors that are being studied nowadays [Schimetta.2000; Dohyuk.2012;
Tushar.2012; Xin.2012]. During this thesis work, a high frequency SAW based pressure
sensor with a high sensitivity has been developed on an AlN/free-standing diamond
structure.
In industrial applications, many systems that operate at higher pressure also
suffer from other harsh situations, such as high radiation, chemical corrosion, etc.
Therefore, an appropriate material for those environments with a high sensitivity, that
increases for higher frequencies [Venema.1986], is desirable.
87
In a SAW resonator, equations (23) and (24) show the linear approximation of
the frequency dependence on pressure when the temperature is kept constant:
𝑓 𝑝 = 𝑓0 + 𝑆∗(𝑝 − 𝑝0) (38)
where
𝑆∗ = 𝑓0𝑆, (39)
𝑝0 is a reference value, and 𝑆 is the first order pressure coefficient [Buff.1997].
In a SAW pressure sensor, therefore, the higher the central frequency, 𝑓0, the higher the
shift in the frequency for the same pressure change. Thus, our high frequency devices
fabricated on AlN/diamond structures are excellent candidates for high precision SAW
pressure sensors. Furthermore, as already stated, CVD diamond is an appropriate
material to work in harsh environments, and the AlN/diamond system has a low
temperature coefficient of delay (TCD) (between -21 and -9 ppm/ºC [Assouar.2007] ).
Thus, apart from the enhanced sensitivity at higher frequencies, the use of the
AlN/diamond bi-layer system is also justified by the possibility of using this structure in
harsh environments and by its low TCD.
In order to optimize the pressure sensitivity to the structure, we have used a 20
µm thick NCD free standing membrane, which silicon substrate was removed by
chemical etching. This structure allows the deposition of highly c-axis oriented AlN due
to the low surface roughness of the nucleation side. The procedure to obtain the cavity
is summarized in figure 6.1. The rough side of the NCD was glued with an epoxy
adhesive to an alumina substrate with a hole. The silicon substrate was removed by
HF:HNO3 (2:1) chemical etching at room temperature. Different experiments were done
with NCD films from 2 µm- to 20 µm-thick in order to obtain free-standing diamond.
The 20 µm-thick NCD is the only film that didn‟t break during the processing to obtain
the free-standing diamond.
88
Figure 6.1. Schematic process followed to obtain the free-standing diamond membrane used to deposit
AlN and process SAW device pressure sensors.
Once the diamond membrane was obtained, it was cleaned using a standard
procedure. The sputtering conditions used to deposit the AlN film are summarized in
Table 4.1. The thickness used of this piezoelectric film was 300 nm (see section 5.3),
the optimum one for free-standing diamond films. The design of the SAW resonator
devices was the same as that used in previous sections.
In order to characterize these devices, nitrogen gas in a pressure range of 2-5
bars was used to measure the pressure sensitivity of the SAW sensor. A schematic
illustration of our pressure measurement set-up is shown in figure 6.2. The pressure
controlled nitrogen gas flow was located approximately 5 mm above the SAW
resonator. After analyzing different positions, the device was located on the edge of a 5
mm2 diaphragm. The pressure underneath the AlN/diamond structure was kept constant
at atmospheric pressure. The pressure applied to the membrane results from the
difference (Δp) between the pressures applied on top (P1) and the atmospheric pressure
underneath the AlN/diamond membrane (P2).
89
Figure 6.2. Schematic of the set-up used to characterize the SAW based pressure sensor.
The pressure was applied using compressed nitrogen, and its value was modified
both upwards (from 2 until 5 bar) and downwards (from 5 until 2 bar). As the pressure
underneath the hole was kept constant at atmospheric pressure, the real range of P
applied to the membrane in both directions was between 1 and 4 bar, which was
modified in steps of 0.5 bar. The SAW resonators were characterized in the frequency
range [10.2 GHz - 11.2 GHz] at room temperature.
Figure 6.3 shows the first Sezawa mode of the reflection coefficient and its
frequency shift due to the pressure variation for positive direction (upwards) for the
device fabricated on the etch of the membrane. Figure 6.4 shows the behavior of the
reflection coefficient when the pressure is varied downwards. Other devices were
located in the center of the membrane obtaining non-linear behavior or less sensitivity
as shown in Figure 6.5.
90
10770 10780 10790 10800
-6
-5
-4
-3
S1
1 (
dB
)
Frequency (MHz)
1 bar
1,5 bar
2 bar
2,5 bar
3 bar
3,5 bar
4 bar
a)
10784 10786 10788
-6,0
-5,7
-5,4
S1
1 (
dB
)
Frequency (MHz)
b)
Figure 6.3. a) Frequency response of AlN/Diamond SAW resonators located on the etch of the membrane
under upwards pressure variation and b) zoom in.
91
10770 10780 10790 10800
-6
-5
-4
-3
S1
1 (
dB
)
Frequency (MHz)
1 bar
1,5 bar
2 bar
2,5 bar
3 bar
3,5 bar
4 bar
a)
10784 10786 10788
-5,95
-5,60
-5,25
S1
1 (
dB
)
Frequency (MHz)
b)
Figure 6.4. a) Frequency response of AlN/Diamond SAW resonators located on the etch of the membrane
under downwards pressure variation and b) zoom in.
92
10,76 10,77 10,78
-0,6
S1
1 (
dB
)
Frequency (GHz)
1 bar
1,5 bar
2 bar
2,5 bar
3 bar
3,5 bar
4 bar
a)
10,762 10,764 10,766 10,768
-0,78
-0,76
-0,74
-0,72
-0,70
-0,68
S1
1 (
dB
)
Frequency (GHz)
b)
Figure 6.5. a) Frequency response of AlN/Diamond SAW resonators located on the center of the
membrane under pressure variation and b) zoom in.
Figure 6.6 shows the behavior of the frequency shift of the device for different
pressures for the upwards and downwards measurements. Besides experimental errors
in the measurement of the pressure or its distribution across the membrane (the
93
manometer error is ± 0.1 bar), the small differences between both behaviors may be
related to hysteresis effects inherent to piezoelectric phenomena; in fact, this hysteresis
has been reported to lead to a decrease of the piezoelectric coefficient d33
[Kholkin.1996; Setter.2005].
A non linear behavior was observed when a pressure variation in the range ΔP =
0 to 1 bar was applied. Differences between air and N2 density and mainly the
dimensions of the AlN/diamond membrane, as well as the location SAW device on top
of it are considered to be the most critical factors determining this kind of behavior.
Moreover, a temperature variation takes place caused by the N2 flow, leading to a shift
of the centre frequency as studied elsewhere [Buff.1997]. A dual oscillator temperature
compensation technique could be used in order to overcome the non linear behavior and
eliminate those temperature effects. To achieve this compensation, a configuration with
two SAW resonators in close proximity [Cullen.1980], or two devices with and without
the membrane in the same substrate, can be used.
The sensitivity obtained by linear fitting is 0.33 ± 0.02 MHz/bar, as shown in
Figure 6.6. The high sensitivity obtained here is not as high as the observed for similar
configuration in bibliography [Nabipoor.2006], nevertheless it can utterly be tuned up
by optimizing the device location within the diaphragm [Cullen.1980].
Figure 6.6 Experimental frequency shift when pressure is applied in positive (up) and negative (down)
direction (symbols) and linear regression for both (solid line).
94
6.2 Temperature study
As seen in the previous section, several factors like environmental temperatures,
fabrication tolerances, and circuit designs affect the range of use of SAW based sensors.
Typically, temperature is an uncontrollable variable, so that in the design of a high
performance SAW device temperature stability or close-to-zero temperature coefficient
of frequency (TCF), or a simultaneous measurement of the temperature to calibrate the
sensitivity vs. T, is a requirement.
Concerning to the communications field, a satellite just outside the earth
atmosphere is in a critical situation with respect to temperature control. The satellite is
heated by solar radiation during daytime, and internal electronic equipment is another
source of heat, which initially is neglected compared to solar heating. However, at dark
conditions, the temperature is much lower. Therefore, satellites may operate in extreme
temperatures from −150°C to 150°C [Maini.2007]. Other applications in measuring and
manufacturing also require SAW sensors capable of operating in a wide-temperature
range and harsh environments, like tire pressure monitoring systems [Dixon.2006;
Cook.2009], automotive [Kalinin.2007], power, nuclear, steel industry, chemical and
petrochemical processes plants [Hamsch.2004; Fachberger.2006; Hamidon. 2006].
Several research groups are working to achieve SAW devices capable of
operating in high temperatures conditions up to 1000ºC with low TCF [Aubert.2012a;
Aubert.2012b; Bartasyte.2012]. A new generation of piezoelectric materials and
structures [Damjanovic.1998] stable in harsh conditions are investigated in order to
replace conventional substrates as quartz [Hornsteiner.1998] or lithium niobate
[Aubert.2009], that are quite limited in temperature. To overcome these limitations and
to obtain higher frequencies, layered structures such as AlN/SiC, AlN/sapphire or the
studied during this thesis work AlN/diamond are very promising [Benetti.2008;
Aubert.2012b; Caliendo.2012]. AlN can be used in high temperature conditions; it melts
at 3214ºC but starts to decompose chemically in vacuum at 1040ºC [Ambacher.1996]
and it oxidizes under air conditions at 950ºC [Lin.2008]. In addition, the AlN/diamond
bilayer structure is robust under harsh conditions [Werner.2001]. Moreover, thin film
technology makes possible the sandwich of the metalized interdigital transducers (IDT)
95
between the AlN film and the substrate, offering them a natural protection against
external aggressions.
Despite all these potential benefits, AlN has rarely been experimentally studied
as a piezomaterial for high and low temperature SAW applications. Not even the rest of
materials mentioned above have been studied in a wide temperature range including
cryogenic temperatures. Most of the works are focused on high temperature. During this
section the thermal stability of the SAW resonators fabricated on AlN/diamond
heterostructure for a temperature range between -250°C to 250°C is studied.
The TCF of a layered SAW structure depends on three parameters. The
temperature variation of the acoustical properties for each material, the film normalized
thickness (kh), and the substrate coefficient of thermal expansion (α). The two first
parameters yield to the temperature coefficient of velocity (TCV), and the third one
provides directly the amount of delay generated by the thermal expansion. The TCF is
given by the following equation:
TCF = 1
f0
δf
δT= TCV − CTE, (40)
TCV =1
V
δ𝑣
δT, (41)
where 𝑣 is SAW velocity, T is temperature, f0 the frequency at room temperature and
CTE is the SAW velocity and the thermal expansion coefficient of the substrate to the
SAW propagation direction.
For this temperature study, a 600 nm thick (0002)-textured AlN was deposited
on polished microcrystalline diamond substrate. The fabricated SAW devices on this
structure have been measured in a Janis ST-500 probe station together with an Agilent
N5230A vector network analyzer. Coplanar-waveguide probe tips for cryogenic and
high temperatures have been used with standard calibration techniques. This
combination allows to measure SAW devices in 10 MHz-20 GHz range under vacuum
condition in a temperature range between -250ºC and 250ºC.
Figure 6.7 shows the reflection spectra of the fabricated one port SAW resonator
at room temperature. The main resonance corresponds to the first-order Sezawa mode
96
S1 at 10.57 GHz. In addition, the quality factor Q and out-of-band rejection for the S1
mode is 1158 and 17.5 dB, respectively.
10 11 12 13 14 15 16-25
-20
-15
-10
-5
0
S1
1 (
dB
)
Frequency (GHz)
S1
S2
S3
S4
Figure 6.7.Reflection coefficient (S11) of a one-port SAW resonator fabricated on AlN film on diamond.
Si denotes the Sezawa modes.
Figure 6.8 shows the behavior of the measured main resonance (Fig. 6.8a) and
second order Sezawa-mode (Fig. 6.8b). There are two different regions of the shift
frequency for both modes: in the lower temperature range, from -250ºC to -100ºC, the
obtained TCF is -5.06 ppm/ºC for the first Sezawa mode (S1) and -8.53 ppm/ºC for the
second mode (S2); the other range, observed for temperatures from -100ºC to 250ºC, has
TFC of -22.90 for S1 and -18.90 ppm/ºC for S2. This slight difference between the
values of the TCF for the two modes is likely due to propagation at different depths.
The wide temperature range used in this experimental study, and the high
working frequencies of the SAW resonator devices studied here, have not been reported
so far. Moreover, the TCF obtained for this IDT/AlN/diamond structure is lower than
the one obtained for similar piezoelectric/substrate bilayers [Nakahata1995; Aubert.
2012b] at much lower working frequencies.
In order to obtain a SAW device with zero TCF, a material with positive TCF
could be deposited on top of the structure. Silicon dioxide, for example, could be used
to offset the TCF observed for the structure studied here [Dubois.1999b; Lakin.2000].
In collaboration with Ashu Wang, from ISOM-UPM, simulations of this
structure under the same temperature conditions were done using the commercial
97
software COMSOL, in order to better understand the experimental results. Only one
behavior has been observed for both modes in the simulations, see figure 6.9. Because
of the AlN thickness consider in the simulation little differences in frequency are
observed between experimental and simulation data. For the first Sezawa mode, the
simulated TFC is -16.13 ppm/ºC and -14.15 ppm/ºC for the second mode. These values
are similar to the ones obtained for the second range on the measured devices response.
The different values observed on the frequency responses can be related to a different
behavior of the elastic stiffness constant at low temperature [Varshni.1970;
Ledbetter.2006] which was not considered in the simulation.
10,40
10,48
10,56
-300 -200 -100 0 100 200 300
12,60
12,67
12,74
TCF2= -18.90 ppm/ºC
TCF2= -22.52 ppm/ºC
S1
linear regretion 1
linear regretion 2
Fre
qu
en
cy
(G
Hz)
TCF1= -5.06 ppm/ºC
a)
TCF1= -8.53 ppm/ºC
S2
linear regresion 1
linear regresion 2
Temperature (ºC)
b)
Figure 6.8. Frequency behavior of the first (a) and second (b) Sezawa modes with temperature.
98
10,52
10,56
10,60
b)
Sezawa 1
Fre
qu
en
cy
(G
Hz)
a)
-300 -200 -100 0 100 200 300
12,52
12,56
12,60
Temperature (ºC)
Sezawa 2
Figure 6.9. COMSOL simulation of frequency behavior of the first (a) and second (b) Sezawa modes with
temperature.
6.3 Summary and conclusions
In this chapter, a free-standing NCD layer is used as a substrate to sputter AlN
and fabricate SAW devices for sensor applications, which are improved at the high
frequency operating range. The main resonance of a device obtained using this structure
with an IDT period of 800 nm is at 12.5 GHz, corresponding to the first Sezawa mode,
and it shows a Q factor of 643 and an out-of-band rejection of 11 dB. As the membrane
of the structure is free to move, these devices have been used as pressure sensors, which
have shown a sensitivity of 0.33 MHz/bar. This sensitivity could be improved by
optimizing the localization of the sensor in the free-standing membrane. Also, its value
can be affected by temperature changes, so a low TFC is desired for SAW application
under harsh temperature conditions.
99
The second part of this chapter shows how the AlN/diamond is an attractive
structure for SAW devices operating in a wide range of temperatures. Both materials
AlN and diamond offer an outstanding natural advantage for harsh conditions. We have
experimentally demonstrated a low TCF in a wide temperature range for SAW one port
resonators working at 10 GHz. Different values of TCF were determined for two
temperature ranges, namely of -5.81 and -8.53 ppm/ºC (-22.52 and -18.90 ppm/ºC) for
the -250ºC to -100ºC (-100ºC to 250ºC) range for the first and second Sezawa modes,
respectively. COMSOL simulations have also been carried out, and the values obtained
for -100ºC to 250ºC range are similar to the experimental ones in the same temperature
range.
100
7 GENERAL CONCLUSIONS AND FUTURE
WORK
In this chapter the main conclusions and results regarding the material deposition
and properties, the fabrication and assessment of SAW devices, and their applications as
high frequency resonators and pressure sensors are presented. The future work is
principally related to the CMOS integration that arises from these results.
The development of CMOS compatible, high frequency and high performance
SAW devices requires the deposit by reactive sputtering of AlN on a fast substrate.
Growth conditions have been studied in order to get c-axis oriented AlN. The
parameters that affect this growth have been varied independently in order to obtain
piezoelectric AlN on an arbitrary substrate. But not only system parameters determine
the quality of the film; the substrate roughness and AlN film thickness affect on the
growth of it.
Regarding the system parameters involved during the sputtering process, AlN
oriented on the c-axis has been obtained for different target to substrate distances and
different Ar/N2 flow ratios, provided that the stoichiometry is guaranteed. A target-
substrate distance of 45 mm and 3/9 sccm for Ar/N2 respectively have been used to
optimize the most critical deposition parameters regarding the c-axis orientation of the
layers, namely the process pressure and the discharge power: the optimized values
derived from our study are 3 mTorr and 700 W, respectively.
Concerning to the substrate characteristics, the most important ones to achieve
highly c-axis oriented polycrystalline AlN are surface roughness and grain size. For the
analyzed substrates, with different materials as well as crystallographic properties, the
surface roughness should be kept below 4 nm rms to guarantee the desired orientation
of the film.
MCD and NCD have been the principal substrates used during this thesis. In
order to optimize the performance of the SAW devices, different AlN thicknesses have
been tested. AlN films from 150 nm to 1200 nm were deposited to study the influence
of the thickness on the piezoelectric film orientation. Using the optimized sputtering
parameters, highly c-axis AlN were obtained for all the thicknesses on different
101
substrates. The FWHM of the (0002)-rocking curve, which was chosen as the figure-of-
merit for the material quality, increases when the thickness decreases.
The most critical issue of the synthetic diamond for SAW devices is its high
surface roughness. Polished MCD have been used to overcome it; also, a “free-
standing” structure has been developed from NCD substrates for the same purpose.
The AlN thin films deposited on polished MCD and “free-standing” NCD have
been used to develop one-port SAW resonators and to study the influence of the film
thickness on their response. For device fabrication, e-beam lithography has been used in
order to obtain finger widths of = 800 nm and 400 nm.
For the design of the AlN/diamond structure, tradeoffs must be made to optimize
very-high-frequency and low-loss SAW devices. The piezoelectric film thickness
should be kept well below the SAW wavelength, but it has to be large enough to ensure
a rocking curve FWHM value of the AlN (0002) reflection of at least 4º as measured by
X-ray diffraction.
The fabricated devices work in the 10-14 GHz range ( = 800 nm) and 16-28
GHz range ( = 400 nm) with excellent out-of-band rejection. Those frequencies are the
highest obtained for this material system. Similar devices have been fabricated on
AlN/”free-standing” NCD structures. This is proposed as an alternative to the
AlN/MCD bilayer system for the processing of SAW one port resonator; similar results
have been obtained for both structures.
The free-standing NCD structure with a movable membrane has also been
developed for pressure sensor applications. The devices fabricated on this structure have
a main resonance at 12.5 GHz and show a Q factor of 643 and an out-of-band rejection
of 11 dB. Under pressure conditions, these devices show a sensitivity of 0.33 MHz/bar.
A study of these devices working under temperature conditions has been done
since it affects the device performance. We have experimentally demonstrated a low
TCF in a wide temperature range for SAW one port resonators working at 10 GHz.
Different values of TCF were determined for two temperature ranges, namely of -5.81
and -8.53 ppm/ºC (-22.52 and -18.90 ppm/ºC) for the -250ºC to -100ºC (-100ºC to
250ºC) range for the first and second Sezawa modes, respectively. This study shows
102
how the AlN/diamond is an attractive structure for SAW devices operating in a wide
range of temperatures.
In conclusion, this technology is very promising to achieve CMOS compatible
devices with small size and thermal stability for high frequency oscillators, filters and
sensors applications.
Different improvements and new studies can be done for future work after
analyzing the results obtained during this thesis. Concerning material aspects, the
temperature growth of diamond substrates is around 800ºC; as this temperature is
decreased the surface roughness increases and crystal quality decreases. AlN has been
obtained at room temperature and thus CMOS compatibility is guaranteed. As future
work, the optimization of diamond growth at low temperature and the deposit of c-axis
oriented AlN at room temperature on these substrates have to be done for a total CMOS
integration.
The incorporation of carbon in the AlN/diamond interlayer should be studied by
Rutherford backscattering spectrometry (RBS) in order to confirm the X-ray diffraction
and photoluminescence results.
Regarding fabrication aspects, the IDT metallization could be improved studying
the influence of different metals and thicknesses on the device response. The
AlN/diamond structure studied here can be used for fabrication of other devices such as
filters working at high frequency. Besides, the AlN/”free-standing” structure could be
used for the fabrication of MEMS. All these devices, including the one-port SAW
resonators, could be used for chemical sensor applications. Moreover, biochemical
sensors can be developed with this technology because of the biocompatibility of the
structure. For the particular case of the pressure sensor studied in this thesis, its
sensitivity could be improved by increasing the size of the free-standing membrane, and
by localizing the sensor farther away from the membrane edges.
Finally, the thermal behavior can also be optimized. Despite the low TFC
obtained, some materials with a positive TFC could be deposited on top of the structure
in order to reduce this value.
103
Acknowledgments
En primer lugar me gustaría agradecer a Fernando Calle y Gonzalo Fuentes la
oportunidad que me han dado al poder formar parte de su equipo para realizar esta tesis.
A Fernando por la dirección y el tiempo que ha dedicado. A Gonzalo por la dirección,
por su confianza, optimismo y por encontrar siempre las palabras adecuadas en los
peores momentos.
Muchas son las personas que han formado parte de este largo camino y que de
una forma u otra han contribuido a que nunca olvide estos años de mi vida.
En el terreno técnico agradecer su dedicación más allá de lo profesional a Maika,
sin olvidarme de los sabios consejos de David ni por supuesto del tiempo invertido por
Alicia y Juan Ramón (“Jota”).
Destacar las colaboraciones con los diferentes grupos esenciales para el
desarrollo de esta tesis. A Oliver Williams por los substratos de diamante, Daniel
Araujo por las medidas TEM, Jesús Grajal por las medidas a muy alta frecuencia,
Jimena Olivares por las medidas FTIR, Ruyen Ro por la extracción de los parámetros
COM, Jorge Pedrós por sus consejos, discusión de resultados y ayuda en la simulación
de la velocidad de dispersión y Neuschl por las medidas de PL.
Gracias a todas y cada una de las personas del ISOM que me transmitieron su
conocimiento para poder utilizar todas las técnicas que han sido necesarias para llevar a
cabo esta tesis. Destacar a Rocío por montar el equipo de sputtering y transmitirme todo
lo que aprendió durante su proyecto fin de carrera, a Maika por su paciencia al
enseñarme las técnicas de la sala blanca, a Raquel, Fernando Posadas, Javier Martínez y
Francesca por enseñarme a medir AFM. A Zarko y Javier Grandal por enseñarme a
medir Rayos-X. A Ashu por las simulaciones de temperatura. A todos aquellos que
aunque no estén mencionados específicamente contribuyeron a la discusión de
resultados, resolución de problemas e intercambio de ideas.
Agradecer la oportunidad que se me ha dado al realizar esta tesis no sólo a la
hora de formarme si no de conocer a gente que ya es parte de mi vida. A esos
compañeros que ahora son mis amigos, Gema, Sebastian, Francesca, Ana Pérez, Manu,
Mariajo, Ana Bengoechea, Verónica, Alberto y Tommaso. Especialmente a Sara y
104
Maika, este tiempo juntos nos ha servido para unirnos más aún, gracias por esas risas y
momentos de desconexión. Sin ninguna duda sois una parte muy importante para que
esta tesis haya salido hacia adelante.
A todas las personas que formaban parte de mi vida antes de empezar este
camino, a las que han pasado a formar parte de ella y aún están ahí y aquellos que se
quedaron atrás. A Nuno, José y mis amigos portugueses; a María, Eva, Estefa y amigos
de Valladolid; a María Orgaz y mis amigos del pueblo; a Diego y Víctor; a Francesca,
Arancha, Piero y Maite. Todos vosotros, a veces sin ser conscientes de ello, habéis
contribuido a que este trabajo haya llegado a su fin.
Y por último, agradecer a mi familia la fuerza que me ha transmitido a pesar de
la distancia, en especial a mis padres y hermanos Mar, Azu y Antonio. Gracias mamá,
eres un ejemplo a seguir, por tu fuerza y por apoyarme en TODO momento.
“Todo concluye, nada perece”
(Lucio Anneo Séneca)
105
References
[Aigner.2003] Aigner, R. (2003). MEMS in RF filter applications: Thin-film
bulk acoustic wave technology. (pp. 175-210) WILEY-VCH
Verlag.
[Ambacher.1996] Ambacher, O., Brandt, M. S., Dimitrov, R., Metzger, T.,
Stutzmann, M., Fischer, R. A., Miehr, A., Bergmaier, A. &
Dollinger, G. (1996). Thermal stability and desorption of group
III nitrides prepared by metal organic chemical vapor deposition.
Journal of Vacuum Science & Technology B: Microelectronics
and Nanometer Structures, 14(6), 3532-3542.
[Assouar.2007] Assouar, M. B., Elmazria, O., Kirsch, P., Alnot, P., Mortet, V., &
Tiusan, C. (2007). High-frequency surface acoustic wave devices
based on AlN/diamond layered structure realized using e-beam
lithography. Journal of Applied Physics, 101, 114507-114512.
[Aubert.2012a] Aubert, T., Bardong, J., Elmazria, O., Bruckner, G. & Assouar, B.
(2012a). Iridium interdigital transducers for high-temperature
surface acoustic wave applications. IEEE Transactions on
Ultrasonics Ferroelectrics and Frequency Control, 59(2), 194-
197.
[Aubert.2012b] Aubert, T., Elmazria, O., Assouar, B., Blampain, E., Hamdan, A.,
Geneve, D. & Weber, S. (2012b). Investigations on AlN/sapphire
piezoelectric bilayer structure for high-temperature SAW
applications. IEEE Transactions on Ultrasonics, Ferroelectrics
and Frequency Control, 59(5), 999-1005.
[Aubert.2009] Aubert, T., Elmazria, O., & Assouar, M. (2009). Wireless and
batteryless surface acoustic wave sensors for high temperature
environments. 9th International Conference on Electronic
Measurement & Instruments, pp. 2-890-2-898.
[Ballantine.1997] Ballantine, D., White, R., Martin, S., Ricco, A., Zellers, E., Frye,
G., & Wohltjen, H. (1997). Acoustic wave sensors: Theory,
design, & physico-chemical applications Academic press.
[Bartasyte.2012] Bartasyte, A., Elmazria, O., Gonzalez, M., Bouvot, L., Blampain,
E., & Boulet, P. (2012). Reduction of temperature coefficient of
frequency in LiTaO3 single crystals for surface acoustic wave
applications. Intl Symp Applications of Ferroelectrics Held
Jointly with 2012 European Conference on the Applications of
Polar Dielectrics and 2012 International Symp Piezoresponse
Force Microscopy and Nanoscale Phenomena in Polar Materials
(ISAF/ECAPD/PFM), pp. 1-3.
106
[Belonoshko.2010] Belonoshko, AB.;Arapan, S.;Martonak, R. & Rosengren, A,
(2010). MgO phase diagram from first principles in a wide
pressure-temperature range. Physical Review B 81, 054110-9
[Bénédic.2004] Bénédic, F., Assouar, M., Mohasseb, F., Elmazria, O., Alnot, P.,
& Gicquel, A. (2004). Surface acoustic wave devices based on
nanocrystalline diamond and aluminium nitride. Diamond and
Related Materials, 13(2), 347-353.
[Benetti.2008] Benetti, M., Cannata, D., Di Pietrantonio, F., Verona, E.,
Almaviva, S., Prestopino, G., Verona, C. & Verona-Rinati, G.
(2008). Surface acoustic wave devices on AlN/single-crystal
diamond for high frequency and high performances operation.
IEEE Ultrasonics Symposium, pp. 1924-1927.
[Benetti.2005] Benetti, M., Cannata, D., Di Pictrantonio, F. & Verona, E. (2005).
Growth of AlN piezoelectric film on diamond for high-frequency
surface acoustic wave devices. IEEE Transactions on Ultrasonics,
Ferroelectrics, and Frequency Control, 52(10), 1806-1811.
[Binder.2013] Binder, A., Bruckner, G., Schobernig, N. & Schmitt, D. (2013).
Wireless SAW pressure and temperature sensor with unique
identification based on LiNbO3. Sensors Journal, IEEE, PP(99),
1.
[Brizoual.2007] Brizoual, L. & Elmazria, O. (2007). FEM modeling of
AlN/diamond surface acoustic waves device. Diamond and
Related Materials, 16(4–7), 987-990.
[Bu.2003] Bu, G., Ciplys, D., Shur, M., Schowalter, L., Schujman, S. &
Gaska, R. (2003). Temperature coefficient of SAW frequency in
single crystal bulk AlN. Electronics Letters, 39(9), 755-757.
[Buff.1997] Buff, W., Rusko, M., Goroll, E., Ehrenpfordt, J. & Vandahl, T.
(1997). Universal pressure and temperature SAW sensor for
wireless applications. IEEE Proceedings Ultrasonics Symposium,
1. pp. 359-362.
[Caliendo.2012] Caliendo, C. (2012). Theoretical investigation of high velocity,
temperature compensated rayleigh waves along AlN/SiC
substrates for high sensitivity mass sensors. Applied Physics
Letters, 100(2), 021905-021905-3.
[Campbell.1998] Campbell, C. (1998). Surface acoustic wave devices for mobile
and wireless communications. Academic Press.
107
[Campbell.1968] Campbell, J., & Jones, W. (1968). A method for estimating
optimal crystal cuts and propagation directions for excitation of
piezoelectric surface waves. IEEE Transactions on Sonics and
Ultrasonics, 15(4), 209-217.
[Chan.2012] Chan, K., Zhang, W., Windt, D., Hong, M., Saha, T., McClelland,
R., Sharpe, M. & Vivek, H. (2012). Reflective coating for
lightweight x-ray optics. Procceding Space Telescopes and
Instrumentation, 8443,
[Chandran.2012] Chandran, M., Tiwari, B., Kumaran, C., Samji, S., Bhattacharya,
S. & Ramachandra-Rao, M. (2012). Integration of perovskite PZT
thin films on diamond substrate without buffer layer. Journal of
Physics D: Applied Physics, 45(20), 202001.
[Changbao.2011] Changbao, W., Yong, F., Wan, L., Wen, Z., Xin, X., Yan, S.,
Yan, M., & Li, M. (2011). Carbon dioxide gas sensor using SAW
device based on ZnO film. Applied Mechanics and Materials, 135
- 136, 347-352.
[Chen.2005] Chen, Y., Wang, R., Wang, B., Xing, T., Song, X., Zhu, M. &
Yan, H. (2005). Effects of mean free path on the preferentially
orientated growth of AlN thin films. Journal of Crystal Growth,
283(3-4), 315-319.
[Chen.1985] Chen, D. & Haus, H. (1985). Analysis of metal-strip SAW
gratings and transducers. IEEE Transactions on Sonics and
Ultrasonics, 32(3), 395-408.
[Cheng.2003] Cheng, H., Sun, Y. & Hing, P. (2003). The influence of
deposition conditions on structure and morphology of aluminum
nitride films deposited by radio frequency reactive sputtering.
Thin Solid Films, 434(1–2), 112-120.
[Chou.2006] Chou, C., Lin, Y., Huang, J., Tai, N. & Lin, I. (2006). Growth of
high quality AlN thin films on diamond using TiN/Ti buffer layer.
Diamond and Related Materials, 15(2-3), 404-409.
[Chung.2009] Chung, G., & Hong, H. (2009). Effect of a 3C-SiC buffer layer on
the SAW properties of AlN films grown on si substrates. Journal
of the Korean Physical Society, 55(4), 1446-1450.
[Clement.2003] Clement, M., Iborra, E., Sangrador, J., Sanz-Hervas, A., Vergara,
L., & Aguilar, M. (2003). Influence of sputtering mechanisms on
the preferred orientation of aluminum nitride thin films. Journal
of Applied Physics, 94(3), 1495-1500.
108
[Collazo.2012] Collazo, R., Xie, J., Gaddy, B., Bryan, Z., Kirste, R., Hoffmann,
M., Xie, J., Moody, B., Kumagai, Y., Nagashima, T., Kubota, Y.,
Kinoshita, T., Koukitu, A., Irving, D., & Sitar, Z. (2012). On the
origin of the 265 nm absorption band in AlN bulk crystals.
Applied Physics Letters, 100(19), 191914-191914-5.
[Cook.2009] Cook, JD., Dixon, Kumar, S., Marsh, BJ. & Qasimi, MAJ. (2009).
Wireless and batteryless pressure sensor apparatus for tire
pressure monitoring system (TPMS) in vehicle has surface
acoustic wave (SAW) sensor button package which is adapted as
base for mounting and containing SAW sensor and antenna.
PatentUS20090167503Al(2009),US20090165546Al(2009),US
7,726,184 B2.
[Cullen.1980] Cullen, D. & Montress, G. (1980). Progress in the development of
SAW resonator pressure transducers. Ultrasonics Symposium, pp.
696-701.
[Damjanovic.1998] Damjanovic, D. (1998). Materials for high temperature
piezoelectric transducers. Current Opinion in Solid State and
Materials Science, 3(5), 469-473.
[Defosseux.2012] Defosseux M., Allain M., Defay E. & Basrour, S. (2012). Highly
efficient piezoelectric micro harvester for low level of
acceleration fabricated with a CMOS compatible process. Sensors
and Actuators A: Physical, 188(0), 489-494.
[Didenko.2000] Didenko IS., Hickernell, FS. & Naumenko, B. (2000). The
experimental and theoretical characterization of the SAW
propagation properties for zinc oxide films on silicon carbide,
IEEE Transations on Ultrasonic Ferroelectric Frequency Control,
47, 1, 179–187.
[Dixon.2006] Dixon, B., Kalinin, V., Beckley, J., & Lohr, R. (2006). A second
generation in-car tire pressure monitoring system based on
wireless passive SAW sensors. IEEE International Frequency
Control Symposium and Exposition, pp. 374-380.
[Dohyuk.2012] Dohyuk, H., Vries, W., John-Simon, W., Irazoqui, P., & William,
J. (2012). Polymer-based miniature flexible capacitive pressure
sensor for intraocular pressure (IOP) monitoring inside a mouse
eye. Biomedical Microdevices, 14(1), 207-215.
[Dubois.2001] Dubois, M., & Dubois, M. (2001). Stress and piezoelectric
properties of aluminum nitride thin films deposited onto metal
electrodes by pulsed direct current reactive sputtering. Journal of
Applied Physics, 89(11), 6389-95.
109
[Dubois.1999] Dubois, M., Muralt, P., & Plessky, V. (1999). BAW resonators
based on aluminum nitride thin films. IEEE Proceedings
Ultrasonics Symposium , 2. pp. 907-910.
[Dubois.1999b] Dubois, M. & Muralt, P. (1999b) Properties of aluminum nitride
thin films for piezoelectric transducers and microwave filter
applications. Applied Physics Letters, 74, 20, 3032-3034.
[Eaton.1997] Eaton, W., & Smith, J. (1997). Micromachined pressure sensors:
Review and recent developments. Smart Materials and
Structures, 6(5), 530.
[Elmazria.2003] Elmazria, O., Mortet, V., El Hakiki, M., Nesladek, M., & Alnot,
P. (2003). High velocity SAW using aluminum nitride film on
unpolished nucleation side of free-standing CVD diamond. IEEE
Transactions on Ultrasonics, Ferroelectrics, and Frequency
Control, 50(6), 710.
[Emanetoglu.1999] Emanetoglu, N., Gorla, C., Liu, Y., Liang, S., & Lu, Y. (1999).
Epitaxial ZnO piezoelectric thin films for saw filters. Materials
Science in Semiconductor Processing, 2(3), 247-252.
[Engelmark.2002] Engelmark, F. (2002). AlN and high-k thin films for IC and
electroacoustic applications. Doctoral Thesis. Acta Universitatis
Upsaliensis. Upsala University,
[Epcos.2013] Epcos Inc. (2013). Surface acoustic waves component for RF
control systems. www.epcos.com.
[Every.2002] Every, A. (2002). Measurement of the near-surface elastic
properties of solids and thin supported films. Measurement
Science Technology, 13, R21.
[Fachberger.2006] Fachberger, R., Bruckner, R., Hauser, R., & Reindl, L. (2006).
Wireless SAW based high-temperature measurement systems.
IEEE International Frequency Control Symposium and
Exposition, pp. 358-367.
[Fahmy.1973] Fahmy, A., & Adler, E. (1973). Propagation of acoustic surface
waves in multilayers: A matrix description. Applied Physics
Letters, 22(10), 495-497.
[Fujii.2012] Fujii, S. & Jian, C. (2012). High-Frequency SAW filters based on
diamond films. IEEE Transations on ultrasonics, ferroelectrics
and frequency control, 59, 12, 2758-2764.
110
[Fujii.1997] Fujii, S., Seki, Y., Yoshida, K., Nakahata, H., Higaki, K.,
Kitabayashi, H., & Shikata, S. (1997). Diamond wafer for SAW
application. IEEE Proceedings Ultrasonics Symposium, 1. pp.
183-186 vol.1.
[George.2009] George, T., Islam, M., & Dutta, A. (2009). Large-area low-
temperature ultrananocrystalline diamond (UNCD) films and
integration with CMOS devices for monolithically integrated
diamond MEMS/NEMS-CMOS systems. Micro- and
Nanotechnology Sensors, Systems, and Applications, 7318,
731817-1-731817-7.
[Gerfers.2010] Gerfers, F., Kohlstadt, P., Ginsburg, E., He, M., Rubio, D.,
Manoli, Y., & PengWang, L. (2010). Sputtered AlN thin films for
piezoelectric MEMS devices - FBAR resonators and
accelerometers. Solid state circuits technologies InTech.
[Giacovazzo.2002] Giacovazzo, C., Monaco, H., Artioli, G., Viterbo, D., Ferraris, G.,
Gilli, G., Zanotti, G., & Catti, M. (2002). Fundamentals of
crystallography Oxford University Press.
[Goleniewski.2006] Goleniewski, L., (2006), Telecommunications essentials. Pearson
education Inc.
[Grove.1852] Grove, W. (1852). On the electro-chemical polarity of gases.
Philosophical Transactions of the Royal Society of London, 142,
87-101.
[Hamidon.2006] Hamidon, M., Skarda, V., White, N., Krispel, F., Krempl, P.,
Binhack, M., & Buff, W. (2006). High-temperature 434 MHz
surface acoustic wave devices based on GaPO/sub 4/. IEEE
Transactions on Ultrasonics, Ferroelectrics and Frequency
Control, 53(12), 2465-2470.
[Hamsch.2004] Hamsch, M., Hoffmann, R., Buff, W., Binhack, M., & Klett, S.
(2004). An interrogation unit for passive wireless SAW sensors
based on fourier transform. IEEE Transactions on Ultrasonics,
Ferroelectrics and Frequency Control , 51(11), 1449-1456.
[Hara.2005] Hara, M., Kuypers, J., Abe, T., & Esashi, M. (2005). Surface
micromachined AlN thin film 2 GHz resonator for CMOS
integration. Sensors and Actuators A: Physical, 117(2), 211-216.
[Hashimoto.2013] Hashimoto, K., Sato, S., Teshigahara, A., Nakamura, T., & Kano,
K. (2013). High-performance surface acoustic wave resonators in
the 1 to 3 GHz range using a ScAlN/6H-SiC structure. IEEE
111
Transactions on Ultrasonics, Ferroelectrics and Frequency
Contro, 60(3), 637-642.
[Hashimoto.2000] Hashimoto, K. (2000). Surface acoustic wave devices in
telecommunications, modelling and simulation. Berlin: Springer.
[Hickernell.1996] Hickernell, F. (1996). Measurement techniques for evaluating
piezoelectric thin films. IEEE Proceedings Ultrasonics
Symposium , 1. pp. 235-242.
[Hornsteiner.1998] Hornsteiner, J., Born, E., Fischerauer, G., & Riha, E. (1998).
Surface acoustic wave sensors for high-temperature applications.
Proceedings of the IEEE International Frequency Control
Symposium, 1998., pp. 615-620.
[Ingebrigtsen.1969] Ingebrigtsen, K. (1969). Surface waves in piezoelectrics. Journal
of Applied Physics, 40(7), 2681-2686.
[Iriarte.2012] Iriarte, G., Rodriguez-Madrid, J., & Calle, F. (2012). Fabrication
of sub-100 nm IDT SAW devices on insulating, semiconducting
and conductive substrates. Journal of Materials Processing
Technology, 212(3), 707-712.
[Iriarte.2011] Iriarte, G., Rodriguez, J., & Calle, F. (2011b). Effect of substrate
target distance and sputtering pressure in the synthesis of AlN
thin films. Microsystem Technologies, 17(3), 381-386.
[Iriarte.2010] Iriarte, G., Rodríguez, J., & Calle, F. (2010). Synthesis of c-axis
oriented AlN thin films on different substrates: A review.
Materials Research Bulletin, 45(9), 1039-1045.
[Iriarte.2003a] Iriarte, G. (2003a). AlN thin film electroacustic devices. Doctoral
thesis. Uppsala University,
[Iriarte.2003b] Iriarte, G. (2003b). Surface acoustic wave propagation
characteristics of aluminum nitride thin films grown on
polycrystalline diamond. Journal of Applied Physics, 93(12),
9604-9609.
[Iriarte.2003c] Iriarte, G., Engelmark, F., Katardjiev, I., Plessky, V., & Yantchev,
V. (2003c). SAW COM-parameter extraction in AlN/diamond
layered structures. IEEE Transactions on Ultrasonics,
Ferroelectrics and Frequency Control, 50(11), 1542-1547.
[Iriarte.2003d] Iriarte, G., Engelmark, F., Ottosson, M., & Katardjiev, I. (2003d).
Influence of deposition parameters on the stress of magnetron
112
sputter-deposited AlN thin films on Si(100) substrates. Journal of
Materials Research, 18(7), 1733-1733.
[Iriarte.2002] Iriarte, G., Engelmark, F., & Katardjiev, L. (2002). Reactive
sputter deposition of highly oriented AIN films at room
temperature. Journal of Materials Research, 17(6), 1469-1475.
[Ishihara.1998] Ishihara, M., Li, S., Yumoto, H., Akashi, K., & Ide, Y. (1998).
Control of preferential orientation of AlN films prepared by the
reactive sputtering method. Thin Solid Films, 316(1–2), 152-157.
[Jinyi.2013] Jinyi, M., Jing, Y., Bo, D., Lu, W., Hongmin, J., & Jie, H. (2013).
Design of SAW specific gas sensor with high sensitivity. Applied
Mechanics and Materials, 303-306, 137-142.
[Jonsson.1999] Jonsson, L., Hedlund, C., Katardjiev, I., & Berg, S. (1999).
Compositional variations of sputter deposited Ti/W barrier layers
on substrates with pronounced surface topography. Thin Solid
Films, 348(1–2), 227-232.
[Kalinin.2007] Kalinin, V., Lohr, R., Leigh, A., & Bown, G. (2007). Application
of passive SAW resonant sensors to contactless measurement of
the output engine torque in passenger cars. IEEE International
Frequency Control Symposium, Joint with the 21st European
Frequency and Time Forum. pp. 499-504.
[Kang.2012] Kang, A., Lin, J., Ji, X., Wang, W., Li, H., Zhang, C., & Han, T.
(2012). A high-sensitivity pressure sensor based on surface
transverse wave. Sensors and Actuators A: Physical, 187(0), 141-
146.
[Kar.2008] Kar, J., Mukherjee, S., Bose, G., & Tuli, S. (2008). Effect of
inter-electrode spacing on structural and electrical properties of
RF sputtered AlN films. Journal of Materials Science.Materials
in Electronics, 19(3), 261.
[Kholkin.1996] Kholkin, A., Colla, E., Tagantsev, A., Taylor, D., & Setter, N.
(1996). Fatigue of piezoelectric properties in pb(zr,ti)O3 films.
[null] Applied Physics Letters, 68, 2577-2579.
[Kline.1983] Kline, G., & Lakin, K. (1983). 1.0‐GHz thin‐film bulk acoustic
wave resonators on GaAs. Applied Physics Letters, 43(8), 750-
751.
[Lakin.2005] Lakin, K. (2005). Thin film resonator technology. IEEE
Transactions on Ultrasonics, Ferroelectrics and Frequency
Control, 52(5), 707-716.
113
[Lakin.2000] Lakin, K., McCarron, KT. & McDonald, JF.
(2000).Temperature compensated bulk acoustic thin film
resonators. IEEE Ultrasonic Symposium. 855
[Lakin.1992] Lakin, K., Kline, G., & McCarron, K. (1992). Thin film bulk
acoustic wave filters for GPS. IEEE Proceedings Ultrasonics
Symposium, 1. pp. 471-476.
[Ledbetter.2006] Ledbetter, H. (2006). Sound velocities, elastic constants:
Temperature dependence. Materials Science and Engineering: A,
442(1–2), 31-34.
[Lin.2008] Lin, C., & Lu, F. (2008). Oxidation behavior of AIN films at high
temperature under controlled atmosphere. Journal of the
European Ceramic Society, 28(3), 691-698.
[Maini.2007] Maini, A., & Agrawa, V. (2007). Satellite technology: Principles
and applications. Wiley.
[Malshe.1999] Malshe, A., Park, B. S., Brown, W., & Naseem, H. (1999). A
review of techniques for polishing and planarizing chemically
vapor-deposited (CVD) diamond films and substrates. Diamond
and Related Materials, 8(7), 1198-1213.
[Mauder.1995] Mauder, A. (1995). SAW gas sensors: Comparison between delay
line and two port resonator. Sensors and Actuators B: Chemical,
26(1–3), 187-190.
[Morgan.2000] Morgan, D. (2000). A HISTORY OF SURFACE ACOUSTIC
WAVE DEVICES. World Scientific Publishing Co.
[Morgan.1998] Morgan, D. P. (1998). History of SAW devices. IEEE Intl.
Frequency Control Symposium,
[Mortet.2003] Mortet, V., Elmazria, O., Nesladek, M., D'Haen, J., Vanhoyland,
G., Elhakiki, M., Tajani, A., Bustarret, E., Gheeraert, E.,
D'Olieslaeger, M., & Alnot, P. (2003). Study of aluminium
nitride/freestanding diamond surface acoustic waves filters.
Diamond and Related Materials, 12(3-7), 723-727.
[Mortet.2002] Mortet, V., Elmazria, O., Nesladek, M., Assouar, M.,
Vanhoyland, G., DHaen, J., DOlieslaeger, M., & Alnot, P. (2002).
Surface acoustic wave propagation in aluminum nitride-
unpolished freestanding diamond structures. Applied Physics
Letters, 81(9), 1720.
114
[Nabipoor.2006] Nabipoor, M. & Majlis, BY. (2006). A new passive telemetry LC
pressure and temperature sensor optimized for TPMS. Journal of
Physics: International MEMS Conference. 34, 770–775
[Naik.1999] Naik, R., Reif, R., Lutsky, J., & Sodini, C. (1999). Low-
temperature deposition of highly textured aluminum nitride by
direct current magnetron sputtering for applications in thin-film
resonators. Journal of the Electrochemical Society, 146(2), 691-
696.
[Nakahata.2003] Nakahata, H., Fujii, S., Higaki, K., Hachigo, A., Kitabayashi, H.,
Shikata, S., & Fujimori, N. (2003). Diamond-based surface
acoustic wave devices. Semiconductor Science and Technology,
18, S96.
[Nakahata.1995] Nakahata, H., Higaki, K., Fujii, S., Hachigo, A., Kitabayashi, H.,
Tanabe, K., Seki, Y., & Shikata, S. (1995). SAW devices on
diamond. IEEE Proceedings Ultrasonics Symposium , 1. pp. 361-
370.
[Nishimura.2005] Nishimura, K., Shigekawa, N., Yokoyama, H., & Hohkawa, K.
(2005). Temperature dependence of surface acoustic wave
characteristics of GaN layers on sapphire substrates. Japanese
Journal of Applied Physics, 44, 564-565.
[Nye.1957] Nye, J. (1957). Physical properties of crystals: Their
representation by tensors and matrices. Oxford science
publications.
[Okamoto.2001] Okamoto, M., Khin-Yap, Y., Yoshimura, M., Mori, Y. & Sasaki,
T. (2001). The ohmic character of doped AlN films. Diamond and
Related Materials. 10, 1322-1325.
[Okano.1993] Okano, H., Tanaka, N., Shibata, K., & Nakano, S. (1993). Ghz-
band surface-acoustic-wave devices using aluminum nitride thin-
films deposited by electron-cyclotron-resonance dual ion-beam
sputtering. Japanese Journal of Applied Physics, 32(9B), 4052-
4056.
[Pedros.2011] Pedros, J., Garcia-Gancedo, L., Ford, C., Barnes, C., Griffiths, J.,
Jones, G., & Flewitt, A. (2011). Guided propagation of surface
acoustic waves and piezoelectric field enhancement in ZnO/GaAs
systems. Journal of Applied Physics, 110(10), 103501-103501-7.
[Pedros.2007] Pedros, J. (2007). Dispositivos SAW en estrucutras de nitruros del
grupo III: Propagación y modulación óptica y electrónica.
Doctoral Thesis. Universidad Politécnica De Madrid,
115
[Pedros.2005] Pedrós, J., Calle, F., Grajal, J., Jiménez-Riobóo, R., Takagaki, Y.,
Ploog, K., & Bougrioua, Z. (2005). Anisotropy-induced
polarization mixture of surface acoustic waves in GaN∕ c-sapphire
heterostructures. Physical Review.B, Condensed Matter, 72(7),
075306.
[Plessky.2000] Plessky, V., & Koskela, J. (2000). Coupling-of-modes analysis of
SAW devices. International Journal of High Speed Electronics &
Systems, 10(4), 867.
[Pohl.1998] Pohl, A., & Seifert, F. (1998). New applications of wirelessly
interrogable passive SAW sensors. IEEE Transactions on
Microwave Theory and Techniques, 46(12), 2208-2212.
[Rajamani.2001] Rajamani, A., Beresford, R., & Sheldon, B. (2001). Intrinsic
stress evolution in aluminum nitride thin films and the influence
of multistep processing. Applied Physics Letters, 79(23), 3776-8.
[Rayleigh.1885] Rayleigh, L. (1885). On waves propagated along the plane surface
of an elastic solid. Proc.London Math.Soc., 17, 4-11.
[Rfsaw.2013] Rfsaw Inc., (2013). The global SAW tag- a new technical
approach to RFID. www.rfsaw.com .
[Ricco.1985] Ricco, A., Martin, S., & Zipperian, T. (1985). Surface acoustic
wave gas sensor based on film conductivity changes. Sensors and
Actuators, 8(4), 319-333.
[Ro.2010] Ro, R., Chiang, Y., Sung, C., Lee, R., & Wu, S. (2010).
Theoretical analysis of SAW propagation characteristics in (100)
oriented AlN/diamond structure. IEEE Transactions on
Ultrasonics, Ferroelectrics and Frequency Control, 57(1), 46-51.
[Rodríguez-Madrid.2013]
Rodríguez-Madrid, J. G., Iriarte, G. F., Williams, O. A., & Calle,
F. (2013). High precision pressure sensors based on SAW devices
in the GHz range. Sensors and Actuators A: Physical, 189(0),
364-369.
[Rodríguez-Madrid.2012]
Rodríguez-Madrid, J., Iriarte, G., Araujo, D., Villar, M.,
Williams, O., Müller-Sebert, W., & Calle, F. (2012). Optimization
of AlN thin layers on diamond substrates for high frequency
SAW resonators. Materials Letters, 66(1), 339-342.
116
[Rodriguez-Madrid.2012b]
Rodriguez-Madrid, J., Iriarte, G., Pedros, J., Williams, O. A.,
Brink, D., & Calle, F. (2012). Super-high-frequency SAW
resonators on AlN/Diamond. IEEE Electron Device Letters,
33(4), 495-497.
[Rosenbaum.1988] Rosenbaum, J. (1988). Bulk acoustic wave theory and devices
Artech House.
[Ruby.2004] Ruby, R., Bradley, P., Clark, D., Feld, D., Jamneala, T., & Wang,
K. (2004). Acoustic FBAR for filters, duplexers and front end
modules. IEEE International Microwave Symposium, 2. pp. 931-
934.
[Ruby.2001] Ruby, R., Bradley, P., Oshmyansky, Y., Chien, A., & Larson, J.
(2001). Thin film bulk wave acoustic resonators (FBAR) for
wireless applications. IEEE Ultrasonics Symposium , 1. pp. 813-
821.
[Ruby.1994] Ruby, R., & Merchant, P. (1994). Micromachined thin film bulk
acoustic resonators. Proceedings of the IEEE International
Frequency Control Symposium, pp. 135-138.
[Saravanan.2005] Saravanan, S., Keim, E. G., Krijnen, G., & Elwenspoek, M.
(2005). Growth and surface characterization of piezoelectric AlN
thin films on silicon (100) and (110) substrates. Microscopy of
Semiconducting Materials, 107, 75-78.
[Schimetta.2000] Schimetta, G., Dollinger, F., & Weigel, R. (2000). A wireless
pressure-measurement system using a SAW hybrid sensor. IEEE
Transactions on Microwave Theory and Techniques, 48(12),
2730-2735.
[Schmidt.2000] Schmidt, F., & Scholl, G. (2000). Wireless SAW identification
and sensor systems. International Journal of High Speed
Electronics and Systems, 10(04), 1143-1191.
[Scintag.1999] Scintag, I. (1999). Basics of X-ray diffraction. (pp. 4-25)
http://www.scintag.com/.
[Sedhain.2009] Sedhain, A., Du, L., Edgar, J., Lin, J., & Jiang, H. (2009). The
origin of 2.78 eV emission and yellow coloration in bulk AlN
substrates. Applied Physics Letters, 95(26), 262104-262104-3.
117
[Seifert.1994] Seifert, F., Bulst, W., & Ruppel, C. (1994). Mechanical sensors
based on surface acoustic waves. Sensors and Actuators A:
Physical, 44(3), 231-239.
[Setter.2005] Setter, N. (2005). Electroceramic-based MEMS: Fabrication-
technology and applications (Chapter 10) Springer Science and
Business Media.
[Sezawa.1927] Sezawa, K. (1927). Dispersion of elastic waves propagated on the
surface of stratified bodies and on curved surfaces. Bulletin
Earthquake Research InstituteTokyo, 3, 1-18.
[Shih.2010] Shih WC., Su HY. & Wu MS., (2010). Surface acoustic wave
properties of aluminum oxide films on lithium niobate. Thin Solid
Films, 518, 7143.
[Shiosaki.1982] Shiosaki, T., Yamamoto, T., Oda, T., & Kawabata, A. (1980).
Low‐temperature growth of piezoelectric AlN film by rf reactive
planar magnetron sputtering. Applied Physics Letters, 36(8), 643-
645.
[Stoney.2012] Stoney, R., Donohoe, B., Geraghty, D., & O‟Donnell, G. (2012).
The development of surface acoustic wave sensors (SAWs) for
process monitoring. Procedia CIRP, 1(0), 569-574.
[Strite.1993] Strite, S., Lin, M., & Morkoç, H. (1993). Progress and prospects
for GaN and the III–V nitride semiconductors. Thin Solid Films,
231(1–2), 197-210.
[Takagaki.2004] Takagaki, Y., Hesjedal, T., Brandt, O., & Ploog, K. (2004).
Surface-acoustic-wave transducers for the extremely-high-
frequency range using AlN/SiC (0001). Semiconductor Science
and Technology, 19, 256.
[Takagaki.2002] Takagaki, Y., Santos, P., Wiebicke, E., Brandt, O., Schönherr, H.,
& Ploog, K. (2002). Superhigh-frequency surface-acoustic-wave
transducers using AlN layers grown on SiC substrates. Applied
Physics Letters, 81(14), 2538-2540.
[Tjiu.2004] Tjiu, W., Ahanchian, A., & Majlis, B. (2004). Development of
tire condition monitoring system (TCMS) based on MEMS
sensors. IEEE International Conference on Semiconductor
Electronics.
[Tonisch.2006] Tonisch, K., Cimalla, V., Foerster, C., Romanus, H., Ambacher,
O., & Dontsov, D. (2006). Piezoelectric properties of
118
polycrystalline AlN thin films for MEMS application. Sensors
and Actuators A: Physical, 132(2), 658-663.
[Tran.2013] Tran, A., Wunnicke, O., Pandraud, G., Nguyen, M., Schellevis,
H., & Sarro, P. (2013). Slender piezoelectric cantilevers of high
quality AlN layers sputtered on ti thin film for MEMS actuators.
Sensors and Actuators A: Physical, accepted and pending to be
published(0), 1-18.
[Tsubouchi.1981] Tsubouchi, K., Sugai, K., & Mikoshiba, N. (1981). AlN material
constants evaluation and SAW properties on AlN/Al2O3and
AlN/Si. Ultrasonics Symposium, pp. 375-380.
[Tushar.2012] Tushar, S., Sang-Soo, J., Brijesh, G., & John, Z. (2012).
Patterning piezoelectric thin film PVDF–TrFE based pressure
sensor for catheter application. Sensors and Actuators A:
Physical, 177(0), 87-92.
[Varshni.1970] Varshni, Y. (1970). Temperature dependence of the elastic
constants. Physical Review B, 2, 3952-3958.
[Venema.1986] Venema, A., Nieuwkoop, E., Vellekoop, M., Nieuwenhuizen, M.,
& Barendsz, A. (1986). Design aspects of saw gas sensors.
Sensors and Actuators, 10(1–2), 47-64.
[Vossen.1991] Vossen, J., & Kern, W. (1991). Thin film processes II. Academic
Press Limited.
[Wang.2013] Wang, C., Mao, Y., & Zeng, X. (2013). Properties of ITO-AZO
bilayer thin films prepared by magnetron sputtering for
applications in thin-film silicon solar cells - Springer-Verlag.
[Weigel.2002] Weigel, R., Morgan, D., Owens, J., Ballato, A., Lakin, K.,
Hashimoto, K., & Ruppel, C. (2002). Microwave acoustic
materials, devices, and applications. IEEE Transactions on
Microwave Theory and Techniques, 50(3), 738-749.
[Werner.2001] Werner, M., & Fahrner, W. (2001). Review on materials,
microsensors, systems and devices for high-temperature and
harsh-environment applications. IEEE Transactions on Industrial
Electronics, 48(2), 249-257.
[Wessling.2006] Wessling, B., Mokwa, W., & Schnakenberg, U. (2006). RF-
sputtering of iridium oxide to be used as stimulation material in
functional medical implants. Journal of Micromechanics and
Microengineering, 16(6), 142.
119
[White.1965] White, R., & Voltmer, F. (1965). Direct piezoelectric coupling to
surface elastic waves. Applied Physics Letters, 7(12), 314-317.
[Wu.2001] Wu, P., Emanetoglu, N., Tong, X., & Lu, Y. (2001). Temperature
compensation of SAW in ZnO/SiO2/Si structure. IEEE
Ultrasonics Symposium , 1. pp. 211-214.
[Xin.2012] Xin, L., Qin, L., Shixin, P., Kaixian, X., Hui, T., & Chensong, S.
(2012). High-temperature piezoresistive pressure sensor based on
implantation of oxygen into silicon wafer. Sensors and Actuators
A: Physical, 179(0), 277-282.
[Yim.1973] Yim, W., Stofko, E., Zanzucchi, P., Pankove, J., Ettenberg, M., &
Gilbert, S. (1973). Epitaxially grown AlN and its optical band
gap. Journal of Applied Physics, 44(1), 292-296.
[You.2007] You, Y., & Kim, D. (2007). Influence of incidence angle and
distance on the structure of aluminium nitride films prepared by
reactive magnetron sputtering. Thin Solid Films, 515(5), 2860-
2863.
[Zhang.2005] Zhang, J., Cheng, H., Chen, Y., Uddin, A., Yuan, S., Geng, S., &
Zhang, S. (2005). Growth of AIN films on si(100) and si(111)
substrates by reactive magnetron sputtering. Surface & Coatings
Technology, 198(1-3), 68-73.
[Zukowski.2012] Zukowski, E., Fellner, T., Wilde, J., & Berndt, M. (2012).
Parameter optimization of torque wireless sensors based on
surface acoustic waves (SAW). International Conference on
Thermal, Mechanical and Multi-Physics Simulation and
Experiments in Microelectronics and Microsystems (EuroSimE),
pp. 1-6.
