Pro Gradu

Technical background, applications and implementation of quartz crystal

microbalance systems

Han Jie

15. September 2006

UNIVERSITY OF JYVÄSKYLÄ DEPARTMENT OF PHYSICS

Acknowledgements

First, I would like to sincerely thank God’s blessing. His strength has allowed

me to endure, and I remain eternally grateful. Then, I would like to sincerely

thank the following individuals who helped me along this journey and made

possible for me to achieve my goal.

I would like to sincerely thank Professor Päivi Törmä, PhD, for giving me an

opportunity to study at the Department of Physics, University of Jyväskylä,

Finland. I appreciate the confidence she has placed on me. Next I would like

to thank research manager Osmo Miinalainen, MSc, (Savonia Polytechnic,

Kuopio, Finland) for giving me a chance to work at the Information

Technology R&D Unit of Savonia Polytechnic, Engineering, Kuopio. I

appreciate that he saw the potential in myself.

I would like to express sincerely thanks to my supervisors, project manager

Pasi Kivinen, PhD, and Mikko Laasanen, PhD, (Savonia Polytechnic), senior

researcher Jussi Toppari, PhD, (University of Jyväskylä), for their inspiring

supervision, patient guidance and friendship throughout the course of this

work. I am very grateful to researcher Asmo Jakorinne, MSc, (Savonia

Polytechnic) for invaluable help with QCM sensing system and data

interpretation. I would like to thank Timo Ollikainen, Eng, (Savonia

Polytechnic) for making the electronic

circuit for the QCM system. I am also grateful to other members of Savonia

Polytechnic and University of Jyväskylä for their guidance. During this study,

they have given me great help in many ways. I sincerely thank all my friends

in Finland and China. I believe that along the road of life every person has

potential to teach us some lesson. To all of you, I really want to express my

dearest thanks.

Finally, I would like to sincerely thank my great parents and younger brother

for providing me many moments of instruction, advice, and encouragement

during studying time. I love them dearly, and I truly believe that I could not

have achieved my dream without their great support, love and understanding.

This thesis is dedicated to them.

Jyväskylä, September 2006 Han Jie

Abstract Jie Han, Technical background, applications and implementation of

quartz crystal microbalance systems, 59 pages, September 2006.

Supervisors: Mikko Laasanen, PhD, Savonia Polytechnic, Kuopio, Finland Pasi Kivinen, PhD, Savonia Polytechnic, Kuopio, Finland Jussi Toppari, PhD, University of Jyväskylä, Jyväskylä,

Finland _______________________________________________________________

Quartz crystal microbalance (QCM) is a system for measuring mass changes

in the nanogram level. The quartz crystal microbalance is based on the fact

any the foreign mass on the oscillating quartz crystal changes the resonant

frequency of it. The quartz crystal is piezoelectric and the cutting angle with

respect to crystal orientation determines the performance of QCM for specific

needs. QCM applications have been introduced for many fields of science and

industry. In this work, especially gas sensing system, bio-sensing system, and

micromachined cantilever sensing system are introduced. The basic

microfabrication techniques for sensors (deposition, patterning, and etching)

are identical with those used in IC fabrication.

In the final part of the work, a custom QCM system was developed and

implemented. It was seen that the decrease of QCM oscillation frequency

correlated linearly with a deposition of gold, with R2=0.989 as an obtained

Pearson's correlation coefficient for 11 measurement points. The 8 MHz

crystal failed to oscillate when the excess mass was over 2000 ng/cm2.

Reproducibility of the measurements, as indicated by the coefficient of

variation, was high (0.003%, RMS average). In many applications the cleaning

of the QCM sensor surface is highly challenging and limits its commercial

use. Therefore, several recent studies have been introduced about this issue.

1

Table of contents

Glossary……………………………………………………………………..1

1. Introduction of QCM..........…………………………………………2

1.1 History of QCM…………………………………………………….…2

1.2 Structure of QCM sensor………………………………………….…3

1.3 Electronic properties of quartz and electromechanical

Coupling………………………………………………………………..7

1.4 Calculation of circuit parameter and load impedance………….…11

2. QCM applications…………………….……………………………...13

2.1 Gas sensing system……………………………………….………….15

2.2 Bio-sensing system...............................................................................18

2.3 Alternative microbalance systems: Micromachined

Cantilevers...........................................................................................20

3. Fabrication technology of sensors………......................................25

3.1 Mircofabrication process...................................................................26

3.1.1 Substrate materials....................................................................27

3.1.2 Thin film deposition...................................................................28

3.1.3 Patterning....................................................................................33

3.1.4 Etching........................................................................................38

2

4. Materials and methods.......................................................................41

4.1 QCM measurement setup...................................................................41

4.2 Light microscopy and scanning electron microscopy......................46

4.3 Analyses of measurement reproducibility........................................46

5. Results......................................................................................................48

6. Discussion................................................................................................53

6.1 Conclusion...........................................................................................53

6.2 Future works.......................................................................................54

Reference.....................................................................................................56

1

Glossary ALD Atomic Layer Deposition

ANN Artificial Neural Networks

BVD Butterworth van-Dyke

BAW Bulk Acoustic Resonator

CVD Chemical Vapor Deposition

DRIE Deep Reactive Ion Etching

GC Gas Chromatography

LEM Lumped-Element Model

LPCVD Low Pressure Chemical Vapor Deposion

PECVD Plasma Enhanced Chemical Vapor Deposion

PVD Physical Vapor Deposition

QCM Quartz Crystal Microbalance

REVS Rupture Event Scanning

RIE Reactive Ion Etching

SAM Self-Assembled Monolayer

TSM Thickness Shear Mode

VOC Volatile Organic Compound

2

1. Introduction to QCM

1.1 History of QCM

In 1880, the brothers Pierre and Jacques Curie did the first experiment which

was related to the piezoelectric effect. They applied pressure to the certain

crystallographic direction of the so called Rochelle salt (sodium potassium

tartrate tetrahydrate) and noticed that by doing that the crystals of Rochelle

salt were able to produce electricity. Quartz, tourmaline topaz and cane sugar

were tested as well. However, in Rochelle salt (and in quartz) the electrical

polarization was strongest. The inverse effect, i.e. production of strain by

application of electricity, was shown one year later.

In 1917, piezoelectric quartz crystals were used as transducers and receivers of

ultrasound in water. The actual application was sonar for detecting

submarines. In 1919, loudspeakers, microphones and sound pick-ups based on

piezoelectricity of Rochelle salt were developed. In 1921, the first X-cut

quartz crystal controlled oscillators were described, which unfortunately had

the drawback of being very temperature sensitive. Nowadays the X-cut

crystals are used only in places where the large temperature coefficient is of

little importance, such as transducers in space sonar [1].

AT-cut quartz crystal was developed in 1934. The major advantage of the

AT-cut quartz is that it has nearly zero frequency drift around the room

temperature. AT cut quartz crystal came up in all kind of frequency control

3

applications. In the very beginning the application was just a frequency control

element [1].

In 1959, Sauerbrey showed one very important property about the quartz

crystal. He showed that the resonance frequency shift of the crystal, observed

earlier by many researchers, is proportional to addition of the mass on the

quartz crystal [2]. That was the first step towards a new quantitative tool for

measuring very tiny mass changes by quartz crystal microbalance system.

In this study, the operating principle and application of the QCM are

introduced. Today the quartz crystal sensors can measure mass changes,

pressure, force, acceleration, temperature and viscosity.

1.2 Structure of QCM sensor

Quartz is a crystalline form of silicon dioxide (SiO2). It is a hard, brittle,

transparent material with a density of 2649 kg/m3 and a melting point of

around 1650 ° C. Quartz is insoluble in ordinary acids, but soluble in

hydrofluoric acid and in hot alkalis. When quartz is heated to 573° C, its

crystalline form changes. The stable form above this transition temperature is

known as high-quartz or beta-quartz, while the stable form below 573° C is

known as low-quartz or alpha-quartz. For resonator applications, only alpha-

quartz is of interest and unless stated otherwise the term quartz in the sequel

always refers to alpha-quartz [3]. Alpha-quartz crystals are employed because

of their superior mechanical and piezoelectric properties.

4

Quartz crystal microbalance, also known as working in thickness shear mode

(TSM), is a typical bulk acoustic resonator (BAW) type ultra sensitive mass

sensor, and traditionally used for thin film deposition control. In quartz crystal,

the cutting angle with respect to crystal orientation determines the mode of

oscillation. Different angles of orientation can be machined for enhancing

performance for the specific needs [4]. Typical special cuts would include AT-

cut crystals, BT-cut crystals, SC-cut crystals, IT-cut crystals, and FC-cut

crystals, etc. as shown in Fig. 1. For instance, AT-cut crystal is fabricated by

slicing through a quartz rod with a cutting angle 35°10' with respect to the

optical axis, and this crystal performs shear displacement perpendicular to the

resonator surface. BT-cut quartz wafer is cut at an angle of 49°00’ with

respect to the optical axis. SC-cut crystal is cut at an oblique angle with

respect to two crystal axes (doubly rotated), etc [5].

Fig. 1 In the photo on left, the natural quartz crystal is shown. In the illustration on right, different cutting angles for quartz crystal are shown. In quartz crystal, the cutting angle with respect to crystal orientation determines the mode of oscillation. Different angles of orientation are applied for enhancing performance in the specific needs [6].

5

In practice, QCM sensor consists of a thin disk, sliced from a single crystal of

an alpha-quartz. The crystal is sandwiched between two metal electrodes

which are vapor deposited on either side of the crystal (Fig. 2).

Fig. 2 Structure of the QCM sensor. Quartz crystal is sandwiched between two metal electrodes that are vapor deposited on either side of the crystal. When the AC voltage is applied, quartz crystal will oscillate. Analyte molecules will be attached on the coated surface of QCM and the mass of the crystal increases, and therefore the resonance frequency of the quartz crystal changes [7]. If measuring the impedance of the crystal as a function of frequency, one

observes that the impedance is lowest at a certain resonant frequency. Due to

this, when used in an electronic oscillator circuit, a quartz crystal’s natural

resonant frequency determines the frequency of oscillation of the circuit. Any

mass addition or adsorption to the quartz crystal surface will change the

resonance frequency. This oscillation is generally very stable due to the high

quality of the resonance (high Q factor) [1]. Gold electrodes have been the

most commonly used in QCM studies, because of the ease evaporation of

Au. However, Cu, Ni, Pt and other metals have also been employed.

Theoretical background for using quartz crystals as mass sensors take origin in

the Sauerbrey equation [2, 8]. It expresses the correlation between the change

6

of resonant frequency (∆f) of a piezoelectric quartz crystal with the change of

mass attached to the crystal. Sauerbrey equation is expressed as follows:

21

20

)(

2

qqA

mff

μρ

Δ−=Δ . (1)

where f0 is the initial resonant frequency of the crystal, ∆m is

deposited/absorbed mass, A is the active area of the crystal (between the

electrodes), ρq = 2.648g/cm3 is the density of quartz and μq =

2,947×1011g/(cms2) is the shear modulus of quartz.

The Sauerbrey equation was developed for oscillation in air and only applies

to rigid mass attached to the crystal. The QCM was just regarded as gas-phase

mass detector for many years until 1980, when Kanazawa and co-workers did

liquid phase QCM measurements and realized that a quartz crystal could be

excited to a stable oscillation when it was completely immersed in a liquid

[9]. They showed that the change in resonant frequency of a QCM taken from

air into a liquid is proportional to the square root of the liquid’s density-

viscosity product:

21

23

0 )(qq

LLffμπρμρ

−=Δ . (2)

where fΔ is a measured frequency shift, 0f is the resonant frequency of the

unloaded crystal, Lρ is the density of liquid in contact with the crystal, and Lμ

is the viscosity of liquid in contact with the crystal.

7

jXRZ +=

It was found out that an excessive viscous loading would not prohibit the use

of the QCM in liquids, and that the response of the QCM is still extremely

sensitive to the mass changes. QCMs have been used in direct contact with

liquids and/or visco-elastic films to assess changes in mass and visco-elastic

properties. Even in air or vacuum, where the damping of layers has been

considered to be negligible or small, the QCM has been used to probe

dissipative processes on top of the quartz crystal. This is especially true for

soft condensed matters such as thick polymer layers deposited on the quartz

surface [1].

1.3 Electrical properties of quartz and electromechanical coupling

The electrical properties of a quartz crystal can be described by means of a

Butterworth van-Dyke (BVD) model of lumped electrical parameters also

known as the Lumped-Element Model (LEM) and shown in Fig. 3. R1

represents the losses in the crystal, frictional, or thermal; L1 represents the

inertial mass of the crystal; C1 is a measure of the energy stored in the crystal

during each oscillation, and C0 is the static plus stray capacitance. The

complex impedance of a quartz crystal can be represented by the following

relationship [10]:

. (3)

8

11

11

12

12C

LfC

fLXω

ωπ

π −=−=

Fig. 3 The Butterworth van-Dyke (BVD) or lumped element model (LEM) for a quartz crystal showing the individual electrical components. R1 represents the losses, L1 is inertial mass, C1 is measurement of energy stored, C0 is static plus stray capacitance [10].

a the reactive or the imaginary part of the complex crystal impedance (X) is

formed by frequency dependent components, i.e., inductive and capacitive

impedance together [10]:

. (4)

where ω is the angular frequency (radians/sec). The ratio of the reactive part to

the resistive part is defined as the phase angle (θ). At a particular frequency,

the inductive and the capacitive contributions of the motional arm are equal,

(the motional arm is composed of the L1, C1 and R1 values of the crystal and

are referred to as motional parameters [11] ), and the reactive part reduces to

zero, thereby minimizing the impedance of the crystal just to the resistive

losses. This frequency is defined as the series resonant frequency fs as at this

frequency the phase angle is zero and impedance of the motional arm is at a

minimum (Z = R1). The voltage and the current are in phase at this frequency.

Similar to crystal impedance, its inverse, admittance, can also be used to

describe the quartz crystal. The series resonant frequency is characterized by a

9

jBGY +=

maximum in the frequency-admittance plot, where admittance can be written

as [10]

. (5)

where Y is admittance, G is conductance, and B is susceptance. At series

resonant frequency just as both phase and reactance reduce to zero, so does the

susceptance. At this point the conductance equals 1/R1 .

By means of impedance analysis, one can determine quantitatively how shifts

in the circuit parameters, upon loading, relate to the physical properties of the

load. As a result of loading the quartz crystal electrode with liquid, the

electrical parameters of the crystal are modified and can be represented as a

modified BVD model (Fig. 4). The changes in the crystal capacitance on

loading are extremely small and are not considered for liquid loadings [12].

Fig. 4 Modified Butterworth van-Dyke model for a loaded quartz crystal showing additional impedance elements imposed due to placement of load on the surface of the quartz crystal electrodes. Zm

U refers to the electrical impedance of the motional arm

under condition of no load and Zm L

refers to the electrical impedance imposed due to the load [10].

The principle of using a quartz crystal for measurement of fluid rheology is

based on equation (2) via the viscosity and the density of the liquid. The

10

⎟⎟⎠

⎞⎜⎜⎝

⎛=+=

Q

Ls

s

Lm Z

ZCK

NjXRZ0

222 4 ωπ

equation (2) is obtained from a simple physical model which couples the shear

wave in the quartz to a damped shear wave in the fluid [13, 14]. R2 and L2

obtained after impedance analysis are used to calculate the mechanical

impedance of the load (ZsL), which is defined as the ratio of the surface stress

to particle velocity at the surface with units of gm/(scm2). Just like electrical

impedance of the motional arm (ZmL

), ZsL is also a complex quantity consisting

of a real lossy (viscous) part and an imaginary storage (inertial mass/storage)

part. ZmL and Zs

L (See Fig. 4) are related through a constant (A) comprising of

the electromechanical coupling constants, the crystal fundamental resonance

frequency and mode, piezoelectrically active area, and unloaded quartz

impedance as follows [15]:

.

(6)

where R2 is the real component of the imposed electrical impedance due to the

surface load, X2 = ωL2 is the imaginary component of that, N is the overtone

number, K2 = 7.74 × 10-3 is the piezoelectric coupling constant

(dimensionless), ωs is the series resonant frequency (radians/sec), C0 is the

static plus stray capacitance, and ZQL is the quartz mechanical impedance

which equals (µqρq)1/2 =8.839 × 105 gm/(scm2). The constant term for a

specific frequency crystal, i.e., 0

24 CKN

sωπ is grouped together to give the

constant A. Thus, R2 is proportional to the real part of the mechanical

11

2/12

220

2

22 )1(

1

)1(⎥⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢⎢

⎣

⎡

⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜

⎝

⎛

−+

−−+

⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜

⎝

⎛

−+=

mmm

mm

mmm

m

CLR

CL

C

CLR

RY

ωω

ωω

ω

ωω

impedance ( )Re(2 sZR ∝ )and X2 is proportional to the imaginary part of

mechanical impedance ( )Im(2LsZX ∝ ) through the constant

QZA [16].

1.4 Calculation of circuit parameters and load impedance

The total admittance of the circuit, with or without excess load, is presented as

mZCjY /10 += ω . (7)

and its magnitude is given by the following relationship [12]:

. (8)

where Rm = R1, Lm = L1, and Cm = C1 for unloaded crystal and Rm = R1 + R2,

Lm = L1 + L2, and Cm = C1 for loaded crystal. The modified BVD model (Fig.

4, and Equation 5) can be fitted to the admittance data to obtain the

parameters. The parameters C1 and C0 calculated for unloaded crystal can be

used as such during fitting under loaded condition as they are assumed to

undergo no change [17].

The surface electrical impedance arising due to deposition of a rigid mass

layer is given by [15]

12

2/1

22 2 ⎟⎟⎠

⎞⎜⎜⎝

⎛== LiqLiq

QZAXR

ηωρ

sQ

Lm Z

AfLXZ ωρπ === 22 2 . (9)

where ρs is density of coating, and the surface electrical impedance parameters

for a crystal loaded with a Newtonian liquid are given by [10]

. (10)

where ρLiq is density of liquid and ηLiq is viscosity of liquid. The decay length

or the penetration depth (boundary layer) of the shear wave in a liquid media,

δ, defined as the length at which the amplitude reduces to 1/e of the initial

amplitude, is proportional to the fluid kinematics viscosity [17].

13

2. QCM applications

QCM has so high sensitivity that it can measure single layer of atoms or even

mass changes related to a fraction of the mono-layer. The high sensitivity and

the real-time monitoring of mass changes on the sensor crystal makes QCM a

very attractive technique for a large range of applications. Especially, the

development of QCM systems for fluids or visco-elastic deposits has

dramatically increased the interest towards this technique. A major advantage

of the QCM technique used for liquid systems is that it allows a direct

electrical detection of small molecules [18]. A partial list of the application

areas of the QCM is shown below, and it seems that the application areas are

only limited by one’s imagination (See in Table 1).

14

Table 1: Some applications and example of QCM

Application area Examples

Thin Film thickness monitoring

e-beam evaporation, sputtering, magnetron, ion and laser deposition.

Electrochemistry of interfacial processes

electrode surfaces

Biotechnology 1. Interactions of DNA and RNA with complementary strand. 2. Specific recognition of protein ligands by immobilized receptors, immunological reactions. 3. Detection of virus capsids, bacteria, mammalian cells. Adhesion of cells, liposomes and proteins. 4. Biocompatibility of surfaces. 5. Formation and prevention of formation of biofilms.

Functionalized surfaces 1. Creation of selective surface. 2. Lipid membrabces. 3. Polymer coatings. 4. Reactive surfaces. 5. Gas sensors. 6. Immunosensors.

Thin film formation 1. Langmuir and Langmuir-Blodgett films. 2. Self-assembled monolayers. 3. Polyelectrolyte adsorption. 4. Spin coating. 5. Bilayer formation. 6. Adsorbed monolayers. 7. ALD

Surfactant research 1. Surfactant interactions with surface. 2. Effectiveness of surfactants.

Drug Research 1. Dissolution of polymer coatings. 2. Molecular interaction of drugs. 3. Cell response to pharmacological substances. 4. Drug delivery.

Liquid Plating & Etching Plating, Etching In situ monitoring

lubricant and petroleum properties

15

2.1 Gas sensing systems

Fig. 5 Electronic nose consists of quartz crystal microbalance (QCM) sensor array. The idea is based on human olfactory system. The crystals were exposed to a stream of nitrogen gas until obtained stable frequencies. Then a liquid sample was injected into the chamber. Gas chromatography (GC) triggered to start its sampling. The responses from the crystals were recorded by the software and compared to GC [19].

Fig. 6 Construction of QCM sensor used e.g. for detecting of volatile organic compound (VOC). A polymer coating is applied to the sensor surface to serve as a sensing material. The coating layer can then bind a VOC of interest [20]. The principle of electronic noses is based on human olfactory system [21] and

the system consists of QCM sensor array (Fig. 5). A sensitive polymer film is

applied to the resonating disk of QCM (Fig. 6 and Fig. 7). The coating layer

can bind a specific volatile organic compound (VOC) of interest. Gas

molecules adsorbed to the surface of the polymer coating increase the mass of

16

the disk, thereby reducing its resonance frequency. The reduction is

proportional to odorant mass absorbed by the polymer via equation (1).

Fig. 7 Process of QCM polymer coating, Coating is sprayed on the disk in order to serve as the active sensing material [22].

The sensor typically consists of an array of several crystals, each coated with a

different polymer. This design is aimed to improve identification, although it

is hampered by the limited selectivity of individual films. But employing more

than one crystal, and coating each with a different partially selective polymer,

different responses can be obtained for different gases or compositions of

gases. The combined response of these crystals can then be used as a signature

pattern of the VOC detected [23].

In addition, QCM sensors are remarkably linear over a wide dynamic range.

Their response to water is dependent upon the absorbent materials employed.

Also their sensitivity to changes in temperature can be made negligible [20].

Tailoring the QCM for specific applications is done by adjusting its polymer

coating. Fortunately, a large number of coatings are available from those

developed for Gas Chromatography (GC). The response and recovery times of

17

the resonant structure are minimized by reducing the size and mass of the

quartz crystal along with the thickness of the polymer coating. Batch-to-batch

variability is not a problem because these devices measure the normalized

frequency change [20].

When the dimensions are scaled down to the micrometer level, the surface-to-

volume ratio increases. The larger the surface-to-volume ratio is the noisier the

devices get, because of the surface processes that cause instabilities. Hence,

signal-to-noise ratios degrade with increasing surface-to-volume ratio, thereby

limiting the measurement accuracy. It should be noted that this phenomenon

applies to most of the microfabricated devices [20].

18

2.2 Bio-sensing systems

Analytical devices, which incorporate biological materials or are biomimic

(e.g. tissue, micro-organisms, organelles, cell receptors, enzymes, antibodies,

nucleic acids etc.), are intimately associated with a physicochemical

transducer or transducing microsystem, or are even integrated within these.

These transducers can be optical, calorimetric, acoustic, electrochemical, or

magnetic type [24]. When a biological material layer is exposed to an analyte

and produces a physicochemical change, the transducer detects the changes

and generates an electronic signal proportional to the concentration of the

analyte (Fig. 8).

Fig. 8 A generalized structure of a biosensor. When biological material layer is exposed to an analyte and produces a physicochemical change, the transducer detects the change and generates an electronic signal proportional to the concentration of the analyte [24].

For example, the QCM can be used to detect the acoustic noise when the

interactions are broken. This process is termed Rupture Event Scanning

(REVS) and it can be used to measure the adhesion forces between a surface

and small particles. Sensitivity of the REVS approaches detection of a single

virus particle [6].

19

Fig. 9 Representation of the Rupture Event Scanning (REVS). REVS can be used to measure the adhesion forces between a surface and small particles. Sensitivity of the REVS approaches detection of a single virus particle [25].

In Fig. 9 an example of REVS system is shown. The chromium layer, gold

layer and Self-Assembled Monolayer (SAM) are coated onto a QCM disk, and

at the top of that antibody layer that mediates specific attachment of the virus.

The disk is then transversely oscillated by applying an alternating voltage to

the gold on either side of the disk. The alternating voltage is then

monotonously increased, resulting in a transverse oscillation of greater

amplitude, which, in principle, leads to greater inertial forces between the

virus and the surface, and concomitant deformation of the surface and the

virus. Ultimately this leads to a bond breakage, and the elastic energy, due to

deformation of the surface and of the virus, is then partly converted to acoustic

energy. Some of this acoustic energy propagates along the surface of the QCM

20

disk, and the resultant acoustic emission can be detected by the QCM acting as

a sensitive microphone [25].

2.3 An alternative microbalance system: micromachined cantilevers

QCM is not the only microbalance system. For comparison we briefly discuss

the cantilever based microbalance system. Micromechanical cantilever arrays

can be used as chemical sensors which can simultaneously detect the resonance-

frequency and bending. Information on cantilever bending and resonance-

frequency shifts during exposure to analyte vapor can be used for qualitative and

quantitative characterization and recognition of a variety of chemical

substances, such as water, primary alcohols, and alkanes. Based on the pattern

of cantilever responses, the setup can be used for complex analytes. The

discrimination power can be greatly enhanced by the use of neural network

techniques. Application areas of such sensors are mainly in quality and process

control. Recently the large potential of micromachined cantilever sensors to

study protein adsorption, anti-body-antigen recognition, DNA hybridization, and

rapid medical diagnostics in liquids, has been pointed out [26, 27, 28].

Instead of detecting changes in electrical properties of sensor materials, one

measures the mechanical response of thin beams of silicon, microcantilevers,

which are arranged in a micro fabricated array (Fig. 10). Each of these one-

side coated cantilevers with a sensor layer coated on one side, shows an

individual response to analyte molecules. The sensor layers can transduce

changes of physical properties or energy transfers related to chemical reaction

21

into a mechanical response. When the sensor layer is exposed to an analyte,

the cantilever mechanically responses by bending on the nanometer scale

because of surface stress change or heat transfer and by mass change, leading

to a shift in the resonance frequency (dynamic mode). During the absorption

of molecules, also static deflection can occur due to surface stress change at

the interface between cantilever and polymer layer (static mode) [26]. Also

heat transfer due to e.g. chemical reaction or phase transitions or bimetallic

effect can occur in addition to the static deflection (Fig. 11) [26].

Fig. 10 Scanning electron micrograph of a cantilever array, i.e., microfabricated sensor consisting of 8 cantilevers The length of each cantilever is 500 μm, thickness 0.5 μm, and the pitch between two cantilevers is 250 μm. The sensor was made by IBM’s Zurich Research Laboratory [29].

22

Fig. 11 a) Bimetallic effect occurs in addition to the static deflection. b) Dynamic mode, absorption of analyte molecules in a sensor layer leads to a shift in the resonance frequency. c) static mode, the cantilever bends owing to adsorption of analyte molecules and change of surface stress at the cantilever surface [29]. In general, the cantilever-based sensor device can operate in vacuum, gas, or

liquid environment with high speed and sensitivity. Geometrical dimensions

characterize a spring constant and resonance frequency of cantilever. The

spring constant can be calculated in the following way [30]:

133 )4( −= CantCantCant lwEtk . (11)

where Cantt , Cantw and Cantl are the thickness, width and length of the cantilever,

respectively, and E is the Young’s modulus of the material used. Eq. (11) is

only valid for Cantl >> Cantt . The resonance frequency of a long and thin

rectangular can be calculated as [30]

ρπE

lt

fCant

Cant20 2

.21

= . (12)

where ρ is the density of the cantilever material.

Absorption of analytical vapor produces stress at the interface between the

cantilever and the sensor layer, thus leading to the bending of cantilever.

Stoney’s law gives the dependence of the surface stress (ơ) on the bending

radius of the cantilever [31]:

23

12 ))1(6( −−= vREtCantσ

)()4( 20

21

1 −−− −=Δ ffnkm π

. (13)

where v is the Poisson’s ratio of the cantilever material and R is the bending

radius of the cantilever. The following material constants have been used for

silicon, ,/107.1 211 mNE ×= 33 /1033.2 mkg×=ρ [26], and v = 0.25. Small

changes in cantilever mass can be detected via the oscillation of the cantilever

at its resonance frequency by using an external piezoelectric crystal. The mass

change depends on the resonance frequency of the oscillation cantilever in the

following way [26]:

. (14)

where k denotes the spring constant of the cantilever, n is a geometry

dependent correction factor (n = 0.24 for rectangular cantilever), and 0f is the

resonance frequency prior to the experiment. Eq. (14) is only valid if the

spring constant k does not change during the experiment. Diffusion of analyte

molecules through the cantilever coating may change the elastic constants of

the cantilever.

At the Micromechanics Department of IBM’s Zurich Research Laboratory,

cantilever sensor arrays have been fabricated from silicon using combined dry

and wet etching techniques. A series of eight cantilevers has been assembled

on the chip. Cantilevers have been optimized for different applications by

choosing different cantilever dimensions. The length of the cantilevers was

selected to be 500 µm, and width 100 µm. For operation in the dynamic mode,

24

a thickness of 8.6 µm was chosen, resulting in a resonance frequency of

approximately 50 kHz. For operation in the static mode, the thickness was

selected to be approximately 1.2 µm. Eight cantilever sensors are linearly

arranged at a pitch of 250 µm [26].

Mirocantilever sensors have been shown to exhibit over two orders of

magnitude greater absolute sensitivity compared to other currently available

sensors such as quartz crystal microbalance. This increase in sensitivity can be

attributed largely to the extremely small size of the sensing element. With an

active area of 1000 μm2, the microcantilevers used in this work have

approximately five orders of magnitude smaller area than, e.g. QCM devices.

The potential advantages due to smaller size include also faster response time,

the possibility to fabricate sensor arrays with small overall dimensions, the

ability to explore microenvironments, and or an improved portability for field

applications [32]. But the manufacturing of microcantilever costs more than

QCM sensors, and therefore, QCMs are better for certain applications.

25

3. Fabrication technology of the sensors

Microfabrication processes are used to produce devices with dimensions from

micrometer to millimeter range, and these processes can be effectively applied

to construct a single device or thousands of devices. The microfabrication

processes hence significantly differ from conventional, sequential machining

processes, such as drilling or milling with mechanical tools. The

micromechanical components are fabricated by using compatible processes

that selectively etch away parts of the silicon wafer or add new structural

layers to form mechanical and electromechanical devices [33].

Standard microfabrication processing steps originating from semiconductor

technology can be used in combination with dedicated micromachining steps

to fabricate three-dimensional (3-D) mechanical structures, which form the

basis for the sensors. Microfabrication techniques can also be used to either

significantly improve sensor characteristics or to develop devices with new

functionality, which cannot be realized in conventional fabrication technology

[34].

The quality of a QCM resonator can be influenced by altering surface

roughness of the quartz crystal. Utilizing the etching technology one can

achieve flat, smooth, and ideal geometrical shape for the QCM sensor. It is

well known that a cross sectional shape etched by wet chemical etching [18] is

dependent on the crystal orientation (especially for small diameter) and the

surface flatness is not so good. Deep Reactive Ion Etching (DRIE) technology

26

can etch the quartz crystal independently of the crystal orientation.

Combination of photolithography technology and dry etch process can provide

high Q factors and smaller diameter of QCM and, thus makes it possible to

miniaturize the QCM [35].

3.1 Microfabrication of Sensors

Three basic microfabrication techniques for sensors are identical with those

used in IC fabrication: deposition, patterning and etching [34].

Fig. 12 Flow diagram of an IC fabrication process using the four basic microfabrication techniques: deposition, photolithography, etching, and doping [34].

Fig. 12 shows a typical flow diagram for IC fabrication. First, a thin layer of

e.g., insulating silicon dioxide film, is deposited on a substrate. A light-

sensitive photoresist layer is then deposited on top of it and patterned using the

UV-light. Finally, the pattern is transferred from the photoresist layer to the

silicon dioxide layer by an etching process. After removing the remaining

photoresist, the next layer is deposited and structured, and so on. Doping of a

semiconductor material by ion implantation can be done directly after

27

photolithography or after patterning an implantation mask e.g., a patterned

(sacrificial) silicon dioxide layer [34].

3.1.1 Substrate Materials

Silicon is the standard substrate material for the IC fabrication and thus the

most common substrate material in microfabrication. It is supplied as single

crystal wafers with diameters from 100 to 300 mm. The use of silicon

substrate material enables the integration of transducers and circuitry, an

advantage which is explored, e.g., in CMOS-based Microsystems [36]. With

favorable electrical properties and excellent mechanical properties, the crystal

silicon enable the design of micromechanical structures [37] and a large

number of micromachining techniques have been developed to structure

silicon substrates [34, 38, 39].

The character of glass is transparency to visible light, so it is therefore suited

for devices with optical detection principle. It can be supplied in wafer form in

different compositions and diameters. A number of micromachining

techniques such as isotropic wet etching or anisotropic dry etching, have been

developed to structure glass. Glasses are chemically inert and suitable for

high-temperature applications [34].

Ceramics have been used as substrate for hybrid microelectronics and are

common in microelectronics packaging [40]. The standard material is alumina

(Al2O3); other materials include BeO and AlN. Most microfabrication

28

techniques for ceramic materials have been adapted from microelectronics

packaging processes [34].

Nowadays, polymers have been more and more explored as an inexpensive

substrate material. Special processes have been developed to structure polymer

materials even in the micrometer range. Due to the cost advantage, disposable

devices, such as microfluidic arrays or microstructured immuno or DNA

assays, are often based on polymers [34].

3.1.2 Thin film deposition

The two most common thin-film deposition methods in microfabrication are

Chemical Vapor Deposition (CVD) and Physical Vapor Deposition (PVD),

such as sputtering and evaporating. Typical CVD and PVD film thicknesses

are tens of nanometers up to a few micrometers. Other techniques include

electroplating of metal films and spin- or spray-coating of polymeric films

such as photoresist. Both processes can yield film thicknesses from less than 1

μm to several hundred micrometers [34].

CVD process is a chemical reaction which transforms gaseous molecules,

called precursors into a solid material, in the form of thin film or powder on

the surface of the substrate. The process is widely used in fabricating of

semiconductor devices, especially to fabricate high quality insulator films, e.g.

SiO2 [41]. Two most important CVD technologies in microfabrication of

sensors are the Low Pressure Chemical Vapor Deposion (LPCVD) and Plasma

Enhanced Chemical Vapor Deposion (PECVD). Comparing LPCVD and

29

PECVD process, LPCVD has such advantages: It produces excellent

uniformity of the thin film and material characteristics, and can deposit films

on both sides of at least 25 wafers at a time. The disadvantages are operation

at high temperature (higher than 600 degree) and the relatively slow

deposition rate [42]. PECVD has advantage: the process can operate at lower

temperatures (down to 300°C) thanks to the extra energy supplied to the gas

molecules by the plasma in the reactor. The disadvantage: most PECVD

deposition systems can only deposit the material on one side of the wafers on

1 to 4 wafers at a time [42]. A schematic diagram of a typical LPCVD reactor

is shown in Fig. 13 below.

Fig. 13 Typical hot-wall LPCVD reactor. The reactor consists of a quartz tube heated by a resistance heated in 3 zone furnace to maintain a uniform temperature along the reactor. Gases are introduced in one end and pumped out from the other end of the reactor. A large number of wafers (~ 100) are stacked vertically, perpendicular to the gas flow, in a quartz holder. The inlet gases may undergo homogeneous gas phase reactions to produce the deposition precursors which are transported to the wafer surface by gas-phase diffusion. Excellent film uniformities are obtained in these reactors [42].

PVD covers a number of deposition technologies in which material is released

from a source and transferred to the substrate. The two most important

technologies are evaporation and sputtering.

30

In evaporation the substrate is placed inside a vacuum chamber, in which a

block (source) of the material to be deposited is also located. The source

material is then heated to the point where it starts to boil and evaporate. The

vacuum is required to allow the molecules to evaporate freely in the chamber,

and they subsequently condense on all surfaces. This principle is the same for

all evaporation technologies, only the method used to the heating of the source

material differs. There are two popular evaporation technologies, which are e-

beam evaporation and resistive evaporation each referring to the heating

method. A schematic diagram of a typical system for e-beam evaporation is

shown in Fig. 14 below [42].

Fig. 14 Typical system for e-beam evaporation of materials. An electron beam is aimed

at the source material causing local heating and evaporation [42].

Sputtering is a technology in which the material is released from the source at

much lower temperature than during evaporation. The substrate is placed in a

vacuum chamber with the source material, named a target, and an inert gas

(such as argon) is introduced at low pressure. A gas plasma is struck using an

RF power source, causing the gas to become ionized. The ions are accelerated

towards the surface of the target, causing atoms of the source material to break

31

off from the target in vapor form and condense on all surfaces including the

substrate. As for evaporation, the basic principle of sputtering is the same for

all sputtering technologies. The differences typically relate to the manor in

which the ion bombardment of the target is realized. A schematic diagram of a

typical RF sputtering system is shown in Fig. 15 below [42].

Fig. 15 Typical RF sputtering system [41].

There are basically two different kind of film types used in microfabrication.

Dielectric layers, predominantly silicon dioxide and silicon nitride, are used as

insulating materials, as mask materials, or for passivation and are usually

deposited by CVD methods [34]. Metal layers are used, e.g., for electrical

interconnections, as an electrode material, for resistive temperature sensors

(thermistors), or as mirror surfaces. Metals such as aluminum, titanium, and

tungsten are routinely deposited by sputtering or by electron-beam

evaporation. A large number of other metals, including gold, palladium,

platinum, silver, or alloys, can be deposited with PVD methods. A number of

metals and metal compounds, such as Cu, WSi, TiSi, TiN, and W, can also be

32

deposited by CVD. Metal CVD processes are less common, but can provide

improved step coverage or local deposition of metals [34, 43].

Highly doped polycrystalline silicon (polysilicon) is used as a gate material for

MOSFETs and for piezoresistive sensing structures. Polysilicon acts as an

electrode and resistor material. For thermistors the polysilicon acts as

thermoelectric material. Polysilicon is usually deposited in a LPCVD process

using silane (SiH4) as gaseous precursor [34].

Polymers such as photoresist are commonly deposited by spin- or spray-

coating. Similar techniques are used to coat, e.g., chemical sensors with

sensitive polymer films [44]. Alternatively, chemical or biological films can

be deposited by dispensing or microcontact printing [34].

33

3.1.3 Patterning

Photolithography can transfer a pattern onto a certain material. The process

sequence is illustrated in Fig. 16. During photolithographic process, first the

electron-beam lithography is used to write the desired mask pattern. The mask

is a glass plate with a patterned opaque layer (typically chromium) on the

surface [34].

Fig. 16 Schematic of a photolithographic process sequence for structuring a thin-film layer. a) In the photolithographic process, a photoresist layer (photostructurable polymer) is spin-coated onto the material to be patterned. (b) Photoresist layer is exposed to UV light through the mask. (c) After development the resist covers only the wanted areas. (d) The remaining photoresist transfers the pattern onto the underlying material [34]. Through the mask, the spinned photoresist layer is exposed to UV light by a

mask aligner. Depending on the mask aligner generation, mask and substrate

are brought either in contact or close proximity, or the image of the mask is

projected onto the photoresist-coated substrate. Depending on whether a

positive or negative photoresist was used (Fig. 17), the exposed or the

34

unexposed photoresist areas are removed during the resist development

process, respectively. The remaining photoresist acts as a protective mask

during the etching process, which transfers the pattern onto the underlying

material. Alternatively, the patterned photoresist can be used as a mask for a

subsequent ion implantation. After the etching or ion implantation step, the

remaining photoresist is removed, and the next layer can be deposited and

patterned.

The so-called liftoff technique (Fig. 18) is a way to fabricate a thin-film

material, which would be difficult to etch. Here, the thin-film material is

deposited on top of the patterned photoresist layer. In order to avoid a

continuous film, the thickness of the deposited film must be less than the resist

thickness. By removing the underneath photoresist, the thin-film material on

top is also removed by “lifting it off,” thus leaving only a structured thin film

on the substrate.

35

Fig. 17 a) Pattern definition in positive resist. If the exposed material is etched away by the developer and the unexposed region is resilient, the material is considered to be a positive resist. b) Pattern definition in negative resist. If the exposed material is resilient to the developer and the unexposed region is etched away, it is considered to be a negative resist [42]. Thick photostructurable polymer layers, such as SU-8 [45], can also be used as

a mold for electroplating metal structures. The plating process sequence is

illustrated in Fig. 19. Recently, microcontact printing or soft lithography has

been introduced as an additional method for pattern transfer [46].

Fig. 18 a) Pattern transfer from patterned photoresist to underlying layer by etching. b) Pattern transfer from patterned photoresist to overlying layer by lift-off [42].

36

Fig. 19 Typical setup for electrodeposition. The substrate is placed in a liquid solution (electrolyte). When an electrical potential is applied between a conducting area on the substrate and a counter electrode (usually platinum) in the liquid, a chemical redox process takes place resulting in the formation of a layer of material on the substrate and usually some gas generation at the counter electrode [42].

Soft lithography (Fig. 20) represents a non-photolithographic strategy based

on the self assembly and replica molding for carrying out micro- and

nanofabrication. In soft lithography, an elastomeric stamp with patterned relief

structures on its surface is used to generate patterns and structures with feature

sizes ranging from 30 nm to 100 μm. Five techniques have been demonstrated:

microcontact printing (μCP), replica molding (REM), microtransfer molding

(μTM), micromolding in capillaries (MIMIC), and solvent-assisted

micromolding (SAMIM) [47]. The elastomeric stamp is often made from

poly(dimethylsiloxane) (PDMS). The chemical formula for PDMS (Fig. 21) is

(CH3)3SiO[SiO(CH3)2]nSi(CH3)3, where n is the number of repeating

monomer[SiO(CH3)2] units. Industrial synthesis can begin from

37

dimethylchlorosilane and water by the following net reaction [48]:

n[Si(CH3)2Cl2] + n [H2O] → [Si(CH3)2O]n + 2n HCl

Elastomeric stamp is formed by a molding process using a master fabricated

using conventional microfabrication techniques. After “inking” the stamp with

the material to be printed, the stamp is brought into contact with the substrate

material, and the pattern of the stamp is reproduced. Surface properties of the

substrate can thus be modified to, e.g., locally promote or prevent molecule

adhesion. Soft lithography has been specifically developed for biological

applications such as patterning cells or proteins with the help of, e.g., self-

assembled monolayers (SAM) [34, 46].

Fig. 20 Chemical structure of PDMS. Its other names are Polydimethylsiloxane, dimethicone, and E900 [48].

38

Fig. 21 Schematic of soft lithography. a) "Inking" a stamp. PDMS stamp with pattern is placed in Ethanol and Octadecanethiol (ODT) solution. b) ODT from the solution settles down onto the PDMS stamp. Stamp now has ODT attached to it which acts as the ink. c) The PDMS stamp with the ODT is placed on the gold substrate. When the stamp is removed, the ODT in contact with the gold stays attached to the gold. Thus the pattern from the stamp is transferred to the gold via the ODT "ink" [48]. 3.1.4 Etching

Wet etching uses liquid chemicals and dry etching uses gas-phase chemistry.

Both methods can be either isotropic, i.e., provide the same etch rate in all

directions, or anisotropic, i.e., provide different etch rates in different

directions (see Fig. 22). The material etch rate is important criteria for

selecting a particular etching process, especially, the selectivity to the material

to be etched versus other materials. Also the isotropy/anisotropy of the etching

process is important criteria [34].

39

Fig. 22 Schematics of isotropic and anisotropic thin-film etching. a) Dry etching is often anisotropic, resulting in a better pattern transfer, as mask underetching is avoided. b) Wet etching is usually isotropic. It typically provides a better etch selectivity for the material to be etched in comparison to accompanying other materials [34].

Wet etching is usually isotropic with the important exception of anisotropic

silicon wet etching in, e.g., alkaline solution (Fig. 23). Moreover, wet etching

typically provides a better etch selectivity for the material to be etched in

comparison to accompanying other materials [34].

Fig. 23 Difference between the anisotropic and the isotropic wet etching. Anisotropic etching in contrast to isotropic etching means different etch rates in different directions in the material [42].

Dry etching is often anisotropic, resulting in a better pattern transfer, as mask

under etching is avoided. In RIE system, reactive ions are generated in plasma

40

and are accelerated toward the surface to be etched, thus providing directional

etching characteristics (Fig. 24). Higher ion energies typically result in more

anisotropic etching characteristics, but also lead to reduced etching selectivity.

Whereas a large number of dry etching recipes exist, mainly fluorine- or

chlorine based etching chemistry is used in semiconductor microfabrication

[34, 43, 49].

Fig. 24 Typical parallel-plate reactive ion etching system. A plasma is generated in the gas mixture using an RF power source, breaking the gas molecules into ions. The ions are accelerated towards the surface of the material being etched, where it reacts forming another gaseous material while consuming the solid [42].

41

4. Materials and methods

The quartz monitor crystal is created by applying electrodes on a thin quartz

disc. When the electrodes are connected to an oscillator circuit built around

the QCM, the quartz crystal starts to oscillate at its resonance frequency. In

this study gold particles are used for testing of a custom made QCM. Observed

frequency changes of the oscillator correlate to the amount of added mass.

And if the amount of material deposited on the crystal was too large, the

crystal failed to oscillate.

4.1 QCM measurement setup

Altogether four easily available, AT-cut quartz crystal oscillators (T5330,

8MHz, 49 TQG02) (Fig. 25) were used. Housing of the oscillator was

removed by using a mini-drill. The testing liquid included gold particles in

Borate liquid + NaN3 solution. Optical concentration of the testing liquid was

10, as determined by the manufacturer (Reagena Ltd, Siilinjärvi, Finland).

During the testing, concentration was diluted by using distilled water. A

custom-made amplifier circuit was developed for the testing (Fig. 26a, Fig.

26b). In Fig. 26b, on the left side there are two similar oscillator circuits, on

the right side two uppermost circuits are buffers, and below them is a

regulated power supply (5 V). The oscillator circuit is a common emitter

amplifier circuit, where the crystal (Y1) acts as the feedback element. Two

diodes (D3) maintain a constant voltage (1.4 V) and therefore, with R3, the

transistor Q1 is fed a constant bias current. The varistors U2 and U4 are only

for the electromagnetic shielding. As the transistor opens, the left side of the

42

potential of the crystal Y1 is at 0.7 V and the right side at 0.2 V (the collector-

emitter voltage of the transistor). Therefore the current flows through the

crystal and the transistor closes. As the transistor is closed, the current

direction through the crystal is reversed and the transistor opens again. The

capacitors C1 and C3 affect on the amplitude and shape of the oscillation. The

output is sinusoidal if taken from the left side (as in this circuit) or square if

taken from the right side of the crystal. The buffer on the right side is for

buffering and adjusting the output signal for the proper output impedance level

(50 Ohm). The circuit had a double structure to enable the differential

measurement. However, in this study, only one sensor (oscillator) was used.

Fig. 25 T5330 is an easily available and cheap crystal oscillator. Housing of the oscillator was removed by using a mini-drill.

43

Fig. 26a A custom made amplifier circuit was used for connecting the crystal oscillator with a frequency counter.

44

Fig. 26b The custom-made amplifier circuit, for connecting the quartz crystal oscillator with a frequency counter and an external power source.

Subsequently, the quartz crystal oscillator (T5330) was connected into the

amplifier circuit (Fig. 26a, Fig. 26b), and the output of the circuit was

45

connected with BK Precision 1856A 2.4GHz frequency counter (Maxtek

International Corp., Chicago, IL, U.S.A) (Fig. 27). 50g of distilled water was

used for dilution of the test solution. HR 200 micro-scale-balance (A&D

Company, Ltd., Tokyo, Japan) was used for determining the exact amount of

water. Then, 1 µl of testing liquid (Au solution) was taken into a microliter

syringe (Hamilton 701N,9a 26S/51mm/pst 2, 10μl, Hamilton Bonaduz AG,

Bonaduz, Switzerland) and mixed with distill water in flask by using a

magnetic stirrer (Yellow-line MSH basic, IKA.Werke, GmbH&co. KG,

Staufen, Germany). The final concentration of Au solution was 0.002%.

Fig. 27 BK Precision 1856A 2.4GHz frequency counter was used for determining the QCM frequency changes.

First, the frequency of the oscillator was measured without Au deposition.

This frequency was the reference frequency. Then 10 µl 0.002% Au solution

was taken into the microliter syringe and the solution was dropped on the

surface of crystal oscillator. The droplet was dried by using a hairdryer. After

drying the surface of the oscillator, frequency of the oscillator was measured

again. Addition of the Au solution and the frequency measurement were

repeated with the same procedure until amount of Au deposition on the

46

∑=

−=N

ii xx

N 1

2)(1δ

∑=

+++==

N

i

Ni N

xxxxN

x1

21 ...1

oscillator was so large that crystal failed to oscillate. The measurement was

repeated for all four sensors.

4.2. Light microscopy and scanning electron microscopy

When the oscillation failed, we observed the surface with a light microscope

Zeiss Axioskop (20x/0.40 objective) (Carl Zeiss Werk Göttingen, Göttingen,

Germany) and took a micrograph with a installed camera named Zeiss

AxioCam MRc. Furthermore, a scanning electron microscope FEI XL30

ESEM TMP, (FEI Company, 5600 KA Eindhoven, The Netherland) was used

to observe the Au particles in unit area of the surface.

4.3 Analyses of measurement reproducibility

Standard deviation (SD), mean and coefficient variation of the frequency were

calculated for each deposited Au layer. Standard deviation (δ) of the frequency

was calculated by using the equation (15):

. (15)

The mean frequency ( x ) was obtained by equation (16):

. (16)

47

%100×=x

CV σ

Coefficient variation (CV) was calculated by using the equation (17):

. (17)

The lower the CV value, the better the reproducibility.

48

5. Results

The main results are listed in the Table 2. The frequency of the oscillator

decreased as the number of Au layers increased. The individual frequencies of

the repeated measurements with the four oscillators are shown in Fig. 28. It

was seen that the first layer was not spread on the whole sensor surface.

Therefore, the frequency change was not adequate between the reference

frequency (= no layers on the surface) and the frequency of the first layer.

Further, in Fig. 29 the mean results from the four measurements are presented

beginning from the layer number 2. It was seen that the frequency of the

oscillator correlated highly linearly and negatively with the mass increase (R2

= 0.989, Fig. 29). Mass change between the layers was calculated using the

Sauerbrey’s equation (Fig. 30). Mean weight of one layer was 100.8(±49.0)

ng/cm2. Reproducibility of the measurements decreased as the number of

layers increased on the sensor (Fig. 31). Mean reproducibility of the

measurement was extremely high 0.003% (root mean squared coefficient of

variation). The layers were clearly visible on the sensor surface, as determined

by light microscope (Fig. 32). The scanning electron microscope image of

dried Au testing liquid is shown in Fig. 33. Unfortunately, the image quality

did not enable the calculation of Au particles per unit area.

49

Table 2: All results from the frequency measurements with four same type crystal resonators and 12 layers.

Number of

Layers

Oscillator1 Frequency

(Hz)

Oscillator 2 Frequency

(Hz)

Oscillator3 Frequency

(Hz)

Oscillator 4 Frequency

(Hz) Mean (Hz)

SD (Hz)

CV (%)

- 8000658 8000658 8000961 8001012 8000822 191 0.00238 1 8000230 8000203 8000546 8000564 8000386 196 0.00245 2 7999805 8000060 8000378 8000315 8000140 262 0.00327 3 7999734 7999977 8000181 8000198 8000023 217 0.00271 4 7999759 7999814 8000121 8000153 7999962 204 0.00255 5 7999663 7999747 8000161 8000029 7999900 234 0.00293 6 7999505 7999675 8000019 7999990 7999797 249 0.00312 7 7999403 7999625 8000036 8000002 7999767 306 0.00382 8 7999407 7999703 7999874 7999888 7999718 224 0.00280 9 7999351 7999424 7999957 7999679 7999603 275 0.00344

10 7999267 7999511 7999770 7999412 7999490 212 0.00265 11 7999212 7999534 7999638 7999280 7999416 203 0.00253 12 7999051 7999517 7999693 7999326 7999397 275 0.00344

7998500

7999000

7999500

8000000

8000500

8001000

8001500

-1 0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number of layers

Freq

uenc

y (H

z)

Crystal 1Crystal 2Crystal 3Crystal 4

Fig. 28 Frequencies of oscillators decreased when the amount of Au deposition increased on the surface of oscillators. It was concluded that the frequency drop between the reference frequency (= no layers on the sensor surface) and frequency of the first layer is not adequate since it was seen that the first layer was not spread on the whole sensor surface. Error bars represent the standard derivation.

50

y = -74.807x + 8E+06R2 = 0.9891

7999000

7999200

7999400

7999600

7999800

8000000

8000200

8000400

1 2 3 4 5 6 7 8 9 10 11 12 13

Number of layers

Freq

uenc

y (H

z)

550

750

950

1150

1350

1550

1750

1950

2150

2350

Approxim

ated mass (ng/cm

2)

Fig. 29 Mean of frequencies for 4 same type oscillators (without the first 2 measured frequencies, see Fig 31). Correlation between the frequency and mass change is highly linear. The approximated mass is calculated using the Sauerbrey’s equation.

y = -0.3921x + 12.236R2 = 0.0588Mean: 100.8 ng/cm2

SD: 49.0 ng/cm2

0

20

40

60

80

100

120

140

160

180

2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12

Number of layers

Mas

s ch

ange

(ng/

cm2)

Fig. 30 Mass change between the layers, as calculated using the Sauerbrey’s equation.

51

y = 4E-05x + 0.0027R2 = 0.1373

0.00200

0.00220

0.00240

0.00260

0.00280

0.00300

0.00320

0.00340

0.00360

0.00380

0.00400

-1 0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number of layers

Coe

ffici

ent o

f var

iatio

n (%

)

Fig. 31 Reproducibility of the measurements, as indicated by the coefficient of variation (CV). Mean reproducibility of the measurement was 0.003% (Root Mean Square average CV). The lower the CV-value, the better the reproducibility.

Fig. 32 Surface of the sensor as imaged using the light microscope. The different layers on the sensor surface are clearly visible (the black lines).

700 µm

52

Fig. 33 Scanning electron microscope image of dryed 100% Au testing liquid. As seen, the image quality did not enable the calculation of Au particles per unit area.

53

6. Discussion

6.1 Conclusions

The QCM sensors have been used in a variety of applications and could help

to solve problems in many fields, including food product quality assurance,

health care, environmental monitoring, pharmaceuticals, electrochemistry of

interfacial processes at electrode surfaces, surfactant research, drug research,

indoor air quality, safety and security, and the military. In this study the

principle and applications of QCM sensors were discussed. Furthermore, the

basic microfabrication techniques of the sensors were introduced. Some of the

microfabrication techniques are identical with those used in IC fabrication:

deposition, patterning and etching.

In addition, a basic QCM system was developed and implemented. The AT cut

crystal oscillators were used as a mass sensor and it was demonstrated that the

change of oscillation frequency correlates to the amount of added mass. The

Au deposition on the surface of crystal oscillator was studied. It was seen that

surface of the crystal oscillator is very challenging to clean completely. Four

of the same type crystal oscillators were used, and it was seen that the

reproducibility of the measurement is high (CV = 0.003%). The study of QCM

crystal oscillator gives a wealth of information about not only the mass, but

also the formed film.

54

6.2 Future Work

Future work could be focused on cleaning of the oscillator surface, in order to

enable using of the same sensor several times. In this study, when number of

Au particles deposited on the surface of oscillator were too much, the crystal

failed to oscillate. Au multilayer was formed on the surface and it was hard to

be cleaned totally. In many commercial applications, it is difficult if the sensor

has to be changed after the use.

Presently, a laser-cleaning tool capable of removing surface contaminants such

as submicron-sized particulates and organic films has been constructed and

implemented in practical use. The nanosecond-pulsed laser irradiation, shortly

after the deposition of a thin liquid film on the surface, induces explosive

vaporization of the liquid and removal of particulates (“steam cleaning”). The

ultraviolet (UV) irradiation also causes ablative photodecomposition of

organic film contaminants on the surface (“dry cleaning”). Hence, cleaning

can best be done with the combination of these two laser-cleaning modes. The

cleaning strategy and the probability of thermal damage due to laser cleaning

are discussed in [50].

Ultrasonic waves generate and evenly distribute cavitation implosions in a

liquid medium. The released energies reach and penetrate deep into crevices,

blind holes and areas that are inaccessible to other cleaning methods. The

removal of contaminants is consistent and uniform, regardless of the

complexity and the geometry of the substrates. Ultrasonic waves are

55

mechanical pressure waves formed by actuating the ultrasonic transducers

with high frequency. A typical industrial high power generator produces

ultrasonic frequencies ranging from 20-120 kHz [51, 52]. It is possible that if

the sensor is coated with some active material to collect specific particles,

laser cleaning or ultrasonic wave cleaning techniques can be harmful for the

coating, too.

Quartz crystal oscillator based microbalancing system is relatively

straightforward and inexpensive to take into use, especially if compared with

more advanced system, such as micromachined cantilevers. Main limitation of

the QCM is related to cleaning of the surface after the use.

56

References

[1] Viitala T. “What is Quartz crystal microbalance”, KSV instruments Ltd, technical note. October 2004. [2] Sauerbrey G. “Verwendung von Schwingquarzen zur Wägung dünner Schichten und zur Mikrowägung”, Z. Phys. 1959, Volume 155, 206-222; Sauerbrey G. Phys. Verh. 1957, Volume 8, 113-114. [3] “Quartz crystal, the timing material”, technical note, QUARTZ TECH. http://www.4timing.com/techquartz.htm. [4] “Quartz Crystal Microbalance (QCM)”, technical note. http://www.tau.ac.il/~phchlab/experiments/QCM/QCM.html. [5] Isotemp research, INC. “Understanding Ovenized Oscillators”, technical note. http://www.isotemp.com/146-005.html. [6] Ruf H.H. “Biosensor and biochip”, lecture notes on “Mikrofabrikation und Sensorapplikationen”-course, Fraunhofer IBMT und Universität des Saarlandes, St. Ingbert. 25.1.2005. [7] Yuwono A.S. and Lammers P.S. “Performance Test of a Sensor Array - Based Odor Detection Instrument”, Agricultural Engineering International: the CIGR Journal of Scientific Research and Development. 2004, manuscript Number BC 03 009, 1-16. [8] “Sauerbrey equation”, technical note. http://www.answers.com/topic/sauerbrey- equation. [9] Kanazawa K.K. and Gordon J.G. “Frequency of a quartz microbalance in contact with liquid”, Anal. Chem. 1985, Volume 57, 1770–1771. [10] Atul S.A.; Devendra S.; Kalonia. “Measurement of Fluid Viscosity at Microliter Volumes Using Quartz Impedance Analysis”, AAPS. PharmSciTech. 2004, Volume 5, Article 47, 1-14. [11] “Crystal information”, technical notes. http://www.electrodynamics.com/catalog/technicalnotes.pdf#search= %22motional%20arm%22. [12] Nwankwo E.; Durning C.J. “Impedance analysis of thickness-shear mode quartz crystal resonators in contact with linear viscoelastic media”, Rev Sci. Instrum. 1998, Volume 69, 2375-2384. [13] Reed C.E.; Kanazawa K.K.; Kaufman J.H. “Physical description of a viscoelastically loaded AT-cut quartz resonator”, J Appl. Phys. 1990, Volume 68, 1993-2001. [14] Kanazawa K.K.; Gordon J.G. “The oscillation frequency of a quartz resonator in contact with a liquid”, Anl Chem. 1985, Volume 175, 99-105. [15] Bandey H.L.; Martin S.J.; Cernosek R.W.; Hillman A.R. “Modeling the responses of thickness-shear mode resonators under various loading conditions”, Anal Chem. 1999, Volume 71, 2205-2214.

57

[16] Martin S.J.; Frye G.C.; Ricco A.J.; Senturia S.D. “Effect of surface roughness on the response of thickness-shear mode resonators in liquids”, Anal Chem. 1993, Volume 65, 2910-2922. [17] Behrends R.; Kaatze U. “A high frequency shear wave impedance spectrometer for low viscosity liquids”, Meas Sci. Technol. 2001, Volume 12, 519-524. [18] Zimmermann B.; Lucklum R.; Hauptmann P.; Rabe J. and Buttgenbach S. “Electrical characterisation of high-frequency thickness-shear mode resonators by impedance analysis”, Sens. Actuators B 76. 2001, 47-57. [19] Byström E. “Construction of a Portable Piezoelectric Quartz Crystal Sensor Array for Determination of Petroleum Compounds in Contaminated Soil”, Department of Chemistry, UMEA University. 2003. Master thesis in Environmental Analytical Chemistry. [20] Nagle H.T.; Schiffman S.S.; Osuna1 R.G. “The How and Why of Electronic Noses”, IEEE Spectrum. 1998, Volume 35, 22-34. [21] Couto A. “Current Status of Electronic Nose: Research and Applications”, technical note, http://www.eng.auburn.edu/~troppel/courses/elec663spring2000/WrittenReports/ Couto_Aroldo_report.doc. [22] “Sensors overview III, polymer based transducers”, Institute of Physical Chemistry, University of Tübingen, lecture note. 2000. [23] Polikar R.; Shinar R.; Honavar V.; Udpa L.; Porter M.D. “Detection and identification of odorants using an electronic nose”, IEEE. 2001, Volume 4, 7041-7803. [24] Mascini M. “Immobilization of Bionmolecules”, Europeanschool on Nanoseciences and Nanotechnologies, Grenoble, France, lecture note. 2004, http://www.esonn.inpg.fr/0oldweb/ESONN04/Lectures/esonn_2004_Mascini3.pdf. [25] Matthew A.C.; Fedor N.D.; Minson T.; Victor P.O.; Abell C. and Klenerman D. “Direct and sensitive detection of a human virus by rupture event scanning”, Nature biotechnology. 2001, Volume 19. [26] Battiston F.M.; Ramseyer J.P.; Lang H.P.; Baller M.K.; Gerber C.; Gimzewski J.K.; Meyer E.; Guntherodt H.J. “A chemical sensor based on a micrfabricated cantilever array with ultaneous resonance-frequency and bending readout ”, Sensor and Actuators B 77. 2001, 122-131. [27] Fritz J.; Baller M.K.; Lang H.P.; Rothuizen H.; Vettiger P.; Meyer E.; Guntherodt H.J.; Gerber Ch.; Gimzewski J.K. “Translating biomolecular recognition into nanomechanics”, Science. 2000, Volume 288, 316-318. [28] Baller M.K.; Lang H.P.; Fritz J.; Gerber Ch.; Gimzewski J.K.; Drechsler U.; Rothuizen H.; Despont M.; Vettiger P.; Battiston F.M.; Ramseyer J.P.; Fomaro P.; Meyer E.; Guntherodt H.J. “A cantilever array-based artificial nose”, Ultramicroscopy. 2000, Volume 82, 1-9. [29] Berger R.; Lang H.P.; Delamarche E.; Gerber Ch.; Gimzewski J.K.; Andreoli C.; Brugger J.; Despont M.; Vettiger P. “Integration of silicon micromechanical arrays with molecular monolayers for miniaturized sensor systems”, Physical Sciences, RZ 2926 (# 91521). 05.05.1997.

58

[30] Chen G.Y.; Thundat T.; Wachter E.A.; Warmack R.J. “Adsorption induced surface stress and its effects on resonance-frequency of microcantilevers”, J Appl. Phys. 1995, Volume 77, 3618-3622. [31] Stoney G.G. “The tension of metallic films deposited by electrolysis”, Proc. R. Soc, Lond. Ser, A82. 1909, 172-175. [32] Fagan B.C.; Tipple C.A.; Xue Z.L.; Sepaniak M.J.; Datskos P.G. “Modification of

micro-cantilever sensors with sol-gels to enhance performance and immobilize chemically selective phases”, Talanta. 2000, Volume 53, 599–608.

[33] “What is MEMS”, technical note, MEMS and Nanotechnology Clearinghouse. http://www.memsnet.org/mems/what is.html. [34] Hierlemann A.; Brand O.; Hagleitner C. and Baltes H. “Microfabrication Techniques For Chemical/Biosensors”, IEEE. 2003, Volume 03, 0018-9219. [35] Takashi A.B.; Li L.; Hung V.N.; Esashi M. “Energy dissipation in small-diameter quartz crystal microbalance experimentally studied for ultra-high sensitive gravimetry”, IEEE. 2003, Volume 3, 7744-7803. [36] Brand O. and Baltes H. “CMOS-based microsensors and packaging”, Sens. Actuators A. 2001, Volume. 92, 1–9. [37] Petersen K.E. “Silicon as a mechanical material”, IEEE. 1982, Volume. 70, 420–457. [38] Menz W.; Mohr J. and Paul O. “Microsystems Technology”. Wiley-VCH. 2001. [39] Madou M. “Fundamentals of Microfabrication”, FL: CRC. 1997. [40] Tummala R.R. “Fundamentals of Microsystems Packaging”, New York: McGraw- Hill. 2001. [41] “Chemical Vapour Deposition”, technical article. http://chiuserv.ac.nctu.edu.tw/~htchiu/cvd/home.html. [42] Huff M. “MEMS fabrication”, Sensor Review. 2002, Volume 22 , 18-33. [43] Chang C.Y and Sze S.M. Eds, ULSI Technology. New York: McGraw-Hill. 1996. [44] Hierlemann A.; Lange D.; Hagleitner C.; Kerness N.; Koll A.; Brand O. and Baltes H, “Application-specific sensor systems based on CMOS chemical microsensors”, Sens. Actuators B. 2000, volume. 70, 2–11. [45] Lorenz H.; Despont M.; Fahrni N.; Brugger J.; Vettiger P.and Renaud P. “High- aspect-ratio, ultrathick, negative-tone near-UV photoresist and its applications for MEMS”, Sens. Actuators A. 1998, Volume. 64, 33–39. [46] Whitesides G.M.; Ostuni E.; Takayama S.; Jiang X. and Ingber D.E. “Soft lithography in biology and biochemistry”, Annu. Rev Biomed Eng. 2001, Volume. 3, 335–373. [47] Xia Y.N. and Whitesides G.M. “Soft Lithography”, Annual Review of Materials Science. 1998, Volume 28, 153-184.

59

[48] “Soft Lithography”, technical note, http://en.wikipedia.org/wiki/Soft_lithography. [49] Sze S.M. VLSI Technology, 2nd. New York: McGraw-Hill. 1988. [50] Grigoropoulos C.P.; Leung W.P.; Tam A.C. “A practical excimer laser-based cleaning tool for removal of surface contaminants”, IEEE.1994,Volume 17, 631-643. [51] Awads S. “Ultrasonic Cavitations and Precision Cleaning”, technical notes CREST ULTRASONICS CORP. http://www.crest-ultrasonics.com/clean.pdf. [52] Kanegsberg B. “Handbook of Critical Clean”, ISBN:0849316553, Publisher: CRCPress, Publication Date: 26 Dec 2000, Subject: Technology & Industrial Arts.