21EE742 Module - 2

21EE742 – MICRO & NANO SENSORS &
Regulation – 2021 TRANSDUCERS

MODULE - 2
MODULE – 2
Ultrahigh Sensitivity, Wide Dynamic Range Sensors - Structure, Theory, Experimental Results,
Conclusion, Other Motion and Acceleration Microsensors
A CO Gas Sensor Based on Nanotechnology - Structure, Theory, Assembly of Sensors,
Experimental results, Auxillary Experimental Results, Conclusion
Smoke Detectors - Structure, Theory, Assembly of Sensors, Experimental results, Auxillary
Experimental Results, Conclusion
ULTRAHIGH SENSITIVITY, WIDE DYNAMIC RANGE SENSORS
 A number of new applications, such as self-guided small projectiles and autonomous

surveillance airplanes, have created a requirement for a highly sensitive yet very small
acceleration sensor.
 The fundamental problem in acceleration sensors is the size–sensitivity tradeoff.
Specifically, a highly sensitive acceleration sensor must necessarily be a large sensor.
 A new miniature acceleration sensor with a radical new design was recently introduced. The
sensor offers a very large dynamic range and high sensitivity.
 This new type of acceleration sensor will be available commercially within a time span of 1–
2 years.
STRUCTURE
 The new acceleration sensor is based on the concept of creating a variable ultra-capacitor
structure that consists of one small droplet of electrolyte that is positioned between two
carbon nanotube (CNT) electrodes.
 At rest, the CNT electrodes remain outside of the electrolyte due to their hydrophobic
nature. Under acceleration, however, the inertial forces push the CNT electrodes into the
electrolyte and the typical capacitance of an ultra-capacitor is obtained.
 The prototype described here has shown a capacitance variation of approximately 5 μF
under an acceleration increase from zero to 2200 g. The sensitivity is therefore 2.27nF/g.
The concept behind the new sensor is shown in Figure 2.1.
 As shown in Figure 2.1(a), a small droplet of electrolyte (fluid with high ionic conductivity)
is placed between two electrodes on which CNTs of a length of 20 μm are grown.
 One electrode is fixed while the other is movable, as shown. Since the weights of the droplet
and the electrodes are very small, and since CNTs are super hydrophobic, the CNTs do not
penetrate the electrolyte when the mechanism is at rest.
 The capacitance between the two electrodes is hence approximately equal to zero in that
configuration.
 In Figure 2.1(b), as acceleration occurs along the axis of the device (here, vertically,
upward), the inertial forces created by the droplet and the moving electrode cause the CNTs
to penetrate the electrolyte, as shown.
Prepared by: Dr. RAMYA. K, Dept. of EEE Sri Sairam College of Engineering Anekal.
MODULE - 2
Figure 2.1: Principle of operation of the new acceleration sensor
 Accordingly, the capacitance between the two electrodes (which is proportional to the
immersed area of the CNTs) increases. In Figure 2.1(c), the forces created due to inertia
further increase at higher accelerations, and the penetration of the CNTs into the electrolyte
increase to a maximum.
 The capacitance of the ultra-capacitor increases substantially high value (approximately 5
μF in the present prototype).
 It is to be pointed out that the mechanism shown in Figure 2.1 is essentially two ultra-
capacitors in series, as capacitance forms at the interface between the electrolyte and the
surface of the CNTs.
Figure 2.2 shows a photograph of the actual prototype that was assembled by the author. As the
figure shows, an SOIC chip that measures 6 mm × 8 mm has an open cavity in which the mechanism
is embedded. The figure shows the electrolyte droplet (with a diameter of 2 mm) and the upper
electrode (removed and placed next to the chip). The lower electrode is underneath the droplet and
is not shown in the figure.
Figure 2.2: Photograph of the sensor. The SOIC package measures 6 mm × 8 mm. The diameter of
the electrolyte droplet is 2 mm, and the diameter of the moving electrode is 3 mm.
The sensor, as described, is a one-axis sensor, and can be used to measure acceleration along the X,
Y, or Z axes. Of course, if a 3-axis acceleration sensor is desired, then 3 units of the sensor can be
packaged and used together. The electrodes used in the sensor are stainless steel electrodes, where,
as indicated, a layer of CNTs is grown on one side of each electrode. These CNTs were deposited by
MODULE - 2
a specialized commercial supplier that utilizes the process of chemical vapor deposition. Figure 2.3
is a scanning electron microscope (SEM) micrograph that shows the CNT layer.
Figure 2.3: Multiwalled carbon nanotubes of an average diameter of 250 nm and a length of about
20 μm, grown on a stainless steel electrode.
Figure 2.4 shows a mechanical diagram of the main components of the sensor. As the figure shows,
the electrolyte droplet, the fixed electrode, and the moving electrode are placed inside the cavity in
the SOIC chip.
Figure 2.4: Mechanical structure of the sensor (scale = 16:1)
 A very thin aluminum cylinder with a height of approximately one-half of the cavity is also
inserted in the cavity.
 The purpose of the cylinder is to act as a contact terminal for the moving electrode. As
shown in the figure, a small trace of a conductive liquid that is known commercially as
“Liquid Metal” is placed on top of the moving electrode and serves to establish electrical
contact between the electrode and the aluminum cylinder.
 The aluminum cylinder is, in turn, connected to an external terminal.
 The “Liquid Metal” is a gallium alloy that is normally liquid at temperatures above −20°C. As
also shown in Figure 2.4, a steel ring is placed on top of the cavity and serves to retain the
moving electrode, and hence the electrolyte droplet, inside the cavity (the retaining ring
appears in Figure 2.2).
MODULE - 2
 The electrolyte used in the present application is propylene carbonate (solvent) in which an
ionic salt is dissolved (a typical ultra-capacitor electrolyte).
 Two important observations must now be made. First, it is to be pointed out that propylene
carbonate has a boiling point of 240°C.
 Accordingly, any possible evaporation of the electrolyte will occur only at elevated
temperatures.
 The sensor was indeed tested at temperatures of up to 80◦ C, and no evaporation of any kind
was observed.
 The second issue concerns the possible effect of the inclination angle on the measurement
provided by the sensor. It is to be pointed out that the weight of the electrolyte droplet is
only 5 mg, and hence the surface tension forces are far larger than the deformation forces
that exist due to the weight.
 Furthermore, the droplet is confined within a restricted space, as shown in Figure 2.4, and
therefore it is always in contact with the CNT electrodes. These facts indicate that the
inclination angle has absolutely no effect on the measurement provided by the sensor. This
was indeed confirmed through the experiments.
THEORY
Figure 2.5 depicts an array of CNTs that is penetrating a liquid electrolyte. For each CNT, the
hydrophobic force (or force that repels the fluid away from the CNT) will be given by
𝐹 = 2𝜋𝑟𝛾 cos(180° − 𝜃) (2.1)
Figure 2.5: Array of CNTs penetrating a liquid electrolyte
Where r is the radius of one CNT, g is the surface tension of the fluid, q is the liquid–solid contact
angle, and the product 2prg represents the surface tension force. The total upward force that is
acting on the fluid will be given by NF, where N is the total number of CNTs penetrating the fluid.
We shall first develop a formulation for the capacitance at the interface between the moving
electrode and the electrolyte (see Figure 2.1). Given an electrode with a mass m that is moving with
a uniform acceleration a, the Newtonian force F = ma must be balanced by the total hydrophobic
force, that is,
𝐹𝑇𝑜𝑡𝑎𝑙 = 𝑚𝑎 = 2𝜋𝑟𝑁𝛾 cos(180° − 𝜃) (2.2)
The surface area A of the CNTs that is immersed in the electrolyte will be given by
MODULE - 2
𝐴 = 2𝜋𝑟 ∑ 𝑥𝑖 (2.3)
𝑖=1
Where 𝑥𝑖 is the immersion depth of any given CNT. The capacitance of the electrode – electrolyte
interface will be now given by
𝜀0 𝜀𝑟 𝐴
𝐶= (2.4)
𝑑
Where 𝜀0 the permittivity of free space is, 𝜀𝑟 is the relative permittivity of the electrolyte, and d is
the separation between the positive and negative charges at the interface (this separation is about 1
nm in ultra-capacitors). The above expression is the well-known equation of the capacitance of a
parallel-plate capacitor. It is important to note here that the capacitance of a cylindrical capacitor is
very nearly the same as the capacitance of a parallel-plate capacitor when the interelectrode
separation is much smaller than the radius of the cylinder. From the previous two equations, and by
taking er ≈ 1, we have
𝑁
𝜀0
𝐶= 2𝜋𝑟 ∑ 𝑥𝑖 (2.5)
𝑑
𝑖=1
In the present application, the capacitance between the moving electrode and the electrolyte is not
directly measurable, however. What is measurable, in fact, is the series combination of the
capacitances C1 and C2 at the two interfaces. This is the observed capacitance of the ultra-capacitor
(Figure 2.6).
Figure 2.6: The total capacitance of the ultra-capacitor is the series combination of the capacitances
at the two interfaces C1 and C2.
In general, C2 will be larger than C 1, since the lower interface is acted upon by both the upper
electrode and the electrolyte droplet. Now, given the general form of Eq. (2.5), the series
combination of C1 and C2 can be expressed as
MODULE - 2
𝑁
𝜀0
𝐶𝑇𝑜𝑡𝑎𝑙 = 𝐾 2𝜋𝑟 ∑ 𝑥𝑖 (2.6)
𝑑
𝑖=1
Where K is constant that ranges from 1 at very small accelerations (whereC2≫C1) to 0.5 at very
high accelerations (where C2 ≈C1). From equation (2.2) and (2.6) we now have
𝑑𝐶𝑇𝑜𝑡𝑎𝑙
𝑚𝑎 = 𝑁𝛾 cos(180 − 𝜃) ( ) (2.7)
𝐾𝜀0 ∑𝑁
𝑖=1 𝑥𝑖
Hence,
𝑑𝑁𝛾 cos 𝜃
𝑎= 𝐶𝑇𝑜𝑡𝑎𝑙 𝑓𝑜𝑟 𝜃 > 90° (2.8)
𝐾𝑚𝜀0 ∑𝑁
𝑖=1 𝑥𝑖
Equation (2.8) can be greatly simplified by observing the following fact: the length of the CNTs is
very short (20 μm, or less than the thickness of a human hair). Then, to a very good approximation
𝑖=1 𝑥𝑖 ≈ 𝑁𝑥, 𝑤ℎ𝑒𝑟𝑒 𝑥 = 20𝜇𝑚 is the full length of the CNTs. Equation (2.8) now simplifies as
∑𝑁
𝑑𝛾 cos 𝜃
𝑎= 𝐶 𝑓𝑜𝑟 𝜃 > 90° (2.9)
𝐾𝑚𝜀0 𝑥 𝑇𝑜𝑡𝑎𝑙
It is to be pointed out that the length of the CNTs is extremely small (20 μm), and hence no
“threshold” for the penetration was observed in the operation of the sensor. As pointed out, the
operation depends mainly on the droplet’s spreading rather than on the droplet’s penetration into
the CNTs. The only purpose of using CNTs in this application is to provide an extremely large
surface area and hence a very large capacitance.
EXPERIMENTAL RESULTS
Physical parameters of the sensor and the dynamic range:

The following are the physical parameters that characterize the present sensor:
 d ≈ 1 nm
 g = 41×10−3
 q ≈ 160° (according to a number of published physical studies on the wetting of CNTs by organic
liquids similar to propylene carbonate. It is important here to point out that the force equation can
be written in terms of cos 𝜃 or cos(180 − 𝜃). The contact angle θ will be substantially different in the
two cases)
 m = 0.1 g
 x = 20 μm
 Area of the moving electrode = 0.071 cm2
 N = 109 CNTs per cm2; hence N on the electrode = 71×106 CNTs
 r (radius of one CNT) ≈ 125 nm
By substituting with the above parameters in Eq. (2.9), we obtain
MODULE - 2
𝐶𝑇𝑜𝑡𝑎𝑙 𝑚
𝑎 = 2.18 × 109 ⁄𝑆 2 (2.10)
𝐾
Substitution in Eq. (2.5) and assuming full use of the CNTs (i.e., N = 71×10 6) gives the maximum
capacitance per interface as 9.87 μF. Hence, the maximum capacitance of the device (C Total, max) is
approximately 4.94 μF. With the constant K ranging from 0.5 at high accelerations to approximately
1 at low accelerations, we can conclude from Eq. (2.10) that the maximum measurable acceleration
will be approximately 21517 m/s2, or 2193 g (a substantially high acceleration). Theoretically, the
minimum acceleration that can be measured is 0; however, measurement of a capacitance that is
less than 1 pF is not usually reliable. If we now assume that C total (min) ≈ 1 pF and that K ≈ 1, Eq.
(2.10) shows that the minimum measurable acceleration is approximately 2×10−3 m/s 2 or 2×10−4
g. It is therefore clear that the dynamic range of this sensor is substantially large.
Measurement of the capacitance constant K
The capacitance constant K was determined by mounting the sensor on an electrodynamic shaker
and measuring the capacitance Ctotal as a function of acceleration.
Three different types of measurements were performed:
 For very low accelerations (less than 1 g), an electrodynamic shaker from Data Physics Corp;
was used in the testing.
 For small to medium accelerations (1 g to 100 g), an electrodynamic shaker from Dynamic
Solutions, Inc. (Northridge, CA) was used.
 For high accelerations (greater than 100 g) the sensor was actually tested by mounting the
sensor and its associated circuitry inside a low-speed 0.45-caliber bullet. When tested on the
electrodynamic shaker, the sensor was subjected to shock pulses of duration of about 0.2 s
and different known accelerations.
 The acceleration in each case was also measured simultaneously with an off-the-shelf sensor
from Freescale Semiconductors, Inc. To record the acceleration and the associated
capacitance in each test, both physical variables were converted to signals and observed on
digital storage oscilloscopes. The value of the unknown constant K was then calculated from
Eq. (2.10) for each measurement performed. The results are shown in Figure 2.7.
Figure 2.7: The capacitance constant K as a function of acceleration (0.0002 g to 2200 g).
MODULE - 2
As Figure 2.7 shows, K does indeed vary linearly from approximately 1 at very low accelerations to
approximately 0.5 for accelerations above 2000 g. As the figure also shows, for accelerations that
are less than about 100 g, K is very nearly equal to 1, and hence the acceleration a in Eq. (2.10) will
be simply proportional to Ctotal for that wide range of accelerations.
Measurement of acceleration
 By using the same experimental setup, Eq. (2.10) was tested directly by measuring the
capacitance of the sensor during the acceleration pulse and calculating the value of the
acceleration predicted by the equation.
 The value of the acceleration was also measured simultaneously with the capacitance by
using an off-the-shelf sensor. The range of the accelerations used in this test was the low
range of 2×10−4 g to 1g, and K was taken to be equal to 1. The results are shown in Figure
2.8. Clearly, the sensor is quite linear in that range.
Figure 2.8: Measured and theoretical acceleration in the range of 0.0002 g to 1 g as a function of the
measured capacitance.
 The sensor was subsequently tested in the small to medium range of accelerations (1 g to
100 g). Figure 2.9 shows the actual acceleration that was measured versus the value
predicted by Eq. (2.10) without correction for the value of K.
 The maximum error is approximately 3%, near the higher values of acceleration. Of course,
if corrected for the variation in K, the acceleration predicted by Eq. (2.10) would precisely
match the actual value.
 The sensor was finally tested in the acceleration range of 100 g to 2200 g.
 No mechanical shaker is capable of achieving such high values of acceleration, and for that
reason the sensor was tested by actually mounting it inside a 0.45-caliber bullet.
 Figure 2.10 shows the type of bullet that was used, the circuit board on which the sensor
was mounted, the 1.5 V battery that was used for powering the board, and the handgun that
was used for firing the bullets. It is to be pointed out that the acceleration of a normal bullet
inside a gun’s barrel is typically on the order of 10 6 g or higher. To get the lower range of
MODULE - 2
100 g to 2200 g that is applicable to the sensor, the chemical propellant inside the bullet’s
case was reduced considerably.
Figure 2.9: Acceleration in the range of 1 g to 100 g as a function of the measured capacitance.
Without correction for the value of K, the maximum error is approximately 3%.
Figure 2.10: The circuit board and the 0.45-caliber bullet that were used for testing the sensor at
accelerations up to 2200 g.
 On the circuit board shown in Figure 2.10, a second, off-the-shelf acceleration sensor is also
present.
 A block diagram of the circuit used is shown in Figure 2.11. As shown, the sensor
(represented by the capacitance C), is connected to a well-known 555 oscillator circuit that
generates a square wave with a frequency that is dependent on the capacitance.
MODULE - 2
 A frequency counter then converts the frequency to a digital readout. The digital readout is
finally processed by an on-board 8 bit microprocessor. The microprocessor simply tracks
the highest number and stores that number in a nonvolatile memory.
 In the circuit shown in Figure 2.10, the microprocessor actually processes two signals: one
from the new sensor described here and the other from the off-the-shelf sensor that is also
present on board.
Figure 2.11: Block diagram of the interface circuit used for measuring the capacitance C of the
sensor in real time.
The testing results for the acceleration range of 100 g to 2200 g are shown in Figure 2.12. As
expected, the correct value of K must be used in Eq. (2.10), or otherwise serious errors will result in
that high acceleration range.
Figure 2.12: Acceleration in the range of 100 g to 2200 g as a function of the measured capacitance.
MODULE - 2
Conclusion
 The new MEMS acceleration sensor described in this chapter has an ultra-wide dynamic
rage and high sensitivity. Specifically, the sensitivity is 2.27nF/g, and the dynamic range
extends from 2×10−4 g to 2193 g.
 It will therefore be useful in a wide array of new applications, including self-guided
projectiles and autonomous surveillance aircraft.
 The prototype device that was fabricated and tested has exhibited a change in capacitance of
approximately 5 μF over the dynamic range of the sensor.
 The sensor is linear for accelerations up to approximately 100 g and nonlinear for higher
accelerations.
 In the nonlinear region, a capacitance constant K must be fully included in the acceleration–
capacitance relationship to avoid large errors.
 The equation that relates the acceleration to the measured capacitance is fairly simple, and
testing has shown a very good agreement between theory and experiment.
OTHER MOTION AND ACCELERATION MICROSENSORS
Other approaches for motion and acceleration sensors have also been introduced recently. Figure
2.13 shows a concept for a monolithic integration of pressure plus acceleration composite Tire
Pressure Monitoring Sensor (TPMS) with a single-sided micromachining technology. Figure 2.14
shows a photo of the actual sensor.
Figure 2.13: Concept for monolithic Tire Pressure Monitoring Sensor (TPMS)
MODULE - 2
Figure 2.14: Photo of actual sensor
MODULE - 2
A CO Gas Sensor Based on Nanotechnology - Structure, Theory, Assembly of Sensors,

Experimental results, Auxillary Experimental Results, Conclusion
INTRODUCTION
Gas and smoke sensors have also benefited greatly from the recent advances in nanotechnology.
This chapter introduces a carbon monoxide (CO) gas detector that uses gold nanoparticles
deposited on an array of carbon nanotubes as a sensing element, in addition to a smoke detector
that achieves miniaturization by replacing a bulky alpha particle detection chamber with a silicon
nanostructure. These types of detectors will be available commercially within a time span of 1 to 2
years.
A CO Gas Sensor Based on Nanotechnology

 Carbon monoxide (CO) detectors are used extensively in the household and the industrial
sector. Usually, a CO concentration in excess of 100 parts-per-million (ppm) is considered to
be harmful for humans.
 To detect such a small concentration, detectors have traditionally been large devices.
 The necessary sensitivity is achieved by using a large electrochemical cell or a group of
heated semiconductor wires where in either case the oxidation reaction of CO (to form CO2)
results in a small electric potential that can be detected.
 Very recently, a miniature, highly sensitive CO gas sensor that makes use of the very high
catalytic activity of gold nanoparticles was introduced.
 Gold nanoparticles are deposited on the surfaces of an array of carbon nanotubes (CNTs).
The CO gas is transformed into CO2 as it comes in contact with the nanoparticles, and free
electrons are released.
 The extra free electrons result in a measurable reduction in the resistivity of the CNT array.
The detector has proved to be very sensitive at room temperature to CO gas concentrations
as low as 100 ppm. It is characterized by a very small size and can be easily powered with a
battery.
Structure
The fundamental principle behind the new CO gas sensor is shown in Figure 3.1. As the figure
shows, gold nanoparticles are attached to an array of multi-walled carbon nanotubes.
Figure 3.1: Fundamental principle of operation of the new CO detector
MODULE - 2
Gold nanoparticles have a large affinity for oxygen, and oxygen molecules (from the ambient
atmosphere) become adsorbed on the large surface presented by the nanoparticles, where they
easily disassociate. CO, a reducing gas, easily combines with atomic (or molecular) oxygen and
releases free electrons according to a well-known reaction. The presence of extra free electrons
reduces the overall resistivity of the carbon nanotube array, which can be detected with an external
circuit. The two reactions that can be synthesized with catalysts are
2𝐶𝑂 + 𝑂2− → 2𝑐𝑜2 + 𝑒 − (3.1)
𝐶𝑂 + 𝑂− → 𝑐𝑜2 + 𝑒 − (3.1)
According to the principle shown in Figure 3.1, these well-known chemical reactions are
synthesized in the new solid-state detector by using the excellent catalytic activity of gold
nanoparticles. Recent density functional theory studies have shown that such reactions have a
barrier of less than 0.4eV when a gold catalyst is used and are therefore synthesizable at
temperatures below room temperature.
The miniature solid-state CO detector described in this section has the following
characteristic advantages:
 Sensitivity to CO concentrations of less than 100 ppm at room temperature

 Can be battery operated, with a very long battery life (the power required is quite negligible
when the audio alarm is not activated)
 Being a miniature detector, it can be placed virtually anywhere.
The new technology reported here can be understood in proper context when compared with other
CO sensing technologies that have appeared in the literature during the past decade. There are
essentially two groups of new technologies.
 The first group, which requires temperatures significantly higher than room temperature,
includes technologies such as field effect transistors and Schottky diodes that are coated
with various metal oxides, sensors based on cobalt and rubidium catalysts, sensors based on
combinations of catalysts and polymers and infrared detection techniques.
 The second group of technologies offers operation at room temperature in addition to high
sensitivity. It includes optical techniques and microwave techniques. This latter group of
technologies, however, has the disadvantage of very high cost and the requirement for bulky
detection equipment.
 As far as the author is aware, the new technology reported here is the first technology that
offers the combined features of high sensitivity, relatively low cost, very small size, and
operation at room temperature.
Theory
The rate of the catalytic reactions in (3.1) can be described by the well-known reaction rate
equation
𝑑𝑃 𝐸𝑎
− = 𝑘0 𝑒𝑥𝑝 (− ) 𝑃 𝑚 (3.2)
𝑑𝑡 𝑅𝑇
MODULE - 2
Where P is the partial gas pressure, k0 is the pre-exponential factor, Ea is the activation energy, and
m is the pseudo-kinetic order of the reaction. The catalytic oxidation of CO by various catalysts
shows a pseudo-kinetic order m that is usually 0 or 1. A useful measure of the performance of a gas
sensor is the conversion percentage which, in this application, is the percentage of CO molecules
that are converted to CO2. The conversion percentage is usually expressed as
𝑃(𝑡)
𝛼 (𝑡) = 1 − (3.3)
𝑃0
Where P0 is the partial pressure of CO at time t =0. If m=1, the solution of Eq. (3.2) is
𝐸𝑎
𝑃 = 𝑃0 𝑒𝑥𝑝 (−𝑘0 𝑒𝑥𝑝 (− ) 𝑡) , 𝑚=1 (3.4)
𝑅𝑇
and hence the conversion percentage a is found to be
𝐸𝑎
𝛼 = 1 − 𝑒𝑥𝑝 (−𝑘0 𝑒𝑥𝑝 (− ) 𝑡) , 𝑚=1 (3.5)
𝑅𝑇
If m = 0, however, the solution of Eq. (3.2) is
𝐸𝑎
𝑃 = 𝑃0 − (𝑘0 𝑒𝑥𝑝 (− ) 𝑡) , 𝑚=0 (3.6)
𝑅𝑇
and 𝛼 is found to be
𝑘0 𝐸𝑎
𝛼= 𝑒𝑥𝑝 (− ) 𝑡, 𝑚=0 (3.7)
𝑃0 𝑅𝑇
Hence, a measurement of the conversion percentage of the reaction as a function of time and a
determination of whether the rate is exponential as in Eq. (3.5) or linear as in Eq. (3.7) can lead to
the determination of the value of m conclusively (note that Eq. (3.7) is a linear equation in time).
Furthermore, a numerical solution of either Eq. (3.5) or Eq. (3.7) based on the data provided by
such a measurement can lead to the determination of the activation energy E a of the reaction. Most
important, a measurement of the conversion percentage of the reaction, together with resistivity
measurements, can lead to the characterization of the performance of the sensor at room
temperature and at other temperatures.
Assembly of the Sensor

Multiwalled carbon nanotubes (CNTs) were obtained from a commercial supplier. The CNTs used
have a diameter of approximately 50 nm, a length of approximately 20 μm, and were grown on a
sheet of stainless steel. Two different methods were followed to attach gold nanoparticles to the
CNTs:
(1) Electrodeposition
(2) Electroless deposition
These methods were done by means of a spontaneous redox reaction of gold ions in a solution.
Figure 3.2 shows a scanning electron microscope (SEM) photograph of gold nanoparticles attached
to the lateral surfaces of the CNTs after the electrodeposition process. The electrodeposition was
carried out with the CNTs acting as the cathode and with an electrolyte consisting of an HAuCl 4
(Au3+) solution. Electric field intensities ranging from 5 V/cm to 100 V/cm were used. Lower field
MODULE - 2
intensity and a longer exposure time generally resulted in better coatings and more uniform
particle size distributions (the size of the deposited gold nanoparticles in this process ranged from
10 nm to 50 nm).
Figure 3.2: SEM micrograph of gold nanoparticles deposited on multi-walled carbon nanotubes by
electrodeposition. The CNTs were exposed for 20 min under electric field intensity of 5 V/cm.
Figure 3.3: SEM micrograph of gold nanoparticles deposited on multi-walled carbon nanotubes by
direct redox reaction.
Figure 3.4: Photograph of the assembled detector (interface circuit not included).
MODULE - 2
 Figure 3.3 shows an SEM photograph of gold nanoparticles deposited with the direct redox
reaction process. By comparing Figures 3.2 and 3.3, it is clear that the second process results
in a substantially better coating and substantially more uniform particle size distribution.
 The deposition process was performed by immersing the CNTs in a solution of HAuCl 4
dissolved in equal volumes of ethanol and distilled water, at a concentration of 5 mM. After
the CNTs were immersed for a period of 3 min, they were rinsed with distilled water and
dried in a nitrogen atmosphere.
 The CNT array used in the present detector measures 20 mm × 20 mm. After the deposition
of the gold nanoparticles, a second electrode was formed on top of the CNT array by
sputtering. The second electrode is aluminum, and has sufficient porosity due to the
deposition by sputtering.
 The pores in the aluminum electrode thus allow the CO gas to easily penetrate and reach the
gold nanoparticles attached to the CNT array.
 The assembly of the steel electrode, the aluminum electrode, and the CNT array sandwiched
between the two electrodes, constitutes the complete detector.
 The resistivity measurements described in Section 3.4 were taken between the two
electrodes of the detector. Figure 3.4 shows a photograph of the finished detector.
Experimental Results
Measurement of the conversion percentage as a function of time and temperature

Activity tests were performed to determine the conversion percentage of the reaction as a function
of time, and for different values of temperature.
The sensor was placed inside a small stainless steel reaction chamber with a volume of 3 cc, which
is slightly larger than the volume of the sensor.
The reaction chamber was in turn placed inside a programmable temperature container. CO gas
mixed with air, at concentrations ranging from 100 ppm to 600 ppm, was then pumped into the
reaction chamber.
The rate of flow of the gas/air mixture was measured with a flowmeter and was adjusted to
different values, the maximum rate being 3 cc/10 ms.
Conversion percentage as a function of time for temperatures ranging from 230 K to 300 K
MODULE - 2
The outlet of the chamber was connected to a hydrocarbon gas analyzer (model NGA 2000, from
Emerson Process Management Corp.) for monitoring the residual concentration of CO after passage
through the sensor.
It should be pointed out that the air used in the tests was ordinary room air that was filtered with a
charcoal filter in order to remove moisture and contaminants (more results related to the moisture
content are given in the following section).
Figure 3.5 shows three different plots of the conversion percentage a as a function of time. The tests
were conducted at temperatures of 300 K, 270 K, and 230 K. The data in Figure 3.5 is associated
with a CO concentration of 100 ppm.
It is clear that the plots in Figure 3.5 show a linear, rather than exponential, relationship.
Accordingly, the value of the constant m is 0. It is also clear that total conversion (from CO to CO 2)
occurs very rapidly at room temperature (within 50 ms). The value of the slope for any of the linear
plots in Figure 3.5 is given by k0/P0 exp(−Ea/RT). By comparing the slopes of two different plots,
the value of the ratio k0/P0 as well as of the activation energy Ea can be determined. From the data
shown in the figure, Ea was estimated to be approximately 15 kJ.mol −1.
Figure 3.6 shows the effect of the CO concentration in the gas/air mixture. The measurements in
Figure 3.6 were all taken at a fixed temperature of 300 K.
Clearly, the effect of the concentration on the performance of the sensor is minimal. It can be
noticed, however, that the reaction will require a slightly longer time for completion at higher
concentrations, as it is obvious that more CO molecules will be competing for access to a fixed
number of catalyst sites.
Figure 3.6: Conversion percentage as a function of time at a fixed temperature of 300 K and for 3
different CO gas concentrations
Measurement of the resistance of the CNT array as a function of CO concentration
 The resistance of the CNT array was measured with a highly sensitive ohmmeter as the CO
gas was repeatedly turned ON for a duration of 10 sec and OFF for a duration of 10 sec.
 Two multimeters were used to perform and verify the measurements: a specially modified
Keithley model 610C multimeter, and an Agilent model 34420A multimeter.
 Both multimeters are capable of directly measuring a resistance of 100nW or higher. For
measurement of resistances that are less than 100nW, however, the two mentioned
multimeters were used in a 4-wire configuration, where one multimeter acts as a current
MODULE - 2
source and the other as a voltmeter. The first test was conducted with a CO gas
concentration of 100 ppm.
 The result is shown in Figure 3.7. As shown, the resistance drops from a nominal value of 0.5
mW to approximately 1 μW when the gas is present.
Figure 3.7: Resistance of the CNT array (log scale) as a function of time, as the CO gas
(concentration: 100 ppm) is turned ON and OFF. The measurement was performed at room
temperature. The nominal (upper) value of the resistance is 0.5 mW, and the lower value
(measured when the gas was present) is approximately 1 μW.
Figure 3.8: Resistance of the CNT array (log scale) as a function of time, as the CO gas
(concentration: 600 ppm) is turned ON and OFF. The measurement was performed at room
temperature. The nominal (upper) value of the resistance is 0.5 mW, and the lower value
(measured when the gas was present) has an average value
of 10 nW.
MODULE - 2
 Hence, a good sensitivity to gas concentrations of 100 ppm or less is evident. The second test
was conducted with a gas concentration of 600 ppm. The result is shown in Figure 3.8. It is
evident that the response of the sensor is far greater, with the overall resistance dropping to
an average value of 10nW in that case.
 It is to be pointed out that the area of the sensor (that is, of the CNT array) can be reduced to
make the sensor smaller; however, the variation in the resistance of the sensor assembly will
be less in this case (as can be concluded from the well-known relationship between
resistance and cross-sectional area).
Auxiliary Experimental Results

Effect of temperature on the performance of the sensor
The effect of temperature on the reaction rate is shown in Figure 3.4. The reaction rate (and
hence the sensor’s response) is quite acceptable at room temperature, but can be much worse at
substantially lower temperatures. The sensor was also tested at elevated temperatures (up to
80◦ C), and the reaction rate was found to be generally higher at such temperatures.
Effect of moisture on the performance of the sensor

 Moisture was added to the gas/air mixture (by circulating the mixture through a valve-
controlled container filled with distilled water), and the relative humidity in the mixture was
monitored with a moisture sensor.
 Figure 3.9 shows the result of a test that was conducted at room temperature and with a CO
concentration of 100 ppm. As the figure shows, the effect of moisture was negligible for the
relative humidity (RH) range of 5% to 40%. For the range of RH from 40% to 80%, however,
the reaction rate slowed significantly.
 This can be understood on the basis of the fact that the hydrogen present in the moisture is
itself a reducing gas and will compete with the CO molecules for access to the available
catalyst sites.
Figure 3.9: Effect of relative humidity on the reaction rate at room temperature (CO gas
concentration = 100 ppm)
MODULE - 2
Other performance parameters

 Tests to determine the repeatability of the measurements were conducted. For the same gas
concentration, temperature, and relative humidity conditions, the measured resistance of
the sensor was found to be generally repeatable within the bounds of the measurements
shown in Figures 3.7 and 3.8.
 Other tests were conducted to determine the sensitivity of the sensor to contaminants. Small
amounts of contaminants such as hydrogen, ethanol, and benzene were introduced into the
gas/air mixture flowing into the sensor.
 When the contaminant was a reducing gas, such as hydrogen, the effect on the performance
of the sensor was generally proportional to the amount of contaminant present. Otherwise,
the contaminant had no effect on the CO→ CO2 reaction rate.
Conclusion
 The miniature, highly sensitive CO gas sensor presented in this section has the smallest size
and the highest sensitivity at room temperature among all the known types of CO gas
detectors.
 Furthermore, the detector requires only minimal power (in comparison with conventional
CO sensors, which require heating of the sensing element) and can be easily powered with a
battery.
 The effective power consumption of these circuits is normally very low, and hence they are
suitable for battery operation. One disadvantage of this new sensor is the cost. CNT arrays
are expensive to produce at the present time, and gold nanoparticles must be deposited on
the CNT array (although the mass of gold used in the sensor is only about 1 μg).
 With mass production, however, the cost can be substantially reduced. Regardless, the
sensor can be very useful in applications where a small size and high sensitivity are
required.
Smoke Detectors
 In this section we introduce a novel new smoke detector that is characterized by a very
small size and very high sensitivity.
 The new detector is fundamentally based on a-particle radiation source like the well-known
a-particle smoke detector; however, instead of utilizing the principle of ionization of the air,
the particles are made to strike the gate of an n-channel MOSFET.
 This results in a net positive charge on the gate of the transistor. The current through the
MOSFET will be proportional to the charge on the gate and hence to the intensity of the
particles.
 If particles of smoke enter the detector and screen the particles, the positive charge on the
gate drops, which leads to a reduced current or total cut-off of the MOSFET.
Structure
 The ubiquitous alpha-particle smoke detector has been in wide use for the past four decades.
Figure 3.10 (a) shows the basic structure of that popular smoke detector: an a-particle
emitter (such as Americium 241) emits a particle into an ionization chamber that consists of
two metal electrodes separated by air.
 The particles ionize the air molecules, which results in the release of a large number of free
electrons inside the chamber. Accordingly, the resistivity of the air inside the chamber
MODULE - 2
decreases, and a small current can be circulated between the two metal electrodes. When
smoke particles (which contain a large number of carbon atoms), however, enter the
chamber, they quickly attach to the particles and hence reduce the degree of ionization of
the air.
 The momentary increase in resistivity results in a lower current between the electrodes,
which can be detected with the external circuit. Unfortunately, the change in the resistivity
of the air in the traditional smoke detector is quite small in practice, which necessitates the
use of a large ionization chamber and an elaborate external circuit to detect the small
changes in resistivity that occur when smoke particles enter the detector.
Figure 3.10: (a) Principle of operation of the traditional a-particle smoke detector. (b) Principle of
operation of the new detector
 Accordingly, the smoke detector has been a traditionally large device. In recent years,
however, there has been an interest in a very small smoke detector that can be hidden close
to a passenger seat in an aircraft, train, or bus. Such a detector can be useful not only for
detecting cigarette smokers, but also for detecting potential terrorists (who have attempted
in recent years to use explosive/combustible materials inside public transportation
systems).
MODULE - 2
 Such a miniature smoke detector was recently introduced by the author. The fundamental
principle of the new detector is shown in Figure 3.10 (b). Instead of relying on the principle
of ionization of the air, the particles emitted from the particle source are made to strike
directly the gate of an n-channel MOSFET transistor.
 The particles, being positively charged helium nuclei, will deposit their positive charges on
the gate. The MOSFET is known to be highly sensitive to a charge on the gate and will
immediately start conducting when the particles reach the gate.
 The current I through the MOSFET will be proportional to the charge on the gate, and hence
to the number of a particles reaching the gate. If smoke particles enter the detector and get
attached to the particles, a lower positive charge will be present on the gate, which will lead
to a lower current or total cut-off of the MOSFET.
 Theoretical and experimental investigations will show that this principle is highly sensitive
in comparison with the old technique of detecting smoke, and hence this new device can be
made substantially smaller than the traditional ionization-based smoke detector (in fact, the
new detector is much smaller than the audio buzzer that is typically connected to the
detector as an alarming mechanism). The advantage of radiation-based detectors in
comparison with optical detectors is their capability to detect very minute (invisible)
amounts of smoke.
Qualitative Description of the Detector
Figure 3.11 shows a schematic of the actual circuit used in the present prototype. As shown, an a-
particle source is arranged such that the emitted a particles strike a small plate that is directly
connected to the gate of an n-channel, enhancement mode MOSFET.
The a-particle source is a commercially available thin foil containing the Am isotope. An optional 10
MW resistor can be connected to the gate, as shown, to drain the positive charge from the gate
when the particles are not present.
Figure 3.11: Circuit used in the present prototype. A DC buzzer can be connected to the output of
the driving transistor to obtain an audible alarm
The voltage on the drain terminal of the MOSFET is sensed with a voltage-follower circuit
consisting of a high-impedance opamp and an NPN transistor. The purpose of the NPN transistor is
MODULE - 2
to act as a load driver (for driving a DC buzzer or other means for indicating the presence of
voltage).
When the particles are striking the gate of the MOSFET and a positive charge is present, the
MOSFET conducts and hence the output voltage will be low. If, however, the particles are screened
by the presence of smoke and no positive charge is present on the gate, the MOSFET shuts off. But
since the input to the opamp is connected to a pull-up resistor, the output voltage in that case will
be high (indicating an alarm).
Figure 3.12 shows a photograph of the new detector next to a traditional a-particle smoke detector
(for size comparison). The prototype shown is complete except for the optional DC buzzer, which is
not included in the photograph.
Figure 3.12: Photograph of the new detector (small circuit board on right) next to a traditional a-
particle smoke detector
Theory
Determination of the gate voltage VGS
 The radioactive foil used in the present detector contains approximately 0.1 μCi of the
particle emitter 241Am (one-tenth of the quantity that is typically used in household smoke
detectors). 0.1 μCi is equivalent to 3700 emissions/s. Since the helium nucleus contains two
protons, the equivalent of 7400 positive electron charges will be striking the gate of the
MOSFET each second.
 To determine the gate-source voltage VGS, consider the simplified cross-sectional drawing of
an n-channel MOSFET shown in Figure 3.13. There are two capacitances of interest in this
application.
 The first is the gate-source capacitance, which can be measured and is usually given in the
datasheet of the device.
 The second capacitance is the gate-channel capacitance which is a substantially smaller and
much harder to determine capacitance (that capacitance must be calculated from an
elaborate mathematical model of the device).
MODULE - 2
Figure 3.13: Structure of the n-channel MOSFET
 Due to the bombardment of the gate with a particles and the deposition of positive charges
on the gate, electrons will be attracted to the channel by induction.
 Since there is no well-defined voltage source between the gate and the source terminals, the
only capacitance that must be accounted for when attempting to calculate the voltage rise on
the gate will be the gate-channel capacitance.
 The normal approach would be to use the well-known equation Q = CV where Q is the charge
on the gate and C is the capacitance, to determine the unknown voltage. Unfortunately, such
a calculation cannot be performed since the gate-channel capacitance cannot be determined
with reasonable accuracy.
An alternative approach for determining the steady-state gate-source voltage VGS is the following:
the steady-state current that will be flowing into the gate terminal will be given by
𝑑𝑄 7400 × 1.6 × 10−19

𝐼= = = 1.18 × 10−15 𝐴𝑚𝑝𝑠 (3.8)
𝑑𝑡 1
Under steady-state conditions, this very small current of about 1 fA will constitute a leakage current
that will flow to the source terminal, through the insulating silicon dioxide (SiO 2) layer that isolates
the gate terminal. The leakage current density J through the SiO2 layer will be related to the electric
field intensity E between the gate and the source terminals by the well-known relationship
𝐽 = 𝜎𝐸 (3.9)
Where s is the conductivity of SiO2 (this value is approximately 10 −16 W−1 m−1). The above equation
can be written as
𝐼 𝑉𝐺𝑆
= 𝜎 (3.10)
𝐴 ∆𝑥
Where A is the surface area of the gate electrode and ∆x is the distance, in general, between the gate
and the source. VGS can be calculated by using the expression
𝐼∆𝑥
𝑉𝐺𝑆 = (3.11)
𝜎𝐴
MODULE - 2
The ratio ∆x/A can now be determined by knowledge of the gate-source capacitance. According to
the datasheet of the MMBF170 transistor used in the present prototype, C ≈ 24 pF
𝐴
𝐶 = 𝜀0 𝜀𝑟 (3.12)
∆𝑥
where 𝜀0 is the permittivity of free space and 𝜀𝑟 is the relative permittivity of the insulating
material (𝜀𝑟 = 3.9 for SiO2), the ratio ∆𝑥 /A can be calculated to be
∆𝑥 𝜀0 𝜀𝑟
= ≈ 1.438 𝑚 −1 (3.13)
𝐴 𝑐
By substituting with the quantities determined above into Eq. (3.11), the steady state VGS is found to
be approximately equal to 17V. Direct measurement of the gate voltage with a Keithley high-
impedance electrometer has confirmed this estimated voltage. It is to be pointed out that the
steady-state gate voltage is typically reached within a fraction of a second in the present
application.
Determination of the ON-OFF operating points of the detector
Figure 3.14 shows the load line for the circuit in Figure 3.11 and the two MOSFET characteristic
curves that are of interest in this application. Because the load attached to the MOSFET (the 10 kΩ
pull-up resistor) is large, the MOSFET will be operating in the ohmic region, as Figure 3.14 shows.
When VGS = 0 (no charge on the gate), the MOSFET will be cut off and V DS ≈ Vcc, which is 6 V in the
present prototype. This is the lower operating point in Figure 3.14. To determine the upper
operating point (where VGS = 17 V), we must solve the following two simultaneous equations
Figure 3.14: Load line for the 10 kΩ pull-up resistor, and the two operating curves of the MOSFET.
The drain current ID is plotted versus the drain-source voltage VDS, with an arbitrary scale for ID.
𝑉𝑐𝑐 1
𝐼𝐷 = − 𝑉𝐷𝑆 (3.14)
𝑅 𝑅
𝑉𝐷𝑆
𝐼𝐷 =
𝑅𝑂𝑁
MODULE - 2
Where R is the value of the load resistance (10 kΩ) and Ron is the ON resistance of the MOSFET (the
inverse of the slope of the characteristic curve shown in Figure 3.14). The first of the above
equations is the load-line equation, and the second is the equation representing the characteristic
curve. The solution of the two simultaneous equations is
𝑉𝐶𝐶
𝑉𝐷𝑆 = (3.13)
1 + 𝑅 ⁄𝑅𝑂𝑁
With R = 10 kΩ, and the MOSFET resistance Ron being typically equal to 1 Ω, it is clear that VDS ≈ 0
when the MOSFET is fully ON.
Experimental Results
Sensor output as a function of obscuration
When smoke enters the detector with various densities and screens the a particles, the operating
point of the sensor will move from point A in Figure 3.14 to point B. Accordingly, the output voltage
(VDS) will increase from 0 to approximately Vcc (which is 6 V in the present prototype). Figure 3.15
shows the measured output voltage as a function of the obscuration (in percentage). The level of
obscuration was determined at each point in the graph of Figure 3.15 in accordance with standard
NFPA-270, where a laser detector was used in a conical radiant source configuration to measure the
smoke obscuration. The results in Figure 3.15 were obtained by testing the sensor in still air.
Figure 3.15: MOSFET drain-source voltage (output of the circuit in Figure 3.11) as a function of the
obscuration.
The tests were also conducted by using moving air, and a very small improvement in the sensitivity
was observed. Since the improvement is incremental, it will not be reported here. Figure 3.16
shows the measured output of a conventional household smoke detector (as percentage of full-
scale output) as a function of obscuration.
MODULE - 2
Figure 3.16: Output of a conventional household ionization-based smoke detector (as percentage of
full-scale output), as a function of obscuration.
By comparing Figures 3.15 and 3.16, it is clear that the new sensor has substantially higher
sensitivity at low levels of obscuration. The sensitivity is, however, the same for levels of
obscuration higher than 35%.
Sensor output as a function of the distance between the a-particle source and the MOSFET
gate
The output of the sensor was measured as a function of the distance between the a-particle source
and the gate of the MOSFET, at a fixed obscuration of 30%. The result is shown in Figure 3.17. For
the present prototype, the distance between the a-particle source and the gate is 5 mm.
Figure 3.17: Output of the sensor as a function of the distance between α-particle source and the
gate, at a fixed obscuration of 30%.
MODULE - 2
Effect of the ambient temperature
The MOSFET is known to be slightly sensitive to temperature. More specifically the drain current ID,
decreases at a temperature higher than room temperature and increases at lower temperature. For
the circuit in Figure 3.11, the result is that the output voltage VDS drifts higher and lowers with
temperature. To test the effect of temperature variation, the sensor was placed inside a variable
temperature chamber and the tests shown in Figure 3.15 were repeated. Figure 3.18 shows the
results for temperatures of +80°C and −40°C. Clearly, the effect of the ambient temperature on the
performance of the sensor is minimal.
Figure 3.18: MOSFET drain-source voltage (output of the circuit in Figure 3.11) as a function of the
obscuration, at temperatures of +80°C, +25°C, and -40°C.
Conclusion
The new α-particle smoke detector described in this section is substantially smaller and
substantially more sensitive than conventional smoke detectors that depend on the principle of air
ionization. The principle used in the new detector, namely, the transfer of the charge of the α
particles to the gate of a MOSFET, allows the construction of a detector that is characterized by
small size and high sensitivity in comparison with the ionization chamber that is used in
conventional α-particle-based detectors. As Figure 3.15 shows, the sensitivity of the new detector is
substantially high at low levels of smoke obscuration. The sensor will be very useful for detecting
smoke in tight places, such as next to a passenger’s seat in a bus or airplane. As indicated earlier in
this section, the advantage of radiation-based detectors in comparison with optical detectors is
their capability of detecting very minute amounts of smoke.

21EE742 Module - 2

Uploaded by

Copyright:

Available Formats

21EE742 Module - 2

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

21EE742 Module - 2

Uploaded by

Copyright:

Available Formats

21EE742 – MICRO & NANO SENSORS &

Regulation – 2021 TRANSDUCERS

ULTRAHIGH SENSITIVITY, WIDE DYNAMIC RANGE SENSORS

 A number of new applications, such as self-guided small projectiles and autonomous

Figure 2.1: Principle of operation of the new acceleration sensor

Figure 2.4: Mechanical structure of the sensor (scale = 16:1)

𝐹 = 2𝜋𝑟𝛾 cos(180° − 𝜃) (2.1)

Figure 2.5: Array of CNTs penetrating a liquid electrolyte

𝐹𝑇𝑜𝑡𝑎𝑙 = 𝑚𝑎 = 2𝜋𝑟𝑁𝛾 cos(180° − 𝜃) (2.2)

Physical parameters of the sensor and the dynamic range:

By substituting with the above parameters in Eq. (2.9), we obtain

Measurement of the capacitance constant K

OTHER MOTION AND ACCELERATION MICROSENSORS

Figure 2.14: Photo of actual sensor

A CO Gas Sensor Based on Nanotechnology - Structure, Theory, Assembly of Sensors,

A CO Gas Sensor Based on Nanotechnology

Figure 3.1: Fundamental principle of operation of the new CO detector

2𝐶𝑂 + 𝑂2− → 2𝑐𝑜2 + 𝑒 − (3.1)

 Sensitivity to CO concentrations of less than 100 ppm at room temperature

and hence the conversion percentage a is found to be

If m = 0, however, the solution of Eq. (3.2) is

Assembly of the Sensor

Measurement of the conversion percentage as a function of time and temperature

Measurement of the resistance of the CNT array as a function of CO concentration

Auxiliary Experimental Results

Effect of moisture on the performance of the sensor

Other performance parameters

Qualitative Description of the Detector

Figure 3.13: Structure of the n-channel MOSFET

𝑑𝑄 7400 × 1.6 × 10−19

Determination of the ON-OFF operating points of the detector

Effect of the ambient temperature

You might also like