Comparison of Three MEMS Thermal Conductivity Sensors

Tu Xiang Zheng

MEMS sensors have greatly matured and proliferated in use amongst many disciplines. There has been great interest in MEMS thermal conductivity sensors due to the benefits of miniaturization: low cost, small device footprint, low power consumption, greater sensitivity, integration with on-chip circuitry, etc. which can be used to analyze a whole range of gas mixtures, provided that there are only two gases present and that the two gases have significantly different thermal conductivities. 

A single silicon wafer MEMS thermal conduction sensor has developed by the present author. The sensor consists of a heat transfer cavity with a flat bottom and an arbitrary plane shape, which is created in a silicon substrate. A heated resistor with a temperature dependence resistance is deposed on a thin film bridge, which is the top of the cavity. A heat sink is the flat bottom of the cavity and parallel to the bridge completely. The heat transfer from the heated resistor to the heat sink is modulated by the change of the thermal conductivity of the gas or gas mixture filled in the cavity. This change can be measured to determine the composition concentration of the gas mixture or the pressure of the air in a vacuum system.

In order to know the advantages of the present MEMS thermal conductivity sensors a comparison of several MEMS thermal conductivity sensors is made. The others MEMS thermal conductivity sensors were developed by Arndt et al’s and Hsiai et al’s.

 

1. Comparison and contrast between the Arndt et al.’ sensors and the present sensors are shown in Fig.1. Several significant differences between them can be found from the comparison and contrast. These significant differences include the follows:

(1) The Arndt et al.’ sensors are not single silicon wafer MEMS thermal conduction sensors. Actually it is constructed by three wafer including one base plate and two porous cover plates. In contrast, the present sensors are single wafer MEMS thermal conduction sensors that are constructed from single silicon wafers.

(2) The heat transfer cavity of the Arndt et al.’ sensors can not be arbitrary shape because it is formed by KOH etching that limits the shape of the cavity to be bounded by the (111) crystal planes of the silicon wafer. Compared with the Arndt et al.’ sensors, the heat transfer cavity of the present sensors can be arbitrary shape because it is by anodization in HF solution which is not governed by the crystal structure of the silicon wafer.

(3) The side wall of the heat transfer cavity of the Arndt et al.’ sensors can not be curved due to limitation of the crystal structure of the silicon wafer. On the contrary, the side wall of the heat transfer cavity of the present sensors can be curved due to its formation without limitation of the crystal structure of the silicon wafer.

(4) The Arndt et al.’ sensors have two heat transfer cavities with one on the top of the base plate and the other on the bottom of the base plate. Unlike the Arndt et al.’ sensors, the present sensors only have one heat transfer cavity.

(5) The two heat transfer cavities of the Arndt et al.’ sensors with each is formed by bonding a silicon base plate and a porous plate together. Contrasted with the Arndt et al.’ sensors, the heat transfer cavity of the present sensors are formed inside the silicon wafer without using an additional wafer.

(6) The Arndt et al.’ sensors have two adhesive layers that are used for bonding three different plates together. Conversely the present sensors have no adhesive layers because it is a single wafer structure without wafer bonding.

In summery, almost no similarities between the Arndt et al.’ sensors and the present sensors have been found. There is impossible for the Arndt et al.’ sensors to teach the present sensors.

2. Comparison and contrast between the Hsiai et al.’ sensors and the present sensors are shown in Fig.2. It can be found that the Hsiai et al.’ sensors and the present sensors are related to different types of sensors according to the following reasons:

The Hsiai et al.’ sensors are fluid shear stress sensors that are based on heat transfer by fluid convection. But the present sensors are gas thermal conduction sensors that are based on heat transfer by gas conduction.

The Hsiai et al.’ sensors do not need a heat transfer cavity because fluid convection occurs over the diaphragm. However the present sensors do need a heat transfer cavity because gas conduction occurs inside the cavity.

The cavity of the Hsiai et al.’ sensors have no bottom used for heat sink because they have no heat transfer function. Nevertheless the heat transfer cavities of the present sensors have a bottom so that the heat generated by the heater on the top of the cavity can be transferred to the bottom through the gas filled in the cavity.

In view above mentioned three major differences between the Hsiai et al.’ sensors and the present sensors, it should be very clear that the present thermal conduction sensors are not able to learn from the Hsiai et al.’ fluid shear stress sensors.

3. The present sensors are fabricated using porous silicon micromachining technique instead of KOH etching micromachining technique that is used by Arndt et al. and Hsiai et al.

(1) KOH etching of silicon is an anisotropic etching. The main characteristics of this micromachining technique can be listed as:

Etch rates of crystallographic planes are: (100) > (110) > (111)

(111) family of crystallographic planes are the “stop” planes

producing standard anisotropic etching structures: V-grooves and pyramidal cavities

From these characteristics it can conclude that the cavities of the Arndt et al.’ sensors and the Hsiai et al.’s sensors are pyramidal cavities with (111) sidewall planes, because they all use KOH etching to create their cavities. It also can be conclude that the heat transfer cavity with curved sidewalls provided by the present sensors can not be created using KOH etching technique.

(2) In addition to KOH micromachining technique, a porous silicon based micromachining technique is more and more widely used for MEMS devices fabrication. Porous silicon layers can be fabricated by partially electrochemical dissolution of silicon wafers with etching masks covered on the surface of the silicon wafers. Then the porous layers are etched away to form microstructures including cavities. The shape of the cavities is determined by the etching masks without limitation based on the silicon crystallographic structure. Like the present sensors the cavities of the sensors can have arbitrary shape and curved sidewalls.

4. Conclusion

Comparing and contrasting between the Arndt et al.’s sensors and the present sensors and between the Hsiai et al.’s sensors and the present sensors have been made respectively. Six significant differences between the Arndt et al.’s sensors and the present sensors have been found. It can be seen that it is impossible for the present inventor to learn from the Arndt et al.’s sensors for designing and fabricating the present sensors. Three major differences have been found between the Hsiai et al.’s sensors and the present sensors. Actually, the Hsiai et al.’s sensors and the present sensors belong to different sensor types. It is not able for the present inventor to learn how to design the present sensors from the Hsiai et al.’s sensors. Moreover the present sensors use porous silicon based micromachining technique for fabricating the present sensors, which is different from KOH micromachining technique used by the Arndt et al.’s sensors and the Hsiai et al.’s sensors. Hwy the heat transfer cavity with arbitrary shape and curved sidewalls provided by the present sensors can not been created in the Arndt et al.’s sensors and the Hsiai et al.’s sensors, because they all use KOH micromachining technique instead of porous silicon based micromachining technique.

Measuring Bike and Wind Velocities Using Thermal Motion Sensors

Xiang Zheng Tu

It is widely acknowledged that cycling is one of the best ways for people to achieve good health and fitness. There are more than a billion bicycles in the world, twice as many as automobiles. In recent years bike production had climbed to over 100 million per year, which is compared to 50 million cars. Bicycles were introduced in the 19th century and since then have been and are employed for many uses: recreation, work, military, show, sport etc. In the USA, people use bikes for slimming and better feeling, but other countries people use bikes mostly for transportation needs. For these reasons in some countries bikes are especially popular.

How many calories burn when cycling?  It is best to known that the energy consumed by cycling which can be expressed as:

P = C_m * V_b* [ c_d  * A * ρ/2 * ( V_b + V_w )² + F_rg  + V_b * C_rvn ],        (1)

where P is rider's power, V_b is velocity of the bicycle, Vw is wind speed, C_m  is coefficient for power transmission losses and losses due to tire slippage, C_d  is air drag coefficient, A is  total frontal area (bicycle + rider),  ρ is air density, C_rVn is coefficient for the dynamic rolling resistance, normalized to road inclination; C_rVn = C_rV*cos(β), C_rV  is coefficient for velocity-dependent dynamic rolling resistance, β  is  inclination angle, F_rg  is rolling friction (normalized on inclined plane) plus slope pulling force on inclined plane. As seen from equation (1), the energy consumption depends on many factors including two variables: cycling velocity and wind velocity. The other factors are coefficients relating to cycling conditions.

It is worth to point out that the equation (1) is only for rough calculations, because some coefficients usually are not available. Fortunately, bicycles are used mostly in the flatlands. In this case most of the coefficients can be assumed as constant and the energy consumption of a cycling can be simply calculated using measured bicycle velocities and wind velocities.

As shown in the above figure, a thermal motion sensing system is used to measure the bicycle velocity and the wind velocity respectively when a man is riding a bicycle. The sensing system is installed in a concentric tube with an inner tube and an outer tube. Several small holes are drilled around the outside of the two ends of the outer tube and two end holes are communicated through the outer tube. A large hole is drilled through the axis of the inner tube, which is separated from the outer tube by the tube wall. The concentric tube is pointed in the direction of the bicycle traveling. According to the principle of relative motion, an air flow caused by the cycling passes through the large hole. The air flow comprises two components: a cycling flow and a wind flow. The cycling also causes another air flow which passes through the outer tube. Since the outside holes are perpendicular to the direction of the bicycle traveling any head wind or tail wind can not enter the outer tube. So this air flow only comprises the cycling flow.

Referencing to the above figure, there are two thermal motion sensors in the concentric tube. One of them is disposed on the wall of the outer tube, which is used to measure the total velocity V_t of the air flow passing through the inner tube. The other is disposed on the wall of the inner tube, which is used to measure the cyclist velocity V_c of the air flow passing through the outer tube. The total velocity V_{t  is} related to the cycling in a head wind or a tail wind and is equal to the sum of the cyclist velocity V_c and the head wind velocity V_hw  or the tail wind velocity V_tw. So the wind velocity can be calculated by

V_hw = V_t - V_{c ,}       V_tw = V_t + V_c         (2)

The thermal motion sensor is made using both microfabrication techniques and Micro-Electro-Mechanical Systems, or MEMS technologies and operated based on gas convective heat transfer. The sensor comprises of a thermal insulated base recessed into and surrounded by a silicon chip. A heating resistor and the hot junctions of a thermopile display on the surface of the base. The cold junctions of the thermopile display on the surface of the silicon chip near the base edge. The heating resistor is heated to maintain a continuous overheat between the resistor and the air flowing over the surface of the sensors. The thermopile is functioned as a temperature sensor. Both the actuations are performed by a POSIFA proprietary integrated circuitry.

The POSIFA proprietary integrated circuitry is an 8 bit MCU based microcontroller, since the thermal motion sensors applications do not involve a lot of number crunching, just waking up periodically to check for sensors input and making a decision on it. A low noise programmable operation amplifier is embedded into the microcontroller. This is done in order to match the thermal motion sensors with very low inherent noise.  The offset of the thermal motion sensors can be completely canceled by programmable hearting of the sensor resistor. This can be realized using a modulated DAC output of the microcontroller. Actually the DAC output here is used as a controllable precision voltage source. The DAC resolution provided by the microcontroller is required to be reasonable high. 

It is widely acknowledged that cycling is one of the best ways for people to achieve good health and fitness. There are more than a billion bicycles in the world, twice as many as automobiles. In recent years bike production had climbed to over 100 million per year, which is compared to 50 million cars. Bicycles were introduced in the 19th century and since then have been and are employed for many uses: recreation, work, military, show, sport etc. In the USA, people use bikes for slimming and better feeling, but other countries people use bikes mostly for transportation needs. For these reasons in some countries bikes are especially popular.

How many calories burn when cycling?  It is best to known that the energy consumed by cycling which can be expressed as:

P = C_m * V_b* [ c_d  * A * ρ/2 * ( V_b + V_w )² + F_rg  + V_b * C_rvn ],        (1)

where P is rider's power, V_b is velocity of the bicycle, Vw is wind speed, C_m  is coefficient for power transmission losses and losses due to tire slippage, C_d  is air drag coefficient, A is  total frontal area (bicycle + rider),  ρ is air density, C_rVn is coefficient for the dynamic rolling resistance, normalized to road inclination; C_rVn = C_rV*cos(β), C_rV  is coefficient for velocity-dependent dynamic rolling resistance, β  is  inclination angle, F_rg  is rolling friction (normalized on inclined plane) plus slope pulling force on inclined plane. As seen from equation (1), the energy consumption depends on many factors including two variables: cycling velocity and wind velocity. The other factors are coefficients relating to cycling conditions.

It is worth to point out that the equation (1) is only for rough calculations, because some coefficients usually are not available. Fortunately, bicycles are used mostly in the flatlands. In this case most of the coefficients can be assumed as constant and the energy consumption of a cycling can be simply calculated using measured bicycle velocities and wind velocities.

As shown in the above figure, a thermal motion sensing system is used to measure the bicycle velocity and the wind velocity respectively when a man is riding a bicycle. The sensing system is installed in a concentric tube with an inner tube and an outer tube. Several small holes are drilled around the outside of the two ends of the outer tube and two end holes are communicated through the outer tube. A large hole is drilled through the axis of the inner tube, which is separated from the outer tube by the tube wall. The concentric tube is pointed in the direction of the bicycle traveling. According to the principle of relative motion, an air flow caused by the cycling passes through the large hole. The air flow comprises two components: a cycling flow and a wind flow. The cycling also causes another air flow which passes through the outer tube. Since the outside holes are perpendicular to the direction of the bicycle traveling any head wind or tail wind can not enter the outer tube. So this air flow only comprises the cycling flow.

Referencing to the above figure, there are two thermal motion sensors in the concentric tube. One of them is disposed on the wall of the outer tube, which is used to measure the total velocity V_t of the air flow passing through the inner tube. The other is disposed on the wall of the inner tube, which is used to measure the cyclist velocity V_c of the air flow passing through the outer tube. The total velocity V_{t  is} related to the cycling in a head wind or a tail wind and is equal to the sum of the cyclist velocity V_c and the head wind velocity V_hw  or the tail wind velocity V_tw. So the wind velocity can be calculated by

V_hw = V_t - V_{c ,}       V_tw = V_t + V_c         (2)

The thermal motion sensor is made using both microfabrication techniques and Micro-Electro-Mechanical Systems, or MEMS technologies and operated based on gas convective heat transfer. The sensor comprises of a thermal insulated base recessed into and surrounded by a silicon chip. A heating resistor and the hot junctions of a thermopile display on the surface of the base. The cold junctions of the thermopile display on the surface of the silicon chip near the base edge. The heating resistor is heated to maintain a continuous overheat between the resistor and the air flowing over the surface of the sensors. The thermopile is functioned as a temperature sensor. Both the actuations are performed by a POSIFA proprietary integrated circuitry.

The POSIFA proprietary integrated circuitry is an 8 bit MCU based microcontroller, since the thermal motion sensors applications do not involve a lot of number crunching, just waking up periodically to check for sensors input and making a decision on it. A low noise programmable operation amplifier is embedded into the microcontroller. This is done in order to match the thermal motion sensors with very low inherent noise.  The offset of the thermal motion sensors can be completely canceled by programmable hearting of the sensor resistor. This can be realized using a modulated DAC output of the microcontroller. Actually the DAC output here is used as a controllable precision voltage source. The DAC resolution provided by the microcontroller is required to be reasonable high.
 

Walking Step-By-Step Swing Leg Velocity Measurement

Xiang Zheng Tu

More and more people believe that they never avoid becoming completely old, but they may delay the time becoming old. They may look younger when they are 70 and may live into their nineties. This may be true if they pay attention to personal health care and increase the amount of physical activities such as walking and running in daily life. Research shows getting up and walking around for two minutes out of every hour can increase your lifespan by 33 percent, compared to those who do not. According to the UK’s National Health Service (NHS), the average person only walks between 3,000 and 4,000 steps per day, but aiming for 10,000 steps is a better goal.

Therefore, it necessary for the people to know how many steps they walk or more practical how much calories they consume. So many pedometers have been developed. A sophisticated pedometer likes a swinging pendulum-hammer and measures steps with two or three accelerometers. A microchip is arranged at right angles that detect minute changes in force as a people move his legs. Since accelerometers are often built into gadgets like cell phones, it's increasingly common to find these sorts of pedometers. 

Counting steps with a pedometer based on inertial measurement units (IMUs) sounds super-scientific, but you need to remember that it's only an approximate measurement. Not all steps will be correctly counted and some false movements such as jolts in the road as people ride in a car might be counted as steps too. Don't take the count too seriously; assume that it's in error by least 10 percent.

IMUs can only be used for counting step number but not for measuring step length. Weiwei et al subjected that the swing leg moves at a constant speed, the energy cost of swing foot movement is inversely proportional to the step length. The fact is that step length varies from person to person and also varies for the same person when walking or running at different speeds. So without step length it is impossible to calculate the calories consumption precisely.

According to Kuo et al’s assumptions, the energy or calories E consumed in a leg swing during human walking has a component corresponding to the cost of pushing off with the stance leg (i.e., toe-off) E_toe, and a component corresponding to forced motion of the swing leg, E_sming. The energy E can be expressed as

E = E_toe + E_swing ~ω^-1 υ³ +c ω^k-λ υ^2λ,   (1)

where ω is the swing leg natural frequency, υ is the swing leg velocity, c is the relative proportional between E_toe and E_swing, κ and λ are the metabolic cost exponents. The swing leg natural frequency ω related to the swing leg speed υ by a simple approximation as

υ ~ ω^1/2 p^1/2,    (2)

where p is the toe-off impulse which is applied heel strike. From equations (1) and (2), the calories consumption E at a given walking step is mainly determined by the swing leg velocity υ. It seems that to get more accurate calories consumption E needs to measure the swing led velocity directly.

Direct calories consumption measurement can be done using a new type of pedometers provided by the present author. The new pedometer uses three thermal motion sensors to measure the linear velocity of each swing leg trajectory at human walking or running. As shown in the above figure, a man is walking across the paper. His legs swing through the space resulting in two curved paths which are called swing leg trajectories. The blue one is the right leg swing trajectory and the red one is the left leg swing trajectory. Each trajectory can be divided into two sections: the first or up section and the second or down section. Three thermal motion sensors are attached to the right shoe of the walking man and arranged to be along with y+, x+ and y- axis of a sensor coordinate system respectively. The sensor coordinate system is tracking the right swing leg trajectory. In the up section of the trajectory the linear velocity υ_up of the swing right leg motion is located in the first quadrant of the sensor coordinate system and its y+ component υ^y+_up

and x+ component υ^x+_up can be measured by y+ motion sensor and x+ motion sensor , respectively. In the down section of the trajectory the linear velocity υ_down of the swing right leg motion is located in the forth quadrant of the sensor coordinate system and its y- component υ^y-_down and x+ component υ^x+_down can be measured by y- motion sensor and x+ motion sensor, respectively. Understood that for any right triangle the side lengths and hypotenuse are related by the Pythagorean theorem:

υ​²​​ = υ​_x​​​²​​ + υ​_y​​​²​​.     (3)

The magnitude of the linear velocity υ can be found simply by taking the square root.

The measured linear velocity is an instant velocity of a swing leg movement. A walking step length can be calculated by integration of the measured instant velocity. Each walking step is performed with the following actions:

·      Lift one leg off of the ground,

·      Using the leg in contact with the ground, push your body forward,

·      Swing your lifted leg forward until it is in front of your body, and

·      Fall forward to allow your lifted leg to contact the ground.

The start and the end velocities are always zero, since at these points the leg should be in contact with the ground. So the integration is a definite integral witch is limited by an interval between the start zero and the end zero. In the definite integral the answer does not have any constants such as an infinite integral. This means that the walking step length can be determined by integration of the linear velocity without need extra information.

The thermal motion sensors measure the linear velocity of the swing leg movement is based upon the concept of convective head transfer. Each thermal motion sensor comprises of a thermal insulated base recessed into and surrounded by a silicon chip. A heating resistor and the hot junctions of a thermopile display on the surface of the base. The cold junctions of the thermopile display on the surface of the silicon chip near the base edge. The sensors are attached to a shoe and move forward with the swing leg of a walking man. The heating resistor is heated to maintain a continuous overheat between the resistor and the flowing air over the surface of the sensors. The thermopile is functioned as a temperature sensor. Both the actuations are performed by a POSIFA proprietary integrated circuitry. As air flows by the heated sensors, flowing air molecules transport heat away from the sensors and as a result, the sensors cool and the heat is lost. The circuit balance is disrupted, and the temperature difference between the heated resistor and the air is produced. The temperature difference is measured and converted to the data representing the velocity of the sensors movement.

An academic Speech Trailer

Xiang Zheng Tu

The following trailer was posted by School of Computer & Communication Engineering, University of Science and Technology Beijing on April 21, 2016. The topic of the speech: Our MEMS sensors based on porous silicon micromachining technology and their applications in IOT. The speech was given by the present author.

Temperature Modulated MEMS Metal Oxide Gas Sensors

Tu Xiang Zheng

Metal oxide sensors are very popular as a consequence of their reasonable price and good durability.  However, they are lack of selectivity and response drift, which is why are used in low cost alarm-level gas monitors for domestic and industrial applications. It is important to propose new methods, which are able to improve the state-of-the-art in gas sensing. In order to do so, the temperature modulation of sensors has been proposed.

From electron spin resonance (ESR) measurements it has been proposed that adsorbed oxygen can be present in various chemical species transferring electrons from gas sensing oxide to the chemisorbed oxygen according to the following process:

O₂(gas)⇔O₂(ad)⇔O₂⁻⇔O⁻(ad)⇔O₂⁻(ad)⇔O₂⁻(lattice)

The temperature dependence of the different species has been examined. It states that an oxygen transition temperature is at 150 °C. Below 150 °C oxygen is mainly present as O₂^- and above chemisorbed oxygen in the forms of O^- or O^2- is present. This change in chemistry was correlated to a decrease in sample conductivity that occurred at around 160 °C. From these dependences the following basic mechanism for detection of combustible gases seems plausible: If the gas sensor is operated under ambient conditions it can be assumed that chemically adsorbed oxygen species are present at the surface. Combustible gases may react with these oxygen species and thus result in depletion of charged surface oxygen which in turn increases the conductivity of the gas sensing material. In this way combustible gases may not directly interact with the gas sensing material but its presence controls the concentration of pre-adsorbed oxygen, which controls the surface charge and thus the conductivity of the gas sensor.

According to this mechanism, it can be concluded that the modulation of a metal oxide sensor working temperature alters the kinetics of adsorption and reaction that occur at the sensor surface in the presence of atmospheric oxygen and other reducing or oxidizing species. It is reasonable to inference that sensor response patterns are characteristic of the species present in the gas mixture. Actually, many works have demonstrated that modulating the operating temperature of the sensors can achieve a high degree of selectivity. As an example, two components in a mixture of CO and NO₂ in air has been  simultaneously and accurately quantified by processing the response dynamics of a single micromachined tin oxide sensor operated in a temperature-modulated mode.

Similar to the above example, the MEMS metal oxide gas sensors proposed by the present author have their operating temperature modulated in a more efficient way. As well known, the thermal time constant of screen-printed sensors is quite large. As a result, up to now the temperature modulation frequency (20 MHz) has been too low and the corresponding principle-related response time (50 s) has been too high for many applications. With a special design, the thermal response of the MEMS metal oxide gas sensors is as low as 0.8 ms, as shown in the above figure. It compares much favorably with the thermal response of seconds found in conventional sensors.

The MEMS metal oxide gas sensor is based on a silicon wafer and fabricated utilizing CMOS technologies. Since the sensor is required to be operated at an elevated temperature a thermal insulating base is formed in the silicon wafer which is used to support the sensor body. Both a resistor for heating and a thermopile for temperature sensing are formed on the thermal insulating pad. Then depositing an electrical insulating layer and laying a tin dioxide layer is formed thereon. By employing such device structure with good thermal insulation to the silicon wafer, the sensor presents a series of advantages such as miniaturized size, low power consumption, and fast response.

In operation, the MEMS metal oxide sensor is exposed to a gas mixture, using a fully automated test setup, which consisted of computer driven mass flow controllers, a sensor chamber, and a data acquisition system for measurements in the millisecond range. The temperature of the sensor is varied by applying a modulate voltage to the heating resistor. Temperature range and frequency have been optimized.

Important features can be extracted from the sensor responses in two ways: the fast Fourier transform (FFT) and the discrete wavelet transform (DWT). Principal component regression (PCR), partial least squares (PLS), and multilayer perception neural networks (MLP) can be used to build quantitative predictive models. Then the different components of the mixture can be quantified precisely.

Portable Smell Generators

Tu Xiang Zheng

A portable smell generator is proposed by the present author. The smell generator is constructed simply by using at least one MEMS vaporizer. The MEMS vaporizer converts an essential oil into its vapor. If the essential oil is a rose essential oil you will be reminded of a bouquet of beautiful flowers send to you by your friend. The smell generator may comprise a lot of MEMS vaporizers each of them converts a certain essential oil into its vapor. Then any favorable smell may be produced in accordance with the will of the people.

Actually, smells have been classified by a variety of methods, depending on the application. In the food industry, a convenient method is classification by the identity of the edible material of which they are reminiscent. The food smells have been grouped into caramel, honey, vanilla, citrus, and butter. In the cosmetic industry, smells are more likely classified by floral and herbal groupings, such as jasmine, rose, balsam, or pine. The smalls of the majority of foods and perfume can be produced by a combination of these primary smells, just as colors are produced by a combination of three primary colors. 

In order to do so all necessary essential oils need be produced. Essential oils are volatile and liquid smell compounds from natural sources, usually plants. Essential oils are prepared by fragrance extraction techniques such as distillation, pressing, or maceration. There are more than 90 essential oils, such as allspice essential oil, angelica essential oil, bay essential oil, benzoin essential oil, and  bergamot essential oil, each has its own health benefits. There also have synthetic fragrance oils which are primarily made from petrochemicals and attempt to duplicate the smell of a specific plant.    

The MEMS vaporizer is a silicon-based device.  It is fabricated using CMOS processing techniques and MEMS processing techniques. A MEMS vaporizer composes: a silicon substrate, a micro-channel array created in the silicon substrate, a membrane suspending, at least a resistance heater disposed on one the membrane, a resistance temperature sensor disposed on the membrane, two cavities are created in the silicon substrate, which all are integrated to form a vaporizer chip. A printed circuit board is for packaging the vaporizer chip and a reservoir is for inserting the printed circuit board with the vaporizer chip. An essential oil liquid is stored in the reservoir, and an air filter disposed on the top of the reservoir which allows air entering the reservoir and a same volume of the liquid in the reservoir entering the micro-channel array. 

In the operation, each of MEMS vaporizers is applied with a voltage and connected with a power field-effect transistor (FET). The power field-effect transistor functions as a switch for heating the vaporizer on demand. The temperature sensors are used to measure the temperatures of the essential oil entering the micro-channel and vaporizing respectively. A signals produced by the temperature sensors are amplified by a pre-amplifier and then send to a microcontroller for digital processing. After digital processing, the microcontroller will send a pulse-width modulation (PWM) to the gate of the power field-effect transistor which allows the voltage being applied to the vaporizer.

The portable smell generators can achieve small size, light weight, low power consumption, and low cost without sacrificing performance or features. Due to these advantages it become handy to carry it anywhere alongside. While it has some advantages in the smartphone and gaming market, it offers an attractive path of continued innovation especially for designers of wearables and smart clothing.

Single Crystal Silicon Micromachined Capacitive Microphone

Tu Xiang Zheng

 

In 2006, the present author designed a Micro-Electro-Mechanical System, or MEMS microphone as shown in the above picture. The MEMS microphone is capacitive sensing device. In essence, it operates like a high frequency pressure sensor. It is comprised of two capacitor plates that, under the influence of the sound wave, vibrate with respect to each other. The resulted variation of the capacitance is then amplified by an interface circuit to produce either an analog or digital output signal.

The market for MEMS microphones has been growing steadily. It was reported to reach approximately $1.2 billion in 2015. Of the total market, almost $900 million is coming from smart phone, tablets, and wearable platforms. Additional applications include hearing aids, automotive, virtual reality, headsets, smart home, and internet of thing. As the quality of MEMS microphones continues to improve, new applications such as far field and directional audio are emerging.

Many efforts have been made to fabricate acoustic capacitive microphones. W. Kuhnel et al. have reported a micromachined subminiature capacitive microphone. The described capacitive microphone consists of a membrane chip and a back plate chip. The membrane chip has a silicon nitride thickness of 150 nm and a metallization layer thickness of 100 nm. The back plate chip has an electrode on a silicon bridge. Both the chips are fabricated respectively and then bonded together to form a capacitor.

J. J. Bernstein et al. have reported the fabrication and results of very high sensitivity acoustic transducers fabricated using surface and bulk silicon micro-machining techniques in a manufacturing environment. The silicon microphone described here is a capacitive microphone. The basic movable element is a thin (.about.3 micron thick) diaphragm made from p+ silicon. The p+ silicon is one side of an air gap capacitor. The p+ regions are formed using boron solid source diffusion at high temperatures. The other plate of the capacitor is a 20 micron thick perforated gold back plate formed using electroplating. The air gap is defined using a 2.2 micron thick sacrificial photoresist.

Altti Torkkeli et al. have reported a capacitive silicon microphone. The reported capacitive silicon microphone consists of two freestanding polysilicon membranes, a low-stress bending membrane and a high-stress back plate, which are separated by an air gap. A back chamber is arranged by encapsulation and static pressure changes are prevented with small equalization holes in the bending membrane. The device is fabricated combining bulk and surface micromachining techniques. Silicon substrates are etched in TMAH and sacrificial oxide between the membranes is etched in PSG-etch followed by freeze drying to prevent sticking.

The microphone design has gone through a number of iterations since the fabrication of the first batch of working devices. The most notable efforts have been made to reduce the thickness of the flexible plate and the air gap and lower the bias voltage of the capacitor.

However, it should be pointed out that difficulties have frequently been encountered with such efforts. In a thin plate there are two kinds of forces which resist deflection in response to acoustic signals. The first kind of force includes plate bending forces which are proportional to the thickness of the plate. These forces can be reduced by using a very thin plate. The second kind of force, which resists deflection, includes membrane forces which are proportional to the tension applied to the plate. In the case of a thin plate, tension is generally a result of the fabrication technique and of mismatches in thermal expansion coefficients between the plate and the particular means utilized to hold the plate in place. The thermal mismatched tension lowers the flatness of the plate. Reducing the thickness of the plate and air gap may mean the capacitor plates pulling together under a lower bias voltage.

The present design provided a single crystal silicon micromachined capacitive microphone whose capacitor structure comprises a single crystal silicon substrate, an acoustic cavity recessed from the back side of the substrate, a flexible single crystal silicon plate with the edge clamped to the inside of the substrate and the rear side facing the cavity, a single crystal silicon contained supporting frame having the top surface coated with a thin insulating layer, a stiff and perforated single crystal silicon plate supported at the edge by the supporting frame, an air gap sandwiched by the flexible plate and the stiff plate and surrounded by the supporting frame, and two electrodes disposed around the stiff and perforated plate and interconnecting to the flexible plate and the stiff and perforated plate, respectively.

Compared with the prior art capacitive microphone, it is easy to find that the single crystal silicon micromachined capacitive microphone has the following outstanding features:

Firstly, the single crystal silicon microphone is made from a three layer structure consisting of a single crystal silicon substrate, a thinner epitaxial single crystal silicon layer, and a thicker epitaxial single crystal layer and the prior art microphone is made from a five layer structure consisting of a single crystal silicon substrate, a thin insulating layer, a thin single crystal silicon layer, a thicker oxide layer, and a thicker polysilicon layer. The three layer structure of the single crystal silicon microphone is composed of a same kind of material. In this structure there is no thermal mismatched tension to reside therein. All thermal mismatched tension related problems are able to cancel forever. The five layer structure of the prior art microphone is composed of three different kinds of materials. Due to having different thermal expansion coefficient, thermal mismatched tension always exists between each two different material layers. As is well known, lower tension may result in lowering the sensitivity of the devices and higher tension may result in damage of the devices. Furthermore, a released thin plate with a strong tension often bucks up so that the achievable thickness of the flexible plate and the air gap of the microphone are severely limited.

Secondly, the acoustic cavity of the all single crystal silicon microphone has an opening area smaller than the area of the flexible plate and the acoustic cavity of the prior art microphone has an opening area larger than the area of the flexible plate. A small opening area means less losing mechanical strength and enables to further shrink the microphone size.

Thirdly, the epitaxial single crystal silicon layer for making the stiff and perforated plate has a rest portion with high quality, which can be used to fabricate an electronic circuit, such as a CMOS circuit for conditioning the electronic signals generated by the microphone. For the prior art microphone the top layer is a polysilicon layer that cannot be used to fabricate the CMOS circuit.