Thermal MEMS Technology Associated with Spirometer Applications
Abstract This electronic document is a live template. The various components of your paper [title, text, heads, etc.] are already defined on the style sheet, as illustrated by the portions given in this document. Keywords: Silicon, MEMS, thermal, flow-sensor, spirometer. I.
INTRODUCTION
A spirometer is a device used to measure the ventilation
capacity of the lungs, i.e. the volume of air inspired and expired by a person. The ability to quantify the inspiration and expiration of people has many beneficial applications in the healthcare industry. The diagnosis of lung diseases such as asthma, emphysema and bronchitis are only possible with an apparatus capable of performing this task. Other applications include: (1) assessing the cause of breath shortage, (2) assessing the effect on certain medications and the implications they have, (3) determining the implications of being exposed to chemicals, and (4) checking lung function before surgery is undertaken. MEMS devices are becoming more frequently used in healthcare applications due to the micro scale in which they can operate. This paper explores current research into the use of MEMS devices in spirometer applications and how they can be used to further develop spirometer devices. II. IMPLEMENTATION OF MEMS DEVICES IN SPIROMETER APPLICATIONS A.Thermal Flow Sensor 1 Silicon Nitride Base An enhanced silicon base is used for fabrication of this implementation [1]. The thermal sensor is based on thermal properties of silicon nitride making it very sensitive to physical properties of gas. The total dimensions used for the sensitive silicon nitride membrane were 300m 600m. Line heaters were also placed adjacent with the membrane in 3 places to act as a thermometer, leaving the sensor to operate as a calorimeter which can detect the flow direction.
Figure 2: Photography of the sensitive element of the HSG-IMIT fabricated MEMS
[1]. Figuresensor 1:Photography of the sensitive element of the HSG-IMIT fabricated
Due to its very thin membrane and small thermal mass,
the sensor exhibits a very short response time of
have a sensitivity of approximately 150V/K, and when
considered along with the small response time it implies a sensor which is very effective in detecting small changes in flow. Figure 2 shows the relationship between sensor output voltage and time during the rapid increase in volumetric flow rate.
Figure 2: Output of sensor to a rapid flow increase induced by an impact on
an
One very important aspect that was considered in this
implementation was how to accurately measure the flow velocity of a human breath when the flow is considered turbulent. The sensors ability to sense flow direction would be obsolete when considering turbulent flow, therefore this issue was tackled by implementing mesh screens to induce a uniform behavior of the flow direction over the sensor effective area. ANSYS Workbench was used in the design process of this MEMS device to determine an appropriate model of the flow channel. This was used in attempt to eliminate any turbulent flow in the effective sensing area. The simulation results yielded a flow velocity which remained less than 16 meters per second and a Reynolds number of 2320. Internal flow dynamics describe laminar flow as having a Reynolds number of less than 2100. 4000 defines the Reynolds number at which flow becomes fully turbulent. Therefore the flow was simulated to have a small degree of turbulence still present, however it was deemed to be negligible. Figure 3 shows the ANSYS solution to the fluid flow modelled through the main channel.
signals such as ambient temperature and mechanical stress.
The choice of temperature detector also allows these common mode signals to be reduced. The area covered by the substrate in this design is 2.3m 2.8m, which is much larger than the aforementioned design. However this membrane area is split into seven different active regions, all isolated from each other and the surrounding s using trench isolation areas as depicted in figure 5.
Figure 3: Exemplified simulation result of the fluid velocity in the main
channel [1].
The implementation has other serious flaws which needed
to be addressed before giving reliable readings. Thermal sensors are based on the fluid properties flowing over it and the presence of humidity and saliva can vary the fluid mixture extensively giving false readings. The sensor design utilizes heating elements to evaporate the saliva on the micro scale over the dew point. Figure 5: The structure of this silicon thermal flow sensor.
This particular implementation uses seven p-n junction
diodes as the temperature detectors. Connecting these in series is intended to increase the overall sensitivity of the temperature detector [2].
Figure 4: Evaporation time of a 1 l water drop in dependence of the
attained heating [1].
The final results of this sensor yielded a higher sensitivity
to flow at lower flow velocities. The two thermopiles symmetrical about the heater which allow the flow direction to be found becomes accurate in the range of flow velocities expected in both inspiration and expiration allowing the transition time to be pinpointed. Due to the sensor being a MEMS device it has a very low power consumption. B. Thermal Flow Sensor 2 Finite Element Modelling Another silicon based thermal sensor was developed using Finite Element Modelling to analyze the flow patterns [2]. In many ways this design is very similar to the aforementioned design however the temperature detectors werent necessarily thermopiles and could be replaced with resistors, diodes or transistors. This allows a different selection common mode
ANSYS Workbench is also used in this design. It is
utilized to create a mesh by discretizing the domain into approximately 30,000 elements [2]. This study compares results taken from a simulated normal airway, resistive airway and an obtruded airway. Figure 6 shows simulation results pulled from ANSYS describing the spirometry volume over time, air flow vs volume and flow velocity over time for all three cases. The results for the normal airway in comparison to the resistive and obtruded cases is somewhat expected, with the spirometer volume reaching capacity much quicker, therefore a higher average flow velocity.
distribution, however the residual heat is carried further
downstream. This residual heat has an effect on the temperature detectors adjacent to the heater therefore yielding varied results from that detector. The study points this out however the authors do not go into depth on how this inaccuracy was accounted for. It does however point out one crucial point which needs considering for future designs of thermal based spirometers. The placement of the temperature detectors is critical to the accuracy of the sensor and without giving this full consideration, false readings would occur. In a healthcare application such as a spirometer it is vital to ensure the results are reliable if it were to be used as a diagnostic tool. III. COMPARISONS BETWEEN IMPLEMENTATIONS The first design explored focuses many of its efforts minimizing the size and ensuring the flow focused over the sensor is laminar. The design achieves this however the results only explore the case of an open airway as opposed to the second design. Testing for different cases has the advantage of creating a design which is applicable to more than one specialized task, for example a design which incorporates different percentages of resistance to flow could be used to diagnose for bronchitis, emphysema and asthma whilst also determining shortage of breath in different scenarios. IV. CONCLUSIONS As devices are increasingly becoming smaller, the need for development in MEMS technology is imperative. The two cases explored in this paper both provide important research which should be considered when designing future sensors which can be used in a spirometer application. Future research could be conducted in decreasing the cost of the device and the total power consumption as these two factors werent thoroughly addressed in either study. If portable spirometers were to be made available for users to take home, these two factors become very important. Figure 6: The air exhalation curves used in Spirometry (a) The volume as a function of time and (b) Air Flow vs Volume (c) Air flow velocity as a function of time for normal, resistive, and obstructive, respectively [2].
ANSYS was also able to determine the temperature
distribution at different flow velocities. Interestingly enough, the maximum simulated flow velocity for a normal airway had a similar temperature distribution in the immediate surroundings of the sensor to the zero flow temperature
REFERENCES [1]
[2]
F. Hedrich, K. Kliche, M. Storz, S. Billat, M. Ashauer, R. Zengerle,
Thermal flow sensors for MEMS spirometric devices, Sensors and Actuators A: Physical., vol. 162, Aug. 2010, pp. 373378 doi:10.1016/j.sna.2010.03.019 I. Hariadi, Sensitivity analysis of silicon MEMS Thermal Flow Sensor for spirometer application, Instrumentation, Communications, Information Technology, and Biomedical Engineering (ICICI-BME), Nov. 2011, doi:10.1109/ICICI-BME.2011.6108619