Sensors and MEMS

Summer semester 2011

Artikel auf Deutsch

The increasing age of the building stock surrounding us and the ever increasing demand for load capacity makes monitoring of existing buildings necessary. As it is not efficient to permanently have knowledgeable engineers on site and that cable-based monitoring generates significant costs, methods for wireless construction monitoring have been developed over the last few years. For new constructions, which are sometimes very slim, a constant monitoring makes sense as early detection of damage can significantly lower repair costs instead of the damage not being discovered for many years and “spreading”.

Wireless construction monitoring uses sensor nodes (also known as motes) which are attached onto the construction. These motes assess the data of their environment (e.g. Temperatures, accelerations, pressures), process them and then send them in a wireless network to a central computer. There the data can be read and interpreted by someone. It is possible to re-programme the motes via the central computer.

It is of interest to place the motes in such a way that they are as small as possible and require the least possible amount of energy. This is why in a mote one finds not only a sensor but a small computer that is able to process the assessed data and carries out an initial interpretation so that only relevant data is sent to a central computer. This process is important as the sending of data consumes the most energy and saving energy is one of the main goals in mote development.

Definitions

Systems engineering is a relatively young discipline and the term’s definitions are not always easy to understand. Thus, the following explains the most important terms.

Sensor: In general, a sensor is a detector. A sensor can qualitatively and/or quantitatively collect the physical and/or chemical properties of its surroundings. In contrast, Glaser ^[1] is of the opinion that sensors are nowadays no longer only a detector. Through advances in micro-systems engineering, it has become possible to connect several sensors to a measurement unit and additionally include data processing and communication units in a single sensor.

MEMS: Summary of the term micro electro-mechanical systems. These are very small mechanical systems, which transform the energy or signal of a domain (mechanical, electrical, magnetic...) into energy or signals of another domain. ^[2] The production of MEMS is identical to the production of semiconductor components, through etching techniques

Mote: Motes are sensor nodes in which sensors, energy supply, communication units and arithmetic-logic units are combined.

Mote structure

The components of a mote should always be selected depending on the specific application. In particular, their size, price and energy consumption have to suit the specific purpose. However, it is not necessary to produce a completely unique mote for each application. Generally, motes can at least be partially created from commercially available components.

Optimising a mote should not focus on any single component. Also, how a mote should finally be deployed is an important decision which should be focused on during the planning process of wireless construction monitoring. The mote can operate either as event-driven (the mote is switched off until a certain event is triggered), periodically (the mote switches on at certain intervals, determines measurements and switches itself off again) or with function approximation. With the latter, the detected data is not only collected by the mote but also initially interpreted. This process could be utilized to detect isotherms as a fire develops in a building.

With today’s technology, it is possible to build a mote the volume of which does not exceed 1 cm³, weights less than 100 g yet costs less than 1 US$. ^[3] It is already possible to build significantly smaller motes. But it is questionable if these can be sensibly deployed in wireless construction monitoring.

The schematic structure of a mote is shown in figure 2.

CPU: (Central Processing Unit) The CPU forwards the relevant data and can run programmes. With this and the programmes saved in memory, an initial data analysis and reduction already occurs on the mote.

Memory: Data and smaller evaluation programmes are stored here. If there are different sensors the various data are often stored in different memories.

Communication unit: This part of the mote is responsible for enabling communication between the motes. Only with this is a wireless sensor network formed. Whether the communication unit is a radio module or any other wireless data transmission possibility (e.g. Laser) depends on the specific application: The sending of data is the process that consumes the most energy. Thus, data should be sent as quickly as possible.

Energy supply: Ensuring energy supply is one of the greatest problems of wireless sensor networks. This can be managed by batteries or devices that harvest the energy by themselves via Energy Harvesting (e.g. with solar cells).

Sensors/Actuators: These build the actual connection to the external environment. Sensors transform the measured quantities into electric current. Sometimes it is sensible to combine different sensors to obtain measurements. As for example, bending a component generates thermal energy. When this energy is large enough to be measured, using thermal sensors it is possible to learn about the physical stress on the component. Actuators transform electric current into mechanical energy.

Sensors and MEMS

Sensors are components of a mote that recognize the physical and chemical properties of the mote’s environment and can convert processable signals (e.g. electrical signals). The recording of measurands can as a result be determined qualitatively or quantitatively.

Common sensor units are generally relatively large and consume a lot of energy. That is why, in hybrid motes they are only applicable to a certain extent. Today most common sensors have been replaced by MEMS. These micro electro-mechanical systems feature both electrical and also mechanical components and can be integrated onto chips and thus be inserted into motes as sensors.

The partitioning of the sensor units is according to their energy consumption (active and passive sensors) or the type of measurement. In micro system engineering many sensors are also called converters as they convert physical quantities into electric current.

Partitioning according to energy consumption

Active sensors

Active sensors need electricity to deliver measurement results. Those applied on the converter are passive. With a change of environmental conditions, their parameters change as well. To show this as an electrical signal, an auxiliary current has to be applied, so that a primary electronic can convert the changes of the measurements. With these sensors it is possible to collect static and quasi-static facts as the delivery of a measurement is not linked to the change of a physical size.

Compared with passive sensors, these sensors have enhanced properties regarding sensitivity, linearity and the range of the collectable measurements. As an energy supply for the sensors is already called for, partial signal conditioning is often already done in MEMS-based active sensors. ^[4]

An example of active sensors are strain gauges. Those integrated into materials change their electrical resistance with strain deformation. To measure this, the sensor has to be connected with a power source.

Passive sensors

Passive sensors do not require external energy to deliver measurements. The converters inside passive sensors are active which means they deliver energy when the collected measurements change. They cannot record static ratios. piezoelectric converters can, for example, be integrated into passive sensors.

The only example of passive sensors with the ability to record static ratios are temperature sensors. They also deliver with constant temperature electric voltage using the Seebeck-Effect.

Partitioning according to the measurement principle

Conventional sensors are often too large and thus too expensive to carry out a cost-effective wireless construction monitoring. That is why the opportunities provided by micro systems engineering is increasingly used to build very small sensors. With these it is now possible to detect, for example, temperatures, accelerations, pressures, forces, radiation, chemical substances and biological structures. ^[2] The following explains how individual measurements can be realised in MEMS. The focus is on the various principles with which the individual sensors operate.

Thermal sensors

There are two fundamentally different possibilities to measure temperature. With one, the measurement can be made at the moment of connection between sensor and the component to be measured. Yet there is also the option of non-contact temperature measurement. When there is contact between a sensor and medium, a measurement error can occur if the temperature of the sensor influences that of the component. This effect can be avoided by the fact that the mass and the specific heat capacity of the sensor are significantly lower than those of the component. According to Schwesinger ^[2] the conversion effects shown in Table 1 can be used for temperature measurement.

Table 1: Conversion-effects for temperature measurement according to Schwesinger ^[2]

Thermal sensors that function as a result of contact with the component, use the fact that the electrical conductivity of a metal changes with its temperature. With increasing temperature the structure of a metal starts to oscillate more strongly. This obstructs the movement of the free charge carrier, thus the electrical conductivity decreases with increasing temperature. In micro systems engineering this effect can be used by applying stratified arrangements of resistances on a substrate and then determine the resistance via a bridge circuit.

Non-contact arrangements are known as bolometers or measuring devices utilizing the Seebeck-effect. Bolometers use a two level conversion process in which arriving thermal radiation is initially absorbed. With the help of the absorbed radiation, the resistor material is heated and its electrical conductivity changes due to the higher temperature. The Seebeck-effect is based on the fact that within an electrical circuit comprising two conductors with different temperatures, the generated voltage can be measured.

Piezoelectric sensors

The function of these sensors relies on the piezoelectric effect. This states that with physical stress on piezoelectric materials, the charge carriers are moved in such a way that if the electrodes are in the correct position, a voltage can be measured. As a physical impact on the piezo-material can directly generate electrical energy, a piezoelectric sensor can be assigned to the group of passive sensors. In micro systems engineering, zinc oxide or aluminium nitride are typically used for piezoelectric thin layers.

Piezoelectric sensors can be used for the determination of pressure, acceleration, voltage and force. They can also be used to analyse vibration behaviour of constructions and to detect possible damage of a construction through the change of this behaviour over time.

Piezoresistive sensors

These sensors are predominantly used in strain gauges. They incorporate piezo-resistive materials. With strain on such a material, its electrical resistance changes. To measure the strain, an electrical current must be applied to the strain gauge. Depending on whether the current flows in parallel or diagonally to the strain direction, with increasing strain, the electrical resistance decreases or increases.

In semiconductor technology, germanium, polycrystaline silicon or amorph silicon are chiefly used as piezo-resistive materials.

The employment of piezo-resistive sensors in wireless construction monitoring is not easy, as a voltage always has to be applied to achieve a measurement and the available energy for wireless sensors is not unlimited.

Capacitive sensors

In the case of capacitive sensors, the arrangement of components is distorted by the impact of mechanical energy. These are typically mass-spring systems which are displaced at high acceleration. With this displacement, the distance of the shock-mounted and the fixed system components changes. These system parts are capacitors. Due to the movement of the capacitor bank, the capacity changes, which although small is nevertheless measurable.

Advantages and disadvantages

Advantages

• Due to their small size, MEMS can also be deployed in confined spaces and in large scale.
• Due to the use of very common components and the production methods of chip manufacture, MEMS can be produced very cost-effectively.
• Due to various sensor technologies, a suitable sensor network can be designed for almost every application.
• Constructions can evaluate their own condition and when in doubt, sound the alarm.

Disadvantages

• The energy supply of a wireless MEMS is often problematic.
• The transfer of data has to be secure and loss-free.
• MEMS have to handle rough environmental conditions.
• The conversion of physical into processable electrical measurements can lead loss of data.
• A large amount of data is generated and has to be interpreted.

Lifespan of MEMS

MEMS should not only be used for the monitoring of the interiors of buildings. To monitor the behaviour of components, MEMS should also be applied to the outer walls of a building. As a result, they have to withstand adverse environmental conditions such as solar radiation, frost and heavy moisture.

Yet, it is not only the environment that can be troublesome for MEMS. An additional factor that can limit the lifespan of a micro electro-mechanical system is fatigue cracks in silicone components and high abrasion between the micro-mechanical components. Even slight air moisture in the environment of a MEMS can promote abrasion between the micro components.

Applications of of MEMS

Wireless construction monitoring is still in its infancy. However, the application of sensors and MEMS in fire detection, acoustic-emission analysis for the detection of cracks in a component or as strain sensors is conceivable. MEMS could also be further applied in the field of thermal sensor technology for the control of air conditioning systems in buildings.

Literature

1. Glaser, S.; Shoureshi, R.; Pescovitz, D.: Frontiers in sensors and sensing systems. Smart Structures and Systems, Vol. 1, No. 1 (2005). 103 - 120.
2. Schwesinger, N.; Dehne, C,; Adler, F.: Lehrbuch Mikrosystemtechnik. Oldenbourg Verlag München. 2009.
3. Karl, H.; Willig, A.: Protocols and Architectures for wireless sensor networks. John Wiley and Sons, Ltd. 2005.
4. Große, C.; Glaser, S.; Krüger, M.: Initial developement of wireless acoustic emission sensor Motes for civil infrastructure state monitoring. Smart Structures and Systems, Vol. 6, No. 3 (2010) 197 - 209.
5. Allameh, S.M.; Gally, B.; Brown, S.; Soboyejo, W.O.: Surface Topology and Fatigue in Si MEMS Structures. Mechanical Properties of Structural Films, STP 1413, C. Muhlstein, and S. Brown, Eds., American Society for Testing and Materials, West Conshohocken, PA. 2000.
6. Große, C.: Grundlagen der Zerstörungsfreien Prüfung. TUM, Sommersemester 2011.
7. Große, C.; Gehlen, C.; Glaser, S.: Sensing methods in civil engineering for an efficient constructional management.
8. Große, C.: Monitoring of structures using wireless sensors and acoustic emission techniques. CCC 2008 - Challenges for Civil Construction.
9. Krüger, M.; Große, C.: Preiswerte Dauerüberwachung von Bauwerken mit drahtloser Sensorik. Fachtagung Bauwerksdiagnose 2006, Vortrag 16.
10. Lynch, J.: An overview of wireless structural healthmonitoring for civil structures. Phil. Trans. R. Soc. A (2007) 365, 345 - 372.
11. Tanner, D., Walraven, J.; Irvin, L.; Dugger, M.; Smith, N.; Eaton, W.; Miller, W.; Miller, S.: The Effect of Humidity on the Reliability of a Surface Micromachined Microengine. Presented at the 1999 IEEE International Reliability Physics Symposium, March 21 - 25, San Diego CA, pp. 189 - 197.
12. Warneke, B.: Miniaturizing Sensor Networks with MEMS. Handbook of Sensor Networks: Compact Wireless and Wired Sensing Systems; Edited by: Mohammad Ilyas and Imad Mahgoub.
13. Warneke, B.; Pister, K.: MEMS for Distributed Wireless Sensor Networks. Berkeley Sensor and Actuator Center, University of California at Berkeley.