Forest fires cause significiant financial damage every year. Unfortunately, fire desasters - like the one in February 2009 in Australia - also often claim human lives. Due to global warming, the number of forest fires is expected to increase continuously within the next years. To ensure that forest fires do not get out of control, it is important to locate the fire as early as possible. Fire watchtowers are still in use for this purpose; however, they are not permanently manned. In search of better early-warning systems that can automatically monitor large forest areas around-the-clock, turning to nature is worthwhile.
A "Cool" Heat Sensor
A number of insects represent perfect remote-tracking systems for forest fires. The most well-known insect associated with fires is the species of Buprestid, the jewel beetle Melanophila (Figure 1). Twelve species of this beetle can be found in the Northern Hemisphere.
Figure 1: The pyrophilous jewel beetle Melanophila acuminata can detect forest fires from large distances.
The tracking performance of the beetle surprises even the experts. In 1925, a swarm of Melanophila beetles was observed flying to a burning oil tank farm in Coalinga (California). The tank farm was located in the middle of a desert area. The nearest forest in which the beetles were located was 80 kilometers away. Therefore, the beetles must have detected the fire from this long distance. Phenomena like this attract the interest of bionic research.
Melanophila beetles use a small infrared (IR) sensory receptor, located next to their mesothoracic legs, to detect forest fires. They are composed of 50-100 individual sensilla (each approx. 15 microns in diameter) and located at the bottom of a 100 µm deep cavity. The sensilla are packed together tightly. Consequently, the beetle has a highly-miniaturized IR sensor array. This three-dimensional structure was visualized using scanning electron microscopy (Figure 2).
Figure 3 a) shows a magnified scanning electron micrograph of IR receptors and Figure 3 b) a schematic cross-section of a receptor.
Figure 2: Scanning electron micrograph of the IR sensory organ ofMelanophila acuminata. Approximately 70 IR sensilla are located at the bottom of the 100-µm-deep cavity. The dome-shaped receptors are 12-15 µm in diameter.
Figure 3: Dome-shaped IR receptors: a) scanning electron micrograph; b) schematic cross-section. s: sphere's rigid exocuticular outer shell with its web of nanocanals (nc); exo: exocuticle; il: spongy intermediate mesocuticular layer. The tip of the dendrite (d) enclosed in the internal pressure chamber (ipc).
The IR receptors of the beetle are in fact mechanoreceptors that convert infrared radiation into a mechanical stimulus, resulting in a deformation of the sensory cell membrane. The details of the underlying mechanism are unclear.
We presume that this process unfolds in the following manner: The inside of the sphere contains a spongy porous mesocuticular area (il, Figure 3 b). Its hollow spaces are filled with fluid. The shell of the outer sphere consists of a hard exocuticular shell reinforced with chitin fibers (s, Figure 3 b). Because water is incompressible, expansion of the entire intermediate layer will instanteneously produce an increase in pressure in the internal pressure chamber (ipc, Figure 3 b). This increase of pressure can only be compensated by a pliable structure. The membrane of the dendritic tip (d, Figure 3 b) is the only compliant structure inside the sphere, so the resulting deformation of the cell membrane will open stretch-activated ion channels and thus generate an electrical signal. Hence Schmitz et al. (2007) proposed that an IR sensillum most probably acts as a micro fluidic device converting infrared radiation into an increase of pressure inside the sphere. This is registered by the mechanosensitive neuron. Therfore, the beetle has a fast-reacting thermometer on the infrared scale; it feels heat. But why does the sensor not react to slow ambient temperature fluctations?
Scanning electron micrographs offer a still speculative explanation: nanocanals (nc, Figure 3b) are located in the hard exocuticle shell (s, Figure 3b) of the sensor sphere (diameter < 100 nm), through which a slow increase in pressure is immediately compensated. Hence, the beetle has a microfluidic IR sensor in which the fluid in hollow cavities is used to transfer short pressure pulses to the mechanosensor. This principle is described as “photomechanical”. We use it to develop and to design a new technical IR sensor.
We will build the IR sensor using silicon-bulk micro-machining technology. A scheme of the design is shown in Figure 4.
Figure 4: Cross-section of the technological design of the IR sensor (not to scale). Total sensor thickness including cover approximately 0.9 mm; chip width: 5 mm; thickness of the capacitor membrane: 1 µm
The pressure chamber has a window for IR radiation. The fluid in the pressure chamber expands after an incident IR radiation pulse is absorbed. A parallel plate capacitor assumes the role of the biological mechanical receptor: One of the capacitor’s electrodes is coated on a thin flexible membrane, which covers the pressure chamber and is deflected by the expanding fluid. As a result, the capacitance changes. This pulse-induced change in capacitance is registered.
The following problems had to be taken into account:
1. If the membrane is deflected, it changes the pressure in the air-filled volume between the electrodes of the capacitor. In a closed chamber the amplitude of the short deflection pulse would be damped as a result and the measured signal would therefore become weaker. For this reason, we perforated the upper electrode.
2. A change in the ambient temperature, for example between day and night, causes a slow change of pressure in the pressure chamber. Analogous to the biological model, we provided the IR sensor with a pressure balance chamber that is connected to the pressure chamber via a micro capillary. The pressure balance chamber contains a thin soft cover membrane made of silicon (PDMS) which compensates volume changes in the fluid without backpressure.
Figure 5 elucidates how we transferred the essential characteristics of the biological model to the technological sensor.
Figure 5: Technological implementation of the biological model.
Figure 6 shows the arrangement of the sensors ("chips") on a 4'' wafer and a magnified section of a sensor layout.
Figure 6: Layout of an IR sensor. The diagram shows the layout of the capacitor membranes, chambers and pressure balance capillaries.
As shown in figure 6, the length of the pressure balance capillary and the diameter of the capacitor membrane were diversified to obtain the optimum parameters for the sensor. In our layout, there are 408 sensors on a 4'' wafer. The sensor chip is currently 2 x 5 mm2 in size; it is, however, relatively simple to produce considerably smaller ones in the silicon technology.
Tricky Technological Problems
A difficulty in the production of the sensor is the bubble-free filling of the micro chambers and canals with fluid. This is a general problem in micro technology that has not yet been solved. We developed a new method, which leads to promising results and has been submitted for a patent. However, it is still not known whether this micro technology sensor is suitable as fire detector and whether it is as sensitive as its biological role model.
 Schmitz, A., Sehrbrock, A., and Schmitz, H. (2007) "The analysis of the mechanosensory origin of the infrared sensilla in Melanophila acuminata (Coleoptera; Buprestidae) adduces new insight into the transduction mechanism" Anthrop Structure & Development 36, 291-303
 Müller, M., Olek, M., Giersig, M., and Schmitz, H. (2008) "Micromechanical properties of consecutive layers in specialized insect cuticle: the gula of Pachnode marginata (Coleoptera; Scarabaeidae) and the infrared sensilla of Melanophila acuminata (Coleoptera; Buprestidae).“ J. Exp. Biol. 211, 2576-2583
 Bousack, H., Schmitz, H., and Offenhäuser, A. (2008) "Design of a Fluidic Infrared Detector Based on the Infrared Sensilla in the Beetle Melanophila acuminata“ In: Actuator 2008, 11th International Conference in New Actuators, Bremen, 768-789