Fish have a lateral line sense organ which allows them to sense water movement and pressure fluctuations. When other fish swim past they generate water vortices which are detected via the lateral line sense organ. The fish thus obtains information about its immediate surroundings, even in complete darkness. We use this ingenious biological system as a model to develop a micro-engineered flow sensor.
What is micro-biomimetics?
Living beings are equipped with a variety of sensors. A distinction is made between active and passive sensor systems. Active sensors emit signals in order to detect obstacles with the aid of the reflected waves. The ultrasonic waves emitted by bats are a well-known example of this. In contrast to active sensors, passive sensors measure energy fluctuations, and their interaction with their environment is minimal. Many passive sensory systems are based on micro-mechanical principles, such as the vibration detection of spiders, the infrared detection of the fire-beetle or the mechano-sensitive lateral line system of fish. The operating principles of these micro-mechanical sensors are often not fully understood. Many questions still need to be answered, especially in relation to complex systems such as the lateral line sense organ of fish. Scientists are still not certain which information the sensors of the lateral line organ provide to the fish and how, if at all, fish are thus able to form a three-dimensional image of their environment.
There are a number of different approaches to investigating and exploiting the operating principle of passive sensors. The objective: To take nature as the model and use engineering to reproduce biological sensors. Until now, this has mainly been done with the aid of macroscopic methods. The problem: The larger the sensor, the more expensive it becomes. A solution: Miniaturisation with the aid of microsystems technology! We call this approach micro-biomimetics. Micro-biomimetics thus involves scientifically analysing and imitating biological sensors in such a way that industrial production becomes possible.
We would now like to introduce the exciting field of micro-biomimetics in more detail using the example of the lateral line system.
We take the fish as our model
Like all animals, fish sense stimuli with different sense organs. Fish can not only see, smell, taste and hear, they can also sense movements in the water via a lateral line system (Figure 1). This sense organ consists of several thousand highly sensitive sensors, the so-called neuromasts. These are located on the surfaces of head and body (superficial neuromasts) and in liquid-filled canals (canal neuromasts).
Neuromasts consist of a sensory epithelium which contains several thousand sensory hair cells. The sensory hair cells of the lateral line have a long sensory hair at the tip, the so-called kinocilium, and a large number of stereovilli, whose lengths increase uniformly the closer they are to the kinocilium. The stereovilli and kinocilium of the sensory hair cells extend into a surrounding gelatinous substance, the cupula. The cupula is in direct contact with the surrounding water or the canal liquid (Figure 1).
Figure 1: Neuromasts in fish. a. Goldfish (Carassius auratus) with pores of the body’s lateral line canal. b. Arrangement of the sensory hair cells in a sensory epithelium of a neuromast and their connection to the nervous system. c.Schematic drawing of the two different neuromasts. The image on the left shows the superficial neuromast, where the cupula is in direct contact with the surrounding water. The image on the right shows the canal neuromast, which serves as the model in this project.
Shifting the cupula displaces the bundles of stereovilli. The displacement causes a change in the membrane potential of the sensory hair cells. This change in the membrane potential is proportional to the displacement of the bundle of stereovilli. The sensors are highly sensitive: the smallest displacement which is detected - i.e. the threshold sensitivity - is a mere 0.01°.
Each movement of water shifts the cupula of a neuromast - relative to the sensory epithelium below it – by a few ångström. If the flow velocity is constant, the water fluctuations along the longitudinal axis of the cupula trigger a maximum response; fluctuations perpendicular to the longitudinal axis of the cupula, in contrast, trigger no response or only a small one. This allows the fish to accurately locate water fluctuations.
We take this ingenious biological system in fish as our model.
We learned from nature, then reproduced it
We are developing a µ-Biomimetic flow sensor with the support of the German Federal Ministry of Education and Research (BMBF) and in collaboration with Professor Bleckmann from the Institute for Zoology at the University of Bonn, and the company Hydrometer, a specialist in Ansbach, Bavaria. The canal neuromasts of fish serve as our model for the flow sensor: our technical sensor also has something akin to small hairs on its surface - the so-called lamellae. The lamellae are in a flow canal (Figure 2). The flow displaces the lamellae. The displacement can be detected in different ways - optically or electrically. If the flow is laminar, the displacement of the lamellae is constant - the sensor shows no reaction. Fluctuations in the flow, on the other hand, cause the lamellae to vibrate; this vibration triggers a measuring signal.
How can these signals be used to determine the flow velocity? The idea is obvious: we generate artificial vortices in the flow. These vortices cause fluctuations which propagate in the flow canal and are detected at two adjacent lamellae. The flow velocity can be calculated from the temporal difference of the measuring signals and the known separation of the lamellae.
The question as to how the displacement of the lamellae is measured – the exciting question from a technological point of view - remains unanswered. Electrical measuring methods are based on two approaches: one capacitive, the other piezoresistive. Both approaches utilise the same fundamental principle: the lamella is firmly attached to a membrane so that the membrane is deformed by the displacement of the lamella. This membrane deformation causes a change in the electrical measuring signal. With the capacitive approach, the membrane is part of a micro-plate capacitor. The deformation of the membrane changes the capacitance of the capacitor. The piezo-resistive approach makes use of so-called Wheatstone bridges; they measure the electrical changes which the deformation induces in the resistors integrated into a membrane. Promising approaches - but will they lead to success? Simulation calculations for electrical measurement methods are not yet available. The situation is different for optical measuring methods: macroscopic preparatory work has already been successfully undertaken by Professor Bleckmann. This is the reason we have initially concentrated on optical measuring methods. They are introduced in more detail below.
Light used for the measurements
The principle of the optical measuring method can be easily understood: light is guided through optically transparent lamellae. The light falls onto a 2-quadrant diode. A quadrant diode is a diode which is subdivided into several photodiodes. The photodiodes convert light into an electric current and thus allow the intensity of the incident light to be measured. We use a 2-quadrant diode; as the name suggests, it consists of two photodiodes. Fluctuations in the flow displace the lamellae. This changes the intensity of the light incident on the two photodiodes of the 2-quadrant diode (see Figure 2).
Figure 2: The technological principle of the optical flow sensor. Artificial vortices are used to bring about fluctuations in the flow. These fluctuations cause the lamellae to vibrate. The vibration changes the intensity of the light incident on the two photodiodes of the 2-quadrant diode. The propagation time of a fluctuation signal between the two lamellae can be measured and used to calculate the flow velocity. The figure shows the microtechnological materials used, the LEDs and the flow direction in the canal.
The production process is complex and time consuming. Photolithography,
oxidation, etching and deposition processes take turns. Figure 3 illustrates the sequence in a simplified form.
Figure 3: Schematic cross-sectional images of the manufacturing process.The process is shown in a simplified form. The depositions are carried out with the aid of sputtering methods, the etchings with dry and wet chemical methods.
The lamellae consist of polydimethylsiloxane (PDMS). They are manufactured by micro-moulding of etched silicon cavities. What exactly does this mean? A cavity is dry etched into a silicon wafer - the “negative mould” of the lamella. This cavity is under vacuum filled with PDMS (Figure 3, process steps 3 and 6). The method we have chosen – micro-moulding of etched microcavities - is slightly unconventional. Initial results confirm that this procedure promises to be successful. The yield is good and the structures are imaged accurately. The flow canals into which the hair structures are integrated are also produced by chemical dry etching.
Figures 4a and 4b show profilometer images of two PDMS lamellae in flow canals with different cross sections. The smaller lamella is additionally shown in two scanning electron microscope images (Figures 4c and 4d).
Figure 4: Profilometer images of two PDMS lamellae. a. The smaller lamella measuring 50 x 50 µm² is located in a canal constriction of 300 µm. b. The larger lamella with dimensions of 50 x 600 ?m² is integrated into a canal which is 1000 µm wide. The height of the lamella is around 500 µm in both cases. c. and d.Scanning electron microscope images of the PDMS lamella (50 x 50 µm²) in the flow canal.
We have meanwhile succeeded in reducing the surface roughness of the lamellae to below 30 nm. This low surface roughness is important so that the light is not scattered by the surface of the fibre optics.
The PDMS lamellae must be protected before the flow canal is etched (see Figure 3, process step 7). This protection can be achieved by thermal oxidation of the silicon wafer, directly after etching the microcavity before filling with PDMS: the sensitive lamella is thus embedded in silicon dioxide. After the etching process, the protective layer of silicon dioxide is carefully removed by means of a wet chemistry technique which we developed ourselves - without the lamella being attacked (process step 8).
The micro-engineered flow sensor must now be interlinked with the macroscopic environment via an “interface”. We have produced such an interface from polymethyl methacrylate (PMMA). Figure 5 shows the canal structures, without the optical flow sensor on the left, with the sensor on the right.
Figure 5: Channel structures in polymethyl methacrylate (PMMA). a. Image without flow sensor b. Image with flow sensor. The image shows a canal structure with three inlets. If the liquid is allowed to exit from the centre inlet, the behaviour of a single lamella can be studied. If the centre inlet is closed, a system with two lamellae can be investigated and the flow velocity measured.
One aim of micro-biomimetics is to develop bionic sensors where not only the scientific findings, but also the practicality and possibility of low-cost production are taken into account. This is also the case in our project: the intent is to use the sensor to measure flow velocities in water pipes. Changes in the flow velocity can indicate leaks in the drinking water system. These leaks result in up to 40 % of the water being lost. In other words: fish could help us to make better use of our drinking water resources!
It is also conceivable that findings can be fed back into basic biological research: the findings obtained with the aid of the flow sensor may possibly be able to contribute to a better understanding of the model found in nature.
A further micro-biomimetics project is the development of an infrared sensor as presented in the caesar Annual Report 2008; this sensor can provide efficient and low-cost monitoring of forest fires. Enquiries about both bionic microsensors are already being received from actual users and scientific developers across the globe.
 Klein, A. and Bleckmann, H. (2011) “Determination of object position, vortex shedding frequency and flow velocity using artificial lateral line canals” Beilstein J. Nanotechnol. 2, 276-283
 Klein, A., Herzog, H., and Bleckmann, H. (2011) “Lateral line canal morphology and signal to noise ratio” Proc. SPIE 7875, 797507