Microfluidic component having means for determining an external influence factor based on an actuation signal - Patents.com

The microfluidic component monitors its operational status by analyzing control signal time courses, addressing the need for additional sensors in existing systems and enhancing real-time disturbance detection and component reliability.

JP2024544240AActive Publication Date: 2024-11-28FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV

Patent Information

Application Number
JP2024534374
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2021-12-09
Filing Date
2022-12-07
Publication Date
2024-11-28
Estimated Expiration
2042-12-07

AI Technical Summary

Technical Problem

Existing microfluidic components, such as micropumps and microvalves, lack the ability to detect disturbances or monitor their operational status without requiring additional sensor technology or electrodes, which complicates manufacturing and limits real-time monitoring capabilities.

Method used

A microfluidic component with a membrane actuator that generates and processes control signals to determine external influencing factors by analyzing the time course of these signals, allowing self-monitoring during normal operation without additional sensors or electrodes.

Benefits of technology

Enables real-time monitoring of microfluidic components by identifying and classifying external influences such as pressure changes, fluid properties, and mechanical stresses without modifying the component structure, ensuring reliable operation and detection of issues like valve deterioration or fluid changes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024544240000001_ABST
    Figure 2024544240000001_ABST
Patent Text Reader

Abstract

The inventive concept described herein relates to a microfluidic component (1000) comprising a membrane actuator (100) having a membrane element (101) and an actuator element (102) for deflecting the membrane element (101). The signal generating device (103) is configured to generate an electrical control signal (104) having a time-varying signal curve for controlling the membrane actuator (101), whereby the actuator element (102) actuates the membrane element (101). The signal processing device (105) is configured to detect, during operation of the microfluidic component 1000, an influence on the time signal curve of the control signal (104) caused by one or more external influencers, and to identify and / or classify at least one causal external influencer based on the influence on the time signal curve.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The invention described herein relates to devices in the field of microfluidics. For example, a microfluidic component of the invention can be a microvalve or a micropump that can be actuated by a microstructured membrane actuator. The present disclosure proposes an innovative concept that allows a microfluidic component to monitor itself during operation by evaluating the time course of a control signal. [Background technology]

[0002] Fluidic microactuators, such as micropumps or microvalves driven by membrane actuators, can transport liquids or gases or switch fluid paths. In practice, there may be some disturbances or changes in fluid boundary conditions that can impair the intended function of these actuators. These actuator components do not have sensor properties and therefore cannot detect disturbance events.

[0003] For example, microactuators are often used in micro-dosage systems. One possible way to monitor the micro-dosage or detect disturbance events is to use a flow sensor, for example an anemometer, which allows to measure the pressure difference or the change in volume. In addition, a flow sensor or a pressure sensor can be used to detect disturbances. However, this has the disadvantage that separate components must be appropriately integrated into the micro-dosage system.

[0004] A further approach to detect disturbances is to use a sensor system in the actuator membrane, i.e. special sensors are integrated directly into the membrane of the microactuator, making it possible to determine, for example, the voltage state of the membrane. Alternatively, additional electrodes can be attached to the piezoelectric ceramic. However, in this case, additional contacts are required, which increases the overall manufacturing complexity and has the disadvantage that no direct information on the pneumatic or hydraulic state can be determined. In addition, the sensor technology must be housed in the critical membrane actuator area, which involves high voltages and large mechanical stresses.

[0005] A further possible way to detect disturbance events or fault conditions is to use two micropumps connected in series, each containing a piezoelectric element. The two piezoelectric driven micropumps are connected such that one of the two pumps is operating and the other is inactive. The effect of the fluid signal of the operating micropump is then detected by the sensor characteristic of the piezoelectric element of the second micropump, which is turned off.

[0006] In this case, in addition to the microactuator, an additional sensor element is again required, here in the form of a second micropump. Moreover, the fluid signal of the micropump is influenced by the fluid line between the two micropumps and no longer represents the processes in the pump chamber of the active micropump.

[0007] A further possible way to detect disturbances is to use another electrode on the piezoelectric actuation membrane. When a piezoelectric ceramic is used as a pump drive element, it has to be charged with a high voltage within a very short time. This charging current I=C*dU / dt is significantly higher than the expected sensor effect. There are therefore attempts to realize a separate "sensor electrode" on the piezoelectric ceramic. The use of a separate electrode as a sensor electrode offers the possibility of utilizing the sensor effect of the piezoelectric actuation membrane.

[0008] However, this again requires a second electrode and the corresponding wiring. In addition, a part of the piezoelectric ceramic is occupied by the second electrode, which cannot be used for the actuator system. In addition, the interaction is only detected at the location of the piezoelectric ceramic where the separate actuator system is located. [Prior art documents] [Non-patent literature]

[0009] [Non-Patent Document 1] https: / / www.sciencedirect.com / science / article / pii / S0924424710002311 [Non-Patent Document 2] https: / / iopscience.iop.org / article / 10.1088 / 0960-1317 / 4 / 4 / 004 [Non-Patent Document 3] "Simulation of microfluid systems", Zengerle, R., Richter, M., Journal of Micromechanics and Microengineering, 1994, 4(4), pp. 192-204, 004 Summary of the Invention [Problem to be solved by the invention]

[0010] It is therefore desirable to improve existing systems for detecting disturbance events or for monitoring the state of a microfluidic component such that no additional sensor technology or additional electrodes are required. In addition, it is desirable that the microfluidic component does not have to be controlled in a special way, but the state can be monitored during normal operation, i.e. during normal control. [Means for solving the problem]

[0011] This object is achieved by a microfluidic component and a method for operating such a microfluidic component according to the independent claims.

[0012] The microfluidic component of the present invention comprises a membrane actuator having a membrane element and an actuator element for deflecting the membrane element. The microfluidic component of the present invention further comprises a signal generating device configured to generate an electrical control signal having a time-varying signal curve for controlling the membrane actuator for use by the actuator element in actuating the membrane element. The microfluidic component of the present invention additionally comprises a signal processing device configured to determine an influence on the control signal's temporal signal curve caused by one or more external influencers during operation of the microfluidic component. The signal processing device is additionally configured to identify and / or classify at least one causal external influencer based on its influence on the temporal signal curve.

[0013] Thus, a normal control signal of the membrane actuator or its time course in at least one deflection process of the membrane actuator is observed. In this case, different external influence factors have different effects on the time course of the control signal. That is, the control signal varies depending on how the membrane actuator interacts with its environment. These influences of the surroundings cause the membrane actuator (e.g., its voltage state or position) to change in such a way that certain electrical parameters change, which then directly affect the current flow of the control signal. Since these interactions can vary in time, the interaction process can be inferred through a precise determination, or measurement, of the control signal.

[0014] An external influence, for example in the form of a pressure change under a piezoelectric actuated membrane transducer, generates a force action resulting in a current flow on the piezoelectric ceramic caused purely by this pressure change due to the direct piezoelectric effect. In the context of the present disclosure, this current flow is sometimes also referred to as the "sensor current". This sensor current overlaps with the current required to electrically charge the capacitive load. That is, the additional current flow caused due to the external influence overlaps with the control signal (e.g. the charging current) and accordingly causes a deviation in the time course of the control signal.

[0015] Accurate measurement of the control signals (e.g. charging current) of the membrane actuator, which is necessary for normal operation of the membrane actuator, is sufficient to be able to measure fluidic and mechanical time-varying processes at or in the inventive microfluidic component (e.g. pump chamber and / or microvalve of a micropump). This makes the inventive concept described herein highly advantageous, since no changes need to be made with respect to the components that make up the microfluidic component. It is sufficient to integrate the corresponding measurement functions in the driver electronics of the microfluidic component, preferably in combination with suitable means for capturing and evaluating the data.

[0016] Thus, using the inventive concepts described herein, information indicative of how a membrane actuator interacts with its environment hydraulically, pneumatically, mechanically, or piezoelectrically can be extracted from normal control signals in a purely electrical manner, without the use of additional equipment, and in real time. Changes in the associated fluids, such as valve degradation, back pressure fluctuations, actuator fatigue, bubble generation, and the like, can be detected from this information.

[0017] The state of the microfluidic component can be monitored based on the inventive signal evaluation of the time course of the control signal. That is, during operation, in particular during normal operation (i.e. no special calibration modes are required), the inventive microfluidic component can monitor itself. The time course brings information about how the microfluidic component interacts with its environment, i.e. which external influence factors are currently influencing the microfluidic component. Depending on the type of external influence factor, the time course of the control signal is influenced differently. Each individual influence factor leaves its specific trace, so to speak, in the provision of the control signal over time. Through this, for example based on the correspondingly influenced time signal curve, the type of currently prevailing external influence factor can be determined. For example, it can be determined whether there is a defect and, if so, the type of defect.

[0018] The invention further relates to a corresponding method for operating a microfluidic component with a membrane actuator comprising a membrane element and an actuator element for deflecting the membrane element, the method comprising the steps of generating an electrical control signal for controlling the membrane actuator, whereby the actuator element actuates the membrane element, the electrical control signal comprising a time-varying signal curve. In addition, the inventive method comprises the steps of determining a temporal signal curve of the control signal influenced by one or more external influence factors during operation of the microfluidic component, and identifying and / or classifying the at least one external influence factor based on the determined influenced temporal signal curve.

[0019] Further embodiments and advantageous aspects of the inventive microfluidic component and the corresponding method for operating a microfluidic component are set out in the respective dependent claims.

[0020] Some embodiments are exemplarily illustrated in the drawings and are subsequently described in more detail. [Brief description of the drawings]

[0021] [Figure 1] FIG. 1 is a schematic diagram illustrating a microfluidic component with signal evaluation of the present invention according to one embodiment. [Diagram 2] 1A-1D are schematic diagrams illustrating different embodiments of a membrane actuator. [Diagram 3] FIG. 2 is a further schematic diagram illustrating a microfluidic component with signal evaluation of the present invention according to one embodiment. [Figure 4] 1 is a schematic diagram illustrating a signal evaluation device of the present invention according to one embodiment. [Diagram 5] FIG. 1 illustrates an overview of the physical processes within a piezoelectric element. [Figure 6] 1 is a graph illustrating a mathematical function for describing the voltage dependence of the capacitance of a micropump. [Figure 7A] 1 is a graph illustrating a mathematical function for describing the time signal curves of current and voltage curves in the case of sinusoidal control in a micropump. [Figure 7B] 1 is a graph illustrating a mathematical function for describing the time signal curves of current and voltage curves in the case of sinusoidal control in a capacitor. [Figure 8] FIG. 1 shows two Lissajous curves for a visual representation of the hysteresis effect in a micropump compared to a capacitor. [Figure 9A] FIG. 13 shows the Lissajous curves of a micropump delivering liquid with harmonic control and exposed to air bubbles as external influence factors. [Figure 9B] FIG. 1 shows the Lissajous curves of a micropump delivering liquid with harmonic control and subject to blockage as an external influencing factor. [Figure 10] FIG. 1 shows a so-called confusion matrix for determining the system state, or external influencing factors, based on the frequency of correct predictions. [Figure 11] FIG. 13 shows the Lissajous curves of the micropump at different back pressures. [Figure 12]FIG. 1 shows three different time signal curves of control signals with different voltages and curves fitted to these signal curves. [Figure 13] FIG. 13 shows a signal curve of the amplitude of the current term Ip of the micropump during the suction stroke depending on the applied voltage. [Figure 14] FIG. 13 shows the time damping constant of the micropump associated with the current term Ip in the suction stroke depending on the applied voltage. [Figure 15] FIG. 13 shows a comparison of the temporal damping constants of the micropump during the suction stroke and the pressure stroke, respectively, depending on the applied voltage. [Figure 16A] FIG. 13 shows signal curves of the amplitude of the current term ICE of the micropump during the suction stroke depending on the applied voltage. [Figure 16B] FIG. 13 shows the temporal decay constant, associated with the current term ICE, of the micropump during the suction stroke depending on the applied voltage. [Figure 17A] FIG. 13 shows signal curves of the amplitude of the sum of the two current terms IC+ICE of the micropump during a pressure stroke depending on the applied voltage. [Figure 17B] FIG. 13 shows the temporal decay constant associated with the sum of current terms IC+ICE of the micropump during a pressure stroke depending on the applied voltage. [Figure 18A] FIG. 13 shows signal curves of the amplitude of the sum of the two current terms IC+ICE of the micropump in a pressure stroke depending on the applied voltage for small voltages. [Figure 18B] FIG. 13 shows the temporal decay constant associated with the sum of current terms IC+ICE of the micropump in the suction stroke depending on the applied voltage for small voltages. [Figure 19] FIG. 2 is a block circuit diagram of a measurement structure used to perform the measurements shown in the subsequent figures. [Figure 20] 1 shows a time signal curve of the amplitude of the current term Ip of an air conveying micropump in the suction stroke and in the pressure stroke in normal conditions, i.e. without the influence of external influence factors. [Figure 21] FIG. 21 is an enlarged view of the temporal signal curve in the suction stroke according to FIG. 20. [Figure 22] FIG. 21 is an enlarged view of the time signal process in the pressure stroke according to FIG. 20. [Diagram 23] FIG. 2 shows a time signal curve of the amplitude of the current term Ip of a liquid-transporting micropump in the suction stroke and in the pressure stroke under normal conditions, i.e. without the influence of external influence factors. [Figure 24] FIG. 24 is an enlarged view of the temporal signal curves in the suction stroke according to FIG. 23. [Diagram 25] [Fig. 25A] A close-up of the overshoot of the time signal curve in the suction stroke according to Fig. 24. [Fig. 25B] A close-up of the overshoot of the time signal curve in the suction stroke according to Fig. 24. [Fig. 25C] A close-up of the overshoot of the time signal curve in the suction stroke according to Fig. 24. [Figure 26] FIG. 24 shows an enlarged view of the time signalling during the suction stroke according to FIG. 23. [Figure 27A] FIG. 14 shows a temporal signal curve of the amplitude of the current term Ip of a liquid-conveying micropump in the suction stroke under the influence of an external influencer in the form of an air bubble. [Figure 27B] FIG. 27B is a detailed view of the temporal signal curve during the suction stroke according to FIG. 27A. [Figure 27C] FIG. 13 shows a temporal signal curve of the amplitude of the current term Ip of a liquid-transporting micropump in the suction stroke, several pump strokes prior to the generation of an air bubble in the pump chamber. [Figure 27D] FIG. 27B shows the temporal signal curve according to FIG. 27A when an air bubble is present in the pump chamber. [Figure 27E] FIG. 13 shows the temporal signal curve of the amplitude of the current term Ip of a liquid-transporting micropump in the suction stroke, several pump strokes after the air bubble has left the pump chamber. [Figure 28A]FIG. 14 shows the temporal signal curves of the amplitude of the current term Ip of a liquid-conveying micropump in the suction stroke and in the pressure stroke under the influence of an external influence factor in the form of an air bubble. [Figure 28B] FIG. 28B is a detailed view of the temporal signal curve in a pressure stroke according to FIG. 28A. [Figure 29A] FIG. 13 shows a time signal curve of the amplitude of the current term Ip of a liquid-conveying micropump in the suction stroke and in the pressure stroke under the influence of an external influence factor in the form of a continuously decreasing counterpressure. [Figure 29B] FIG. 13 is a diagram showing the time progression of the back pressure in the inlet reservoir and the outlet reservoir. [Figure 29C] FIG. 29B is a detailed view of the temporal signal curve in the suction stroke according to FIG. 29A. [Figure 29D] FIG. 29B is a detailed view of the temporal signal curve in a pressure stroke according to FIG. 29A. [Figure 30A] 1 shows a diagram of the temporal signal curves of the amplitude of the current term Ip of an air-conveying micropump in the suction stroke and in the pressure stroke under the influence of an external influence factor in the form of a continuously decreasing pressure. [Figure 30B] FIG. 30B is a detailed view of the temporal signal curve in the suction stroke according to FIG. 30A. [Figure 30C] FIG. 30C is an enlarged view of the temporal signal curve in the suction stroke according to FIG. 30B. [Figure 30D] FIG. 30B is a detailed view of the temporal signal curve in the suction stroke according to FIG. 30A. [Figure 30E] FIG. 30E is an enlarged view of the temporal signal curve in the suction stroke according to FIG. 30D. [Figure 31A] FIG. 13 shows the Lissajous curves of different micropumps in suction and pressure strokes without external influence factors. [Figure 31B] 13A-13C show Lissajous curves of different micropumps in the suction and pressure strokes under the influence of an external influencing factor in the form of contact of membrane elements. [Diagram 32]FIG. 2 is a schematic block circuit diagram illustrating a system structure according to one embodiment. [Diagram 33] FIG. 2 is a schematic block circuit diagram illustrating possible components of a signal evaluation device according to one embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0022] In the following, the embodiments are described in more detail with reference to the drawings, in which elements having the same or similar functions are presented with the same reference numbers.

[0023] Method steps illustrated or described in the context of the present disclosure may also be performed in orders other than those illustrated or described herein. In addition, method steps relating to particular features of a device may be interchangeable with said features of the device, and vice versa.

[0024] In the context of this disclosure, when reference is made to a fluid, this should be understood as a liquid, a gas, or a mixture of a liquid and a gas.

[0025] Among other things, this disclosure describes a membrane actuator having a membrane element and an actuator element. In the context of this disclosure, when reference is made to pump chamber pressure, this should be understood as the pressure applied to the membrane side of the membrane element that is in contact with the fluid. For example, such pressure may also occur in a valve that includes such a membrane actuator.

[0026] In the context of this disclosure, when reference is made to a variable pressure or variable pump chamber pressure, this should be understood as a time-varying and configuration-varying pressure. Unless otherwise indicated, in the context of this disclosure, the term "variable pressure" should be understood as a time-resolved average value of the time-varying and configuration-varying pressure.

[0027] In the context of the present disclosure, when reference is made to an external influencing factor, this may be understood to be at least one of the following factors: Environmental parameters (e.g. temperature, humidity), Rheological properties of fluids, Disturbance variables (e.g. particles, bubbles in a liquid), The shape or dimensions of the membrane actuator; Material parameters (e.g. Young's modulus, piezoelectric coefficient d31), Operating parameters (e.g. control signals).

[0028] In the context of the present disclosure, where embodiments are described using the example of a piezoelectrically actuated membrane actuator, the corresponding description also applies to an electrostatically actuated membrane actuator, and vice versa.

[0029] In the context of the present disclosure, where an embodiment is described using the example of a micropump, the corresponding description also applies to a microvalve and vice versa.

[0030] In the event of departure from the nomenclature used herein, the following applies. d=CE* I d =I CE =I ce τ h = τ p I A =I a τ A = τ a

[0031] Firstly, and purely diagrammatically, Figure 1 shows a cross-section of a microfluidic component 1000 of the invention having a membrane actuator 100. The membrane actuator 100 comprises a membrane element 101 and an actuator element 102. The actuator element 102 is used to deflect the membrane element 101.

[0032] In addition, the microfluidic component 1000 of the present invention comprises a signal generating device 103 configured to generate an electrical control signal 104 having a time-varying signal curve for controlling the membrane actuator 100, whereby the actuator element 102 actuates the membrane element 101. In this non-limiting example, the signal generating device 103 generates an alternating voltage signal U(t). Similarly, corresponding quantities derived therefrom, such as a corresponding alternating current signal I(t), may be considered.

[0033] Furthermore, the inventive microfluidic component 1000 comprises a signal processing device 105, which is configured to determine an influence on the time-signal curve of the control signal 104 caused by one or more external influence factors during operation of the microfluidic component 1000, and to identify and / or classify at least one causal external influence factor based on this influence on the time-signal curve.

[0034] This external influence factor therefore brings about an influence on the time signal curve of the control signal 104. In this case, different external influence factors cause different deviations in the time signal curve of the control signal 104. Some influence factors cause characteristic deviations in the time signal curve, which correspond to traces, i.e. different external influence factors leave different traces on the time signal curve of the control signal 104.

[0035] Some examples for external influence factors and their detection or identification are given below. In addition, the following is a specific description of how deviations in the time signal curve of the control signal 104 are determined. It should be noted in advance at this point that the signal analysis of the control signal 104 can be performed according to the present invention by considering different signal parts. In the context of the present disclosure, these signal parts are also referred to as "current terms". These individual current terms are based on different external influence factors that can be identified and / or classified using the concept of the present invention.

[0036] However, before going into details, an overview of the physical processes in different membrane actuators 100, e.g., piezoelectrically or electrostatically actuated membrane actuators 100, is given to provide a better understanding of the concept of signal analysis of the control signal 104 described herein. For this purpose, reference is again made to FIG.

[0037] 1 shows that the actuator element 102 functionally interacts with the membrane element 101, i.e. the membrane element 101 can be deflected by the actuator element 102. This causes a volume displacement V below the membrane element 101. The actuator element 102 and the membrane element 101 are electrically controlled using a signal generating device 103. For example, a drive voltage U can be applied to the actuator element 102, while the membrane element 101 is at a ground potential U 0 In this case, a charge Q flows onto the actuator element 103.

[0038] The membrane element 101 in combination with the actuation element 102 essentially has the property of a variable capacitance. In the case of a liquid, a liquid pressure acts on the membrane element 101, or in the case of a gas, a pneumatic pressure acts on the membrane element 101, which takes the form of a pressure p. Furthermore, mechanical forces may act on the membrane element 101, for example caused by the mechanical component 106. In addition to that, surface forces may act on the membrane element 101, for example in the form of a meniscus. A reference pressure p, which is often atmospheric pressure, 0 , or a pressure set above the membrane element 101 acts above the actuator element 102 .

[0039] The fluid, e.g., gas or liquid, below the membrane element 101 is pressurized with a pressure p. The pressure p is time dependent and determined from the interaction between the membrane actuator 100 and the fluid. Upon change of state, which is slow compared to the signal propagation in the membrane element 101 and in the fluid (determined by the speed of sound), the pressure or pressure distribution adapts such that the membrane element 101 is in a state of force equilibrium (action = reaction).

[0040] However, this pressure p applied below the membrane element 101 is equal to the reference pressure p above the membrane element 101. 0 If different from p, then a pneumatic pressure (for a gas as the fluid) or a hydraulic pressure (for a liquid as the fluid) is applied to the membrane element 101 through the membrane surface. Here, the fluid pressure p below the membrane element 101 is also a reference pressure p above the membrane element 101. 0 It does not have to be constant and uniform.

[0041] Thus, the membrane actuator 101 may have a pneumatically or hydraulically induced correlation, i.e., there may be pneumatic or hydraulic external influencers acting on the membrane actuator 100. As explained above, when the mechanical component 106 applies a mechanical force to the membrane actuator 100, the mechanical external influencer acts on the membrane actuator 100 in response.

[0042] According to one embodiment of the present invention, the signal processing device 105 may be appropriately configured to identify and / or classify hydraulic, pneumatic or mechanical forces acting on at least one membrane side of the membrane element 101 as external influencing factors resulting in an effect on the temporal signal curve.

[0043] In some embodiments, the actuator element 102 may include a piezoelectric ceramic, as described in more detail below, in which case piezoelectric external influences may act on the membrane actuator 100 as well.

[0044] In this case, according to the present invention, the signal processing device 105 may be configured to identify and / or classify a piezoelectric force acting on at least one membrane side of the membrane element 101 as an external influencing factor resulting in an effect on the temporal signal curve.

[0045] In all cases, according to the invention, a time-dependent measurement of possibly overlapping pneumatic, hydraulic, piezoelectric or mechanical interrelationships can be performed using a time-precise electrical measurement of the control signal 104 (e.g. a voltage signal U(t) or a quantity associated therewith, such as a current signal I(t)). That is, with the invention described herein, it is possible to detect different external influence factors acting on the membrane actuator 100, such as pneumatic, hydraulic, piezoelectric and / or mechanical external influence factors, using a signal analysis of the control signal 104. In addition, the detected influence factors can be identified and / or classified.

[0046] The detectable or identifiable external influence factors may differ slightly depending on the embodiment of the membrane actuator 100. Figure 2 shows a possible embodiment of the membrane actuator 100 of the present invention.

[0047] The left side of Fig. 2 shows a piezoelectric membrane actuator 100. In this case, the actuator element 102 may comprise a piezo-ceramic or a piezoelectric element fixed to the membrane element 101, for example using a suitable adhesive. As soon as a control signal 104 is applied to the piezo-ceramic, the piezo-ceramic is deformed, causing the membrane element 101 to actuate.

[0048] For example, the piezoelectric membrane actuator 100 of the present invention may be used in a piezoelectrically driven micropump, i.e. in this case the microfluidic component 1000 of the present invention comprises a piezoelectrically driven micropump. In this case the piezoelectric membrane actuator 100 is used as a pump element, i.e. the raising and lowering of the membrane actuator 100 may perform pump strokes, i.e. suction strokes and pressure strokes. In this case the piezoelectrically driven micropump may be configured as a three-chamber membrane pump with active valves or may comprise passive check valves.

[0049] Alternatively, for example, the piezoelectric membrane actuator 100 of the present invention may be used in a piezoelectrically actuated microvalve, i.e., in this case, the microfluidic component 100 of the present invention comprises a piezoelectrically actuated microvalve. In this case, the piezoelectric membrane actuator 100 is used as a means for actuation, i.e., the microvalve may be opened or closed by raising and lowering the membrane actuator 100. The piezoelectrically actuated microvalve may be configured as a NO valve (NO=normally open) or a NC valve (NC=normally closed).

[0050] The right side of Fig. 2 illustrates an electrostatically actuated membrane actuator 100. In this case, the membrane element 101 forms the moving electrode and the actuator element 102 forms the counter electrode. In this case, the actuator element 102 and the membrane element 101 interact capacitively, as will be explained in more detail below. As soon as a control signal 104 is applied to the actuator element 101 (counter electrode), it attracts or repels the membrane element 101 (electrode), generating an up and down movement of the membrane element 101.

[0051] For example, the electrostatic membrane actuator 100 of the present invention may be used in an electrostatically driven micropump, i.e., in this case, the microfluidic component 1000 of the present invention comprises an electrostatically driven micropump. In this case, the electrostatic membrane actuator 100 is used as a pumping element, i.e., by raising and lowering the membrane actuator 100, pumping strokes, which are suction strokes and pressure strokes, respectively, may be performed. Here, the electrostatically driven micropump may also be configured as a three-chamber membrane pump with active valves or may comprise passive check valves.

[0052] Alternatively, for example, the electrostatic membrane actuator 100 of the present invention may be used in an electrostatically actuated microvalve, i.e., in this case, the microfluidic component 1000 of the present invention comprises an electrostatically actuated microvalve. In this case, the electrostatic membrane actuator 100 is used as a means for actuation, i.e., the microvalve may be opened or closed by raising or lowering the membrane actuator 100, respectively. The electrostatically actuated microvalve may be configured as a NO valve (NO=normally open) or a NC valve (NC=normally closed).

[0053] As mentioned at the outset, the inventive concept provides for measurement and / or signal evaluation of the control signal 104, making it possible to determine one or more external influence factors acting on the membrane actuator 100, or on the microfluidic component 1000. In the following, this is explained in more detail exemplarily using the above described possible embodiments of the inventive piezoelectric or electrostatically driven membrane actuator 100.

[0054] 1. Piezoelectric Membrane Actuator 1.1 Piezoelectric micropump A piezoelectrically driven micropump 1000 is typically controlled by a periodic control signal 104, such as a sinusoidal AC voltage or current signal, in which the membrane actuator 100 alternates between upward movement (suction stroke) and downward movement (pressure stroke).

[0055] The control signal 104 is not independent of the way in which the membrane actuator 100 interacts with its environment. Each external influencer (e.g. hydraulic, pneumatic, piezoelectric, mechanical) acting on the membrane actuator 100 changes the mechanical properties of the membrane actuator 100 (e.g. its voltage state or its position), which in turn changes the electrical parameters of the membrane actuator 100 as well. This in turn has a direct effect on the current flow. Since the external influencers are time-varying, the interaction process, or respectively the dominant external influencer, can be inferred using the inventive signal analysis of the control signal 104.

[0056] Thus, for example, a pressure change beneath the piezoelectric membrane actuator 100 induces a force action, which in turn induces a current flow in the piezoelectric membrane actuator 100 due to the direct piezoelectric effect caused solely by the pressure change. In this disclosure, this current flow is also referred to as the "sensor current." This sensor current overlaps with the current required to electrically charge a capacitive load.

[0057] In the case of the piezoelectrically driven micropump 1000, this sensor current can be accurately determined using the signal analysis of the present invention, and other current portions (in particular the charge current I of the capacitance of the piezoelectric element) can be calculated. c) through which, for example, the change in pump chamber pressure can be indirectly measured with respect to time. The change in pump chamber pressure over time is determined, among other things, by the fluid processes in the fluid chamber. However, if correlations as to why the pump chamber pressure changes over time are known, the processes in the pump chamber can be measured by the inventive signal analysis of the sensor current. However, even if the correlations are not completely known in detail, the fluid or mechanical processes in the pump chamber can be inferred, for example, by using so-called "machine learning".

[0058] At this point, it is again clearly emphasized that the inventive concept does not require any additional sensor technology or additional electrodes. In addition, the membrane actuator 100 does not have to be controlled in a special way, for example in a special calibration mode. It is sufficient to control it in a normal way with the control signal 104, i.e. the inventive concept can be implemented during normal operation of the microfluidic component 1000. Time-dependent fluidic and mechanical processes in the microfluidic component 1000 can be measured using only accurate measurements or signal analyses of the control signal 104 (e.g. the charging current I(t)) on the membrane actuator 100, which are necessary anyway to operate the membrane actuator 100. This makes the inventive concept described herein very advantageous, since the components of the microfluidic component 1000 themselves do not have to be modified. It is sufficient to integrate the inventive signal analyses in the driver electronics of the microfluidic component 1000, for example in conjunction with suitable means for data acquisition and data evaluation.

[0059] As mentioned at the beginning, piezoelectric driven micropumps are usually controlled with periodic signals, so the inventive concept additionally allows for continuous time-resolved measurement of suction and pressure strokes during the operation of the micropump. If changes in the sensor current are measured in comparison with previous signal curves, changes in operation can be immediately inferred. In other words, the micropump is continuously and permanently self-monitoring during operation, which is a significant improvement over the prior art.

[0060] 1.2 Piezoelectric microvalve Similarly, for the piezoelectrically actuated microvalve 1000, there is a correlation between the air and liquid pressures and the intrinsic voltage of the membrane actuator 100. These time-dependent pressure changes can be measured using the sensor current (in the same way as for the piezoelectrically actuated micropump), allowing fluid information to be extracted from the valve chamber, or the movement of the valve. In addition, the actuating membrane element 101 (possibly reinforced in the center with a rod-like bulge) is in mechanical contact with the valve seat when the valve closes. This mechanical contact and the interruption of the mechanical movement at high actuation voltages results in a time-dependent current flow, through the piezoelectric effect, that can be detected as a sensor current according to the interruption back pressure in the micropump previously described.

[0061] Note that in this case, if this contact is not damped (e.g. in the case of a gas valve), it will occur within a very short time in the case of a stiff seal of the valve, which is associated with a corresponding high stress and therefore a large but short transient signal in the sensor current. If the valve movement is more strongly damped (e.g. in the case of a liquid), the stress peak will be smaller and the sensor current signal will also be smaller and more spread out in time. This correlation can be derived directly or trained using machine learning methods.

[0062] In principle, the inventive concept can be used not only to detect the closure of the valve, but also to detect which medium (e.g., air or liquid) is located in the valve chamber. In addition, the temporal allocation of the closure time with respect to the actuation cycle can be used to make a statement regarding the change, for example, a drop in the actuator bias, a bulging of a soft valve seat, or a change in an environmental parameter such as pressure above the membrane element or temperature can be detected. In addition, overdriving the membrane actuator 100 and thus excessive mechanical stress of the piezoelectric actuated membrane actuator 100 can be avoided when the valve only operates up to this detectable contact point.

[0063] The following table 1 provides an overview of external influences that may act on an active piezoelectrically actuated microvalve and that may be identified and / or classified using the concepts of the present invention.

[0064] [Table 1]

[0065] 2. Electrostatically Driven Membrane Actuator 2.1 Electrostatically driven micropump In the case of the electrostatically actuated membrane actuator 100 (FIG. 2--right side), compared to the piezoelectric actuation, there is no electromechanical coupling in the sense that the pressure change in the actuator element 102 itself causes an additional current flow. However, there is an analog dependence of the current flow on the membrane actuator 100, which depends on the mechanical and fluidic responses.

[0066] Firstly, this capacitance C depends on the mechanical position of the membrane actuator 100. For example, if the membrane element 101 is relaxed and flat at the beginning of the attraction stroke, the initial capacitance C 0 Let us assume that the control signal (say a voltage U) has a time constant τ = R C 0 , C 0= 100 pF = 1e-10 F, and R = 10 kΩ, then τ = 1 μs. Within this short period of time, it is assumed that the membrane actuator 100 has not yet moved due to the inertia and friction of the materials involved (membrane, fluid).

[0067] Next, the membrane 101 is subjected to an applied voltage U 0 When the membrane element 101 moves towards the counter electrode (actuator element 102) due to the interaction of the electrostatic attraction at and the pump chamber pressure p, the capacitance increases. This causes an additional current flow on the counter electrode (actuator element 102), which is purely correlated with the movement of the membrane element 101, which in turn also depends on the pump chamber pressure p.

[0068] The interaction of the basic movement of the membrane element 101 of the electrostatically driven micropump and the pump chamber pressure p during the suction stroke (i.e. after application of a control signal) is as follows: First, the restoring hook force acts in the opposite direction to the deflection force, which increases non-linearly due to the change in capacitance. When the membrane element is deflected about 1 / 3 towards the counter electrode 102, the membrane element 102 "snaps", causing the membrane element 102 to move very fast towards the counter electrode 102.

[0069] This membrane movement is stopped when the membrane element 102 contacts the counter electrode 102, first at the center of the membrane and then the membrane element 101 nests with the counter electrode 102. After contact and nesting, the capacitance increase slows down. The counter electrode 102 exerts a strong intrinsic counter force until the membrane element 101 finally reaches a force equilibrium state and stops.

[0070] During this entire movement of the membrane element 101, the applied voltage U 0The capacitance C(t) at increases, causing a corresponding current flow. If this current flow is measured in a time-resolved manner by the signal processing device 105 of the present invention, the time-resolved position of the membrane element 101 can be directly inferred (if the relationship between position and capacitance is known).

[0071] Since this "sensor current" dominates in electrostatic actuation, in contrast to piezoelectric actuation, the signal generating device 103 and the signal processing device 105 face significantly lower requirements.

[0072] Furthermore, there is also a difference whether the pump chamber contains air or liquid. In the first case, the membrane element can move faster since air has a significantly smaller mass compared to water. This makes it possible to detect whether the electrostatically driven micropump contains air or liquid. That is, the coagulation state of the fluid placed in the microfluidic component (micropump) 1000 can be detected.

[0073] At the suction stroke, the control signal 104, e.g. voltage, is switched off and short circuited, i.e. U=0V. The charge dissipates with a very short (electrical) time constant and the voltage is no longer applied after this time. If there is an abrupt voltage change, this time constant τ D0 is the capacitance C at the beginning of the pressure stroke together with the charge resistance R D0 depends on τ D0 = R C D0 , which is in the microsecond range. Within this short period of time, liquid cannot yet flow through the valve.

[0074] However, on subsequent movement of the membrane element 101, the capacitance changes, but there is no longer any current flow (due to U=0V) when the membrane element 101 moves with a pressure stroke, and therefore no movement can be measured by the sensor current.

[0075] Alternatively, rather than turning off the voltage completely in the pressure stroke, a small value U1 , for example, U 1 =0.01 U 0 Or even lower, which would allow nearly the entire stroke to be utilized, but still have charge available after the membrane actuator 100 is turned off and residual current flow due to the movement of the membrane actuator 100.

[0076] The membrane element 101 is released from the counter electrode 102 and returns to its initial state. During this movement, the capacitance decreases and a reverse current flow occurs.

[0077] 2.2 Electrostatically actuated microvalve Similar to the piezoelectrically actuated microvalves described above, movement of the membrane actuator 100 and contact with the valve seat can also be detected when the electrostatically actuated microvalve is closed. See 1.2 for a more detailed description.

[0078] In summary, it is noted that the embodiments may provide a piezoelectric actuated membrane actuator 100 or an electrostatically actuated membrane actuator 100. The piezoelectric actuated membrane actuator 100 may be used in a piezoelectric actuated microfluidic component 1000, such as a piezoelectric actuated microfluidic pump or a piezoelectric actuated microfluidic valve. The electrostatically actuated membrane actuator 1000 may be used in an electrostatically actuated microfluidic component 1000, such as an electrostatically actuated microfluidic pump or an electrostatically actuated microfluidic valve.

[0079] In all these embodiments, for example, the signal processing device 104 may be configured to determine the type of fluid used in the microfluidic component 1000 based on the time signal curve of the control signal 104. Alternatively or in addition, the signal processing device 105 may be configured to perform a differentiation of the aggregation state of the fluid based on the time signal curve of the control signal 104, i.e., to distinguish whether the fluid is gas or liquid. Depending on the type of fluid or depending on the aggregation state of the fluid, the mechanical properties of the membrane actuator 100 may change, i.e., the type of fluid and / or its aggregation state may be described as an external influencer that may be identified and / or classified using the concepts of the present invention.

[0080] In an embodiment in which the microfluidic component 1000 comprises a microfluidic pump having a pump chamber, at least one membrane side of the membrane element 101 is in contact with a fluid disposed in the pump chamber, and a variable pump chamber pressure is generated in the pump chamber by actuating the membrane actuator 100. In such an embodiment, the signal processing device 105 may be configured to determine the variable pump chamber pressure based on the time signal curve of the control signal 104 and to use this to identify and / or classify the external influencer. That is, the pump chamber pressure causes a signal portion that results in a time deviation of the control signal 104. The pump chamber pressure then causes a characteristic change due to the presence of the external influencer. Thus, the deviation of the pump chamber pressure causes a change in the time signal curve of the control signal 104, whereby the external influencer causing it can then be determined. In summary, this means that the external influencer changes the pump chamber pressure, which then causes a change in the time signal curve of the control signal 104, whereby the external influencer can be determined. This means that the external influencer can be determined based on the deviation of the pump chamber pressure. This will now be explained in detail based on a specific example.

[0081] In an embodiment in which the microfluidic component 1000 comprises a microfluidic valve capable of opening and closing a fluid path, the signal processing device 105 may be configured to identify and / or classify external influencers based on the temporal signal curve of the control signal 104 and use this to determine time-varying operating parameters of the microfluidic valve.

[0082] At this point, it should again be noted that the external influencer is an external influencer that causes a deviation in the time signal curve of the control signal 104 .

[0083] As an example only, the first three rows of table1 are explained for this purpose.

[0084] Line 1: A change in the state of the valve, e.g. contact or release of a sealing element, causes a characteristic change in the time signal curve of the control signal 104. This change in the time signal curve makes it possible to determine a corresponding operating parameter, e.g. whether the valve is open / closed.

[0085] Line 2: Depending on the viscosity of the fluid used, the damping behavior changes when opening or closing the valve, resulting in a characteristic change in the time signal curve of the control signal 104. This change in the time signal curve makes it possible to determine the corresponding operating parameter, e.g. the presence of air bubbles or condensate, natural aspiration or false aspiration.

[0086] Line 3: For example, when opening a valve, additional forces may act on the actuator (so-called sticking) or the mechanical movement of the valve may be hindered by additional forces, for example if a particle prevents the valve from closing completely. This results in time-distinguishable charge and discharge curves, i.e. a characteristic change in the temporal signal curve of the control signal 104. This change in the temporal signal curve makes it possible to determine the corresponding operating parameter, for example a particle blocking the valve seat or sticking preventing it from opening.

[0087] 3. Measuring the sensor current Next, the concepts of the present invention will be explained using the example of a piezoelectric actuated membrane actuator 100. However, with reference to the above discussion, the examples discussed herein apply to electrostatically actuated membrane actuator 100 as well.

[0088] First of all, it is a challenge to measure the current flow to or from the piezoelectric ceramic (actuator element) 102 without affecting the charging and discharging currents. Also, the measuring circuit (signal processing device) 105 should have high signal quality in order to measure weak currents without errors. It should be possible to realize the entire circuit concept as space- and cost-efficiently as possible.

[0089] FIG. 3 shows a schematic diagram of an embodiment comprising a signal processing device 105 configured to determine the time course of the control signal 104 .

[0090] 3 illustrates the piezoelectric (or electrostatic) actuated membrane actuator 100 described above. In the case of the piezoelectric actuated membrane actuator 100, the actuator element 102 may be a piezoelectric ceramic. The actuator element 102 may be arranged on a carrier substrate 107. In some embodiments, the membrane element 101 may be arranged as a carrier substrate on which the actuator element 102 is arranged. For example, the carrier substrate 107, or the membrane element 101, and the actuator element 102 may be connected by adhesive bonding techniques.

[0091] A control electrode is disposed on the top side of the actuator element 102 and a ground electrode, electrically shorted by a point contact with the carrier substrate 107, or membrane element 101, is disposed on the bottom side.

[0092] The actuator element 102 is connected to a signal generating device (control electronics) 103. The signal generating device 103 generates a control signal 104. The carrier substrate 107, or membrane element 101, is connected to a signal processing device (measurement circuitry) 105 configured to detect a resultant charge flow 108 dissipated from the carrier substrate 107, or membrane element 101.

[0093] In this embodiment, the signal generating device 103 comprises a voltage source that generates the control signal 104 in the form of an alternating voltage signal. Alternatively, the signal generating device 103 may comprise a current source that generates the control signal 104 in the form of an alternating signal.

[0094] The supply voltage V of the voltage source 103 in The output of the voltage source 103 is connected to the control electrode of the piezoelectric ceramic 102. The ground output of the voltage source 103 is connected to the ground of the measurement circuit 105. The ground electrode of the piezoelectric ceramic 102 (or the conductive surface of the carrier substrate 107) is connected to the inverting input of an operational amplifier 109. The non-inverting input is connected to the ground potential of the signal generating device 103. The circuit illustrated here generates a virtual ground potential at the inverting input of the operational amplifier 109. The control signal 104, i.e. the current flow I into the piezoelectric ceramic 102, can be determined by a measurement resistor 110. This current flow I is divided into at least two signal parts: the charging current I C and the sensor current I S Including I C +I S =I=U / R applies.

[0095] By using a voltage controlled op-amp 109, the entire current flow I to or from the capacitive element (membrane actuator) 100 can be measured without loading the current source (signal generating device) 103 through the measurement circuit (signal processing device) 105. This results in a terminal voltage Uk or V at the output of the measurement circuit 105. out For Uk = -R*I we get

[0096] The sensor current I illustrated in FIG. 3 S and the charging current I C For the high-resolution measurement of, a corresponding high amplification is used. However, in the case of a large current, especially when using a rectangular signal, this causes an overdrive of the measurement circuit 105. For these reasons, two proposed solutions (active and passive) to ensure the quality and reliability of the measurement circuit (signal processing device) 105 are presented below.

[0097] In the active solution, the measurement circuit 105 can be extended by a digital potentiometer. Thus, depending on the application and the control signal 104, the amplification of the measurement circuit 105 can be ideally adjusted. In addition, variable amplification is used to amplify the characteristic signal part depending on the application. When using the corresponding high-speed logic circuit, a large charging current I C and a small sensor current I S can be measured.

[0098] Alternatively, FIG. 4 shows a passive solution that can use a fixed resistor and diode network. The maximum current is limited via two Zener diodes 111, 112. The leakage current through the Zener diodes 111, 112 is reduced by additional low-leakage current diodes 113, 114, resulting in improved signal quality. The maximum value of the terminal voltage Uk is determined by the Zener voltage UZ and the forward voltages of the low-leak diodes UF, D and the Zener diodes UF, ZD, and Uk < UZ + UF, ZD + UF, D = 4.3V + 0.7V + 0.5V = 5.5V. Among other things, the advantages of the passive solution are that it is very compact, highly reliable, simple, and easy to integrate into an ASIC.

[0099] Furthermore, the signal processing device 105 may include a measurement resistor 110 connected between the inverting input and the output of the operational amplifier 109. Then, the voltage of the control signal 104 drops across both ends of this measurement resistor 110, and the signal processing device 105 processes this.

[0100] The control signal 104 is accompanied by different temporal signal curves depending on whether external influencers are acting on the microfluidic element 1000. In addition, the temporal signal curve of the control signal 104 is different depending on the prevailing external influencer. That is, in the absence of external influencers, the control signal 104 is accompanied by a first temporal signal curve. However, in the presence of one or more external influencers, the control signal 104 is influenced by them and the influenced control signal 104 comprises a different second temporal signal curve.

[0101] In an embodiment of the present invention, the signal processing device 105 is accordingly configured to distinguish a first temporal signal curve of the control signal 104 corresponding to the operation of the membrane actuator 100 without any external influence from a different second temporal signal curve of the control signal 104 corresponding to the operation of the membrane actuator 100 under the influence of at least one external influence factor.

[0102] 4. Qualitative consideration of external influencing factors of piezoelectric membrane actuator and corresponding interrelationships In the next section, the sensor current I S develop a model for how to represent the fluid interactions of the piezoelectric-driven micropump 1000. The model is subject to simplifications such as the following: Hysteresis ignored - Ignoring piezoelectric creep Assumption of linear valve characteristic curve Separation and superposition of fluid and electrical models Linearization of bubble state changes Description of elastic mechanics by simplification of Kirchhoff plate theory

[0103] Due to these simplifications, this model cannot accurately represent the whole of reality in the sense of a "white box" model (see section 7.1 "White Box"). However, important interrelationships are described, which greatly simplify the invention. In some cases of disturbance detection, "machine learning" using neural networks can be used (see section 7.2 "Black Box"), and the models described subsequently can significantly simplify machine learning (see section 7.3 "Grey Box"). By using the models described subsequently, machine learning training can also be completely omitted in other cases.

[0104] 4.1 Electromechanical coupling of membrane actuators The following sections are again illustratively described based on the piezoelectric actuated membrane actuator 100, and it is clear that all the descriptions are also applicable to the electrostatically actuated membrane actuator 100.

[0105] First, the electromechanical coupling (for small signals) can be described by the following linear approximation:

[0106]

number

[0107] Q: charge on the piezoelectric ceramic, U: voltage at the piezoelectric ceramic, C* E : Coupling coefficient, p: Pressure applied to the membrane actuator. Coupling coefficient C* E describes the change in charge when the pressure difference changes. C el represents the capacitance of the piezoelectric ceramic.

[0108] That is, the charge Q flowing through the membrane actuator 100 depends on the time-varying control signal 104 (voltage U) as well as the time-varying pressure p (FIG. 1) beneath the membrane element 101. The charge Q is determined by the first term C el ·(UU 0 ), which is the capacitance Cel and the potential difference of the time-varying voltage U (UU 0 ) In addition, the charge Q is linked to the second term

[0109]

number

[0110] which includes the piezoelectric coupling coefficient

[0111]

number

[0112] and the time-varying and / or configuration-varying pressure p, or the pressure difference between the bottom and top sides of the membrane element 101 (pp 0 ) Equation [1a] describes the direct piezoelectric effect of the membrane transducer, or actuator 100.

[0113] The corresponding "twin equation" can be formulated as follows: The piezoelectric membrane transducer 100 is 0 ) or potential difference (UU 0 ) (or both). This volumetric displacement V can be calculated as follows:

[0114]

number

[0115] Coefficient C p represents the "fluid capacity" of the membrane transducer, i.e., how much the displaced volume changes when the pressure difference between the top and bottom sides of the membrane changes. The coupling coefficient C E * is the potential difference (UU 0 It describes how much the displaced volume changes when the force (V) changes.

[0116] Coupling coefficient C from equation [1b] E * is the coupling coefficient C according to equation [1a] E It may be emphasized that it is the same as *.

[0117] Under the model requirements of the Kirchhoff plate theory, the coefficient C of the circular piezoelectric membrane (e.g., PZT ceramic) and the circular carrier membrane (e.g., metal or silicone) p (fluid volume) and C E *An analytical solution has been found to calculate the coupling coefficient (Source: https: / / www.sciencedirect.com / science / article / pii / S0924424710002311). Therefore, the fluid volume C p is analytically determined, which is the radius of the piezoelectric ceramic R p , the thickness of the piezoelectric ceramic T p , radius of carrier membrane R M , the thickness of the carrier membrane T M , and depends on the elastic constants and Poisson's number of the two materials (piezoelectric ceramic and carrier substrate). C p and C E The exact expression of the (long) analytical equation for * is given in the appendix to the paper by M. Herz, the lead author of the above publication.

[0118] Coupling coefficient C E * is calculated analytically (with the same derivation), and (C p (as in ) depends on the parameters explained above.

[0119] Coupling coefficient C E *In addition, the piezoelectric charge constant d 31 It is important to note that the piezoceramic resistance is proportional to the piezoceramic capacitance, which is typically given on the piezoelectric ceramic manufacturer's data sheet.

[0120] C ETo illustrate *, consider a counter pressure (p=p 0 The stroke volume ΔV of a micropump without the pump is calculated from Eq. 1b as follows:

[0121]

number

[0122] This is the coefficient C E It serves to illustrate the meaning of *. Coefficient C E * is the coefficient d 31 is directly proportional to and is always negative because it is negative for physical reasons (if an electric field is applied to a piezoelectric ceramic in the polarization direction (z-direction) by applying a voltage, the piezoelectric ceramic will contract laterally). This is why the stroke volume is positive in equation [1c].

[0123] The two coupling equations are symmetric and are illustrated below.

[0124]

number

[0125] This assumes that the electrical and piezoelectric effects add linearly. The current flowing on the piezoelectric membrane actuator 100 is the result of the time derivative of Equation 1:

[0126]

number

[0127] In this time derivative, the capacitance C el and the coupling coefficient C E Take into account (by chain rule) that * is not a constant, but is time-dependent (due to the large-signal behavior of piezoelectric ceramics).

[0128] For microfluidic components 100, such as micropumps and microvalves, the control of the piezoelectric ceramic requires high voltages (or high electric field strengths, where the electric field strength E in the piezoelectric ceramic is proportional to the thickness T of the piezoelectric ceramic) for large signal operation. p So E = (UU 0 ) / T p Since this occurs regularly at the capacitance C el and the coupling coefficient C e * has voltage dependency. During control, the voltage changes with time, so C el and C E The value of * also varies with the large-signal behavior of the piezoelectric material. Again, C E * is the piezoelectric charge constant d 31 Proportional to C E * is increased for large signals (high voltages), C E Note that * is larger at high field strength than at low field strength by a factor of about 1.5 (depending on the piezoelectric material).

[0129] Ground potential U 0 = constant, atmospheric pressure p 0 = constant (within small time intervals), this simplifies to:

[0130]

number

[0131] In the present example, the current I represents the control signal 104. All four terms on the right hand side of Equation 2 have a time dependency, so there are a total of four signal portions, which in the context of this disclosure will be referred to as the “sensor current” I U , I C , I ce , I p The sensor current I U , I C , I ce , I p Adding these together, the total current I, i.e., the control signal 104, is divided into four signal parts I U , I C, I ce , I p Includes.

[0132]

number

[0133] Each of these addends is linked to the evolution of a particular physical quantity over time. After a sudden rectangular disturbance, the electromechanical system goes out of balance and then tries to reach equilibrium again with a typical time constant. More precisely, the next physical subsystem goes out of balance and then tries to reach equilibrium again. 1) Free Charges on Metallization of Piezoelectric Ceramics (I u section) 2) After the voltage is changed, the domain size changes and aims to reach a new equilibrium state (I c and I ce (section 3) The pump chamber pressure seeks to reach a new equilibrium state, which is essentially caused by the flow of the medium through the valve (I p section).

[0134] 4.2. Compensation process after a sudden voltage change is applied After an abrupt voltage change, there are two essential effects (see Figure 5). 1) Electrons flow into the capacitance C of the piezoelectric ceramic 102, which a. Electrons have a time constant τ A =R*C el flow onto the piezoelectric ceramic 102, b. The atoms "stretch infinitely fast" and this stretching is then transmitted at the speed of sound. i. The capacitance changes very quickly (this large signal capacitance is larger than the capacitance at small voltages, which is why more charge flows into the piezoelectric ceramic 102). ii. Coefficient d 31 also increases (large signal behavior).

[0135] 1) All changes under τ A =R*Cel piezoceramic 102. 2) Domain growth leads to an increase in the number of favorably positioned magnetic domains (Weiss domains). These are slowed down by defects and grain boundaries. These changes occur on a significantly slower time scale, τ D Due to domain growth, the polarized volume advantageously increases. This effect is also called piezoelectric creep and accounts for about a few percent of the total effect. It changes: a. Again, capacitance, and b. Coefficient d 31 will also be slightly larger.

[0136] From this consideration, the capacitance C and further the coefficient d 31 (Piezoelectric lateral or transverse effect or d 31 Note that the effect, ω, where the mechanical force acts transverse to the applied field, initially varies very quickly with voltage U and then varies more slowly with domain growth.

[0137] FIG. 6 shows an example for the measurement of the voltage-dependent capacitance C(U) of a piezoelectric ceramic 102. The capacitance in large signal behavior (typical voltage range of micropumps and microvalves) was measured as a function of the applied voltage. The capacitance has a voltage dependence, i.e.

[0138]

number

[0139] Since the voltage has a time dependence, the term

[0140]

number

[0141] also has a time dependence. This is because the term

[0142]

number

[0143] is different from zero. In addition, a slight hysteresis of the capacitance can be determined.

[0144] When there is an abrupt application of a voltage change, each of the above physical subsystems will independently attempt to reach a respective state of equilibrium, which will take a different time, i.e., each physical subsystem will decay with a particular time constant τ as follows: 1) The change in voltage over time, dU / dt, is given by the time constant τA = R*C el occurs exponentially. 2) The change in pressure with time dp / dt occurs for fluid-related reasons. Hence the associated current term or sensor current I p is, so to speak, a typical stroke time τ h , which can be used to "peek" into the pump chamber with a microscope. 3) Change in capacitance dC el / dt is related to the piezoelectric effect. r The capacitance is so large because, among other things, there are so many dipoles in the piezoelectric ceramic 102 after polarization has occurred. While the electrons and atoms of a piezoelectric structure can, so to speak, follow disturbances with "infinite speed", the propagation of magnetic domains (Weiss domains) occurs on a much slower time scale τ C It is known from this so-called "piezoelectric creep" that its time scales can reach the range of several seconds. 4) Change in binding constant dC E * / dt is also related to the piezoelectric effect. This effect is (at least in part) similar to 3). a. The electrons and atoms of the piezoelectric ceramic 102 can follow disturbances with "infinite speed", but the propagation of magnetic domains (Weiss domains) occurs on a slow time scale τ d This occurs at τ CFrom "piezoelectric creep" it is known that this time scale can reach into the range of a few seconds. b.Coupling constant C E * is the piezoelectric coefficient d 31 which depends on the large signal, i.e., for high voltage amplitudes, by a factor d 31 is about 1.5 times larger than for small voltages, and by extension, C E *This may affect the differentiation of

[0145] From the above description, the compensation process can then be described in [Equation 4] as follows:

[0146]

number

[0147] 1) Free charge compensation process:

[0148]

number

[0149] 2) Compensation process of pump chamber pressure (without squeeze film damping, with linearized valve characteristic curve)

[0150]

number

[0151] Fluid time constant τ h is the total fluid flow resistance R fluidic (Definition of flow resistance R: Δp=R*Q) and fluid capacity C fluidic (Definition of fluid capacity C: ΔV=C*Δp) is the product of flow resistance R fluidic can be approximated as the sum of the flow resistance of the pump chamber and the flow resistance of the open and closed valves, whereas the fluid capacity C fluidic is the fluid capacitance of the driving membrane, C p and the (optional) fluid volume of the bubble (C gas), these flow resistances and fluid capacities are not constant but change value during the pumping process. For example, the flow resistance of an open flap valve is much smaller at the beginning of the pumping process when the valve is opened than at the end of the pumping process when the valve is closed. Nevertheless, this compensation process can be approximated by [Equation 6]:

[0152] Thus, the transient current curve I measured by the inventive characteristics of the “intrinsic sensor technology” of the microfluidic component 1000 described herein p provides direct insight into fluid processes in a microfluidic component 1000, for example a micropump.

[0153] 3) Domain Compensation Process a) Coupling coefficient

[0154]

number

[0155] About

[0156]

number

[0157] b) Capacitance C el About

[0158]

number

[0159] Capacitance C el and the coupling coefficient

[0160]

number

[0161] The compensation processes in have the same physical cause, namely the growth of the domains. Therefore, these two compensation processes take place with approximately the same time constant. τ d = τ C = τ piezo [Formula 9]

[0162] Here, the amplitude I C0 and I d0 can be different. This results in: a) Coupling coefficient

[0163]

number

[0164] About

[0165]

number

[0166] b) Capacitance C of the piezoelectric ceramic el About

[0167]

number

[0168] So finally, the approach for the total current I results in:

[0169]

number

[0170] The result is therefore a system with seven coefficients, four current amplitudes and three time constants. The electrical time constants are fast, while the two hydraulic time constants are slower. By fitting the coefficients, the electromechanical coupling can be fully captured during pump operation.

[0171] In interpreting the sensor current, it is essential that these physical compensation processes after a disturbance (e.g. application of a voltage) independently try to reach an equilibrium state. This makes it possible to consider the current flow as a superposition of these effects according to [Equation 12] and to "fit it". Thus, information about the charging of the capacitance, the processes within the piezoelectric ceramic, and the fluid interactions can be determined through a single measurement process.

[0172] 4.3 Electromechanical coupling and compensation processes in control with harmonic AC voltage signals When a sinusoidal AC voltage signal is applied to a piezoelectric actuated membrane actuator 100, such as a piezoelectric driven micro-membrane pump 1000, there is an overlap of time constants, described material effects, and fluid mechanical system responses, which leads to distortion of the current signal.

[0173] Figure 7A shows the current and voltage curves of a sinusoidally controlled micromembrane pump, while Figure 7B shows, in comparison, the time-resolved current curves for a capacitor control.

[0174] To visually illustrate the hysteresis effect, it is appropriate to plot the output signal (here the measured current) against the input signal (the alternating voltage). The resulting diagram (the Lissajous curve) is illustrated in FIG. 8 and provides information about the ratio of the output signal to the input signal. In an ideal capacitor, there is a 90° phase offset between the voltage and the current, resulting in a circular Lissajous curve 181. External influence factors acting on the microfluidic component 1000 or the membrane actuator 100 distort the current signal (FIG. 7a) and result in a Lissajous curve 182 (FIG. 8) that deviates from the circular shape 181.

[0175] With different time constants and overlap of effects, nonlinear differential equations are obtained for which the initial value problem cannot be solved in a straightforward way. Methods from the field of machine learning (see section 7 and also figures 32 and 33) can be used to be able to state different system states (e.g. air bubbles in a liquid agent, closure of a fluid path - blockage) from the recorded current curve. With the help of function approximations from the field of classification, patterns can be reliably derived.

[0176] 9A and 9B show exemplary Lissajous curves for a micropump dispensing a single liquid dose under harmonic control and subjected to the disturbances of an "air bubble" (FIG. 9A) and a "blockage" (FIG. 9B). Both figures show a repeating pattern.

[0177] Through the extraction of significant features, the amount of data can be reduced to a minimum without losing important information. When this reduced data is presented to a classification algorithm, unseen measurement data can be mapped to each state with a certain probability.

[0178] The confusion matrix shown in Figure 10 shows the frequency of correct predictions of unseen data. Higher values ​​on the main diagonal indicate a very good prediction of the state. The first row and first column each describe a normal state with no external influencers. The second row and second column each describe a system state with a blockage. The second row and third column each describe a system state with an air bubble in front of the pump chamber (bubble_up). The fourth row and fourth column each describe a system state with an air bubble located in the pump chamber (bubble_pump). The fifth row and fifth column each describe a system state with an air bubble located behind the pump chamber (bubble_down).

[0179] In addition to classifying the states, machine learning tools also provide the possibility of regression on system size (e.g. backpressure).

[0180] Thus, Figure 11 exemplarily shows the voltage-current curves, or Lissajous curves, of a micropump at different back pressures. Again, there are repeating patterns that can be extracted and classified accordingly.

[0181] 5. Experiments to examine the compensation process A subsequent experiment conducted by the inventors can be used to explain the compensation process described above in more detail. The experiment was designed with the following parameters, thereby avoiding as many interference effects as possible: Structure size 5×5mm 2 A silicon micropump (5 × 5 high flow pump) was used, which has a pump chamber height of 20 μm. This is because the pressure drop in the pump chamber does not play a dominant role compared to the pressure drop over the microvalve. Therefore, as an acceptable approximation, a homogeneous pump chamber pressure p can be assumed. The micropump cycles between U=0V and U=U+V, and considers a stroke that ends with the voltage at U=0V. Through this (when the voltage becomes zero)

[0182]

number

[0183] The term can be ignored. The onset voltage increases in steps from U=1V to U=100V. The transient current data I(t) is measured. The measured data is then fitted with an exponential function.

[0184] All measurements in this section use square wave control, i.e. the voltage change across the piezoelectric ceramic is abrupt at the beginning. The current through the piezoelectric ceramic is measured and the resulting current measurements are fitted with an exponential function.

[0185] Figure 12 shows the raw data of the measurements, i.e. the transient measurement points versus the current through the piezoelectric ceramic. The plotted dots represent the individual measurements, the upper measurement 171 determined at a voltage of U=10V, the middle measurement 171 at a voltage of U=60V and the lower measurement 173 at a voltage of U=120V.

[0186] If the respective measurement data are fitted with a simple exponential function, significant deviations in the measurement results occur. This is shown by the solid curve 171 1 , 172 1 , 173 1 However, when a linear superposition of two exponential functions is fitted, it matches the measured data very well, as shown by the dotted curve 171. 2 , 172 2 , 173 2 These dotted curves correspond to individual measurement points and are almost identical. This indicates that, according to the above explanation, the two compensation processes overlap. Specifically, for each suction stroke and each pressure stroke, I p0 , τ h , I c0 +I d0 , τ piezo Four fitting parameters can be collected from this measurement series:

[0187] 5.1 Sensor Current: Amplitude Measurement and Interpretation In the following, we discuss the dependence of these four fitting parameters on the driving voltages for the suction stroke and the pressure stroke.

[0188] Figure 13 shows the current amplitude I with respect to the voltage during the suction stroke. p0 This is linked to the corresponding time constants which will be explained later with reference to FIG. 14. However, initially, the current amplitude I p0 (y-axis) can be seen as a function of voltage (x-axis) in FIG.

[0189] Notes. current amplitude Ip0 From the compensation process of the pump chamber pressure (without squeeze film damping and with linearized valve characteristic curve), we obtain the following equation according to [Equation 6]:

[0190]

number

[0191] Now the pump chamber pressure is:

[0192]

number

[0193] Among others, formula [13a] was published in 1994 (with different nomenclature) (https: / / iopscience.iop.org / article / 10.1088 / 0960-1317 / 4 / 4 / 004).

[0194] Equation [13a] is important because it describes (with an approximate linearized model) the time behavior of the pump chamber pressure when, after an (exponential) voltage change, the pressure first rises and then decreases due to the flow through the microvalves (through the current resistances REV and RAV, respectively).

[0195] Equation [13a] is a particular solution of a general differential equation with disturbance components for voltage U(t).

[0196] This differential equation can be illustrated as follows:

[0197]

number

[0198] The parameters used are: · p 1 and p 2: Pressure at the inlet of the micropump (p 1 ) and back pressure at the outlet (p 2 ) is used to calculate C fluidic : Capacity of membrane C p , (optional) volume of air bubbles C gas , capacity of the inlet valve (C EV ) and outlet valve capacity (C AV ) as the sum of the last two fluid capacities (C EV and C AV ) is usually very small and can be ignored in most cases. R EV and R AV : these represent the flow resistances of the two passive check valves. In reality, they are not constants but depend on the pressure (especially when the pressure difference across the valve is very small). However, in the analytical solution of the differential equation, they are assumed to be constant. In particular, through this approximation step, the solution p(t) of equation [13b] approximates a real curve only in the sense of a grey-box model. For a certain stroke (say a pressure stroke), the outlet valve opens and the inlet valve closes. Fluid conductance 1 / R AV is 1 / R EV significantly larger than 1 / R EV can be neglected in the solution of p(t) for the pressure stroke (and vice versa for the suction stroke).

[0199] Equation [13a] describes a particular solution of the differential equation [13b] with the following initial conditions: 1) The initial condition is p(0) = 0. 2) The disturbance element is the charge resistance R el The capacitance C of the piezoelectric ceramic across el is the charge of

[0200]

number

[0201] In the case of control via a piezoelectric amplifier, this rise time τ a can be in the range of a few microseconds, which is why τ in standalone battery-operated micropump control electronics is much faster (due to, among other things, the higher electrical internal resistance of the battery). a = 0.5 ms, i.e. a positive voltage is applied to the piezoelectric ceramic, which describes a pressure stroke (the membrane moves towards the bottom of the pump chamber and the stroke volume is displaced through the exhaust valve). 3) In the solution of equation [13a], the pressure and counterpressure are set to zero or to the atmospheric reference pressure p 1 =p 2 =p 0 was set to. 4) The flow resistance of the inlet valve (closed during the pressure stroke) is neglected. The differential equation [13b] changes the sign of the pressure difference across the valve by the parameter 1 / R AV and 1 / R EV Note also that since varies rapidly, it is only partially solvable.

[0202] Further solutions of these differential equations give the general primary pressure p 1 and back pressure p 2 Using this, it can be determined as follows:

[0203] There is an exponential voltage rise (same as equation [13a]), but with the primary pressure p 1 and back pressure p 2 The solution at is as follows:

[0204]

number

[0205] And the solution with sinusoidal voltage excitation is:

[0206]

number

[0207] If the initial condition is p(0) = 0, then:

[0208]

number

[0209] The voltage U(t) can be designed by the developer of the micropump system using the control signal and has a large influence on the mechanical and hydraulic response of the system. If the voltage change U(t) is fast enough that the natural oscillations of the valve (or even of the membrane transducer) are excited, they may oscillate. If the occurrence of the voltage change is too slow (the hydraulic time constant τ h Since the fluid soon flows through the valve at a slow membrane velocity (much slower than ω), significant pressure can no longer build up in the pump chamber. This is especially the case for harmonic excitation with small angular frequency ω.

[0210] These time-dependent equations [13a], [13e], and [13g] are differentiated with respect to time, and C E *When multiplied by, the "sensor current" I p is obtained directly, which, according to the present invention, gives the opportunity to use data science methods and "machine learning" to electrically detect fluidic or mechanical effects "connected" to the pump chamber pressure.

[0211] In summary, with respect to equation [13a] above, among other things, the pump chamber pressure is determined by the time constant τ a with a time constant τ h Note that the pressure amplitude p(t) decreases with the voltage amplitude UU 0 Since it is proportional to, the time derivative dp / dt is UU 0 is proportional to.

[0212] Sensor current I p0 Since is proportional to dp / dt, one would assume that there is a linear increase in amplitude with voltage. However, this increase is non-linear. The non-linear part can be explained as follows:

[0213] amplitude I p0 is the coupling coefficient

[0214]

number

[0215] which in turn is proportional to the coefficient d 31 Figure 13 clearly shows the nonlinear behavior of the voltage amplitude characteristic curve. This is proportional to d 31 This is a clear indication that the nonlinear rise of the characteristic curve between very low voltages (small signal) and large voltages (large signal) corresponds to roughly a factor of 1.5, so the factor 1.5 is the factor d between small and large signals. 31 was assumed as a practical value for increasing

[0216] The coefficient correlating with the pressure amplitude in the pump chamber is the amplitude I p0 can be extracted from, i.e., the coefficient d 31 A measure for the large signal behavior of can also be determined (through the variation of the voltage amplitude). Thus, the coefficient d 31 The large signal behavior of may also be discerned in the inventive concepts disclosed herein.

[0217] 5.2 Sensor Current: Measuring and Interpreting the Time Constant 14 shows the curve of the time constant τh in the suction stroke during the pressure drop in the pump chamber. The time constant (y-axis) is plotted as a function of the applied voltage (x-axis).

[0218] For small voltage values, the noise is still too large and the time constant τ h The experiments described herein show that the "high flow pump" has a typical stroke time τ significantly less than 1 millisecond for air as the medium (for which the pump is designed). hFor gas, this pump exhibits an increase in delivery rate of more than 1 kHz, so the fluid time constant is less than 1 ms, i.e., an exponential fit with a time constant less than 1 ms can be interpreted as the sensor current.

[0219] Time constant τ h increases with increasing voltage amplitude, which is physically consistent because at higher voltages the pump membrane moves closer to the bottom of the pump chamber, resulting in higher flow resistance and slower outflow process.

[0220] According to the experiments described herein, the pressure drop in the pump chamber slit (despite the pump chamber height of 20 μm) is still larger than the pressure drop at the microvalve. Thus, a higher voltage brings the actuator closer to the pump chamber, increasing the slit resistance and decreasing τ h increases.

[0221] In comparison, Figure 15 shows the time constant for the pressure stroke: curve 181 corresponds to the curve for the time constant of the suction stroke previously described with reference to Figure 14. Curve 182 shows the time constant for the pressure stroke.

[0222] First of all, it can be directly deduced from Figure 15 that the time constant for the suction stroke (curve 181) is slightly larger than the pressure stroke, i.e. the pressure drop for the suction stroke is longer than the pressure stroke. In terms of the pump chamber pressure from [Equation 13], this can be explained as follows:

[0223]

number

[0224] Time constant τ h For this, the result is as follows: τ h =R fluidic (CM +C gas ) [Formula 14]

[0225] Fluid capacity of the driving membrane, C p and the air fluid volume C gas Since the sum of is a value in the pump chamber, the typical stroke time is proportional to the fluid volume. In general, fluid volume describes the change in volume as a function of pressure change. In a driven membrane, the fluid volume is essentially constant in the context of Kirchhoff plate theory, but is always larger for gas volumes at smaller absolute pressures (suction strokes) than at larger absolute pressures (pressure strokes) due to the change of state equation (isothermal, adiabatic, or polytropic depending on the rate of state change).

[0226] However, the fluid volume of air, C gas is not a constant, but corresponds to the tangent to the isothermal or adiabatic equation of state and is proportional to the void volume. Pressure stroke: the remaining void volume after compression Suction stroke: remaining void volume after expansion

[0227] That is, the fluid volume of the bubble, C gas is larger in the suction stroke than in the pressure stroke, and therefore the typical stroke time in the suction stroke is larger than in the pressure stroke when pumping a compressible medium such as air.

[0228] Incidentally, this would also mean that 50 / 50 is not the optimal duty factor when pumping a compressible medium such as air, with the suction stroke requiring more time than the pressure stroke.

[0229] 5.3 Extraction of flow resistance in microfluidic components (as in the case of microfluidic valves or in the pumping chamber of a micropump) The sensor current that depends on the pressure underneath the membrane element 101 can be expressed as follows (considering Equation 23):

[0230]

number

[0231] Equation [15a] can be derived via the piezoelectric coupling coefficient k. 2 is a measure for the ratio of mechanical energy output to electrical energy input.

[0232]

number

[0233] The following relations apply:

[0234]

number

[0235] Rearrange the equation and find k 2 When we solve for , we get:

[0236]

number

[0237] Therefore, the sensor current I as a function of k p gives the following result:

[0238]

number

[0239] Therefore, equation [15a] and all subsequent equations can be written without k, and from equation [15a], the time constant τ a The sensor current I after p gives the following result:

[0240]

number

[0241] 5.3.1 Extraction of flow resistance in case of air pump If the microfluidic component 1000 is an air pump, i.e. a micropump configured to pump air, the flow resistance R fluidic can be calculated as follows:

[0242]

number

[0243] In addition, the flow resistance R total can be calculated from the amplitude as follows:

[0244]

number

[0245]

number

[0246] That is, the flow resistance R fluidic can be calculated independently.

[0247] 5.3.2 Extraction of further fluid parameters for the air pump case In addition, the term dI p,0 / dU can be measured in small signals.

[0248]

number

[0249]

number

[0250] In addition, the time constant τ h is the current amplitude Ip can be calculated from

[0251]

number

[0252] 5.4 Consideration of fitted time constants: 1) Exponential function for pump chamber pressure change dp / dt 2) An exponential function for the change in the coupling constant dd / dt.

[0253] The coupling constant d can change for different physical reasons. Firstly, the (static) amplitude of the coupling constant d can change in large signals, and experience shows that the amplitude increases by about 50% between small and large signals. This increase can be due on the one hand to domain growth (this effect indicates slow relaxation) and on the other hand to the nonlinearity of the "force curves" of the atoms (this effect does not indicate slow relaxation).

[0254] Binding constant

[0255]

number

[0256] The time evolution of is also linked to the "pulsation" of the domains. This time constant is defined by the obstacles that the domain growth encounters during propagation or (as in the experiment above) contraction. This pulsation is purely a material property of the ceramic.

[0257] The piezoelectric creep mentioned at the beginning when applying a step function has the same origin. These time constants are slow, ranging from a few milliseconds to a few seconds.

[0258] At small voltages, the entire exponential function seems to disappear. Among other things, this can be explained by the fact that pulsating domain growth hardly occurs any more at small signals. Moreover, the coefficient

[0259]

number

[0260] is multiplied by the pump pressure p, which is also very small at low voltages.

[0261] 5.5 Extraction and consideration of amplitudes In addition to the time constants mentioned above, there are two signal parts: the sensor current I p0 and I d0 The amplitude of is also noteworthy.

[0262] I p The analytical formula for (t) is:

[0263]

number

[0264] The following applies at time t=0:

[0265]

number

[0266]

number

[0267] 5.6 Extraction and consideration of fitting parameters specific to the measured piezoelectric 5.6.1 Domain growth U*dC / dt+p*dd / dt, amplitude, and time constant for suction and pressure strokes Two signal parts or current terms I ce and I C The compensation process in (see Equation 12) is linked to the piezoelectric effect, or to the piezoelectric material used. In this case, there are two parts: a) Coupling coefficient

[0268]

number

[0269] About

[0270]

number

[0271] b) Capacitance C el About

[0272]

number

[0273] 5.6.2 Suction stroke considerations At the suction stroke, the voltage U is switched off, which has the following effect: Since U=0, the term I c is zero. The magnetic domains (Weiss regions) shrink, i.e. by a factor d 31 becomes smaller,

[0274]

number

[0275] becomes negative. pp during suction stroke 0 is less than zero, so this term is also negative. Overall, I ce The calculation of involves multiplication by two negative coefficients, i.e., I ce is positive on the suction stroke.

[0276] The corresponding measurement series for the suction stroke amplitude and time constant are shown in Figures 16A and 16B. Figure 16A shows the piezoelectric induced current term I as a function of voltage U (x-axis). ceFIG. 16B shows the time course of the amplitude of the piezoelectric time constant τ as a function of the voltage U (x-axis). piezo The graph shows the time progression of the graph (y-axis).

[0277] FIG. 16A shows the current term

[0278]

number

[0279] is positive. The pressure amplitude is correlated with the voltage amplitude, so the amplitude increases with voltage. The coefficient d 31 ,or

[0280]

number

[0281] The domain contraction is larger at large voltages than at small voltages due to the large signal effect of the time constant τ piezo Note that for , it is within about 2 milliseconds.

[0282] 5.6.3 Pressure stroke considerations The corresponding series of measurements for the amplitude and time constant of the pressure stroke are shown in Figures 17A and 17B. Figure 17A shows the piezoelectric induced current term I as a function of voltage U (x-axis). ce FIG. 17B then plots the amplitude of the piezoelectric time constant τ as a function of the voltage U (x-axis). piezo The graph shows the time progression of the graph (y-axis).

[0283] In the pressure stroke, the term I ce =pp 0 is positive. The magnetic domains (Weiss domains) expand,

[0284]

number

[0285] Therefore, the coefficient d 31 ,therefore

[0286]

number

[0287] is also increasing, so it is positive. The current term or signal part I ce is therefore positive.

[0288]

number

[0289] The second piezoelectric induced signal portion or current term I is given as follows: C (=capacitance term) is also (UU 0 ) is positive, so the change in capacitance is also positive, which means that U c This is why it is also positive.

[0290]

number

[0291] The amplitude measured in the pressure stroke is positive, but slightly smaller than in the suction stroke. The time constant τ in the pressure stroke piezo is also slightly smaller than the time constant for the suction stroke, i.e., domain growth works faster than domain contraction.

[0292] 5.6.4 Low-voltage considerations 18A and 18B show the corresponding series of measurements of amplitude and time constant at small voltages. FIG. 18a shows the piezoelectric induced current term I as a function of voltage U (x-axis) at small voltages of a pressure stroke. ce FIG. 18B shows the time course of the amplitude of the piezoelectric time constant τ as a function of the voltage U (x-axis) at low voltages during the suction stroke. piezo The graph shows the time progression of the graph (y-axis).

[0293] As can be seen, at very small voltage amplitudes there is no domain growth and no hysteresis for either the pressure or suction strokes.

[0294] 5.7 Further Experiments In addition to the experiments described in Section 5, further experiments may be undertaken to validate and optimize the concepts of the present invention.

[0295] For example, bare piezoelectric ceramics can be measured. This is the pump chamber pressure term I p is completely lost, and hence the signal part or current term I c and I ce This would be the simplest way to find out whether only the parameter I remains without the term dp / dt. ce0 , I c0 , τ piezo A piezoelectric actuated membrane actuator 100 without a valve chip or without a valve flap can be examined to see if dp / dt = 0. Then, dp / dt will always be approximately zero.

[0296] Alternatively or in addition, the bonded and biased piezoelectric ceramic can be measured and a comparison used to determine how much a parameter changes due to bias compared to a bare piezoelectric element.

[0297] Further experiments or measurements may be contemplated, such as the following: 1) Start with U+ = 1 volt and increment the voltage U+ in small steps (thereby ensuring small signals are determined) 2) Consider the pressure stroke, i.e. from 0V to U+. This results in addition to that the term U*dC / dt, i.e. the amplitude I c0 is the amplitude I ce0 should be added to the time constant τ, otherwise the other parameters will be almost unchanged. piezo Instead of the individual time constants τ cand τ d The fitting is performed using, among other things, these parameters τ c and τ d It can be verified that they are equal. 3) A different series resistor (or resistance) R110 that selectively lengthens the electrical time constant can be used to verify the lack of correlation. 4) Besides varying the back pressure (section 6.3 and Figures 30A to 30E), the pressure could also be varied, for example from -15 kPa to 0 kPa.

[0298] 6. Determination of external influence factors based on the time-signal curves using the examples of electrostatically and piezoelectrically driven membrane actuators In summary, the above description of the theory shows that the control signal 104 (e.g., the alternating current signal I(t)) includes a time signal curve that can be influenced by an external influence factor. That is, the time signal curve deviates in the presence of the external influence factor from the time signal curve in the absence of the external influence factor. Different influence factors can cause different deviations in the time signal curve, which correspond to detectable characteristic signatures in the time signal curve.

[0299] In the context of the present disclosure, the time signal curve influenced by the external influence factor is also referred to as the influenced signal curve of the control signal 104. Thus, the control signal 104 itself may also be referred to as the influenced control signal 104.

[0300] The control signal 104 may include several signal portions or current terms depending on whether the microfluidic component 1000 or the membrane actuator 100 is piezoelectrically or electrostatically actuated.

[0301] 6.1 Electrostatically Actuated Membrane Actuator In the case of the electrostatically driven membrane actuator 100, for example, the sensor current may include two different current terms that are linked to the capacitance. The first signal portion or current term I Udescribes the free charge on the actuator element 102. The second signal part or current term I C depends on the domain size, as mentioned above. The second current term I C is linked to the change in capacitance over time between the membrane element 101 and the actuator element 102 due to the movement of the membrane element 101. Thus, in the electrostatically driven membrane actuator 100, the control signal 104 includes two signal parts or current terms:

[0302]

number

[0303] According to an embodiment of the invention, the signal processing device 105 processes at least two different signal portions I U , I C are configured to distinguish from one another, First signal part I U is linked to the change in voltage over time when charging or discharging the capacitance between the membrane element 101 and the actuator element 102, Second signal part I C is linked to the change in capacitance over time between the membrane element 101 and the actuator element 102 due to the movement of the membrane element 101 .

[0304] That is, the membrane actuator 100 may be an electrostatically driven membrane actuator (see FIG. 2 - right side) in which the membrane element 101 forms the movable electrode and the actuator element 102 forms the counter electrode, the actuator element 102 and the membrane element 101 co-operating capacitively, and the control signal 104 induces a flow of charge on the counter electrode 102, which causes the membrane element 101 to move relative to the counter electrode 102. During this movement of the membrane element 101 relative to the counter electrode 102, there is a change in capacitance, which in turn affects the time course of the control signal 104, i.e., a signal portion or current term I C Link to.

[0305] Therefore, the signal portion or current term I of the control signal 104 U and I C has different origins, i.e., separate signal parts or current terms I U and I C There are different external influences that can affect to different degrees the time course of the control signal 104 as a whole. U and I C are correlated with different external influencers that can affect the time course of the control signal 104 differently.

[0306] Therefore, according to an embodiment of the present invention, the inventive signal processing device 105 is configured to perform a signal analysis of the time signal curve of the control signal 104 in order to determine and / or classify at least one causal external influence factor and to extract individual signal portions I of the control signal 104. U , I C are determined and the individual signal parts or current terms I U and I C correlate with different external influencers that have different effects on the time course of the control signal 104. The signal processing device 105 is therefore configured to assign specific signal portions or current terms IU, IC to specific external influencers and use this to identify and / or classify the respective external influencers.

[0307] For example, the signal processing device 105 may detect that the capacitance between the membrane element 101 and the actuator element 102 changes over time while the membrane element 101 moves towards the actuator element 102, thereby generating a signal portion I C changes, so the signal part or current term I C may be configured to determine the membrane deflection, or the time-varying position of the membrane element 101, using

[0308] As such, the time-varying position of the membrane element 101 also depends on the pressure acting on the membrane element 101 (see FIG. 1). Thus, according to an embodiment of the invention, the signal processing device 105 processes the signal portion I C is configured to determine the pressure acting on the membrane element 101 and identify it as an external influencer based on a signal analysis of the time course of the control signal 104, in particular the signal portion or current term I C It can be determined based on:

[0309] Over the entire movement of the membrane element 101, the capacitance C el (t) is the applied voltage U 0 , which increases with capacitance, causing a corresponding current flow. When this current flow is measured in a time-resolved manner with the signal processing device 105, the time-resolved position of the membrane element 104 can thus be directly inferred (given the known relationship between position and capacitance).

[0310] That is, the change in capacitance (= the second signal part I c If the relationship between the pressure, φ, and the membrane deflection is assumed to be known, then deviations from the known signal curve can be interpreted as external influences in the form of additional pressure.

[0311] Correspondingly, according to an embodiment of the present invention, the signal processing device 105 processes the second signal portion I C The signal processing device 105 first knows the time-dependent signal curve of the second signal part I C In contrast to the known time course of the second signal part I caused by pressure C and based on this deviation, determining the pressure acting on the membrane element 101 and identifying this as an external influence factor.

[0312] 6.2 Piezoelectric Membrane Actuators The piezoelectric actuated membrane actuator 100 has already been described in the theory section, where the actuator element 101 comprises at least one piezoelectric element, e.g. in the form of a piezoelectric ceramic, operatively connected to, e.g. a bonded piezoelectric element, e.g.

[0313] For example, when a control signal 104 in the form of an alternating voltage signal is applied to the membrane actuator 100, or actuator element 102, the actuator element 102 (e.g. a piezoelectric ceramic) is deformed due to the inverse piezoelectric effect. Due to this deformation of the actuator element 102, the membrane element 101 bulges upwards or downwards accordingly, i.e. the membrane element 101 is actuated or deflected therethrough. This makes it possible to open or close a valve, or to perform a suction or pressure stroke of a micropump.

[0314] Then, when an external influencer acts on the piezoelectric membrane actuator 100, for example through a particle, a bubble, contact at the valve seat, etc., this causes a force acting on the membrane element 101, which then applies a corresponding counterforce, or counterpressure (action = reaction). Through this, the piezoelectric actuator element 101 generates a corresponding current or voltage signal caused by the direct piezoelectric effect. This signal generated by the direct piezoelectric effect overlaps with the control signal 104, which causes the reverse piezoelectric effect. This overlapping signal caused by the external influencer then changes the time signal curve of the control signal 104. That is, the time course of the control signal 104 is influenced by the external influencer so that there is an influenced control signal 104 in the sense of the present disclosure.

[0315] Correspondingly, according to an embodiment of the present invention, the membrane actuator 100 is a piezoelectric driven membrane actuator in which the actuator element 102 includes at least one piezoelectric element, and the control signal 104 induces a deformation of the piezoelectric element by using the converse piezoelectric effect, which causes the piezoelectric element to apply an actuation force to the membrane element 101. The time signal curve of the control signal 104 is influenced by a signal originating from the piezoelectric element, which is generated due to a reaction force of the membrane element 101 based on the direct piezoelectric effect.

[0316] For example, a pressure change beneath the piezoelectric membrane actuator 100 will cause the application of a force, resulting in a current flow through the piezoelectric ceramic caused solely by the pressure change due to the direct piezoelectric effect. In the context of this disclosure, this current flow is also referred to as the "sensor current."

[0317] In the piezoelectric driven membrane actuator 100, the sensor current may include four different signal portions or current terms (see Equation 4):

[0318]

number

[0319] Thus, in the case of the piezoelectrically driven membrane actuator 100, when compared to the electrostatically driven membrane actuator 100, there are two additional signal parts or current terms I ce , I p are added. The two signal parts or current terms I U , I C is essentially caused by capacitance and occurs in the electrostatically and piezoelectrically driven membrane actuator 100, whereas the two signal parts or current terms I ce , I p is essentially due to the piezoelectric effect and therefore occurs only in the piezoelectric actuated membrane actuator 100. In the piezoelectric actuated membrane actuator 100, the capacitive current term I Cdepends at least in part on the piezoelectric effect since the change in capacitance is caused by the large signal effect of the piezoelectric ceramic (the actuator element 102).

[0320] where, again, this is the individual parts or current terms I U , I C , I ce , I p are correlated with different external influences that have different effects on the time course of the control signal 104.

[0321] Correspondingly, according to an embodiment of the present invention, the signal processing device 105 is configured to perform a signal analysis of the time signal curve of the control signal 104 in order to determine and / or classify at least one causal external influence factor, and to extract individual signal portions I of the control signal 104. U , I C , I ce , I p is determined and the individual signal parts I U , I C , I ce , I p correlates with different external influencers that have different effects on the time course of the control signal 104. The signal processing device 105 is configured to assign specific signal portions to specific external influencers and use it to identify and / or classify the respective external influencers.

[0322] According to one embodiment of the present invention, the signal processing device 105 therefore processes the above-mentioned four different signal portions or current terms I of the control signal 104 to be affected. U , I C , I ce , I p and may be configured to at least distinguish between: First signal part I U is essentially linked to the time variation of the voltage when charging or discharging the capacitance between the membrane element 101 and the piezoelectric actuator element 102, Second signal part I Pis essentially linked to the time variation of the pressure acting on the membrane element 101 and the piezoelectric coupling coefficient

[0323]

number

[0324] Also proportional to Third signal part I C is essentially linked to the time variation of capacitance caused by the large signal effect of the piezoelectric actuator element 102, Fourth signal part I ce is essentially the piezoelectric coefficient d 31 , which is also proportional to the time change of the pressure acting on the membrane element 101.

[0325] As described above in the theory section, these individual signal portions I U , I C , I ce , I p Each of the signals aims to achieve equilibrium through a separate time compensation process, in which each signal portion is compensated for by a separate time amplitude curve I U0 , I p0 , I c0 , I ce0 (See, for example, FIG. 13) and the individual time constants τ at which the time compensation process is carried out. a , τ h , τ d = τ c = τ piezo (See, for example, FIG. 14).

[0326] Correspondingly, according to an embodiment of the present invention, the signal processing device 105 processes the individual signal portions I U , I C , I ce , I p From one or more of the amplitude curves I U0 , I p0 , I c0, I ce0 , and / or the respective time constant τ a , τ h , τ d = τ c = τ piezo The method is configured to determine an external influence based on the external influence.

[0327] 6.3 Evaluation and interpretation of measurements for the determination of external influence factors The following section describes the determination of the external influence factors using a specific series of measurements performed with the piezoelectric actuated membrane actuator 100. From the above discussion, it is clear that these results are also valid for the electrostatically actuated membrane actuator 100.

[0328] Indeed, Fig. 19 shows a schematic of the measurement station used. It comprises a microfluidic component 1000 of the invention in the form of a micro-membrane pump. The micro-membrane pump 1000 comprises a membrane actuator (not illustrated) 100 of the invention having a membrane element 101 and an actuator element 102.

[0329] In the inlet reservoir 201, air and water can be selected as the medium or fluid to be conveyed. In the clockwise direction, two air bubble detectors 202 and one pressure sensor 203 are arranged in front of (upstream) the micropump 1000. Behind (downstream) the micropump 1000 there are also two air bubble detectors 202 and one pressure sensor 203. Through the illustrated conduit 204, the micropump 1000 conveys the fluid into the outlet reservoir 205. Further sensors are arranged between the two reservoirs 201, 205, such as a further pressure sensor 206 and a pressure controller 207 for monitoring and regulating the normalized atmospheric pressure.

[0330] The overall logic is illustrated diagrammatically in the center of Fig. 19. It includes a signal generating device 103 and also a signal processing device 105. The signal generating device 103 may include a function generator configured to generate different input functions (step function, Dirac function, sinusoidal signal, rectangular signal, etc.) as the control signal 104. Furthermore, the logic includes an oscilloscope 208 to determine the (affected) control signal 104. In addition to that, the logic may include a piezoelectric controller 209 for controlling and monitoring the piezoelectric actuator element and also a deflection sensor 210 for monitoring the deflection of the membrane element.

[0331] Using the hardware illustrated in the schematic, the following can be performed: Ambient pressure, temperature and humidity measurement; · The change between water and air as the medium to be conveyed; Injection of air bubbles to simulate external influences (disturbance factors), Air bubble detection by air bubble sensor, Time-dependent detection of the membrane stroke of a micropump, Back pressure sensor for water and air (switchable), (Optional) Measuring conveyor speed in the inlet and outlet paths by air flow sensors; (Optional) Measurement of conveyor speed via differential pressure sensor.

[0332] This allows the following to be controlled: Micropumps using any control or voltage signal (e.g., square or sinusoidal voltages); and / or Switching valves for switching between water and air operation or for injecting different size air bubbles.

[0333] Additionally, primary or counter pressure changes are also possible.

[0334] Using the measurement station illustrated diagrammatically, the following time-resolved measurements can be performed: Sensor currents for suction and pressure strokes via the sensor current electronics, Actuator stroke for suction and pressure strokes, Air bubbles using air bubble sensors, (Optional) Flow rate.

[0335] The following series of measurements, which will now be explained in more detail with reference to the drawings, were carried out. "Normal state" with air, 10Hz, control signal: rectangular, "Normal state" with water, 10Hz, control signal: rectangular, Water with bubbles, 10Hz, control signal: rectangular, Water with counter pressure, 10Hz, control signal: rectangular, Water with bias pressure, 10Hz, control signal: rectangular, and Contact at the bottom of the pump chamber, control signal: sinusoidal voltage, trajectory

[0336] The time course of the control signal 104 was measured and is shown in the following figure. It illustrates how the time course of the control signal 104 changes with the presence of an external influencer (compared to the "normal" state without an external influencer). As explained in the theory section above, the control signal 104 can be divided into up to four different signal parts or current terms I U , I C , I ce , I p Then, the current term I p which is proportional to the pressure underneath the membrane element (or the pump chamber pressure) and also the piezoelectric coefficient

[0337]

number

[0338] This current term I p is also referred to as the sensor current in the following figures. pThe following discussion of the other signal parts or current terms I U , I C , I ce may equally apply to.

[0339] In addition, the measured control signal captures flow effects that cannot be derived from a simplified model of overlapping exponential compensation processes.

[0340] For example, the inlet pressure may be excited to oscillate at its mechanical natural frequency, thereby causing pressure oscillations in the pump chamber. These pressure oscillations are superimposed on the damping pressure, resulting in the term I p and the oscillations are superimposed on the damping process.

[0341] "Fluid resonance" is a further example. It has long been known that the inertia of liquids in the inlet and outlet lines can be coupled to a micropump under certain arrangements. "Simulation of microfluid systems", Zengerle, R., Richter, M., Journal of Micromechanics and Microengineering, 1994, 4(4), pp. 192-204, 004

[0342] When a micropump is working, not only the liquid in the pump chamber but also the liquid column behind the open valve must move, since liquid cannot be compressed. This inertia of the liquid tubes in the inlet and outlet is called the "fluidic inertance" L, which is defined as being completely analogous to electrical inductance. fluidic This fluid inertance can be described by a crude model using the fluid capacitance of the actuating membrane, C p Along with the fluid resonance frequency

[0343]

number

[0344] represents a system that can oscillate at a constant speed, through which the possible oscillations lead to oscillations in the pump chamber pressure, which was shown both theoretically and experimentally in the abovementioned 1994 paper.

[0345] This fluid coupling can be decoupled by providing pressure smoothing elements directly before and after the pump, as already described in 1994. However, these additional elements are only advantageous in some applications, and fluid resonances often do not interfere with the pump operation.

[0346] All these effects overlap with the "normal" exponential compensation process described earlier, thus providing additional information about the state of the valve flaps and thus the outer surroundings of the micropump.

[0347] First of all, Figure 20 shows the case of "normal conditions" with air, with a pump frequency of 10 Hz and a rectangular control signal. Here, a total of 10 measurement curves were recorded and superimposed. The measurement curves look almost identical and characterize normal conditions, i.e. the absence of external influence factors.

[0348] Air was used as the pump medium and the micropump 1000 was operated without back pressure. A <<1ms, and frequency f=10Hz (τ A = τ a ) was used as the control signal 104.

[0349] FIG. 20 shows a suction stroke on the time scale, or between 0 ms and about 10 ms. Following this is a time region where the suction stroke is already fully completed (between about 10 ms and about 15 ms). The pressure stroke starts at 50 ms. The time region where the pressure stroke is performed extends up to about 65 ms on the time scale. After that there is a time region up to 100 ms where the pressure stroke is already fully completed. Then the next suction stroke (not illustrated here) starts at 100 ms. In this normal situation the expected exponential compensation (τh In addition to the fluid resonance (including fluid information about the air), vibration of the valve flaps can be observed. Because air has a density 100 times lower than that of water and a viscosity 50 times lower than that of water, the valve flaps can vibrate. Furthermore, the air in the lines before and after the pump has such a low mass and high fluid capacity that no fluid resonance can be observed when pumping air.

[0350] FIG. 21 is a detailed view of the time signal curve during the suction stroke (with air as pump medium) in normal conditions, i.e. without external influence factors. At the beginning, the signal curve contains a sharp rise resulting from the rectangular control signal 104 together with the capacitance of the piezoelectric charge. This sharp rise is cut off in the signal curve due to the Zener diodes 111, 112 in FIG. 4. The signal curve decays exponentially, i.e. the pump chamber pressure decreases exponentially. At the end of the decay process, there are oscillations resulting from the oscillations of the inlet flap. In the range starting from about 10 ms, the suction stroke is completely completed and the oscillations of the valve are decayed. The decay rate of this oscillation can also provide information about the geometry of the valve flap and the contact ridge.

[0351] FIG. 22 is a detailed view of the time signal curve in a pressure stroke (with air as pump medium) in normal conditions, i.e. without external influences. To begin with, the signal curve contains a sharp (negative) rise, which again results from the rectangular control signal 104 with the capacitance of the piezoelectric discharge. This sharp rise is cut off in the signal curve, caused by the Zener diodes 111, 112 in FIG. 4. What follows is an artifact that only occurs in the pressure stroke and can be ignored for the time being. Then the signal curve also decays exponentially, i.e. the pump chamber pressure increases exponentially. Towards the end of the decay process there are oscillations resulting from the vibrations of the outlet flap. In the range starting from about 56 ms the pressure stroke is completely completed and the valve oscillations decay.

[0352] Figure 23 shows the case of "normal conditions" with bubble-free water, a pump frequency of 10 Hz and a rectangular control signal. A total of 90 measurement curves were recorded and overlaid. The measurement curves look almost identical, characterizing normal conditions, i.e. the absence of external influencing factors.

[0353] Water was used as the pump medium and the micropump 1000 was operated without back pressure. A , and a square voltage with frequency f=10 Hz was used as the control signal 104.

[0354] Figure 23 shows a suction stroke on a time scale between 0 ms and approximately 10 ms. Following this is a time region where the pressure stroke is already fully completed (between approximately 10 ms and 50 ms). The pressure stroke starts at 50 ms. The time region where the pressure stroke is performed extends up to approximately 57 ms on the time scale. Following this is a time region up to 100 ms where the pressure stroke is already fully completed. The next suction stroke starts at 100 ms (not shown here).

[0355] FIG. 24 is a detailed view of the time signal curves in the suction stroke (water is the pump medium) in normal conditions, i.e. without external influence factors. Firstly, the signal curve contains a sharp rise resulting from the rectangular control signal 104 together with the capacitance of the piezoelectric charge. This sharp rise is cut off in the signal curve, caused by the Zener diodes 111, 112 in FIG. 4. The signal curve decays exponentially, i.e. the pump chamber pressure also decreases exponentially, resulting in small oscillations within the region of about 1.5 ms, which will be explained in more detail later. Towards the end of the decay process, there are almost no oscillations of the inlet flap, since the pump medium is water. Water has a viscosity 50 times higher than air, which causes these oscillations to be damped rather strongly. In the region starting from about 6 ms, the suction stroke is completely completed and the oscillations of the valve are damped.

[0356] 25A, 25B, and 25C are detailed views of the previously mentioned overshoot in the exponential decay process, with FIG. 25A corresponding to the previously described FIG. 24, FIG. 25B being an expanded view of the overshoot, and FIG. 25C being an even further expanded view.

[0357] As mentioned at the beginning, 90 curves were superimposed here. The original submitted document of this application was originally color-coded and included a color-coding of the individual measurement curves. This color-coding represents the recording time of the individual measurement curves, i.e. the time when each measurement curve was recorded (with respect to the other measurement curves). According to this color-coding, the measurement time starts with the yellow shade, passes through the green shade, to the blue shade and to the purple shade. That is, the measurement curve with the yellow shade was recorded first, the measurement curve with the green shade was recorded afterwards, the measurement curve with the blue shade was recorded afterwards and the measurement curve with the purple shade was recorded last. In addition, this color-coding was used in all the figures described in this section.

[0358] For example, Figures 25B and 25C show that the measurement curves recorded first (yellow shades) were the furthest away from the actual exponential function (purple) as illustrated in Figure 25. As the measurement time increases, the individual measurement curves approximate the exponential curve, i.e. the slope of the color code moves inwards in the exponential direction of Figure 5 as the measurement time increases. This overshoot is presumed to be due to fluid resonance and / or degassing and / or cavitation.

[0359] FIG. 26 is a detailed view of the time signal curve in a pressure stroke (water is the pump medium) in normal conditions, i.e. without external influences. To begin with, the signal curve again contains a sharp (negative) rise resulting from the rectangular control signal 104 together with the capacitance of the piezoelectric discharge. This sharp rise is cut off in the signal curve, caused by the Zener diodes 111, 112 in FIG. 4. What follows is an artifact that again only occurs in the pressure stroke and can be ignored for the time being. Then the signal curve again decays exponentially, i.e. the pump chamber pressure increases exponentially until equalization with the external pressure (e.g. atmospheric pressure) occurs. Towards the end of the decay process there are only very small oscillations resulting from the oscillations of the outlet flap. In the region starting from about 54 ms, the pressure stroke is now again completely completed and the oscillations of the valve decay.

[0360] Figures 27A and 27B then show the addition of a disturbance variable in the form of an air bubble, which corresponds to an external influencer that modifies the time course of the control signal 104. Figures 27A and 27B show the detection according to the invention of an air bubble in the suction stroke.

[0361] The micropump was operated at a pumping frequency of 10 Hz and a square voltage was used as the control signal. The pump medium was water and an air bubble was introduced into the system. In Fig. 27A and Fig. 27B 30 measurement curves are superimposed, where 20 suction strokes were recorded before the air bubble entered the pump chamber, 3 suction strokes were recorded while the air bubble was located in the pump chamber and 7 further suction strokes were recorded while the air bubble was leaving the pump chamber. Again, the color coding described above was used.

[0362] It can be seen that the temporal signal curve changes significantly when the bubble approaches the pump chamber in the first place. This leads to a significant deviation compared to the exponential decay process (FIG. 24) under normal conditions. Here, (FIG. 27B) a significant overshoot can be seen in the region from about 4 ms to 5 ms, and also in the region of about 8 ms and 11 ms. At this point, it should be noted that since colored drawings and therefore color coding cannot be used in this patent application, all the time indications shown here are purely exemplary to describe the temporal signal curve.

[0363] This overshoot occurs as a result of the water column contacting the pump changing due to air bubbles, which changes the wavelength of the fluid resonance.

[0364] As soon as an air bubble enters the pump chamber, the temporal signal curve changes significantly. In the region of about 4 ms, a very clear spike is seen, i.e. a very clear deviation from the exponential decay function in the normal state (FIG. 24). In this regard, reference is made to FIG. 27b, illustrated by the label "First sensor current measurement after air bubble entry."

[0365] In addition, an increase in the natural vibration of the valve is seen in the subsequent second and third sensor current measurements, where an air bubble is still disposed in the pump chamber. As low viscosity air is temporarily disposed in the pump chamber, the valve flap can then vibrate, which appears in the measured signal.

[0366] If the pump chamber contains air, surface forces act on the pump membrane and the valve flaps. For example, these surface forces may prevent the valve from moving freely any more. Depending on where the meniscus is located, the actuation membrane may also be subjected to asymmetric stresses. All this affects the pressure in the pump chamber. Since it is not possible to predict how exactly the air bubble will move in the pump chamber, the signal behavior appears chaotic.

[0367] As soon as the bubble leaves the pump chamber and moves away, the time signal curve again approximates an exponential decay function as in the normal state (FIG. 24). Now, the further away the bubble moves, the more closely the signal curve approximates an exponential decay function.

[0368] Next, Figures 27C, 27D, and 27E show in addition, again, situations, or signal curves, where there is a long lead time for the arrival and departure of the air bubble. First, Figure 27C shows the temporal signal curve of the control signal during 10 suction strokes before the entry of the air bubble. Next, Figure 27D shows the signal curves just before the entry of the air bubble, during the passage of the air bubble, and also just after the air bubble has left the pump chamber, as previously described with reference to Figures 27A and 27B. Finally, Figure 27E shows the signal curve during 10 suction strokes where the air bubble has moved further away.

[0369] As can be clearly seen from Figures 27A to 27E, the occurrence of air or gas bubbles in the liquid-transporting microfluidic component significantly changes the time signal curve of the control signal, which is an external influence factor that can be detected according to the inventive concepts described herein. Each external influence factor, like the exemplary described air bubble, brings about a characteristic deviation in the time signal curve, i.e. each external influence factor leaves an individual trace in the time course of the control signal, which can be detected using the inventive concepts described herein.

[0370] Figures 28A and 28B, purely for the sake of completeness, show that such characteristic signatures can be determined not only in the suction stroke, as just explained with reference to Figures 27A to 27E, but also in the pressure stroke, even though the signature in the pressure stroke may not be significantly more clearly distinguishable.

[0371] In summary, these experiments show that the ingress of air bubbles into the pump chamber can be detected very clearly. This is of great practical relevance for many micropump applications, for example the size of the air bubble can be fairly easily estimated using the number of pump strokes indicating the ingress of the air bubble and the stroke volume of the micropump. In the present case the stroke volume was about 6 μl and the number of pump cycles was 3, so the volume of the air bubble was about 6 μl*3=18 μl. Above all, the estimation of the amount of air bubbles is useful in drug administration, since very small air bubbles are medically harmless, while large air bubbles must be avoided at all costs.

[0372] In addition, air bubbles are also a significant disturbance in terms of dosing accuracy. When the size of the air bubbles is known, disturbances in dosing accuracy can be detected and compensated for by suitable means. For example, when an air bubble of size 18 μl (which is medically harmless when it reaches the patient in an infusion) is detected and quantified, the dosing may not be canceled with an error message (depending on the medical application) but may continue. Then, under the control of the micropump, three additional strokes may be performed, so that the pre-planned amount of drug reaches the patient. The number and size of the air bubbles may then be recorded in an error protocol.

[0373] 29A to 29D show a further case of detectable external influence factors, which in this case is the back pressure of a liquid-carrying micropump. For example, such a micropump may be used in a drug administration system. Here, a catheter blockage may occur, which has serious consequences for the patient, since the drug will no longer be administered properly. For example, such a catheter blockage may generate the aforementioned back pressure, which corresponds to the occlusion pressure of the micropump (the occlusion pressure is the maximum pressure that the micropump can generate with liquid as pump medium). FIG. 29A firstly shows an overview of the influence of the back pressure on the time signal curves of the suction stroke 291 and the pressure stroke 292.

[0374] The measurement curves illustrated in Fig. 29B show the applied counter pressure. An initial pressure of 120 kPa was used and the pressure was continuously decreased to 0 kPa over a period of about 10 seconds. Measurement curve 293 shows the pressure in the inlet reservoir 201 (Fig. 19), measurement curve 294 shows the processor in the outlet reservoir 205 (Fig. 19), measurement curve 295 shows the pressure at the pump inlet, and measurement curve 296 shows the pressure at the pump outlet.

[0375] Figure 29C then shows an enlarged section of Figure 29A, where the pressure curve of the suction stroke is illustrated. Several measurement curves recorded in the course of a controlled pressure drop (Figure 39B) are now superimposed again. Color coding was used again, i.e. the curves with yellow hues were recorded at the beginning (pressure: 120 kPa) and the curves with blue and purple hues were recorded towards the end (pressure: 0 kPa).

[0376] It can be seen that at an applied starting pressure of 120 kPa, the measurement curve roughly corresponds to an exponential decay process in normal conditions (Figure 24), where there is high back pressure, the inlet and outlet valves are closed and there is no fluid resonance.

[0377] As the pressure drop increases, as illustrated by arrow 299, the measurement curve moves upwards and forms an increasingly pronounced overshoot 300. At such moderate back pressures, the valve begins to open and fluid resonance becomes noticeable, which can be seen in the form of an ever-increasing overshoot 300.

[0378] The overshoot 300 is most clearly visible at the end of the measurement cycle at normal pressure, i.e. at 0 kPa. Thus, if there is no longer any back pressure, the valve can open unhindered and a fluid resonance occurs, which is then expressed by the significant overshoot 300 mentioned above.

[0379] Figure 29D illustrates the measurement curve in the pressure stroke, where the overshoot 300 becomes more and more evident with the decrease in pressure, illustrated by arrow 299. The reasons for this are the same as those mentioned above in the suction stroke.

[0380] As can be clearly seen from Figures 29A-29B, different back pressures in the liquid-transporting microfluidic component lead to significant changes in the time-signal curve of the control signal, which are external influence factors that can be detected according to the inventive concept described herein. Each external influence factor, like the back pressure exemplarily described herein, leads to a characteristic deviation in the time-signal curve, i.e. each external influence factor leaves an individual trace in the time course of the control signal, which can be detected using the inventive concept described herein.

[0381] Measurement of the pressure changes in both the inlet and also the outlet conduits can be detected by suitable means for signal evaluation. This in turn has great practical relevance. For example, a pressure sensor is installed in the drug administration system in the outlet path to detect catheter occlusion. This additional component can be replaced by the inventive "self-sensor system" of the microfluidic component 1000 described herein.

[0382] Furthermore, slow pressure changes can be detected as well. The closure of the patient's access may be gradual, causing a slow pressure rise in the outlet line. In the future, pressure sensors should be utilized to better determine the wearing time of the wearable pump system. Patch pumps, which cannot predict this, are often used only for rather short periods of time to avoid this risk. Prediction would allow longer wearing times, with significant economic advantages. The prediction according to the invention of a slow pressure rise by the inventive "self-sensor system" of the micropump can be performed, for example, by storing the measured signal for larger intervals and comparing it for changes, such as in FIG. 29C.

[0383] Figures 30A to 30E show a further possible case that can be detected using the inventive concept described herein - generated pressure. For example, the same can happen when the patient pushes on the medication bag so that the pressure at the pump inlet becomes higher. This can also have serious consequences for the patient in terms of the dosage of the drug. Usually, this process is avoided with a safety valve, but this can also fail in case of a malfunction.

[0384] Firstly, FIG. 30A again shows an overview of the effect of different pressures on the time signal curves of the suction stroke 301 and the pressure stroke 302. The measurements were performed on an air-delivery micropump driven with a sinusoidal control signal. The pump frequency was 10 Hz. The pressure was varied from a starting pressure of a few kPa and was continuously controlled up to a final pressure of 0 kPa.

[0385] It can be seen that there are large oscillations in the suction stroke 301. In this case the colour coding from yellow to green to blue to purple was used again. It can thus be seen that the measurement curve has a higher amplitude for higher pressures.

[0386] However, there is little oscillation in the pressure stroke 302. This will be explained in more detail below. First, however, refer to Figure 30B, which illustrates a close-up of the temporal signal progression in the suction stroke.

[0387] In Fig. 30B it can be seen that at the beginning of the measurement the measurement curve recorded at high preliminary pressure has a large amplitude (yellow). As the pressure is lowered the amplitude decreases accordingly (blue, green). The amplitude decreases more and more (purple) until the temporal signal curve approximates the normal state signal curve.

[0388] Figure 30C shows a section of Figure 30B. At this point, the suction stroke is long over. However, very clearly distinguishable oscillations are clearly visible in the temporal signal curve. Here, reference number 303 indicates a large amplitude (yellow) recorded at high pressure at the beginning of the measurement. Reference number 304 indicates a small signal amplitude (purple) recorded at the end of the measurement at a barely noticeable pressure.

[0389] Since damping is lower in the air pump than in the water delivery pump, there is a natural resonant valve vibration that causes pressure changes that can then appear in the time signal curve. The valve vibration is excited by the pressure since the inlet valve is open (free flow). Thus, in Figure 30C, it can be seen that the measurement curve has a larger amplitude at higher pressures (yellow) than at lower pressures (purple).

[0390] Figure 30D shows the transition from the suction stroke to the pressure stroke. It shows the last milliseconds of the suction stroke and the first 3 milliseconds of the pressure stroke. This clearly shows the different behavior between the suction stroke and the pressure stroke, i.e. there is significantly less oscillation in the pressure stroke than in the suction stroke.

[0391] FIG. 30E is an enlarged cross-sectional view of FIG. 30D, showing the exponential decay process. During the pressure stroke, an overpressure is generated in the pump chamber, so the inlet valve is closed and cannot oscillate despite the presence of pressure. Therefore, no significant oscillations over the entire duration are visible here. However, during the pressure stroke, there is also a change in the valve oscillations that occurs in the first few milliseconds. Again, a color scheme is used that progresses from yellow through green to purple. It can also be seen here that the amplitude of the signal curve is larger at higher pressures and smaller as the pressure decreases, more and more approximating the exponential decay behavior during the normal state of the pump stroke.

[0392] Thus, as can be seen in Figures 30A to 30E, different pressures cause significant changes in the time signal curve of the control signal of the microfluidic component, which are external influence factors that can be detected according to the inventive concepts described herein. Each external influence detector, like the pressures exemplarily described here, brings about a characteristic deviation in the time signal curve, i.e. each external influence factor leaves an individual signature in the time course of the control signal, which can be detected using the inventive concepts described herein.

[0393] 31A and 31B show Lissajous curves (current vs. voltage) for different control voltages between U=20V and U=120V, based on which it is possible to detect the contact of the membrane element 101 with the counter electrode in the suction stroke or with the bottom of the pump chamber in the pressure stroke. For example, this can be performed for quality measurement in the context of end-of-line testing. Alternatively, this can also be performed during operation, for example over months or years, to determine the degradation of the membrane actuator 100.

[0394] Figures 31A and 31B show the sensor current for two different membrane actuators. Figure 31A shows an actuator where the drive element is free to vibrate. Figure 31B shows an actuator where the drive element contacts the bottom of the pump chamber. Through this contact and the associated effect of the force on the drive element, the sensor current increases.

[0395] The sensor current is plotted not against time but against the voltage applied to the drive element. The time progression of the signal proceeds counterclockwise. The negative amplitude of the suction stroke is the same for all measurements and is -40V. The positive amplitude increases in 5V steps from 20V to 120V.

[0396] Fig. 31A shows the case where the membrane element 101 is not abutted and can oscillate freely during the suction stroke (positive voltage) and pressure stroke (negative voltage). Fig. 31B shows the case where the membrane element 101 contacts the pump chamber bottom when deflected during the suction stroke and is released from the pump chamber bottom during the pressure stroke. During the suction stroke, a clearly distinguishable bending (see arrow 321) is seen, which increases with increasing driving voltage. During the pressure stroke, a bending is also seen in the curved course of the Lissajous curve (see arrow 322), which highlights the release of the membrane element 101, but said bending is smaller.

[0397] 6.4 Further non-limiting examples of detectable external influences The above description with reference to Figures 20 to 31B shows that the respective influence factors can be identified based on the time signal curve of the control signal influenced by the external influence factors. In the next section, based on concrete examples, various ways in which the influence quantities can act on and influence the sensor current are shown.

[0398] As the above description with reference to Figures 20 to 31B shows, each external influencer has a characteristic signature in the time signal curve of the control signal. The above description refers to the signal portion or current term I pThis was done using the example of the pressure acting on the membrane element 101 in time, linked to the piezoelectric coupling coefficient

[0399]

number

[0400] The signal portion or current term I is proportional to p In addition to the above non-limiting examples of detectable or identifiable external influencers, the signal portion or current term I p Based on this, a set of further external influencing factors can be identified.

[0401] Correspondingly, according to an embodiment of the present invention, the signal processing device 105 processes the signal portion or current term I p Based on the current term I, the signal processing device 105 is configured to check the fluid state of a microfluidic component, such as a micropump, for disturbance-free operation. Disturbances can occur in a variety of ways. Thus, the signal processing device 105 calculates a single part or current term I p The method may be configured to determine at least one of the following external influence factors based on: ○ Changes in back pressure (e.g., when the back pressure changes, the sensor current changes significantly in the time domain until the catheter is closed as described above), ○ Change in reserve pressure, ○ Closure of exit ducts such as catheters, the presence of air bubbles, for example in the pump chamber of the micropump (air bubbles can generate an occlusion pressure in the microvalve due to capillary forces, said occlusion pressure being discernible in the sensor current), ○ The size of the bubbles passing through the pump, ○ the arrival of air bubbles in the pump chamber via the inlet line, Air bubbles leaving the pump chamber via the outlet line, ○ Change in the resistance of the pump chamber, ○ (abrupt) changes in stroke volume due to trapped particles, Changes in ambient parameters, such as pressure changes above the membrane element, pressure changes at the inlet of the valve, pressure changes at the outlet of the valve, changes in temperature, etc. ○ Valve status detection, i.e. whether the valve is open or closed, ○ Valve defects, such as broken valves, ○ (rapid) degradation of the valve sealing properties due to particles (for example, if a hard particle blocks a check valve (e.g. the inlet valve), this can be detected using the sensor current in the pressure stroke since the discharge of the stroke volume no longer takes place only via the outlet valve), ○ Settling or aggregation of solids, e.g. sedimentation, (slow) deterioration of the valve sealing properties due to protein denaturation (if the valve gets stuck due to protein denaturation or oil precipitation, this can be detected by the sensor current, as the actuator element 102 presses against the closed valve with a corresponding stroke), o The occurrence of capillary sticking when the meniscus blocks the valve, the occurrence of van der Waals sticking when corresponding molecules are deposited between the valve seat and the supporting web, ○ Blistering or changes in the elastic properties of the sealing element, ○ Change in the adhesive bond between the actuator element and the membrane element, ○ The inertia of the surrounding liquid is coupled to the micropump ("fluid resonance"), so that the micropump (in the case of incompressible liquids as pumping medium) can recognize whether the connected fluid line is long, short, soft or hard. o Change in the viscosity of the medium, which makes it possible, for example, to check whether the rinsing process has rinsed out the entire drug and replaced it with saline.

[0402] Other signal parts or current terms I C , I ce , I U also affect the time course of the control signal, i.e., these current terms I C , I ce , I Ualso affect the control signal depending on which external influence factor is currently dominant. Therefore, according to the invention, these signal parts or current terms I C , I ce , I U can be used to identify several external influencing factors based on a signal analysis of the time course of the corresponding affected control signals.

[0403] Therefore, according to an embodiment of the present invention, the signal processing device 105 is configured to verify that the piezoelectric hysteresis process remains unchanged. In some cases, one or more parameters τ c , I ce or I c changes, providing an indication of the mechanical and / or electrical changes in the piezoelectric ceramic. Thus, the signal processing device 105 processes the signal portion or current term I C and I ce The method may be configured to determine at least one of the following external influence factors based on at least one of the following: mechanical fatigue of the piezoelectric element, such as breakage or the growth of so-called subcritical cracks during very long continuous operation, o Electrical fatigue of the piezoelectric element, such as loss of polarization.

[0404] Therefore, according to an embodiment of the present invention, the signal processing device 105 is configured to verify that the charging process is performed properly. Thus, the signal processing device 105 calculates the signal portion or current term I U The method may be configured to determine at least one of the following external influence factors based on: Breaking of electrical contacts (e.g. broken bond wires), Shorting of the piezoelectric element (e.g. due to condensation of water droplets in high humidity), Damage to the piezoelectric element (e.g. when only parts of the piezoelectric ceramic are in electrical contact, the capacitance and therefore the charging current is reduced).

[0405] When the measurement and evaluation of the present invention is performed in real time, it allows multiple verifications of the state of a microfluidic component 1000, such as a micropump or a microvalve. Obviously, in an electrostatically actuated microfluidic component 1000, or a membrane actuator 100, the signal portion or current term I U and I C can be used for the same purpose. Furthermore, the above-mentioned exemplary amplitude I p , I U , I C , I ce Alternatively or additionally, the respective damping constant or time constant τ a , τ h , τ P , τ d = τ c = τ piezo may be used to detect, i.e. identify and / or classify, the above-mentioned conditions or external influencing factors based thereon.

[0406] These fluidic verifications are very convenient. If there is a significant disturbance, an alarm will be issued and administration may be terminated. In case of a temporary disturbance (e.g. an air bubble), the system controller may, after detection, estimate whether and to what extent the air bubble affects the dose, which must be readjusted. This disturbance may be associated with the fluid specifications of the respective dispensing task of the microfluidic component 1000 (eg a micropump) and an appropriate reaction may follow. In a fluidic network with several microfluidic components 1000, e.g. micropumps (e.g. n micropumps), each micropump may report its status. Failure of an individual micropump may be compensated by other micropumps if there is adequate redundancy.

[0407] The effect of this time-dependent mechanical, pneumatic or hydraulic external influence or disturbance on the membrane actuator 100 affects the sensor current in a "fingerprint" manner.

[0408] For example, fluid resonances of the oscillating liquid columns in the supply and exhaust lines lead to pressure deviations, which can then be influenced by air bubbles transported in the inlet tubes towards the microfluidic component 1000. These resonances are also influenced by the elasticity, diameter and length of the tubing lines used. Furthermore, the presence of air bubbles in the microfluidic component 1000, such as in the pump chamber of a micropump, will cause the sensor current to change more quickly, since air has a viscosity 50 times smaller than water. When an air bubble passes through the microfluidic component 1000, the sensor current can be used to observe the transport in the outlet tube in a similar way. If an air bubble remains partially in the pump chamber, it can be detected as well.

[0409] Alternatively or additionally, using the inventive concepts described herein, the following temporal effects may be detected, i.e., identified and / or classified, as external influencers:

[0410] It is thus possible, for example, to determine the adhesive strength and even the service life of the adhesive layer. The mechanical bias of the membrane actuator 100 influences the sensor current. The position of the actuator is very important not only immediately after assembly, but throughout the entire life cycle of the actuator 100. A loss of bias, i.e. a decrease in the mechanical stress in the assembly (i.e. the assembly of the actuator element 102 and the membrane element 101) and the resulting lowering of the membrane element 101, changes the properties of the microfluidic component 1000 and should therefore be detected during operation. This loss of bias can occur due to failure of the adhesive connection (when polymer chains break due to very high stresses in the adhesive layer over time and the bias decreases if the adhesive is not cured sufficiently). Creep of the adhesive reduces the tension and therefore affects the capacitance of the sensor current.

[0411] In addition to checking the bias, the complete wetting of the interface can also be verified. Trapped air bubbles change the voltage state of the ceramic (i.e. the piezoceramic actuator element 102) and are therefore visible in the sensor current. Sudden delamination or creep delamination of the assembly during operation will cause a change in the sensor current and can therefore be detected early.

[0412] Not only the adhesive layer, but also the ceramic (i.e. the piezoceramic actuator element 102) itself undergoes changes over time. Among other things, with prolonged operation or extreme control voltages, depolarization of the piezoceramic 102 may occur. This leads to loss of stroke and, in extreme cases, failure of the microfluidic component 1000. Early detection allows adaptation of the voltage signal, for example to perform a repolarization or to avoid further depolarization. Such repolarization can be smoothed by extremely short voltage surges that do not cause mechanical deflections and therefore do not affect the transport behavior.

[0413] In addition, the insulation of the actuator element 102 with respect to the membrane element 101 is advantageous for some applications. For example, the use of multi-layer actuators would otherwise cause a short circuit of the membrane element 101. However, in many medical applications it is necessary to also isolate the actuator element 101 from the fluid path. This isolation is then necessarily required for operation and must not change its insulating properties throughout the useful life of the microfluidic component 1000. A slow increase of the leakage current is seen in the sensor current and can therefore be detected early.

[0414] These monitoring possibilities may be applied to microvalves as well.

[0415] 7. Methods for identifying and / or classifying external influencers Due to the inventive concepts described herein, the following compensation processes can be measured and differentiated in the microfluidic component 1000: Charging the capacitance of the piezoelectric membrane actuator 102; Loss processes in piezoelectric ceramics exemplified by the expansion (contraction) of magnetic domains (Weiss domains), and · Changes in pump chamber pressure in the pump chamber.

[0416] this is, In real time, Without having to affect the control signal Without additional sensor elements On every suction stroke, For each pressure stroke, Only by measuring the time dependence of the charging current This can be done.

[0417] In order to put the inventive concept into practice, it is advantageous to provide one or more of the following components: A measurement circuit (as part of the signal processing device 105) for accurate time-resolved measurement of the charge current; A real-time data acquisition means for the time-dependent charge current (as part of the signal processing device 105); means for evaluating the recorded data in real time (as part of the signal processing device 105), for example to determine superimposed amplitudes and / or time constants that can be used to achieve data reduction; and · A means for properly storing the data generated by this evaluation.

[0418] There are in principle two different ways to implement the inventive concept, which, among other things, differ in their concepts for data storage and signal evaluation.

[0419] 7.1 White-box model In so-called white-box models, the connection between the measurement signal (the affected control signal 104 with one or more associated signal parts or current terms) and the causative physical disturbances (external influencers) is known and can be extracted unambiguously from the affected control signal with the respective signal part or current term. Upon evaluation of the data, one already has the desired description of the system, i.e. the external influencers affecting the time-signal curve can be detected and identified.

[0420] In the context of the present disclosure, the phrase "identify" may be understood as indicating that the type of external influencer or its physical context may be clearly indicated, for example, that an air bubble is located in the pump chamber, a valve is clotting, a valve is blocked, there is a certain back pressure, there is a certain pressure, etc.

[0421] Regarding such a white-box model, according to an embodiment of the present invention, the signal processing device 105 comprises a storage device which stores how certain external influence factors affect the time signal curve of the control signal 104. In this case, the signal processing device 105 is configured to identify the respective causal external influence factor based on the time signal curve of the affected control signal 104 and based on the influence factor information stored in the storage device.

[0422] 7.2 Black Box Model In so-called black box models, the relationship between the measurement signal (the affected control signal 104 with one or more associated signal parts or current terms) and the physical disturbances (external influencers) is initially unknown and cannot be extracted directly from the affected control signal 104 at first. In such cases, machine learning methods may be used and neural networks may be utilized. In this case, the membrane actuator 100 may be trained by generating external influencers (e.g. disturbances) and capturing the system response in a statistically significant way. This training results in weighted connections of the neural network. After this training, the membrane actuator 100 only needs these weights to capture each causal external influencer with a relatively high probability and classify it accordingly.

[0423] In the context of the present disclosure, the term "classifying" may be understood such that a weight, or probability, is determined based on which an explanation may be given as to with what probability an external influencer determined based on the time signal curve of the affected control signal may be assigned to a particular class (e.g., valve is blocked, valve is clotting, air bubbles, pressure, back pressure, etc.) For example, such classification may be performed using the aforementioned neural network.

[0424] Regarding such a black-box model, according to an embodiment of the present invention, the signal processing device 105 includes a neural network that is pre-trained by generating different external influencers and determining their respective effects on the time course of the control signal 104. The neural network is configured to classify at least one causal external influencer based on the time signal curve of the affected control signal 104, based on the pre-created training data.

[0425] As described in detail herein, the membrane actuator 100 is fluidly coupled to a system whose interrelationships can be very complex, since on the one hand they overlap and on the other hand they can affect each other. According to the present invention, the sensor current describes an integral over all influencing variables, which also presupposes the use of "machine learning" and "neural networks". Another indication to this is that it is clear that these influences can be detected purely through the observation of the measurement curve of the sensor current, even if no direct physical relationship is evident.

[0426] 7.3 Grey-box model Between the above black-box and white-box models there is a different step change in which the relationship between the measurement signal (the affected control signal 104 including one or more associated signal parts or current terms) and the physical disturbances (external influence factors) can only be approximated by semi-empirical models. This intermediate position can be accordingly called a grey-box model. In this case, training can be significantly simplified and realized with efficient optimization algorithms. Here, disturbance detection is performed with higher reliability than in the case of pure neural networks. 1) The calculations in sections 4 and 5, in which the "step response" of the piezoelectric-driven microfluidic membrane actuator 100 to abrupt voltage changes was analyzed, may be considered as a physical modeling, but it is subject to some simplifications (e.g., not considering the large signal behavior, hysteresis or piezoelectric creep, the phenomenon of fluid resonance). Therefore, this physical modeling cannot accurately represent the reality. However, this model describes some principle features and dependencies of the sensor current on the processes of the membrane actuator 100. Therefore, this concept may be considered as a gray-box description. 2) Another analog example to the grey-box description is the response of the sensor current with respect to harmonic voltage control, as described in Section 4.3.

[0427] By using the corresponding physical model, the relevant signal part or parts or current terms I U , I C , I ce , I p The time signal curve of the affected control signal 104, including, may be determined by fitting a model to the measurements (see Section 5).

[0428] With regard to such a grey-box model, according to an embodiment of the present invention, the signal processing device 105 calculates that a mathematical model with associated amplitude curves and / or associated time constants of each signal portion is generated for each individual signal portion I U , I C , I ce , I p In this case, the signal processing device may be configured to fit a mathematical model to the temporal signal curve of the affected control signal 104 and to identify the respective causal external influencer with the help of the fitted mathematical model.

[0429] Optionally, a neural network may be used to optimize the fitting, in which case the external influencers may be classified based on a fitting performed with a pre-trained neural network.

[0430] 7.4 Possible Realization and System Integration 32 is a schematic block circuit diagram of a possible implementation in hardware and software of the inventive concepts described herein. Firstly, it shows a self-contained system representing the microfluidic component 1000 of the present invention.

[0431] The microfluidic component 1000 comprises a membrane actuator 100, which may be piezoelectrically or electrostatically driven. Furthermore, the microfluidic component 1000 comprises a signal generating device 103 configured to control the membrane actuator 100 by means of a control signal 104. An external influencer 116 may act on the membrane actuator 100, which may have a characteristic effect on the time course of the control signal 104 in the sense of an individual trace.

[0432] The signal processing device 105 measures the control signal 104, which may include an affected time signal curve after the presence of the external influencer 116. The signal processing device 105 is further configured to determine the influence of the control signal 104 on the time signal curve and based thereon identify and / or classify the at least one external influencer 116 responsible therefor.

[0433] As explained in sections 4 and 5, the affected control signal 104 may be divided into several (up to four) different signal parts I, also referred to as current terms 117 in the context of the present disclosure, depending on the embodiment of the actuator element 102. U , I C , I ce , I p The signal processing device 105 may include the individual current terms I U , I C , I ce , Ip Based on this, external influencers 116 may be identified and / or classified.

[0434] The membrane actuator 100 may be controlled by a microcontroller. Optionally, the microcontroller may be configured with a signal generating device 103, as exemplarily shown here in Fig. 32. The charge and sensor current, i.e. the control signal 104, may be pre-processed by a signal processing device 105 (current term measurement circuit) and sampled with the microcontroller. Through this, the time signal curve of the control signal 104 may be discretized.

[0435] The current term data 117' thus discretized can be transferred to a high performance PC 118 via a serial interface. If a neural network is used to classify the external influencers 116, the training of the neural network can be performed on the PC 118. For example, the neural network can be explicitly optimized for a microcontroller and transferred to the microcontroller after training. Optionally, system states can additionally be detected using an external sensor system 119 and also transferred to the PC 118.

[0436] Thus, the microfluidic component 1000 of the present invention can detect its system conditions and react accordingly. For example, the control signal 104 can be adapted or changed (see edge 121) to allow tracking or post-control of the operating point.

[0437] Optionally, a user interface 122 is provided, by means of which the operating point can be manually controlled. Alternatively or additionally, a user can obtain information, such as information on identified and / or classified external influencers, via the user interface 122. If the signal processing device 105 determines a deviation of the operating point from a normal state, for example due to a detected external influencer 116, the detected deviation from the normal state can be indicated to the user by means of the user interface 122 via an optical and / or acoustic signal. For this purpose, for example, an alarm can be triggered.

[0438] FIG. 33 shows a schematic block diagram of the individual possible components of the signal processing device 105, or microcontroller, when a neural network is used.

[0439] Firstly, Fig. 33 shows purely schematic representation of an analog measurement circuit comprising an operational amplifier (OPA) 109, a measurement resistor 110 and Zener diodes 111, 112 as previously described with reference to Fig. 3 and Fig. 4. Also as described therein, the ground electrodes of the microfluidic component 1000 and the ground electrodes of the analog circuitry may be interconnected. This creates a virtual ground potential so that the analog circuitry does not load the signal processing device 103 and therefore the control signal 104 is not distorted.

[0440] The control signal 104 amplified with the OPA 109 can be transferred to a data processing module 123. Here, the data can be discretized, for example with an ADC. Optionally, the signal generator signal (T=1 / f) can be input to the data pre-processing module 123 to determine the duration considered (1 pump cycle: 1 suction stroke, 1 pressure stroke). Optionally, an adjustment of the sampling rate can also be possible. For this purpose, there can be a connection with a microcontroller.

[0441] The input vector 124 containing the discretized raw data 123 may then be transferred to an AI module 125 (AI: Artificial Intelligence). The AI ​​module 125 may include a neural network. The output vector 126 of the AI ​​module 125 may then be transferred to an analog-to-digital converter (ADC) 127. The analog measurement value of an analog measurement circuit may also be transferred to the ADC 127, which is highlighted with an edge 128. The signal digitized by the ADC 127 is then transferred to a defined interface 129 (UART, USB, I 2 C, SPI, ...).

[0442] Thus, possible control of the microfluidic component 1000 could be envisaged using a corresponding microcontroller equipped with an ADC (analog-to-digital converter), a serial interface coupled to a signal processing device 105 (measurement circuitry), and a driver (e.g. a pump driver) equipped with means for generating a high voltage.

[0443] A possible concept for data evaluation could include one or more of the following components: · Classifying stages using time series classification (e.g. anomaly detection with autoencoders). Reaction / control regarding error conditions: Adapting / adjusting the charging current to the current conditions (e.g. by adjusting the PWM signal of the micropump driver) to keep the flow rate constant. Direct integration of algorithms into the pump controller (ASIC) Development of a "disturbance variable IC" (small size, energy efficient), e.g. using FPGA, that evaluates the data from the sensor current electronics and provides the necessary information / closed loop control to the pump driver. If the microfluidic component 1000 includes a pump and a valve, the pump may also respond to an error condition in the valve.

[0444] The inventive concepts described herein make it possible to provide a self-efficiently operating microfluidic component 1000 that includes a membrane actuator 100 (e.g., piezoelectrically or electrostatically driven) that detects system conditions on its own (e.g., using neural networks) and reacts thereto in a problem-solving manner or notifies the user accordingly.

[0445] 8. Supplementary explanation regarding the above explanation The above description assumes prior knowledge or basic theoretical understanding of pump chamber pressure, time behavior after voltage change, and boundary conditions of use of a micropump as an example of the inventive microfluidic component 1000. For completeness, the underlying knowledge is explained below.

[0446] 8.1 Pump chamber pressure The pressure of the fluid beneath the membrane element 101 is a scalar quantity, which in principle can adopt different values ​​at different points in the pump chamber. The pressure can therefore also be different at different points and at different times.

[0447] Secondly, if the height of the pump chamber is so large that there is no significant pressure drop even with flow in the pump chamber, this can be referred to as a homogenous pump chamber pressure p (as early micropumps from the 1990s were constructed). In this case, the pump chamber pressure p has the same size anywhere in the pump chamber at any given time. The theoretical considerations of this disclosure assume such a case.

[0448] However, in modern micropumps, the pump chamber is constructed to be very flat in order to reduce the dead volume and therefore increase the compression ratio. The pressure drop in the pump chamber can no longer be neglected; in many cases it is even higher than the pressure drop in the valve. Therefore, the assumption that the pressure in the pump chamber is homogeneous is not a good approximation, which is why theoretical considerations only represent a reality with some uncertainties (grey box model).

[0449] In reality, when performing suction and pressure strokes, the faster the fluid flows in or out at a given time, the greater the pressure drop. Nevertheless, there is a force summation effect on the membrane element 101, a time-dependent charge transfer dQ / dt=I p This charge transfer correlates with the integral of the configuration-dependent pressure at each time t.

[0450] 8.2 Temporal behavior after voltage change Assuming infinitely fast voltage changes (which do not exist in reality) would have several effects: 1) After applying a voltage infinitely quickly (e.g., within 1 μs if the voltage is generated with a frequency generator and a piezoelectric amplifier that can also provide a corresponding current), an electric field acts within the piezoelectric ceramic (piezoelectric actuator element 102) and the coefficient d 31 which induces a contraction of the piezoelectric ceramic according to the adhesive layer, which applies a bending torque to the membrane element 101 (between the membrane element 101 and the actuator element 102) via the adhesive layer. These mechanical signals propagate at the speed of sound. Assuming a speed of sound in liquid of 1500 m / s and a lateral expansion of the micropump of ~0.015 mm, this signal takes t=0.015 / 1500 =10 μs to propagate through the entire membrane element 101 (hence the faster this is for smaller micropumps). 2) If we assume that an incompressible liquid is placed under the membrane element 101 and that after 10 μs a mechanical force effect appears on the bending actuator 102, a pressure change occurs in the fluid under the membrane element 101 within the same time range (we also assume that the speed of sound is 1500 m / s). Since the fluid does not move in the valve within this short time range, the incompressible liquid generates a reaction force (action = reaction). Since the liquid does not yet flow within this short time range, there is still no pressure drop in the pump chamber and one can even speak of a homogeneous pump chamber pressure. This maximum pressure is the occlusion pressure or "stall pressure". Due to the rather slow time constant of the valve and the flow in the pump chamber, the pressure in the fluid starts to change, i.e. the fluid pressure becomes variable (due to the flow) and decreases with time. If instead of water we use compressible air or partially compressible air (with a volume V gas If there is an air bubble in the pump chamber (through the suction stroke), this air bubble will also be compressed (or expanded in the suction stroke) within a very short time of switching on the voltage according to its equation of state, i.e. before the medium can move through the valve, and the piezoelectric membrane transducer will lose force according to the fluid volume and can no longer generate the occlusion pressure. 3) The piezoelectric membrane interconnect (i.e. membrane element 101 + adhesive + actuator element 102) has a mechanical resonant frequency of several tens of kHz, depending on the design and whether liquid or air is underneath the membrane element 101. The membrane element 101 is excited at its resonant frequency and overshoots (when the voltage change is faster than the resonant frequency), which can be observed in the form of an increased volume, a larger mechanical stress, and a larger transient stroke. These larger strokes can be observed, for example, in gas micropumps when the pump is controlled with a hard rectangular signal. In principle, this behavior can also be detected and studied by the inventive concepts described herein. In practice, however, this phenomenon will not occur, since the increase in voltage levels is usually configured in such a way that these resonant frequencies are not excited. 4) The microvalve in the pump chamber is also a moving element and has a resonant frequency. It is in the order of kHz and can (depending on the design) drop to below 1 kHz when the effective oscillating mass increases, for example due to the presence of liquid. Depending on the implementation of the control electronics, the microvalve can be excited to oscillate in the case of fast voltage changes, which has already been observed experimentally. These oscillations of the valve lead to pressure oscillations in the pump chamber. These pressure oscillations are generated by the sensor current or the current term I p These may have already been demonstrated in experiments with. They occur significantly when there is air on the valve flap (for example, Fig. 21), but they can also be observed in the case of water (Fig. 25). In the case of oscillations of the natural frequency, the surrounding medium must be moved as well, so the effective mass of the oscillating system increases and therefore the observed natural frequency is higher in the oscillations in water (Fig. 25) than in the oscillations in air (Fig. 21). Due to the lower viscosity, the damping of the vibrations is significantly smaller in air than in water, so the vibration amplitude of the valve is significantly larger in air than in water. It must be emphasized that oscillations never occur in all valve designs. If the damping is large enough (for example, if the height of the pump chamber is very low), the asymptotic boundary case can be reached even in the case of rectangular excitation in air, and there are no oscillations.

[0451] 8.3 Monitoring microfluidic components during operation based on sensor current There are many cases where the boundary conditions of use of the microfluidic component 1000 are known and typically do not change. One example of such a case is drug administration using a micropump, where a micropump is mounted close to the body with a drug reservoir and delivers a drug to the patient via a catheter. For these cases, the measured current I (i.e., the control signal 104) can be determined in a time-resolved manner for each suction stroke and each pressure stroke, i.e., I(t). Thus, the suction stroke I of the pump suction (t) and the subsequent pressure stroke I pressure The time-current curves for (t) are obtained. These currents I suction (t) or Ipressure (t) is again the signal portion or current term I U , I C , I ce and I P It consists of:

[0452] In steady state, when the micropump is operating without disturbances, the time-current curve of pump cycle n+1 is very similar to the previous pump cycle n. If a disturbance of the pump operation occurs in pump cycle n+1 (e.g., valve breakage, ingress of a large air bubble, tubing dislodgment, tubing bends, obstruction of patient access, valve blockage by particles), at least one current term I U , I C , I ce , I P The time signal curve of the current signal (control signal 104) I for cycle n+1 changes, and therefore the time signal curve of the control signal 104 also changes, i.e. according to the disturbance in the suction stroke, the pressure stroke, or both strokes. suction (t) n+1 current signal I suction (t) n and any previous current signal (e.g., the current signal I suction (t) n-1 ) changes can be detected.

[0453] According to an embodiment of the present invention, the signal processing device 105 may therefore be configured to compare the temporal signal curve of the control signal 104 of an operating cycle (n+1) of the microfluidic component 1000 with the temporal signal curve of the preceding operating cycle (n) and detect deviations between the two signal curves.

[0454] In case of a detected deviation, for example, the signal processing device 105 may be configured to indicate the detected deviation by optical and / or acoustic signals. In many cases, it is sufficient to indicate that a disturbance has occurred in the dosing process. If the microdosing is safety-related (e.g., life-critical drugs), an alarm may be triggered and corresponding measures may be initiated (e.g., modification of the dosing system). However, in non-critical cases, it may indicate that the dosing module is not working properly, and the pump may indicate this in an optical or acoustic way and report it to the control center.

[0455] Preferably, the control signal 104 and the suction stroke I suction (t) and / or pressure stroke I pressue (t) can be captured in real time. Thus, for example, if the micropump pumps at a pump frequency of f=10 Hz, the control signal I suction (t) or I pressue (t) may be captured every 100 ms. It is possible, but not necessary, to store this raw data in its entirety. Comparing the raw data would be very intensive in terms of storage and computational capacity. For example, there are the following possibilities for reducing the data: The time range of interest of the measured current I(t) is known and measurements are performed only in this range with a suitable sampling rate. In the example of a micropump pumping at 10 Hz and controlled by electronics that generate a voltage within 1 ms, the suction stroke and the pressure stroke (which depend on the flow resistance of the valves and pump chambers and the viscosity of the fluid) are performed in a few milliseconds (e.g. Fig. 23). All current terms I pThe above-mentioned potentially occurring disturbances affecting (t) will only cause significant changes within this time window. It therefore makes sense to sample the time range up to about 5 ms after the start of the suction or pressure stroke and then not take any more measurements. The remaining time window from 45 ms to the next stroke can be used for raw data evaluation and storage of appropriately processed data. According to an embodiment of the invention, the signal processing device 105 can be configured to use a first time portion of the actuation cycle (suction stroke or pressure stroke) of the microfluidic component 1000 to determine the temporal signal curve of the control signal 104 and to use the remaining second time portion of the actuation cycle up to the start of the subsequent actuation cycle to store and / or evaluate the determined signal curve. The amount of data can be further reduced by fitting the measurements appropriately to the time-signal curve and storing only the respective fitting parameters. A model for this fit is already available in the case of grey-box models. In the above example, the total current I(t) is divided into the respective exponentially decaying partial currents I U (t), I C (t), I ce (t), I P The amplitudes and time constants of these partial currents were obtained as fitting parameters. According to an embodiment of the present invention, for example, only these fitting parameters may be stored for each suction stroke and each pressure stroke. Then, a large amount of raw data is reduced to a very small number of values. That is, the signal processing device 105 divides the determined signal curve into at least one signal portion I(t) of the control signal I(t). U , I C , I ce , I p and associated time constants in the form of parameters representing the respective time amplitude curves of the respective signals. Therefore, insofar as the amount of data has been reduced as explained above, the correspondingly reduced values ​​can then be very easily compared. Furthermore, the signal processing device 105 can evaluate the trend evolution of these values ​​over many pump cycles and use this information to evaluate the condition of the micropump.

[0456] · In many cases, it is not necessary to measure each individual pump cycle. For a 10Hz pump, this measurement (and comparison with respect to undisturbed operation) can be performed at any longer interval, for example once per second (i.e. every 10th stroke) or once per hour. This saves a lot of storage and computation capacity, and therefore energy, especially in battery-powered applications or long-lasting implants. Correspondingly, according to an embodiment of the invention, the signal processing device 105 can be configured to capture the temporal signal curve of the control signal 104 during two operating cycles (for example the first and tenth pump strokes), with a number of further operating cycles between which the signal processing device 105 does not capture data. Additionally, it could in principle be envisaged to detect signs of fatigue and degradation. In such cases, the comparison of parameters could be performed by comparing parameters stored over long periods of time (weeks, months or even years) (e.g. in long-lasting implants or in high-quality long-life industrial applications such as the lubrication of fast-rotating bearings or micro-cooling of servers). ○ For example, if there are micro-cracks in the piezo ceramic during very long operation times: The stiffness of the membrane may decrease, reducing the stroke of the micropump (this is due to the current term I p ), Piezoelectric properties (d 31 and capacitance) may change, with the coefficient d 31 teeth

[0457]

number

[0458] Since it is directly proportional to the current term I ce and I C Affects. Or, the adhesive strength between the piezoelectric ceramic (actuator element 102) and the membrane element 101 is reduced. There is a strong shear stress in this adhesive layer, and therefore, if the polymer chains of the adhesive are torn over time, this will reduce the bias of the piezoelectric membrane transducer, lowering the compression ratio and thus the current term I p may have an impact on

[0459] In summary therefore, the inventive concepts described herein provide a membrane actuator 100 for a micropump or microvalve, the membrane actuator 100 being in contact on one side with a fluid (gas or liquid) to which a time-varying electrical control signal 104 U(t) is applied, and U(t) or a quantity I(t) derived therefrom is determined in an electrically accurate manner by a measurement circuit 105 during operation of the membrane actuator, and time-varying influences on the electrical control signal 104 I(t) are captured through hydraulic, pneumatic, piezoelectric or mechanical processes acting on the membrane actuator 100, it being noted that these different processes can be detected and distinguished by evaluation of the control signal 104 I(t).

[0460] The embodiments described above merely provide an illustration of the principles of the innovative concepts described herein. It is clear that modifications and variations of the arrangements and details described herein will be obvious to those skilled in the art. Therefore, it is intended that the concepts described herein be limited only by the scope of protection of the following claims, and not by the specific details presented based on the description and embodiment description herein.

[0461] Some or all of the method steps may be performed by (or using) hardware devices such as a microprocessor, a programmable computer, or electronic circuitry. In some embodiments, some or more of the most important method steps may be performed by such devices.

[0462] Depending on the specific implementation requirements, the embodiments of the present invention can be implemented in hardware or software. The implementation may be performed using a digital storage medium, such as a floppy disk, a DVD, a Blu-ray disk, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, a hard disk, or any other magnetic or optical memory having electronically readable control signals stored thereon that cooperate or can cooperate with a programmable computer system such that the respective method is executed. This is why the digital storage medium can be computer readable.

[0463] Thus, some embodiments according to the present invention include a data carrier comprising electronically readable control signals capable of cooperating with a programmable computer system to cause any of the methods described herein to be performed.

[0464] Generally, embodiments of the present invention may be implemented as a computer program product having program code which is effective to perform any of the methods when the computer program product is run on a computer.

[0465] The program code may for example also be stored on a machine readable carrier.

[0466] Further embodiments comprise the computer program for performing any of the methods described herein, said computer program being stored on a machine readable carrier. In other words, an embodiment of the inventive method is therefore a computer program having a program code for performing any of the methods described herein, when the computer program runs on a computer.

[0467] A further embodiment of the inventive method is therefore a data carrier (or digital storage medium or computer readable medium) on which is recorded a computer program for performing any of the methods described herein. The data carrier, digital storage medium, or computer readable medium is typically tangible or non-volatile.

[0468] A further embodiment of the inventive method is therefore a data stream or a sequence of signals representing the computer program for performing any of the methods described herein. The data stream or the sequence of signals may for example be adapted to be transmitted via a data communication link, for example the Internet.

[0469] A further embodiment comprises a processing unit, for example a computer, or a programmable logic device, configured or adapted to perform any of the methods described herein.

[0470] A further embodiment comprises a computer having installed thereon the computer program for performing any of the methods described herein.

[0471] A further embodiment according to the invention comprises a device or system configured to transmit a computer program for performing at least one of the methods described herein to a receiver. The transmission may be, for example, electronic or optical. The receiver may be, for example, a computer, a mobile device, a memory device or a similar device. The device or system may, for example, comprise a file server for transmitting the computer program to the receiver.

[0472] In some embodiments, a programmable logic device (e.g., a field programmable gate array, FPGA) may be used to perform some or all of the functions of the methods described herein. In some embodiments, a field programmable gate array may cooperate with a microprocessor to perform any of the methods described herein. In general, the methods are performed in some embodiments by any hardware device. The hardware device may be any general-purpose hardware, such as a computer processor (CPU), or method-specific hardware, such as an ASIC. [Explanation of symbols]

[0473] 100 Membrane Actuator 101 Membrane elements 102 Actuator Element 103 Signal Generating Devices 104 Electrical Control Signal 105 Signal Processing Devices 106 Mechanical Components 107 Carrier Board 108 Combined charge flow 109 Voltage Controlled Op-Amp 110 Measuring resistor 111, 112 Zener diode 113, 114 Low leakage diodes 116 External influence factors 117 Current term 117' Current data 118 PC 119 External Sensor System 121 Edge 122 User Interface 123 Discretized raw data 124 input vectors 125 AI Module 126 Output Vectors 127 Analog-to-Digital Converter (ADC) 128 Edge 129 Predefined Interfaces 171 measurements 171 1 , 172 1 , 173 1 curve 171 2 , 172 2 , 173 2 curve 181 Circular Lissajous Curve 182 Lissajous curve 201 Inlet reservoir 202 Air Bubble Detector 203 Pressure Sensor 204 Pipeline 205 Outlet reservoir 206 Pressure Sensor 207 Pressure Controller 208 Oscilloscope 291 Suction Stroke 292 Pressure Stroke 293 Measurement curve 294 Measurement curve 295 Measurement curve 296 Measurement curve 300 Overshoot 301 Suction stroke 302 Pressure Stroke 1000 Microfluidic Components

Claims

1. A microfluidic component (1000), comprising: A membrane actuator (100) having a membrane element (101) and an actuator element (102) for deflecting said membrane element (101); a signal generating device (103) configured to generate an electrical control signal (104) having a time-varying signal curve for controlling the membrane actuator (101), whereby the actuator element (102) actuates the membrane element (101); a signal processing device (105) configured to determine, during operation of the microfluidic component (1000), an influence on the temporal signal curve of the control signal (104) caused by one or more external influence factors, A microfluidic component (1000) for identifying and / or classifying at least one causal external influencer based on said influence on said temporal signal curve.

2. The signal processing device (105) The microfluidic component (1000) of claim 1, configured to distinguish a first temporal signal curve of the control signal (104) corresponding to operation of the membrane actuator (100) without external influence from a different second temporal signal curve of the control signal (104) corresponding to operation of the membrane actuator (100) with at least one dominant external influence factor.

3. The signal processing device (105) The microfluidic component (1000) of claim 1, configured to identify and / or classify hydraulic, pneumatic or mechanical forces acting on at least one membrane side of the membrane element (101) as external influencers responsible for their effect on the temporal signal curve.

4. The membrane actuator (100) is a piezoelectrically driven membrane actuator in which the actuator element (102) includes at least one piezoelectric element; The control signal (104) causes a deformation of the piezoelectric element by using the inverse piezoelectric effect, whereby the piezoelectric element applies an actuation force on the membrane element (101); The microfluidic component (1000) of claim 1, wherein the time signal curve of the control signal (104) is influenced by a signal originating from the piezoelectric element, which generates due to a reaction force of the membrane element (101) based on a direct piezoelectric effect.

5. the signal processing device (105) is configured to perform a signal analysis of the time signal curve of the control signal (104) in order to determine and / or classify the at least one causal external influencer, The individual signal parts (I U , I C , I ce , I p ) is determined, and the individual signal portions (I U , I C , I ce , I p ) correlates with different external influence factors that have different effects on the time course of the control signal (104); The microfluidic component (1000) of claim 4, wherein the signal processing device (105) is configured to assign specific signal portions to specific external influencers and use this to identify and / or classify each of the external influencers.

6. The signal processing device (105) processes at least four different signal parts (I U , I C , I ce , I p ) from one another, The first signal part (I U) is linked to the time variation of the voltage when charging or discharging the capacitance between the membrane element (101) and the piezoelectric element (102), The second signal part (I p ) is linked to the time variation of the pressure acting on said membrane element (101), The third signal part (I C ) is linked to the time variation of the capacitance caused by large signal effects of the piezoelectric element (102), The fourth signal part (I ce ) is the piezoelectric coefficient d that changes due to large signal effects of the piezoelectric element (102) 31 The microfluidic component (1000) of claim 4, linked to a time change of

7. The individual signal parts (I U , I C , I ce , I p ) aims to achieve equilibrium through an individual time compensation process, in which each signal portion has an individual time amplitude curve (I U0 , I p0 , I c0 , I ce0 ) and the respective time constant (τ a , τ h , τ d = τ c = τ piezo ), The signal processing device (105) processes the individual signal portions (I U , I C , I ce , I p ) from one or more of the respective temporal amplitude curves (I U0 , I p0 , I c0 , I ce0 ) and / or the respective time constants (τ a , τ h , τ d = τ c = τ piezoThe microfluidic component (1000) of claim 5 configured to determine the external influence factor based on a

8. The signal processing device (105) processes the second signal portion (I p ) based on the following external influence factors: ○ Change in back pressure, ○ Change in reserve pressure, ○ Closure of exit ducts such as catheters, ○ the presence of air bubbles in the pump chamber, ○ The size of the air bubbles in the pump, the arrival of air bubbles into said pump chamber via an inlet line, air bubbles leaving the pump chamber via the outlet conduit, a change in the resistance of the pump chamber, ○ Change in stroke volume due to particle entrapment, changes in ambient parameters, such as pressure changes above the membrane element, pressure changes at the inlet of the valve, pressure changes at the outlet of the valve, changes in temperature, Valve state detection, i.e. whether said valve is open or closed, ○ Valve defects, such as broken valves, ○ Degradation of valve sealing properties due to particles, Settling or aggregation of solids, e.g. precipitation, deterioration of valve sealing properties due to denaturation of proteins, o The occurrence of capillary sticking when the meniscus blocks the valve, the occurrence of van der Waals sticking when corresponding molecules are deposited between the valve seat and the supporting web, ○ Blistering or changes in the elastic properties of the sealing element, ○ a change in the adhesive bond between the actuator element and the membrane element; the inertia of the surrounding liquid is coupled to the micropump, so that said micropump (in the case of incompressible liquids as pumping medium) knows whether the connected fluid line is long, short, soft or hard; A change in the viscosity of said medium, for example making it possible to check whether the rinsing process has completely rinsed out the drug and replaced it with saline. The microfluidic component (1000) of claim 5 configured to determine at least one of:

9. The signal processing device (105) processes the third and / or fourth signal parts (I C , I ce ) based on the following external influence factors: mechanical fatigue of said piezoelectric element (102), such as breakage or the growth of so-called subcritical cracks on very long continuous operation; Electrical fatigue of the piezoelectric element (102), such as loss of polarization The microfluidic component (1000) of claim 5 configured to determine at least one of:

10. The signal processing device (105) processes the first signal portion (I U ) based on the following external influence factors: - Electrical contact cracks, - shorting the piezoelectric element (102), Breakage of the piezoelectric element (102), e.g. when only a portion of the piezoelectric ceramic is in electrical contact, reduces the capacitance and therefore the charging current; The microfluidic component (1000) of claim 5 configured to determine at least one of:

11. The membrane actuator (100) is an electrostatically driven membrane actuator in which a membrane element (101) forms a moving electrode and an actuator element (102) forms a counter electrode, the actuator element (102) and the membrane element (101) capacitively cooperating; The control signal (104) induces a flow of charge on the counter electrode, which causes the membrane element (101) to move relative to the counter electrode; 2. The microfluidic component (1000) of claim 1, wherein the time course of the control signal (104) is influenced by a change in capacitance that occurs during movement of the membrane element (101) relative to the counter electrode.

12. the signal processing device (105) is configured to perform a signal analysis of the time signal curve of the control signal (104) in order to determine and / or classify the at least one causal external influencer, The individual signal parts (I U , I C ) is determined, and the individual signal portions (I U , I C ) are correlated with different external influence factors each of which has a different effect on the time course of the control signal (104); The signal processing device (105) processes a specific signal portion (I U , I C 12. The microfluidic component (1000) of claim 11, configured to assign a specific external influence factor (a) to a specific external influence factor and use this to identify and / or classify each said external influence factor.

13. The signal processing device (105) processes at least two different signal parts (I U , I C ) from one another, The first signal part (I U ) is linked to the time variation of the voltage when charging or discharging the capacitance between the membrane element (101) and the actuator element (102), The second signal part (I C 12. The microfluidic component (1000) of claim 11, wherein the capacitance between the membrane element (101) and the actuator element (102) is linked to a temporal change in capacitance due to the movement of the membrane element (101).

14. During the movement of the membrane element (101) towards the actuator element (102), the capacitance between the membrane element (101) and the actuator element (102) increases by the second signal portion (I C ) changes over time, The signal processing device (105) outputs a variable second signal portion (I C 14. The microfluidic component (1000) of claim 13, configured to determine the time-variable position of the membrane element (101) relative to the actuator element (102) based on a time-variable position of the membrane element (101) relative to the actuator element (102).

15. the time-variable position of the membrane element (101) being dependent on a pressure acting on the membrane element (101); The signal processing device (105) processes the second signal portion (I C 15. The microfluidic component (1000) of claim 14, configured to determine the pressure acting on the membrane element (101) based on a pressure difference between the membrane element (101) and identify it as the external influencer.

16. The signal processing device (105) processes the second signal portion (I C ) the time signal curve of The signal processing device (105) processes the second signal portion (I C ) compared with the known time course of the second signal portion (I C 16. The microfluidic component (1000) of claim 15, configured to detect a deviation in the temporal signal curve of the membrane element (101) caused by the pressure and, based on this deviation, determine the pressure acting on the membrane element (101) and identify it as the external influence factor.

17. The signal processing device (105) is adapted to process the individual signal portions (I U , I C , I ce , I p a storage device in which each of said one or more of The microfluidic component (1000) of claim 5, wherein the signal processing device (105) is configured to fit the mathematical model to the time signal curve of the affected control signal (104) and to identify, with the help of the fitted mathematical model, each of the causative external influence factors.

18. the signal processing device (105) comprises a storage device for storing how a particular external influence factor affects the time signal curve of the control signal (104); The microfluidic component (1000) of claim 1, wherein the signal processing device (105) is configured to identify each of the causative external influence factors based on the time signal curve of the affected control signal (104) and based on influence factor information stored in the memory device.

19. the signal processing device (105) includes a neural network that has been pre-trained by generating different external influencers and determining their respective effects on the time course of the control signal (104); The microfluidic component (1000) of claim 1, wherein the neural network is configured to classify at least one causal external influencer based on the time signal curve of the affected control signal (104) based on pre-created training data.

20. The microfluidic component (1000) comprises a microfluidic pump having a pump chamber, at least one membrane side of the membrane element (101) being in contact with a fluid disposed in the pump chamber, and a variable pump chamber pressure is generated in the pump chamber by actuating the membrane actuator (100); The microfluidic component (1000) of claim 1, wherein the signal processing device (105) is configured to determine the variable pump chamber pressure based on the temporal signal curve of the control signal (104) and use this to identify and / or classify the external influencer.

21. said microfluidic component (1000) comprising microfluidic valves allowing opening and / or closing of fluid paths, The microfluidic component (1000) of claim 1, wherein the signal processing device (105) is configured to identify and / or classify the external influence factors and / or determine time-varying operating parameters of the microfluidic valve based on the temporal signal curve of the control signal (104).

22. said signal processing device (105) determining the type of fluid used in said microfluidic component (1000) based on the time signal curve of said control signal (104) and performing a differentiation of the aggregation state of said fluid between gas and liquid; The microfluidic component (1000) of claim 1 configured to identify and / or classify the determined fluid and / or its determined aggregation state as the external influencer.

23. The microfluidic component (1000) of claim 1, wherein the signal processing device (105) is configured to compare the time signal curve of the control signal (104) of an operating cycle (n+1) of the microfluidic component (1000) with the time signal curve of a temporally preceding operating cycle (n) and detect deviations between the signal curves.

24. 24. The microfluidic component (1000) of claim 23, wherein the signal processing device (105) is configured to indicate the detected deviation by means of an optical and / or acoustic signal.

25. The signal processing device (105) performs the following steps during an operating cycle of the microfluidic component (1000): determining the time signal curve of the control signal (104) using a first time portion of the operating cycle; The microfluidic component (1000) of claim 1, configured to store and / or evaluate the determined signal curve using a remaining second time portion of the same operating cycle until the start of a subsequent operating cycle.

26. The signal processing device (105) processes the determined signal curve by dividing the determined signal curve by at least one signal portion (I U , I C , I ce , I p 26. The microfluidic component (1000) of claim 25, configured to store in the form of parameters representative of the time amplitude curves of the respective signals and associated time constants.

27. The microfluidic component (1000) of claim 25, wherein the signal processing device (105) is configured to determine a time trend evolution of the temporal signal curve over a number of operating cycles and to use this trend evolution to determine the external influencing factors and / or to evaluate a state of the microfluidic component (1000).

28. The microfluidic component (1000) of claim 25, wherein the signal processing device (105) is configured to capture the temporal signal curve of the control signal (104) during two operating cycles, and between these two operating cycles there are a number of further operating cycles during which the signal processing device (105) does not capture data.

29. The microfluidic component (1000) of claim 1 is configured to detect the external influence factor purely based on a signal analysis of the time signal curve of the electrical control signal (104) without using any additional sensor systems in the membrane element (101) and / or in the actuator element (102).

30. the signal processing device (105) comprises an operational amplifier (109), the inverting input of which is connected to the signal output of the membrane actuator (100) and the non-inverting input of which is connected to the same ground potential as the signal generating device (103), thereby generating a virtual ground potential at the inverting input of the operational amplifier (109); The signal processing device (105) further comprises a measurement resistor (110) connected between the inverting input and the output of the operational amplifier (109); 2. The microfluidic component (1000) of claim 1, wherein the control signal (104) processed by the signal processing device (105) is dropped across the terminals of this measuring resistor (110).

31. A method for operating a microfluidic component (1000) having a membrane actuator (100) comprising a membrane element (101) and an actuator element (102) for deflecting said membrane element (101), said method comprising: generating an electrical control signal (104) for controlling the membrane actuator (100), whereby the actuator element (102) actuates the membrane element (101), the electrical control signal (104) comprising a time-varying signal curve; determining a time signal curve of said electrical control signal (104) under the influence of one or more external influence factors during operation of said microfluidic component (1000); and identifying and / or classifying at least one external influencer based on the determined affected temporal signal curve.

32. 32. A computer readable digital storage medium having stored thereon a program code, the program code being for carrying out the method of claim 31 when the program is run on a computer.

Citation Information

Patent Citations

  • Fluid feeding pump

    JP2013050034A

  • Liquid discharge device, control method for liquid discharge device, and control program for liquid discharge device

    JP2016182800A

  • Valve

    JP2020159475A

  • Inkjet recording device and control method

    JP2021070280A

  • Liquid ejection device and driving waveform control method

    JP6759730B2

Cited By

  • Measuring device

    JP2026163326A