DE10361099B3 - Device for analyzing material samples arranged on a sample plate comprises a support for the sample plate and a contacting unit for electrically contacting the material samples, and a measuring head inserted in a housing support - Google Patents

Abstract

Device for analyzing material samples arranged on a sample plate comprises a support for the sample plate and a contacting unit for electrically contacting the material samples, and a measuring head (26) inserted in a housing support (27) consisting of two measuring wires which lie on the contact surfaces of the sample plate and connected with a measuring and evaluation unit (18). An independent claim is also included for a process for analyzing material samples. Preferred Features: The measuring wires lie on the contact surfaces of the sample plate via melting spheres. The measuring head is connected to a gas supply unit (16) which is connected to a data processing unit (53) of the measuring and evaluation unit.

Description

State of technology

The The invention relates to a device for analyzing a sample plate according to the Preamble of claim 1 further defined type and a Method for analyzing a sample plate.

Especially in the field of materials science, chemistry and pharmacy it is of considerable importance with regard to the respective application find and develop optimized substances and materials. A special field of application here is the sensor technology, the one key technology with a steadily growing number of applications both in industrial as well as in the private sector. This is how sensors become, for example in technical process monitoring, in the field of environmental protection, in the field of medicine or even used in the automotive sector. A considerable amount of development work is currently becoming faster and more sensitive, especially in development Sensors with a low cross sensitivity plugged.

The Limited developments so far usually on an optimization or modification known materials. However, there is the problem that for certain Areas of application of sensor technology a high demand for new materials exists, with conventional methods that are characterized by a production of individual sensors and a subsequent sequential characterization can not be adequately covered.

Especially in the development of novel, sensitive materials or material combinations it may be appropriate Processes in the field of combinatorial chemistry or so-called Use high-throughput methods. In these methods acts it is parallelized synthesis and screening methods, by which can be opened up new materials or material combinations or already known synthesis methods for existing materials can be optimized in a wide parameter field.

From the US 5,985,356 is a general presentation of high-throughput method known, this document in particular the application of the well-known mainly in the field of pharmacy combinatorial chemistry on chemical and materials science applications is proposed.

A device of the type mentioned in the introduction for analyzing a sample plate is known, for example, from US Pat DE 101 31 581 A1 known. This device comprises a sample plate on which 64 material samples are applied in a matrix-like manner, which are each connected to two electrodes, which in turn are provided with contact points, to which a means for reversible and addressable contacting can be brought into contact.

Advantages of invention

The inventive device for analyzing a sample plate having the features of the preamble of claim 1 and with a usable in a housing carrier Measuring head, for electrical connection with the contacting Two measuring wires per material sample includes, which rest with bias on contact surfaces of the sample plate and connected to a measuring and evaluation unit has the Advantage of easy handling, as when using a standardized Sample plate the contact between the material samples and the measuring and evaluation unit by simply inserting the measuring head in the housing carrier produced is.

The Device according to the invention is particularly suitable for development and retrieval of materials or combinations of materials known as Sensor materials are used and their electrical properties can be characterized. For example, the device according to the invention for the development of an optimized sensor material a gas sensor can be used.

With the device according to the invention it is possible to have a high number of potential sensor rates Rialia, which are arranged on the sample plate, under different test gases at different temperatures, which may be up to 800 ° C, for example, to study almost simultaneously. The investigation can be carried out potentiometrically, resistively, capacitively or via the complex impedance spectroscopy.

The term spectroscopy in the present case is to be understood as frequency-dependent measurements, ie the impedance of a sample is examined at different measurement frequencies. For example, the individual material samples are each examined in a frequency range between 10 Hz and 10 ⁷ Hz with a measured data density of 15 measuring points per decade. This means that 180 measured data are determined per material sample with such a measurement data density. A far-reaching data reduction can be achieved, for example, by adapting a suitable circuit equivalent to the measured data.

Around the measuring wires, the particular about Melt balls on the contact surfaces the sample plate abut, to keep under the bias, the measuring wires each connected to a particular gold-plated spring contact be, the constant contact pressure of the respective measuring wire on the respective contact surface guaranteed.

Around To develop a gas sensor, the material samples with a specific Test gas to be able to charge the measuring head can be connected to a gas supply unit.

Around expose the material samples to different test gas or reference gas atmospheres to be able to includes the gas supply unit, which expediently with a data processing unit the measuring and evaluation unit is connected, a gas mixing device. Furthermore, the gas supply unit for moistening the Prüfgas- or Reference atmosphere include a water reservoir.

Of the Furthermore, the measuring head can be designed to be integrated Component includes a gas space above the material samples of Sample plate is arranged and preferably of a substantially bell Distributor device is formed. The gas space is with the gas supply unit connected.

Around a homogeneous distribution of the test or to reach reference gas in the gas space, is at an advantageous embodiment arranged a diffuser of the device according to the invention in the gas space.

Around for a variety of material samples on the sample plate the individual To be able to measure material samples in a simple manner includes the measuring and evaluation unit advantageously has two relay panels, the with the measuring wires and for example at 64 material samples on the sample plate preferably in each case a 3x64 matrix of high frequency suitable Relay have. In this case, the 64 material samples are on the sample plate vermessbar in a measuring cycle, wherein at least three measured quantities are accessible, and for example about an impedance analyzer, the impedance and ent speaking further measuring devices, the DC resistors of the Material samples and their current / voltage characteristics.

The The measuring and evaluation unit is expediently equipped with a measuring and control software that controls the measurement process on the one hand and on the other hand obtained measurement data to a corresponding file or to a relational database, which is provided by an evaluation software can be read.

The Evaluation software preferably operates to provide a fit functionality for calculation theoretical impedance spectra for the individual samples, the calculation preferably on the basis of a circuit equivalent that is at least comprises a virtual or real electronic component. One For example, a virtual device is a Constant Phase element (CPE). The fit functionality So calculates based on the measurement data for a example of a serial RC element existing circuit equivalent to a measured Impedance spectrum best possible approximate, theoretical Impedance spectrum, whereby the adaptation of the theoretical to the measured Spectrum a variation of capacitance and / or resistance the components of the RC element carried out becomes. The best possible adjusted, theoretical impedance spectrum are so in this case assigned a resistance value and a capacitance value for the RC element. From the resistance value, for example, by comparison with a reference value to the sensitivity of the respective material sample getting closed.

Around in a variety of material samples under different measurement conditions to be measured for the individual material samples obtained, for example, the sensitivities representing To make output values accessible for a simple evaluation includes the Evaluation software advantageously a data mining functionality. The data mining functionality is determined for example, using preferably multi-dimensional objective functions the for the respective application optimal material sample on numerical Ways.

alternative or additionally can the data mining functionality also include a visualization functionality. In this case a user can be screen-based the for determine the optimum application of the respective application. For example, the visualization functionality works with one Color spectrum, where a particular color high sensitivity of the material sample and a different color is associated with a low sensitivity.

Around Also find sensor materials that are suitable for applications are where - like for example, in an exhaust - high Temperatures prevail, the device according to the invention preferably a heating device, in which the sample plate preferably is submersible. In this case, the device has particular a high temperature reactor which is limited by the heater is and in which the material sample sample plate with different test or reference gases can be acted upon.

The The invention also has a method for analyzing a sample plate with the features according to claim 19 to the subject. By applying this method, it is possible to have a Sample plate on which a large Number of material samples, for example 64 material samples arranged are to be measured electrically under different conditions and in fully automatic manner for the particular application to select the most suitable material sample. This is done by automatic Selection of starting values for the components of the respective circuit equivalent and the subsequent Fehlerinimierungsrechnung. The starting values are used in the error minimization calculation, starting from the starting values one at each measured Impedance spectrum adapted, theoretical impedance spectrum for the relevant Material sample is calculated under the respective measurement conditions. Through this procedure, a substantial data reduction is possible because starting from the impedance spectra, which has a multitude of measuring points have, representing few derived quantities Fitwerte are determined, which are the individual components of the circuit equivalent describe. The fit values thus represent dimensioning of the components of the circuit equivalent with which the measured impedance spectrum simulates the best possible can be.

at the determination of the circuit equivalent For example, a serial circuit of four RC elements selected be, with the higher RC elements may be set to a starting value, the no influence on the calculation of the theoretical impedance spectrum Has.

The starting values of an RC element required for the error minimization calculation are preferably calculated from the maximum measured, imaginary impedance Z "_ MAX and the corresponding measurement frequency f_Z" _ MAX according to the following formulas:

If the circuit equivalent comprises several RC elements, those required for the higher RC elements Start values for the error minimization calculation preferably from the difference spectra between the measured data and data determined on the basis the for the first RC element calculated or simulated calculated starting values become.

Especially the selection of so-called "good" starting values, the close to the actual Sizes of the components of the circuit equivalent lie, shorten the duration of the subsequent Fehlerinimierungsrechnung crucial. "Bad" starting values, however, can cause that the error minimization calculation performed on the basis of the starting values delivers meaningless values.

In a preferred embodiment of the method according to the invention, the circuit equivalent for simulating impedance spectra is based on an error minimization calculation n values from a serial split of four RC elements. By determining starting values for the individual components, it is then determined by means of a threshold value how many RC elements are taken into account in the simulation calculations. The threshold value is preferably a value presettable by a user. The percentage ratio of the resistance of the first RC element to the resistance of the current RC element n is checked according to the formula RCn_START> value [%] · RC1_START.

The Size "value" is that of the User changeable Size that for example, to 10%. If the argument is not Fulfills is, the starting values for the components of the relevant RC element are set to values that have no Have an influence on a simulation of the impedance spectra. These values are kept constant in the error minimization calculation.

Further In the method, the validation size that determines the match is determined between the calculated, theoretical impedance spectrum and the evaluated respectively associated, measured impedance spectrum.

The Evaluation value the for the respective analysis relevant output value, for example, at the determination of a sensor material for a gas sensor, a sensitivity of the respective Material sample reproduces. The sensitivity is a measure of the goodness of a Sensor.

The sensitivity of a resistive gas sensor can be defined in various ways. Considering, for example, the direction of change of the resistances at a gas admission, a sensitivity S as the quotient of the resistance R_TEST under a test gas atmosphere and the resistance R_0 under reference conditions can be expressed as follows

The Sign of sensitivity S provides information regarding resistance change.

Alternatively, the sensitivity may be described as a change in resistances, according to the formulas

at These formulas result in sensitivities S_Δ between -1 and 1, ie. H. normalized sensitivities.

It is possible, too, the sensitivities to translate both definitions into each other. The sensitivity expressed by the value S is special meaningful in a big change the resistance of a material sample as a result of loading with a test gas. The above S_Δ expressed sensitivity is special meaningful if the resistance of a material sample due to a Prüfgasbeaufschlagung only low changes. However, the sensitivity expressed by S_Δ has the advantage of a big one Tolerance Measurement inaccuracies, a precise Illumination of a range of small changes in resistance and thus a better one Evaluation of cross sensitivities as well as the possibility of an automated visualization and data processing.

at a particularly simple embodiment of the method according to the invention is the circuit equivalent from a virtual arrangement of real electronic components such as Capacitors and resistors. In its simplest version consists of the circuit equivalent from a resistance.

By the error minimization calculation, in which an approximation of the sizes of the components to the measured data takes place, a simulated impedance spectrum for the circuit equivalent for the respective Ma terialprobe determined. From the matched sizes of the devices, one can infer the electrical properties of the processes in the material sample described by the circuit equivalent. If the electrical behavior of the material sample is determined by several processes of different relaxation times, it is necessary to use more complex circuit equivalents, for example consisting of several RC elements connected in series. If the processes react differently to a variation of the measurement conditions, it is possible to assign individual processes to components or component groups in order to be able to analyze the individual processes separately. The error minimization calculation reduces the data volume for the description of the impedance measurement to the arrangement and the sizes of the components.

at a sample plate with, for example, 64 material samples and a Measurement under eleven different gas atmospheres with four different ones Temperatures is the measured data at 2816 impedance spectra. The individual impedance spectra are according to the method of the invention simulated and in each case on the derived measured variables, the by the sizes of Components of the circuit equivalent be reproduced, reduced.

φ = arctan(ωRC)
Y' = cos φ
Y'' = sin φThe calculation of a theoretical impedance spectrum of a circuit equivalent is performed in a preferred embodiment of the method according to the invention such that a complex admittance Y * and a phase shift φ of the individual components of the RC elements at a given angular frequency ω (2Π · measurement frequency) determined according to the following formulas becomes:

φ = arctan (ωRC)
Y '= cos φ
Y '' = sin φ

wobei Z' der Realteil der Impedanz und Z'' der Imaginärteil der Impedanz ist.Through a transformation, the respective impedances can be determined:

where Z 'is the real part of the impedance and Z''is the imaginary part of the impedance.

at A serial circuit of the RC elements can be the impedances of each RC elements are summed directly. A calculation of the impedances for frequencies, which correspond to the measuring frequencies, gives one with the measured data corresponding record, so that an error estimate between the measured record and the theoretically determined record is possible.

The error minimization calculation performed in the method according to the invention is in an advantageous embodiment by varying the sizes of individual components performed by 1%. By analyzing the differences of the theoretically calculated and measured spectrum can be Errors are determined. If the error after a variation shrinks, a renewed variation for the same component of the Circuit equivalent carried out and based on this variation, a theoretical impedance spectrum calculated with which a new error calculation is performed. If the error does not diminish, the variation will occur reversed and the variation of the components varies with opposite signs or another component varies.

The error calculation is preferably performed such that the function for determining the error takes into account the starting value of the resistance of the first RC element. If the value is greater than a nominal measuring resistance of the impedance analyzer used in the measurement, for example, greater than 3 × 10 ⁷ Ω, the error is determined only from the imaginary part of the calculated impedance. In order to be able to weigh the high-frequency range of the determined spectrum, the deviations determined are preferably multiplied by the logarithm of the measuring frequency. This weighting makes it possible to suppress physically meaningless or even incorrect measured data in the low-frequency range.

In order to calculate the validation quantity, ie to validate the quality of the theoretically calculated impedance spectrum for the respective material sample, a corridor about the theoretical impedance spectrum is preferably determined which, for example, contains 90% of the measured data. To minimize the For this purpose, a successive approximation algorithm can be used for calculating the validation quantity, wherein the limit values of the algorithm can be 0 Ω and twice the real part of the summed impedance at the smallest measurement frequency.

The inventive method is preferably used to analyze material samples under different conditions Test gas atmospheres and used in particular at different temperatures. In this Fall for all samples of material placed on the sample plate have impedance spectra measured under the different measuring conditions.

The each measured spectra are then in the method of the Invention in each case by a Fehlerinimierungsrechnung based on a circuit equivalent theoretically simulated. This results in a large amount on evaluation variables, the the target values represent the method.

The Target sizes or Evaluation parameters are in the method according to the invention preferably in a database written and evaluated by a data mining functionality.

at the database that serves to record and provide records, it is expediently to a relational database, in which information according to topics ordered in the form of tables are stored. Relationships between the individual records in the tables are produced by so-called identification key.

The Database can, for example, further properties of the material samples, like their synthesis conditions of the educts, their sample history and the like. These properties are about table relationships linked together.

The Data mining can be done by means of an optionally multi-dimensional objective function and / or performed by means of a visual data mining functionality. The data mining performed by means of an objective function is a numerical one Method based on the individual evaluation quantities stored in the database for example, broken down according to measuring temperature and test gas are stored. It first defines which properties are required of the sought material. For example, when looking for a sensor material for a Gas sensor can be specified with respect to which gas, the sensor material should be sensitive and which cross-sensitivities might disturb. from that Thus, a requirement profile for a fingerprint results Sensitivities.

The visual data mining functionality works advantageously such that the evaluation of the material samples of the Broken down sample plate for example, after test gases and temperatures are shown.

When using the device according to the invention and the method according to the invention, a preferably fully automatic high-throughput impedance system is available, with which new materials can be developed with a high sample throughput and a low expenditure of time and money. For example, by using the system on two days 64 different material samples at four different temperatures and under eleven different test gas atmospheres are examined with regard to their sensory properties and also evaluated.

Especially The system can be used for general material development and specifically for the Development of sensors in the automotive sector and in the field of safety technology be used.

The The invention also has a data processing system with a data processing program to carry out the method according to the invention to the subject.

Further Advantages and advantageous embodiments of the subject to The invention will become apparent from the description, the drawings and the claims.

drawing

An embodiment of the device according to the invention is shown schematically simplified in the drawing and in the following description in connection with a method according to the Er detailed explanation. Show it

1 a measuring assembly of a device according to the invention;

2 a measuring device of the device according to 1 ;

3 a measuring head of the measuring device according to 2 ;

4 a gas distribution device of the measuring head according to 3 ;

5 a measurement procedure based on a flowchart;

6 a diagram showing a time course for material samples at different temperatures and under different gas atmospheres;

7 a diagram in which a run-in behavior of four differently surface-doped material samples is shown;

8th a diagram in which a response time behavior of three different material samples during a Prüfgaspulses is shown;

9 a circuit equivalent;

10 a flowchart of an error minimization calculation;

11 an exemplary course of a sensitivity of a material sample;

12 an exemplary fingerprint of sensitivities of a material sample with respect to different gas atmospheres; and

13 a highly schematized representation of a visual data mining functionality.

description of the embodiment

In the 1 to 4 is a measurement setup of a device 10 for analysis of a sample plate 12 shown on the 64 material samples 13 are arranged. The device 10 includes a high temperature reactor 14 , a gas supply unit 16 and a measuring and evaluation unit 18 ,

The high temperature reactor 14 in particular in the 2 to 4 is shown in more detail, comprises a frame or housing 20 in which a height-adjustable heating block 22 is arranged on a guide rod 23 and on a threaded rod 24 is stored, which with a crank 25 is provided.

The heating block 22 , which consists for example of four a heating chamber limiting heating plates with a heating power of 1100 W, is adjustable in height so that the plate carrier 28 and thus the sample plate 12 into the boiler room is submersible.

Furthermore, the high temperature reactor includes 14 a measuring head 26 in one with the rack 20 connected probe carrier 27 is used and for contacting the 64 on the sample plate 12 arranged material samples 13 with the measuring and evaluation unit 18 serves. The sample plate 12 is on one with the probe carrier 27 connected plate carrier 28 arranged.

The measuring head 26 , in particular in 3 is illustrated comprises a base plate 29 , which is formed for example of a machinable glass ceramic and for holding 128 measuring wires 30A . 30B serves, each consisting of platinum and are sheathed by an aluminum oxide tube. At their lower ends have the measuring wires 30A . 30B each a melting ball 31A . 31B , for contacting the respective measuring wire at a contact surface of the sample plate 12 serves. The sample plate 12 has two contact surfaces per material sample, ie in the present case a total of 128 contact surfaces, each with one of the measuring wires 30A . 30B interact. The measuring wires 30A . 30B are at their upper ends in the area of the base plate 29 each with a gold-plated spring contact 32A . 32B provided with the fixed measuring head 26 a constant contact pressure of the measuring wires 30A . 30B or the melted ball 31A . 31B on the respectively associated contact surfaces of the sample plate 12 guaranteed. The Ummantellungen made of platinum measuring wires 30A . 30B are also on holding plates 33 and 34 fixed on a linkage 35 attached and parallel to the base plate 29 are aligned.

The spring contacts 32A . 32B are over lines 36A . 36B with in the measuring head 26 integrated SMB sockets 37A . 37B connected in turn via shielded SMB lines 38A respectively. 38B with the measuring and evaluation unit 18 the device 10 are connected. All in all 128 BNC connectors 37A . 37B has the measuring head 26 , each with a measuring wire 30A . 30B is connected and to each one to the measurement and evaluation 18 leading SMB line 38A . 38B connected. For the sake of clarity are in the 1 to 3 but only two SMB sockets each 37A . 37B and two SMB lines 38A . 38B as well as the respectively assigned measuring wires 30A . 30B shown.

The measuring head 26 also has a substantially bell-shaped gas distribution device made, for example, of quartz glass 39 , in the 4 is shown in detail and the via a gas supply line 40 made of stainless steel with the gas supply unit 18 connected is. The bell-shaped gas distribution device 39 limits a gas space above that on the sample plate 12 disposed 64 material samples 13 is arranged.

The material samples 13 are formed, for example, from tin oxide SnO ₂ and have different dopants, which are formed, for example, from lanthanides. The material samples 13 are matrix-like in eight rows and eighth columns on the sample plate 12 distributed.

In the 4 detailed gas distribution device shown 39 has in her gas room 41 furthermore a diffuser insert 42 which is formed of a quartz ball and a variety of holes 43 has, each having a diameter of about 1 mm.

For a uniform gas outlet from the gas space 41 To ensure has the bell-shaped gas distribution device 39 Spacers at their edges 43 which has a gap of 0.8 mm width between the gas distribution device 39 and the sample plate 12 establish.

For loading the gas space 41 and thus the material samples 13 with different test gas atmospheres, the gas supply unit 16 two gas cylinder cabinets, not shown here, each with four gas cylinders, each with a gas flow regulator 44A . 44B . 44C . 44D . 44E . 44F . 44G respectively. 44H in one gas bottle moist synthetic air, in the second gas bottle hydrogen, in the third gas falsified methane, in the fourth gas surface synthetic air, in the fifth gas surface nitrogen dioxide, in the sixth gas cylinder nitrogen monoxide, in the seventh gas propene and in the eighth gas cylinders containing carbon monoxide. The capacities of the flow controllers 44A . 44B . 44C . 44D . 44E and 44F each lie between 0 sccm and 100 sccm. The capacities of the flow controllers 44G and 44H each lie between 0 sccm and 10 sccm. By appropriate control of the volumetric flow of the various gases can by means of the battery from the eight gas flow controllers 44A . 44B . 44C . 44D . 44E . 44F . 44G and 44H Test gases of different compositions in a manifold 45 be fed with the gas supply line 40 connected is.

In order to set a relative humidity of the respective test gas, the test gas, a moist carrier gas, which consists for example of synthetic air, are admixed. The humidity of the carrier gas is adjusted by passing it through a water reservoir 46 is directed. A measurement of the moistened carrier gas by means of a moisture sensor not shown here, for example, gives a relative humidity of about 90% at room temperature.

The measuring and evaluation unit 18 includes two relay panels 50 and 51 to each 64 cables 38A respectively. 38B connected to the measuring head 26 of the high temperature reactor 14 to lead. The relay panels 50 and 51 each form a 3x64 matrix of high frequency compatible relays.

The relay panels 50 and 51 are via a digital control line 52 with a measuring and evaluation computer 53 connected. Furthermore, the relay panels 50 and 51 via test leads 54 with an impedance analyzer 64 and a so-called source meter 55 connected. These two gauges are also via the digital control line 52 with the measuring and evaluation computer 53 connected. The addressing of the impedance analyzer 54 and the source meter 55 via the two relay panels 50 and 51 ,

The measuring and evaluation computer 53 is also via another digital control line 56 with a D / AA / D converter 57 connected via an analog control line 58 on the one hand with the heater 22 of the high temperature reactor 14 and on the other with the gas flow controllers 44A . 44B . 44C . 44D . 44E . 44F . 44G and 44H the gas supply unit or gas mixer 16 connected is.

On the measuring and evaluation computer 53 is a modular measurement and control software stored, the script via a complete automation of means of the device 10 carried out measurements. Due to the modular structure of the measuring and control software, a trouble-free extension of the measuring system is possible.

By means of the source meter 55 can be DC resistances, U / I characteristics or voltages of individual material samples 13 be measured. These values as well as those by means of the impedance analyzer 54 Measured values can be determined via the digital control line 52 at the measuring and evaluation computer 53 which includes a database for the measurement data.

To measure the on the sample plate 12 arranged material samples 13 In a high-throughput mode or a high-throughput method, a scripting language is used, whereby a script file is used to pass a list of tasks which consist of keywords and parameters to the software, by means of which the script file is processed and which all Functions of the system controls. The script files ensure continuous checking of control parameters so that measurements under false measurement conditions are excluded by suspending further processing of the script file. The script file is not limited in length.

One by means of the device 1 feasible measurement procedure is in 5 illustrated by a flow chart. For the analysis of resistive gas sensory properties, the material samples 13 the sample plate 12 on the one hand under a reference gas atmosphere, which is formed for example of synthetic air of a relative humidity of, for example, 45%, and electrically characterized under various NEN test gas atmospheres. By modulating the measurement temperature, it is possible to obtain information about the influence of the respective operating temperature on the material samples suitable as sensor material and about the activation energy of conductivity processes.

In order to ensure comparability of measurement data of several sample plates, the script file used is constructed as a standard script, which controls the complete high-throughput screening, whose chronological sequence in the in 5 the flowchart shown is described. Here, in a first method step M1, a measurement temperature T_MESS of the sample plate 12 set. Measurements can be carried out, for example, in a temperature range between 400 ° C and 250 ° C in steps of 50 °. During the cooling phases, the sample plate becomes 12 supplied with a reference gas of synthetic air of relative humidity of 45%.

If the sample temperature changes, a stable base resistance of the material samples only sets in after a certain time, so that a conditioning of the material samples in a step M2 is required. Temperature changes of the material samples lead to a metastable state of intrinsic defects, which may represent oxygen vacancies and whose thermodynamic equilibration requires finite time. In 7 is an example of the conditioning or running-in behavior for four differently surface-doped In ₂ O ₃ samples when reaching a target temperature of 300 ° C shown. The resistances R of the samples approach asymptotically with time a limit value which represents the so-called basic or reference resistance. This limit is extrapolated after about 90 minutes. To ensure a constant reference resistance, it is advantageous to choose a conditioning time of 120 minutes.

Subsequently, the test gas required for the measurement is assembled in a method step M3 and in the above the sample plate 12 arranged gas space 41 initiated. The first test gas includes, for example, hydrogen at a concentration of 25 ppm. The carrier gas is synthetic air. The gas flow is adjusted to 100 sccm with humid synthetic air.

In order to ensure that the individual material samples have reached their basic resistance irrespective of their relative position on the sample plate, a preliminary phase for the test gas, which was previously investigated, then ensues in a step M4. In 8th FIG. 3 shows the resistance profile of three samples of material lying on a diagonal of the sample plate during a 40-minute test gas pulse, the pulse gas comprising propene having a concentration of 50 ppm. The base material of the mate rial samples is at the in 8th illustrated example because of its high sensitivity to hydrocarbons tin oxide SnO ₂ . Regardless of the relative position on the sample plate, the resistance of the material samples drops to a constant value within about 6 minutes due to the test gas pulse. The response time is thus in each case 6 minutes, the response being essentially independent of the position of the respective material sample on the sample plate. After completion of the test gas pulse, the resistances approach the ground resistances asymptotically within about 10 minutes. The response is therefore essentially independent of the position of the material sample on the sample plate. The lead time of each test or reference gas selected for the script file is approximately 15 minutes before a corresponding measurement.

In a subsequent method step M5, the measurement of impedance spectra for the 64 on the sample plate arranged material samples. In the script file for the measurements of the impedance spectra are defined as parameters: Amplitude of the measuring voltage [V] 0.1 Starting frequency [Hz] 10 Final frequency [Hz] 10 ⁷ Measuring points per frequency decade 15 Bias [V] 0 Mode [HS: Highspeed; NO: normal; AV: Average] HS

The measured data obtained are determined by means of the impedance analyzer 64 determined and to the measuring and evaluation computer 53 or the database stored on it.

in the Following this, the measurement can be carried out under a different test gas atmosphere, wherein Here, in turn, the steps M3, M4 and M5 will be traversed. For example, you can as further test gases Carbon monoxide of a concentration of 50 ppm, nitric oxide a concentration of 5 ppm, nitrogen dioxide concentration of 5 ppm or propene of a concentration of 25 ppm used become.

The entire screening can then be done by going back to step M1 be carried out at a different measuring temperature.

In 6 a measurement sequence determined by a standard script is visualized, whereby in this diagram measurements under reference conditions and during conditioning phases are not shown for the sake of clarity. Measurements under reference conditions, ie under a reference gas atmosphere, are always carried out before introducing a new test gas H ₂ , CO, NO, NO ₂ or propene at the relevant temperature T. There is always a so-called test gas-advance phase X and a measurement phase Y before. Furthermore, in 6 the respective concentrations C of the test gases are shown.

The Measurement and evaluation software can have a functionality in which each measured impedance spectrum graphically on a Monitor is displayed. The measurement data, preferably in ASCII format in the database or in the directories associated with the sample plates are stored on the measurement and evaluation computer, are thus directly accessible to a visual inspection.

Of Furthermore, the measuring and evaluation software can be equipped with a functionality be the impedance spectra of the sample materials of a sample plate depending on the position in a picture matrix. This can be both raw data or synonymous derived data. The func onality can also include an evaluation window into which to transfer data of a particular material sample can be. Also, the functionality can combined with other imaging measurement or evaluation systems become. Image data from these systems can then also be position dependent in the image matrix are read in order to provide more information about the To receive samples.

After measuring the impedance spectra for the individual material samples, a theoretical impedance spectrum based on a circuit equivalent is calculated for each material sample. The circuit equivalent consists in the present case of a series circuit of four RC elements of in 9 In each case, the theoretically calculated impedance spectra are optimally adapted to the correspondingly measured impedance spectrum, namely by varying the sizes or dimensions of the individual components of the circuit equivalent.

The error minimization calculation performed will be described below with reference to the in 10 illustrated flowchart explained.

to execution The error minimization calculation requires it for the individual Components of the circuit equivalent To determine starting values. At the same time it is necessary to know the number the RC elements to be considered in the error minimization calculation determine. For these purposes, in a step S1, the maximum imaginary Impedance Z '' _ MAX from the measured data for the determined material sample. In the selection is also judged, whether the maximum measured imaginary impedance is a faulty measurement what happens through an investigation of the adjacent measuring points measured imaginary Impedances, d. H. by examining the local maximum, he follows.

berechnet, wobei f_Z''_MAX die Messfrequenz bei der maximal gemessenen, imaginären Impedanz ist.Subsequently, in a step S2, start values R1_START for the resistance and C1_START for the capacitance of the first RC element are determined on the basis of Z "_MAX and the corresponding measurement frequency according to the formulas

calculated, where f_Z '' _ MAX is the measuring frequency at the maximum measured, imaginary impedance.

Then be in a loop S3 starting values for the resistors and capacities the other three RC elements of the circuit equivalent determined.

For this will be first in a step S4 on the basis of the starting values R1_START and C1_START calculated a theoretical impedance spectrum and in one step S5 a difference spectrum between the theoretical impedance spectrum and the measured impedance spectrum and from this difference spectrum again the maximum of the imaginary ones Impedance Z "_ MAX determined. outgoing from this maximum then in a step S6, the starting values Rn_START and Cn_START (n = 2 to 4) for the considered RC element according to the im Explained in connection with step S2 Calculated formulas.

Subsequently, will in a step S7, a threshold consideration for the calculated Resistor start value Rn START is performed, by means of which it is determined whether the respective RC element in the subsequent simulation calculations considered shall be. The threshold is a user-changeable one Value. Is checked in the comparison, the percentage ratio of the resistance of the first RC element to the resistance of the current component.

If the calculated starting value Rn_START is greater than the threshold, the RC element is taken into account, jumped back to step S3 and made a calculation of the start values Rn_START and Cn_START for the next RC element. If the argument is not satisfied, the sizes of the components of the higher RC elements are set to values which have no influence on the simulation calculation for an impedance spectrum. These values remain constant during subsequent adaptation steps and are, for example, Rn_START = 1 and Cn_START = 10 ^-15 . This determination is made in a step S8.

The Threshold query thus determines the number of RC elements that are in the following steps become. If the threshold is undershot, becomes direct for error minimization calculation based on the number of determined ones RC elements, for example, based on an RC element, in this case a loop S9 via all m were taken into account RC elements performed becomes.

Within In loop S9, another loop S10 run = 1 to 3 is performed, where this is an empirical number of runs, which to increase accuracy of the error minimization calculation.

Within the loop S10, the resistance starting value R1_START of the first RC element is first used as variable VAR in a step S11. The variation of the individual components in the error minimization calculation is 1%, which is expressed in a value FAK = 0.01 in step S11. The error is preset to 10 ⁹⁹ . The variation of the respective component takes place in a step S12 according to the formula VAR = V * FAK. A calculation of a theoretical impedance spectrum is then performed based on the varied value in a step S13.

Subsequently, an error is calculated which is based on a comparison between the impedance spectrum calculated for the pertinent component sizes and the measured impedance spectrum. A function for determining the error depends on the starting value of the resistance of the first RC element. To calculate the error, it is checked in a step S14 whether the respective starting value is greater than the nominal measuring resistance of the impedance analyzer used, for example, greater than 3 × 10 ⁷ Ω. If this is the case, the error in a step S15 only by considering the imaginary parts of the impedances according to the formula Error = | Z '' _ fit - Z '' _ mess | · logf determined, wherein the weighting of the high-frequency range of the spectrum is multiplied by the logarithm of the measurement frequency, whereby physically meaningless or erroneous measurement data in the low-frequency range are suppressed. Otherwise, in a step S16, an error consideration based on both the real parts and the imaginary parts of the impedances according to the formula Error = | Z '' _ fit - Z '' _ mess | + | Z'_fit - Z'_mess |.

In a step S17 is checked whether the Error due to the variation of the device in step S12 smaller has become. If this is the case, the component in question, For example, the resistance of the relevant RC element by jumping back varies again to step S12. If the error has not dropped is, the variation made in step S12 becomes one step Undone S18 and in a step S19 determines whether the sign of the variation changed in a step S20 or in a step S21 and a subsequent step S22 the next Component, in this case the capacitance C, as a variable component chosen becomes. In this case, a return to step S12 takes place again, being the size of the component then again as long as the error decreases.

The Variation of the resistance and the variation of the capacity is through return to the step S10 for every RC element twice repeated. Thereafter, depending on the number of to be considered RC elements a n-fold return to the step S9.

If no further minimization of the error by varying the resistance and the capacity of the RC element possible is, the values giving the smallest error become the resistance and the capacity as evaluation variables of Error minimization calculation issued.

To Completion of the error minimization calculations performed in steps S9 to S22 will for the theoretically determined impedance spectrum in a step S23 Validation function determined by which the quality of the theoretically calculated Spectrum estimated can be. When determining the validation function, the appraisal a goodness the error minimization calculation a corridor around the calculated impedance spectrum which contains 90% of the measured data. The thus determined Error is between 0 and 1, with an error of 0 being an ideal match between the measurement and the simulation calculation and an error from 1 no match between the measurement and the simulation calculation.

When using the method for determining a sensor material of a gas sensor, sensitivities S_Δ are calculated on the basis of the resistances of the circuit equivalents obtained in the error minimization calculation. The sensitivities S_Δ obtained in this way are normalized and lie between -1 and +1, like the diagram in FIG 11 can be seen. In 11 Sensitivities for an assumed reference resistance of 100 Ω are shown with varying test resistance R_test. The sensitivities S_Δ are stored in the database.

Subsequently, on the basis of the ascertained sensitivities S_Δ, a numerical data mining is carried out in order to select an optimum gas sensor for a specific application. The data mining is carried out in particular numerical ways, with access to the individual sensitivities S_Δ, which are stored in the database broken down to measurement temperature and test gas. First, it defines what properties are required of the desired sensor material. In the simplest case it is specified for which test gas the sensor is to be used and which cross-sensitivities disturb. This results in a requirement profile for the fingerprint of the sensitivities, as in 12 is shown. The requirement profile is converted into so-called ">" or "<" requirements, whereby data records with the desired properties are characterized. Other conditions, such as the nature of the base material of the sensor material or its surface doping can be incorporated into the target function be taken. In the example in 12 the requirement profile is such that all test gases A, B, C, E except the test gas D have a sensitivity of less than 0.2 and greater than -0.2 and the test gas D has a sensitivity greater than 0.9. Each X is the reference. The requirement profile is implemented by means of an evaluation functionality as an SQL filter query and passed directly to the database. The results of the SQL statement can then be displayed in tabular form on a screen.

Alternatively or additionally, the sensitivities determined for the different material samples under the different measurement conditions can be evaluated by means of visual data mining functionality. With this functionality, the in 13 is shown, sensitivities of so-called library plates are broken down according to test gases and temperatures displayed on a screen. In the exemplary representation according to 13 are library plates 101 to 112 for four temperatures A, B, C and D and for three different test gases I, II and III. Every library plate 101 to 112 is associated with a temperature A, B, C or D and a test gas I, II or III. The sensitivities S_Δ of the individual materials of the library plates 101 to 112 are each according to their position on the sample plate as circles 120 and shown in false colors. In 13 are for clarity for each library plate 101 to 112 only four of 64 Material samples shown. For example, positive sensitivities S_Δ are represented in colors ranging from black to reds to yellow, while negative sensitivities are represented, for example, in colors ranging from black to blue to turquoise. The diameter of the individual circles 120 is determined by the validation of the error minimization calculation, where a large error represents a small circle and a small error a comparatively large circle. This allows an intuitive assessment of large amounts of data for a user.

Which to the basis 13 Data sets required are extracted directly from the database, so that filter functions can be used to select the measurements to be displayed. The number of displayed data sets is limited only by the working memory of the evaluation computer used.

Of Furthermore, it is possible with the visual data mining functionality that additional Information on the individual material samples through screen-based selection a specific circle.

Claims (29)

Device for analyzing at least two material samples ( 13 ) on my sample plate ( 12 ), comprising a support ( 28 ) for the sample plate ( 12 ) and contacting means for electrically contacting the material samples ( 13 ), characterized by a in a housing carrier ( 27 ) usable measuring head ( 26 ) for electrical connection with the contacting means per material sample ( 13 ) two measuring wires ( 30A . 30B ), which are preloaded on contact surfaces of the sample plate ( 12 ) and with a measuring and evaluation unit ( 18 ) are connected. Device according to claim 1, characterized in that the measuring wires ( 30A . 30B ) over molten spheres ( 31A . 31B ) at the contact surfaces of the sample plate ( 12 ) issue. Device according to claim 1 or 2, characterized in that the measuring wires ( 30A . 30B ) each with a spring contact ( 32A . 32B ), which has a constant contact pressure of the respective measuring wire ( 30A . 30B ) ensured on the respective contact surface. Device according to one of claims 1 to 3, characterized in that the measuring head ( 26 ) with a gas supply unit ( 16 ) connected is. Apparatus according to claim 4, characterized in that the gas supply unit ( 16 ) with a data processing unit ( 53 ) of the measuring and evaluation unit ( 18 ) connected is. Apparatus according to claim 4 or 5, characterized in that the gas supply unit ( 16 ) comprises a gas mixing device. Device according to one of claims 4 to 6, characterized in that the gas supply unit ( 16 ) a water reservoir ( 46 ). Device according to one of claims 4 to 7, characterized in that the measuring head ( 26 ) to Gas loading of the sample plate ( 12 ) one of a substantially bell-shaped distributor device ( 39 ) formed with the gas supply unit ( 16 ) connected is. Apparatus according to claim 8, characterized in that in the gas space a diffuser ( 42 ) is arranged. Apparatus according to claim 8 or 9, characterized in that the gas space is hen with a gas outlet verse, by at least one spacer ( 43 ) formed between the sample plate ( 12 ) and the distribution device ( 39 ) is arranged. Device according to one of claims 1 to 10, characterized in that the measuring and evaluation unit ( 18 ) two relay panels ( 50 . 51 ), which are connected to the measuring wires ( 30A . 30B ) and each having a 3x64 matrix of high frequency suitable relay. Device according to one of claims 1 to 11, characterized in that the measuring and evaluation unit ( 18 ) an impedance analyzer ( 64 ). Device according to one of claims 1 to 12, characterized in that the measuring and evaluation unit ( 18 ) is equipped with a measurement and control software that passes collected and / or derived measurement data to a relational database, which is linked to an evaluation software. Apparatus according to claim 13, characterized in that the evaluation software comprises a fit functionality for calculating theoretical impedance spectra for the individual material samples, wherein the calculation based on a circuit equivalent ( 90 ), which comprises at least one electronic component. Apparatus according to claim 13 or 14, characterized the evaluation software comprises data mining functionality. Device according to claim 15, characterized in that that the data mining functionality works using multi-dimensional objective functions. Device according to one of claims 19 to 16, characterized that the data mining functionality a visualization functionality includes. Device according to one of claims 1 to 17, characterized by a heating device ( 22 ) into which the sample plate ( 12 ) is submersible. Method for analyzing at least two samples of material deposited on a sample plate ( 12 ), comprising the steps of: - measuring an impedance spectrum for each of the material samples - storing the measured impedance spectra in a file or database; Determining a construction of a circuit equivalent as a function of the respective measured impedance spectrum for each of the material samples, wherein the respective circuit equivalent comprises at least one resistor and / or at least one RC element; Determining starting values for the components of the respective circuit equivalent required for an error minimization calculation; Calculating a theoretical impedance spectrum for at least one of the material samples by means of the error minimization calculation on the basis of the impedance spectrum measured for this material sample and the starting values for the components of the relevant circuit equivalent; - determining fit values for the components of the relevant circuit equivalent; - determining a validation quantity for the calculated theoretical impedance spectrum; Determining an evaluation value by comparing at least one of the fit values for the components with a reference value. Method according to claim 19, characterized taking into account a number of RC elements connected in series a preselected Threshold is determined, with a maximum of four RC elements are selected. A method according to claim 19 or 20, characterized in that the starting values for the Bauelemen te of a first RC element of the circuit equivalent in response to a maximum measured, imaginary impedance Z "_ MAX are determined, wherein a starting resistor R1_START and a starting capacity C1_START are calculated. Method according to one of claims 19 to 21, characterized that the error minimization calculation by varying the sizing the individual components of the circuit equivalent is carried out by 1%. Method according to one of claims 19 to 22, characterized that in the error minimization calculation an error of the theoretical Impedanzspektrums by analyzing the difference to the measured Impedance spectrum is determined. Method according to one of claims 19 to 23, characterized that for each of the material samples impedance spectra under different test gas atmospheres and be measured at different temperatures. Method according to one of claims 19 to 24, characterized that the evaluation size for each material sample is written to a database and based on the in the database stored evaluation variables Datamining performed becomes. Method according to claim 25, characterized in that that the data mining is performed by means of a target function. Method according to claim 26, characterized in that that the data mining is performed by means of a visual data mining functionality. Method according to one of claims 19 to 27, characterized that the measured impedance spectra visually checked by means of a control functionality and / or be evaluated. Data processing system with a program for execution A method according to any one of claims 19 to 27.