Dielectric spectroscopic sensor and method for measuring complex dielectric constant

Abstract

A dielectric spectroscopic sensor (1) has a sensing unit (11), a signal separation unit (12), an oscillator (13), and a receiver (14).

Description

Dielectric spectroscopy sensor and method for measuring complex permittivity

[0001] The present invention relates to a technique for measuring the complex permittivity of a liquid, semi-solid material or biological sample.

[0002] As the aging population advances, responding to adult diseases has become a major issue. Tests for blood glucose levels, etc., require blood sampling, which places a heavy burden on patients. For this reason, non-invasive element concentration measuring devices that do not require blood sampling are attracting attention.

[0003] As a non-invasive element concentration measurement device, a method using electromagnetic waves in the microwave to millimeter wave band has been proposed because there is less scattering inside the body compared to optical methods such as near-infrared light, and the energy of one photon is low.

[0004] For example, there is a method using a resonant structure, as shown in Non-Patent Document 1. In this method, a measurement sample is brought into contact with a device with a high Q value, such as an antenna or a resonator, and the frequency characteristics around the resonant frequency are measured. Since the resonant frequency is determined by the complex dielectric constant of the surroundings of the device, the concentration of a component is estimated from the amount of resonant frequency shift by predicting in advance the correlation between the amount of resonant frequency shift and the concentration of the component.

[0005] Another technique using microwave-millimeter wave electromagnetic waves is proposed in Patent Document 1, which discloses dielectric spectroscopy. Dielectric spectroscopy involves irradiating the skin with electromagnetic waves, causing absorption of the electromagnetic waves in accordance with the interaction between the blood component being measured, e.g., glucose molecules and water, and observing the amplitude and phase of the electromagnetic waves. A dielectric relaxation spectrum is calculated from the amplitude and phase of the observed electromagnetic waves at different frequencies. The dielectric relaxation spectrum is generally expressed as a linear combination of relaxation curves based on the Cole-Cole equation, and the complex dielectric constant is calculated. In measuring biological components, the complex dielectric constant correlates with the amount of blood components, such as glucose or cholesterol, contained in the blood, and is measured as an electrical signal (amplitude and phase) corresponding to this change. A calibration model is constructed by measuring the correlation between changes in complex dielectric constant and component concentration in advance, and the component concentration is calibrated from changes in the measured dielectric relaxation spectrum.

[0006] Regardless of which method is used, it is expected that measurement sensitivity will be improved by selecting a frequency band that is highly correlated with the target component, so it is important to measure the change in dielectric constant in advance using broadband dielectric spectroscopy.

[0007] Among dielectric spectroscopy methods, the technique using a coaxial probe (open-ended coaxial probe or open-ended coaxial line) as shown in Non-Patent Documents 2 and 3 and Patent Document 2 allows for the use of easily available samples such as water to calibrate the measuring instrument, and also allows the dielectric constant of the sample to be measured by contacting the sample to the end face of the probe without requiring any special processing of the material. Therefore, this technique is suitable for measuring samples such as living organisms, fruits, soil, or hydrogels, whose electrical properties need to be evaluated without processing.

[0008] To perform these dielectric constant measurements, a reflected signal measurement is used using a reflected signal measuring device such as that shown in Non-Patent Document 1 or a general-purpose measurement device such as a portable vector network analyzer.

[0009] Japanese Patent Publication No. 2013-32933 Japanese Patent No. 6771372

[0010] M. Hofmann, G. Fischer, R. Weigel, and D. Kissinger, "Microwave-Based Noninvasive Concentration Measurements for Biomedical Applications", IEEE Trans. Microwave Theory and Techniques, Vol.61, No.5, pp. 2195-2203, 2013J P. Grant, R N. Clarke, G T. SYymm and N M. Spyrou, "A critical study of the open-ended coaxial line sensor technique for RF and microwave complex permittivity measurements", J. Phys. E: Sci. Instrum, Vol. 22, pp. 757-770, 1989T.P. Marsland, and S. Evans, "Dielectric measurements with an open-ended coaxial probe", IEE Proceedings, Vol. 134, No. 4, 1987

[0011] However, reflected signal measuring devices are designed to be used in environments such as research institutions where the ambient temperature and surrounding environment are constant, and when considering sensing applications for collecting time-series data, there are issues such as a decrease in heat capacity due to the miniaturization of the dielectric spectroscopy sensor system (hereinafter simply referred to as the "dielectric spectroscopy sensor" or "system"), insufficient heat dissipation design, or a decrease in the stability of the measurement signal due to sudden changes in the ambient temperature.

[0012] The dielectric spectroscopy sensor of the present invention is characterized by having a sensing section, a signal separation section, an oscillator, and a receiver.

[0013] The method for measuring the complex dielectric constant of the present invention is a method for measuring the complex dielectric constant of a measurement target sample using the dielectric spectroscopy sensor, and is characterized by including a pre-measurement step of creating an input / output model of the sensing unit without installing the measurement target sample, and a measurement target sample measurement step of installing the measurement target sample and acquiring time-series data.

[0014] The present invention has the advantage that, rather than measuring using only the ratio between the reference signal and the measurement signal, the amount of fluctuation within the dielectric spectroscopy sensor system is estimated from the reference signal and corrected, thereby making it possible to stably measure the amount of signal fluctuation of the substance to be measured (MUT).

[0015] FIG. 1 is a schematic diagram showing one embodiment of the dielectric spectroscopy sensor of the present invention. FIG. 2 is a schematic diagram showing another embodiment of the dielectric spectroscopy sensor of the present invention. FIG. 3 is a schematic diagram showing another embodiment of the dielectric spectroscopy sensor of the present invention. FIG. 4 is a flowchart of a pre-measurement step of the method for measuring a complex dielectric constant of the present invention. FIG. 5 is a flowchart of a measurement target sample measurement step of the method for measuring a complex dielectric constant of the present invention. FIG. 6 is a schematic diagram showing a configuration of a dielectric spectroscopy sensor according to one embodiment of the present invention. FIG. 7 is a graph showing a comparison result of output results at a frequency of 7 GHz of a dielectric spectroscopy sensor according to one embodiment of the present invention. FIG. 8 is a flowchart of output signal measurement in a conventional method.

[0016] The following describes in detail an embodiment of the present invention, but the present invention is not limited to the embodiment described below, and various modifications are possible without departing from the gist of the present invention.

[0017] As a non-invasive element concentration measurement device that does not require blood sampling, microwave-millimeter wave electromagnetic wave techniques have been proposed due to their low scattering in vivo compared to optical techniques such as near-infrared light and the low energy of a single photon. Dielectric spectroscopy, one of the microwave-millimeter wave electromagnetic wave techniques, uses a coaxial probe. It can use readily available samples such as water for calibration and can measure the dielectric constant of a sample by simply contacting the probe end without requiring special processing of the material. This makes it suitable for measuring samples whose electrical properties require minimal processing, such as living organisms, fruit, soil, or hydrogels. These dielectric constant measurements have traditionally been performed using general-purpose reflected signal measurement instruments, such as a portable vector network analyzer. However, reflected signal measuring instruments are designed to be used in environments such as research institutes where the environmental temperature and surrounding environment are constant, and when considering sensing applications for collecting time-series data, there are problems such as a decrease in heat capacity due to the miniaturization of the dielectric spectroscopic sensor system, insufficient heat dissipation design, or a sudden change in environmental temperature, which can reduce the stability of the measurement signal.The dielectric spectroscopic sensor of the present invention is characterized in that the output signal of the sensor is modeled as a signal consisting of a reference signal, an initial value of the measurement signal, and the amount of change from the initial value, and the measurement signal is corrected using a transfer model in which the amount of change in the reference signal calculated separately and the amount of change in the measurement signal inside the dielectric spectroscopic sensor system are the input and output.This feature makes it possible to suppress systematic errors caused by environmental changes and accurately measure the dielectric constant of a sample.

[0018] The basic configuration of the dielectric spectroscopy sensor of the present invention is shown in Fig. 1. The dielectric spectroscopy sensor 1 shown in Fig. 1 has a sensing unit 11, a signal separator 12, an oscillator 13, and a receiver 14.

[0019] The oscillator 13 uses a voltage controlled oscillator, a crystal oscillator, a phase locked loop oscillator, or an oscillator using OE conversion (optical / electrical conversion) with a photomixer, and transmits electromagnetic waves.

[0020] The receiver 14 uses a diode detector, an RMS detector, a mixer, an IQ mixer, or the like to measure the voltage, power, phase, and the like of the radio waves that reach the receiver.

[0021] The signal separator 12 separates the electromagnetic wave transmitted from the oscillator 13 into a reference signal and a measurement signal, and the separation is performed using a directional coupler, a circulator, a switch, or the like. In this case, as in the dielectric spectroscopy sensor 2 shown in FIG. 2A , the reference signal and the measurement signal may be separated using two directional couplers, a first directional coupler 12 a and a second directional coupler 12 b, and two receivers, a first receiver 14 a and a second receiver 14 b. Alternatively, as in the dielectric spectroscopy sensor 3 shown in FIG. 2B , the signal input to one receiver 14 may be separated by switching the propagation path using an RF switch 16, reflecting the reference signal with a reflector 17, and connecting the measurement signal to the sensing unit 11.

[0022] The measurement signal is reflected by the sensing unit 11 and input to the receiver 14. At this time, by placing a measurement target substance 15 (MUT) in the sensing unit, the reflection characteristics of the sensing unit 11 change depending on the dielectric constant of the MUT, and the dielectric constant of the measurement target sample 15 can be measured from the change in reflection characteristics.

[0023] The oscillator 13, receiver 14, and general-purpose measuring instruments such as a signal generator and spectrum analyzer may be used, or a vector network analyzer in which the oscillator 13, signal separator 12, and receiver 14 are integrated may be used.

[0024] At each measurement time, the reference signal and measurement signal are measured, and the output signal is calculated over time from the values ​​of each signal, thereby enabling the acquisition of time-series data of the MUT. Figure 6 shows a flowchart of output signal measurement using the conventional method. The sensor output signal is calculated as shown in Equation (1).

[0025]

[0026] Here, s(ω,t), vdut(ω,t), and vref(ω,t) are the output signal, the measurement signal, and the reference signal, respectively, and ω and t are the frequency of the electromagnetic wave and the measurement time.

[0027] The signal used may be an effective value or instantaneous value of a voltage obtained by a receiver, or a DC voltage after down-conversion by a mixer, etc. If the receiver is also capable of measuring phase characteristics, calculations may be performed using characteristic parameters in a two-terminal circuit network, such as S parameters, V parameters, I parameters, F parameters, and T parameters.

[0028] By acquiring the output signal using equation (1), the influence of fluctuations due to the output of the radio waves transmitted from the oscillator is suppressed, and changes in the characteristics of the sensing unit 11, i.e., changes in the dielectric constant of the material to be measured 15 (MUT), can be measured as time-series data. However, equation (1) assumes use in a constant temperature and surrounding environment, such as a research institute. When considering sensing applications for collecting time-series data, there are issues with the influence of differences in characteristics due to differences in the radio wave propagation paths of vdut(ω,t) and vref(ω,t) caused by a decrease in heat capacity due to system miniaturization, insufficient heat dissipation design, or sudden changes in environmental temperature, resulting in reduced stability when acquiring time-series data. In response to this issue, the present invention is characterized by modeling the sensor output signal as a signal consisting of a reference signal, an initial value of the measurement signal, and the amount of change from the initial value, and correcting the measurement signal using a transfer model in which the separately calculated amount of change in the reference signal and the amount of change in the measurement signal within the system are used as inputs and outputs. This suppresses systematic errors resulting from differences in the propagation paths of the reference signal and the measurement signal, thereby obtaining a stable sensor output signal. The flowcharts of the present invention are shown in Figure 3A (flowchart of pre-measurement) and Figure 3B (flowchart of MUT measurement). In the present invention, the output signal of the sensor is expressed by equation (2).

[0029]

[0030] Here, t0 is the reference time, and Δvn is the amount of change in the reference signal and the measurement signal from the reference time t0 within the system.

[0031] Although the reference signal is generally considered to be unaffected by the characteristic fluctuation of the MUT, the signal of the sensor part of the measurement signal changes due to the characteristic fluctuation of the MUT, and therefore the characteristic fluctuation of the sensor part is defined as Δsensor(t). That is, Δvdut(t) is calculated using equation (3) when Δsensor(t) = 1.

[0032] The difference in the amount of change between the reference signal and the measurement signal is thought to change due to factors such as the temperature characteristics of the propagation path within the system, and since the propagation paths of each signal are constant, the amount of change Δvdut(t) of the measurement signal within the system can be expressed as an input / output system with the amount of change Δvref(t) of the reference signal as an input. That is, Δvdut(t) can be expressed by equation (4).

[0033]

[0034] f(x) is an input / output model. A linear regression model may be used as f(x), or an input / output model such as an ARX model or an ALMA model may be used. This input / output model is identified before the MUT measurement by a preliminary measurement such as warming up the engine, where Δsensor(t) = 1, for example, and measurements are taken while changing the environmental temperature in a state where the sensing unit is in contact only with air and the material properties do not change. This makes it possible to estimate the internal fluctuation Δvdut(t) of the system from Δvref(t). When measuring the MUT, equation (5) is used as the final sensor output.

[0035]

[0036] By correcting the influence of fluctuations within the system, it becomes possible to stably measure signal fluctuations in the sensing unit.

[0037] In addition, the measurement value of high-frequency measurement is generally expressed as the ratio of the measurement signal to the reference signal. However, as shown in (6), for example, the effects of reflections and currency losses within the measuring instrument can be calculated in advance as an error term, and the output signal can be corrected using the error term.

[0038]

[0039] where Sa(ω,τ) and Sm(ω,τ) are the true value of the measurement signal and the raw signal value obtained by the measuring instrument, respectively, and ed, er, and es are error terms derived from the measurement system. In such cases, it is possible to correct for internal system fluctuations by substituting equation (5) into the actual measurement value signal Sm(ω,τ).

[0040] Figure 4 shows a dielectric spectroscopy sensor 4 according to one embodiment of the present invention. The transmitter 13 is a PLL 13a, the receiver 14 is a mixer 14c, the signal separator 12 is a circulator 18 and an RF switch 16, and the sensing unit 11 is a coaxial probe 11a. The reference signal is reflected by the RF switch 16 to a reflector 17 inside the system, and the measurement signal is reflected from the coaxial probe 11a. As a preliminary measurement, the system was continuously operated in a thermostatic chamber without the sample 15 to be measured. An input / output model was created from the reference signal and measurement signal, which change due to heat generation from the system. Then, the sample 15, whose characteristics change little over time, was installed and time-series data was acquired. Figure 5 shows a comparison of output results at a frequency of 7 GHz. In the conventional method, the signal around 180 minutes is affected by the difference in output fluctuation between the reference signal and the measurement signal. In contrast, the present invention successfully suppresses signal fluctuations around 180 minutes, demonstrating stable MUT measurement.

[0041] The dielectric spectroscopy sensor of the present invention is suitable for measuring the complex permittivity of a measurement target sample having a four-layer structure consisting of a stratum corneum, an epidermis, a dermis, and a subcutaneous tissue layer stacked in this order, from the stratum corneum side. The sensing unit 11 is preferably configured to attenuate the electric field strength of a leakage electric field from the sensing unit 11 by 95% or more, and preferably by 99% or more, at the interface between the dermis and the subcutaneous tissue layer at the measurement site of the measurement target sample. The sensing unit 11 preferably uses a copper axial probe having a coaxial structure including an outer conductor, an inner conductor, a dielectric, and a high-frequency coaxial connector. In this case, the attenuation of the electric field strength at the interface between the dermis and the subcutaneous tissue layer can be controlled by adjusting the aperture diameter of the coaxial probe.

[0042] The dielectric spectroscopic sensor of the present invention can be used as a device for measuring the dielectric constant of a solution present in a human or animal, and as a device for measuring the dielectric constant of a solution or semi-solid material collected from a human or animal.

[0043] REFERENCE SIGNS LIST 1, 2, 3, 4... Dielectric spectroscopy sensor 11... Sensing unit 11a... Coaxial probe 12... Signal separator 12a... First directional coupler 12b... Second directional coupler 13... Transmitter 13a... PLL 14... Receiver 14a... First receiver 14b... Second receiver 14c... Mixer 15... Measurement target substance 16... RF switch 17... Reflector 18... Circulator 19... Coaxial probe

Claims

1. A dielectric spectroscopy sensor having a sensing section, a signal separation section, an oscillator, and a receiver.

2. The dielectric spectroscopy sensor according to claim 1, wherein the sensing portion is a copper axial probe having a coaxial structure consisting of an outer conductor, an inner conductor, a dielectric, and a high frequency coaxial connector.

3. A dielectric spectroscopy sensor for measuring the complex dielectric constant of a measurement target sample having a four-layer structure consisting of a stratum corneum, an epidermis layer, a dermis layer, and a subcutaneous tissue layer stacked in that order, from the side of the stratum corneum, characterized in that the dielectric spectroscopy sensor described in claim 2 is structured to attenuate the electric field strength of the leakage electric field from the sensing unit by 95% or more at the interface between the dermis layer and the subcutaneous tissue layer at the measurement site of the measurement target sample.

4. A method for measuring the complex dielectric constant of a measurement target sample using a dielectric spectroscopy sensor according to any one of claims 1 to 3, comprising: a pre-measurement step of creating an input / output model of a sensing unit without installing the measurement target sample; and a measurement target sample measurement step of installing the measurement target sample and acquiring time-series data.

5. A method for measuring a complex dielectric constant as set forth in claim 4, wherein, in the pre-measurement step, the output signal of the sensing unit is modeled as a signal consisting of a reference signal, an initial value of the measurement signal, and an amount of change from the initial value, and, in the measurement target sample measurement step, the measurement signal is corrected using a transfer model in which the amount of change in the reference signal and the amount of change in the measurement signal are input and output.

Images