Patent Grant
10732141
This invention relates to the sensing and identification of gasses by an electrochemical cell in conjunction with a sensing circuit and, in particular, to techniques for reducing flexure of a working electrode and an improved biasing circuit.
Given the dramatic changes in the earth's atmosphere, precipitated by industrialization and natural sources, as well as the dramatically increasing number of household and urban pollution sources, the need for accurate and continuous air quality monitoring has become necessary to both identify the sources and warn consumers of impending danger. Tantamount to making real-time monitoring and exposure assessment a reality is the ability to deliver, low cost, small form factor, and low power devices which can be integrated into the broadest range of platforms and applications.
There are multiple methods of sensing distinct low density materials such as gasses. Common methods include gas chromatography, nondispersive infrared spectroscopy (NDIR), the use of metal oxide sensors, the use chemiresistors, and the use of electrochemical sensors. The present invention pertains to electrochemical sensors. The principle of operation of an electrochemical sensor is well known and is summarized in the following overview: http://www.spec-sensors.com/wp-content/uploads/2016/05/SPEC-Sensor-Operation-Overview.pdf, incorporated herein by reference.
Basically, in an electrochemical sensor, a porous sensor electrode (also known as a working electrode) contacts a suitable electrolyte. The gas permeates the electrode and contacts the electrolyte. The sensor electrode typically comprises a catalytic metal that reacts with the target gas and electrolyte to release or accept electrons, which creates a characteristic current in the electrolyte when the electrode is properly biased and when used in conjunction with an appropriate counter-electrode. The current is generally proportional to the amount of target gas contacting the sensor electrode. By using a sensor electrode material and bias that is appropriate for the particular gas to be detected, the concentration of the target gas in the ambient atmosphere can be determined from the sensing current.
The sensitivity of an electrochemical cell to a particular gas may be impacted by the application of a bias voltage to that cell. Therefore, by applying a set of different biases to an electrochemical cell, and comparing the recorded signal to a library of signals (in a look-up table) corresponding to those signals characteristic of individual known gasses, it is possible to ascertain the presence of, differentiate between, and quantify the occurrence of multiple gasses in the environment of the sensor. By continuously and quickly ramping the bias applied to the cell, a single electrochemical cell may rapidly differentiate between multiple gasses in its environment. Identification of an analyte in a controlled environment, such as that of a laboratory, via a potential ramping scheme is known in the art as “voltammetry”. However, electrochemical sensing systems known in the art are limited with respect to the speed at which voltammetry may run due to a combination of the electrochemical cells having large resistance-capaciotance (RC) time constants and the drive electronics having long settling times, resulting in voltammetry measurements taking up to 20 minutes or even more. Ramping the bias voltage at only a few millivolts per second is typical.
In the event that such a measurement is performed in a highly controlled environment, such a long data collection time is of no consequence since the test conditions may be controlled over that timeframe. However, outside of a controlled laboratory environment, the ambient conditions of the cell may change significantly over the course of such an extended period of data collection. For example, in an everyday consumer use case, the ambient relative humidity, the ambient temperature, and the ambient mixture and concentrations of analyte gasses may change significantly over such an extended data gathering period. All these factors combined with electronic drift inherent to running voltammetry on the electrochemical cell render analysis of the data meaningless. Further, if the gas sensor is intended detect dangerous gasses, the long delay time may result in a harmful effect.
Accordingly, for such a consumer application, an electrochemical gas sensor is required having small form factor, and hence a small time constant, such as described in the Applicant's U.S. patent application Ser. No. 15/598,228, in conjunction with a novel drive circuit enabling short settling times. Such a system enables voltammetry to be performed in a period of a few seconds or less. Over such a timescale, the ambient conditions in a majority of uncontrolled consumer environments in which the cell may be present would be essentially invariant, resulting in the ability to perform voltammetry in an essentially uncontrolled environment whilst providing data of sufficient quality to enable accurate electrochemical analysis.
The varying bias voltage may be sinusoidal or have another waveform. It is common to supply the variable bias voltage via a digital-to-analog converter (DAC) that outputs a varying voltage. Such an output contains discrete steps due to the quantized nature of the DAC output, and thus there is a high dv/dt. The rate at which the bias may be ramped in an electrochemical cell is limited, among other factors, by the capacitance of the cell, and the presence of current spikes occurring within the cell during the step-wise application of the bias ramp to the cell. The impact of the capacitive nature of the cell on the current spikes occurring during a voltage step is described by the equation i=C dv/dt. Excessive transient currents can lead to damage of the cell. In certain schemes, accurate measurement of the cell requires for the transient currents to have mostly decayed so that the cell output is measured while in its steady-state condition. Accordingly, a settling time typically occurs between the point at which a voltage step is applied to the cell and the point at which the current generated at the working electrode of the cell is measured. This limits the rate at which a bias sweep of the cell may be applied. Minimizing the capacitance of the cell and occurrence of current spikes within it allows for maximizing the rate at which voltammetry may be performed, hence maximizing performance of the sensor.
Electronic noise is also an important issue in electrochemical sensors which affects accuracy and speed achievable by the sensors. Electrochemical sensors are known for picking up 60 Hz and a variety of RF noise due to the large electrodes and electrolytes. Commercial systems can minimize the noise through shielded housings around the sensors, powering the sensor using a battery, and using signal processing to filter out the noise. However, noisy environments still generally limit the accuracy of the electrochemical systems and size of the systems due to the extra electronics needed to minimize the noise. This impacts the rate at which the ramps of the voltammetry can be executed at due to higher current spikes and longer settling times which can occur during step transitions due to the additional circuit elements, such as resistors and capacitors, which need to be added to mitigate noise.
Additionally, any flexure of the working electrode changes its electrical characteristics. Flexure may occur due to gas pressure fluctuations. Pressure built up due to small pore size in the electrode can also change the position/formation of the three phase interface inside the electrode which can impact the sensitivity of the sensor. The three phase interface is a critical aspect of a gas diffusion electrode and is formed at the conjunction interface of a gas, solid, and liquid (i.e., gas/electrode/electrolyte). A change in pressure could influence the density of the three-phase interface and where the three-phase interface is formed.
Accordingly, what is needed is an electrochemical sensor system for gasses that employs a varying bias signal that does not have high dv/dt characteristics. What is also needed is an electrochemical sensor that is less susceptible to noise. Additionally, an electrochemical sensor where the working electrode does not substantially flex due to changing gas pressures is required for stability in changing environments. Further, a control method is desired for use with a low capacitance gas sensor that can quickly detect a variety of different gasses in a short time.
A sweeping bias voltage generator is disclosed that applies a varying bias voltage to an electrochemical sensor, where the bias voltage may be sinusoidal or ramping without any steps, so is continuous. Therefore, current spikes are negligible due to i=C dv/dt. As a result, little or no settling time is needed, and the sensor can have a very rapid response to changing ambient gasses.
To minimize flexure of the working electrode, mechanical support features of the working electrode are disclosed that prevent the working electrode being distorted due to changing gas pressures. Additionally, a sensor structure is described that equalizes the gas pressure on the opposing surfaces of the working electrode to prevent flexure.
To minimize electronic noise, conductors and metal shielding are embedded in the walls of the sensor body (e.g., molded ceramic), which prevent impact of noise to the system. Embedding the conductors and shielding protects the metal and prevents corrosion due to electrolyte (e.g., acid) contact and environmental exposure. This also allows for simplification of circuit design, allows for minimal integrated circuit size (so the circuit can be integrated with the sensor), and allows for simplification of post-processing due to less noise.
Uses of the sensor module include detection of air quality (e.g., carbon monoxide), gas exposure control, toxic gas detection, breath analysis, feedback in industrial processes, etc.
Other embodiments and advantages are described.
Elements that are the same or equivalent in the various figures are labeled with the same numeral.
Basically, the cell 12 contains a cavity 16 containing an electrolyte that is in contact with a working electrode (WE) 18, a counter electrode (CE) 20, and a reference electrode (RE) 22. Additional electrodes can be added as additional working and counter electrodes for different variants of the cell for gas detection. The RE 22 provides a means by which a reference potential may be applied to the cell.
The body of the cell 12 includes a gas opening to allow diffusion of the gas or atmosphere being sensed into the porous WE 18. In certain embodiments, the opening is partially or fully filled with a porous material which allows gas to diffuse into the WE 18 but blocks the liquid or paste-like electrolyte from exiting the cavity.
The electrodes 18/20/22 may comprise an electrically conducting material, such as carbon, and a catalyst such as ruthenium, copper, gold, silver, platinum, iron, ruthenium, nickel, palladium, cobalt, rhodium, iridium, osmium, vanadium, or any other suitable transition metal and their alloys. The catalyst is selected so as to react with one or more particular gases. The electrodes 18/20/22 may be partially permeable to both the electrolyte and the gas to be detected so that the electrochemical reaction may occur within the body of the electrodes 18/20/22.
The electrolyte may comprise an ionic material such as an acid. The electrolyte may be viscous such as a gel, or may be a polymer infused with an organic or inorganic acid.
The bias circuit 14 applies electrical potentials between the WE 18, CE 20, and RE 22. A sensing circuit 26 senses electrical currents passing between the WE 18, CE 20, and RE 22, and reports on the sensed signals. The sensing circuit 26 may comprise a trans-impedance amplifier (TIA) in conjunction with an analog-to-digital converter (ADC) capable of converting the sensed signal from the working electrode into a digital representation. The digital signals are processed by a microprocessor on which algorithms may be stored and executed enabling, for example, reporting out of calibrated gas concentrations. The sensing circuit 26 and its connections to the electrodes may be conventional.
In an example, when a toxic gas such as carbon monoxide (CO) comes in contact with the WE 18 (the sensing electrode), oxidation of CO gas will occur on the WE 18 through chemical reaction with water molecules in the air. Connecting the WE 18 and the CE 20 through a potentiostat circuit will allow protons (H+) generated on the WE 18 to flow toward the CE 20 through the electrolyte (an ion conductor). In addition, generated electrons move to the CE 20 through the potentiostat circuit. A reaction with oxygen will occur on the CE 20 to reform water. By measuring the level of current between the WE 18 and the CE 20, the electrochemical cell 12 can detect the concentration of the target gas (e.g., CO) in the air. This is usually done at a single bias point. Different types of electrodes and different bias voltages traditionally have been used to detect different types of gases.
A digital control signal is applied to the input of a current source 28 that converts the digital code to an analog current. The digital code may be constant or vary. A varying digital code may cause the current source 28 to output a substantially sinusoidal current. The current is applied, via an input resistor 30, to the inverting input of an op amp 32. The op amp 32 is connected as an integrator due to the feedback capacitor 34 being connected between the output and the inverting input. The non-inverting input is connected to ground. The feedback operates to cause the potential between the inverting and non-inverting inputs to approximately equal zero volts. The integration of the input current results in a continuous signal at the output of the op amp 32, even if the output of the current source 28 may be stepped. The output of the op amp 32 varies continuously as a function of the current output by the current source 28. Supplying a positive current by the current source 28 decreases the potential at the output of the op amp 32, and providing a negative current increases the potential at the output. Providing zero current maintains a constant output voltage. The output of the op amp 32 is a reverse integral (180° phase shift of the integral) of the input current. Step functions at the input do not cause discontinuities in the output voltage of the op amp 32.
The continuous output of the op amp 32 is then applied to the inverting input of an op amp 38, via a resistor 40, whose output is coupled to the CE 20. The non-inverting input of the op amp 38 is coupled to ground, and the feedback loop tries to keep the potential between the inverting and non-inverting inputs at zero volts.
The RE 22 is coupled to the non-inverting input of another op amp 44, whose inverting input is coupled to its output. The output of the op amp 44 is coupled to the inverting input of the op amp 38 via a resistor 46, the non-inverting input of the op amp 38 is connected to ground, and the feedback loop tries to keep the potential between RE 22 and the inverting input of op amp 38 at zero volts.
The op amp 44 is a unity gain amplifier. To the first order, the op amp 44 and its feedback loop have no impact on the circuit. In reality, op amps are non-ideal and have input leakage current. Input leakage current is typically larger into the inverting input of an op amp than it is into the non-inverting input. The role of op amp 44 is to reduce the input leakage current drawn from the RE 22. The RE 22 ideally applies a bias to the electrochemical cell, but does not draw a current. If it does so, it impacts the performance and state of the cell. The potentiostat circuit (based on op amp 38) works by changing the bias and current being driven through the CE 20 such that: (i) the RE 22 is held at the bias potential being applied at the inverting input to op amp 38, and (ii) the current required to support the electrochemical reaction occurring within the cell between the WE 18 and the CE 20 is provided to the cell.
Accordingly, the RE 22 of the electrochemical cell 12 is properly biased at a varying (e.g., ramping) continuous signal for operating the gas sensor. The RE 22 can be biased at any voltage. The circuit allows the potential of the CE 20 to vary as required to enable the required current to flow between the CE 20 and the WE 18. This is known as voltammetry. Suppression of discontinuities in the applied bias enables the sensor to be accurately controlled with minimized or completely eliminated transient signals that could negatively affect performance. The ramping may be performed rapidly to quickly detect a range of gasses in an uncontrolled environment. No settling time is needed. If a dangerous gas is detected, circuitry may be used to generate an alarm or turn off the source of the gas.
A capacitor (not shown) may optionally be coupled between the RE 22 and the CE 20 to filter high frequency signals.
The op amp 44 may be deleted by connecting a resistor divider between the RE 22 and the inverting input of the op amp 38, with a capacitor connected between the resistor nodes and the CE 20; or in parallel with the resistor divider.
There are various ways to detect the reaction of the sensor to the target gasses. The gas diffuses into the porous WE 18 where it is oxidized or reduced. This electrochemical reaction results in an electric current that passes through the external sense circuit 26. The bias circuit 14 maintains a bias voltage between the WE 18 and RE 22. At the CE 20, a counter half reaction occurs such that the electrons generated or consumed at the WE 18 equals the electrons generated or consumed at the CE 20. Additionally, should the CE 20 be exposed to the environment instead of the WE 18, the CE 20 will run the forward reaction while the WE 18 will run the reverse reaction. In gas sensing, it is rare for the CE 20 to face the environment, though such a configuration is known to be possible.
The magnitude of the current is controlled by how much of the target gas is reacted at the WE 18 so that the output from the sensor is linearly proportional to the gas concentration. Accordingly, the sense circuit 26 measures the current and may perform an analysis to associate the current to the bias voltage and to the target gas being presently measured.
Since the bias circuit 14 is current-controlled rather than voltage controlled, there are no current spikes due to capacitance (so no transient settling time) and the sensor can operate to rapidly detect the presence of various target gasses.
In one embodiment, the current source 28 and the electrical components are selected to vary the applied bias between 1V and −1V. The bias voltages may also be greater than 1V, such as up to 2.5V, and less than −1V, such as down to −2.5V. Some target gases, such as ethylene oxide, may require even higher bias levels for its oxidation depending on the electrode/electrolyte combination chosen, while some gases, such as carbon monoxide, only require a relatively low bias level. Some gases even oxidize at multiple bias levels. The two gases may be differentiated based on the varying bias levels needed to react the gasses at the WE 18. In more convoluted cases where the two gases have very similar bias levels, a more detailed voltammogram can be created by sweeping a number of bias points.
The bias circuit 14 and sensor 26 may be formed as a single integrated circuit.
The entire sensor may use a footprint of less than 10 mm×10 mm.
A novel drive scheme is also disclosed. To identify a single gas, or multiple mixed gasses, multiple biases are applied to the RE 22. Since different reactions are promoted or suppressed at different applied biases, it is possible to identify a particular gas by looking at the “spectrum” of the gas registered during the application of multiple bias points.
In the case of multiple gasses being present, in theory, the recorded spectrum of the gas mixture comprises the linear summation of the spectra of the individual component gasses, scaled by their relative concentrations and sensitivities. By comparing the captured spectrum to a reference library of gas spectra, it should be possible to identify the multiple gasses present.
Similar gasses will have similar peaks in the recorded spectra. For example, methanol and ethanol comprise similar chemical bonds, hence may have similar spectra which may have substantial overlap during any one particular bias sweep. Whereas it is known in the art that changing the waveform of an applied bias to an analyte may change the position of the recorded peaks, it is typically difficult to practically implement due to the limited bias ramp rates which may be successfully applied to typical electrochemical systems. This principle is known for liquid spectroscopy applications, though to the best of the inventors' knowledge, not yet demonstrated for electrochemical spectroscopy for gases due to the limited hardware. Accordingly, in the case of an electrochemical cell enabling fast ramping of an applied bias, it is possible to greatly further differentiate between the two or more similar gasses by changing the applied bias waveform and especially the speed of bias ramp, dV/dt over and above that capability in standard electrochemical systems.
In addition to being able to determine the presence of chemicals, voltammetry enables the determination of the relative humidity, the temperature, and the ionic state of the electrochemical cell. These factors are known to impact the sensitivity of electrochemical cells to gasses the cell is exposed to. In standards and the current state-of-the-art, often one or more of the factors are bundled into the generic parameter of drift resultant of limited cell hardware and ability to generate the necessary waveforms for proper compensation. In the case of voltammetry performed in an uncontrolled environment but in the order of a few seconds or less and in which the ambient and internal state of the cell are therefore essentially constant, it is possible to determine the relative humidity, temperature and ionic state of the electrochemical cell at the time at which the cell is also being exposed to a gas. This data can therefore be used to compensate for the impact of these factors on the sensitivity of the cell to a gas, thereby improving the accuracy with which the system can identify and quantify gasses.
Data collected during calibration of the sensor in a controlled environment is stored in a local look-up table, such as located in the same integrated circuit as the bias circuit and microprocessor. Sensor characteristics (bias levels, sensor current, etc.) corresponding to known target gasses and environmental conditions are also stored. Changes in the detected environment relative to the calibration data are then used to determine the concentration of targeted gasses in the environment and other environmental factors as the bias voltage is swept. The gas readings may be compensated by the detected humidity, temperature, etc. to determine the concentration. Rather than the bias being swept over the prior art cycle time of, for example, 20 minutes or more, the cycle time using the present invention may be less than one minute. This enables the rapid detection of a variety of gasses during a stable environment.
The waveforms corresponding to the detection of target gasses in the environment may comprise overlapping gas signals. A microprocessor then differentiates the different gasses corresponding to the overlapping gas signals based on stored data and algorithms. The microprocessor may perform as an inference engine with a set of equations that identify gasses present.
Another feature of the sensor relates to techniques that reduce or eliminate any flexure of the WE 18. Flexure of the WE 18 in the presence of changing conditions will change the response of the WE 18 to the target gasses. Calibration is initially performed under ideal conditions, and the sensor readings will be skewed if they occur while the WE 18 is flexed relative to it calibration state. One cause of flexure is a gas pressure difference on the front and back surfaces of the WE 18.
The adhesive 63 also acts as a gas seal. The adhesive 63 may be epoxy, silicone, acrylate, or other adhesive, and may be conducting, for example via the inclusion of conducting metal, carbon, or other particles. If the adhesive 63 is conductive, it may serve to electrically connect the WE 60 to conductors that lead to metal pads for bonding to a printed circuit board. The entire front surface of the WE 60 except for the gas opening 64 area is affixed to the rigid body 62, so flexure of the WE 60 due to temperature fluctuations and pressure differentials is prevented. The WE 60 is also formed to be rigid to prevent flexure under normal operating conditions. The gas diffuses laterally and into the WE 60, so it is important that the adhesive 63 does not substantially diffuse into the WE 60. The adhesive 63 may be applied by printing or stamping onto the body 62. The electrolyte may be a liquid or gel contacting the back surface of the WE 60 so does not provide substantial mechanical support to the WE 60.
The flexure of the WE 70 and CE 72 is inhibited by a fairly rigid electrolyte layer 80 that contacts the back surfaces of the WE 70 and the CE 72. The electrolyte layer 80 may be a rigid but porous polymer that is infused with an ionic material, such as sulfuric acid or phosphoric acid. The electrolyte layer 80 is sufficiently thick and rigid to prevent flexure of the WE 70 and CE 72 under normal operating conditions. Charges are free to move between the WE 70 and CE 72 through the electrolyte layer 80 to create the current signature of the target gas under the proper biasing conditions. The electrolyte layer 80 essentially forms an ion bridge between the electrodes. The electrolyte layer 80 may diffuse up to 10 microns or more into the porous WE 70 and CE 72 for good contact. Note that there may be a pressure differential between the gas on opposing surfaces of the WE 70, and such a pressure differential is not sufficient to flex the WE 70 due to the rigidity of the electrolyte layer 80.
In a variation of
Sealing the sides of the electrodes also prevents fraying of the electrodes. The sealant may be any suitable polymer.
In all embodiments, there may be separate cavities in the body 88 which contain at least a working electrode and a counter electrode. Each cavity and associated electrodes is independent and may be biased for a particular gas. All the cavities receive the ambient gas and each cavity may be dedicated to a particular target gas. A current sensor and microprocessor may use multiplexing to detect the various gasses sensed by the different cavities.
The sensing current has a small signal-to-noise ratio, so any noise on the conductors 104 is significant. The conductors 104 are shielded from noise by a metal shield layer 108, which may be grounded or floating. The shield layer 108 is encased in the body 105. The shield layer 108 may be located around the top, bottom, and sides of the body 105, depending on the shielding requirements, to substantially surround the conductors 104. The shield layer 108 greatly reduces external noise being coupled to the bias signals and sensor current signals.
By embedding the conductors 104 within the body 105, they are protected from the electrolyte (e.g., a corrosive acid) and other possible damage. This allows the conductors 104 to be formed of virtually any type of highly conductive metal and allows the conductors 104 to be efficiently routed within the body 105.
While particular embodiments of the present invention have been shown and described, it will be obvious to those skilled in the art that changes and modifications may be made without departing from this invention in its broader aspects and, therefore, the appended claims are to encompass within their scope all such changes and modifications that are within the true spirit and scope of this invention.
This application is based on and claims priority from U.S. provisional patent application Ser. No. 62/459,597, filed on Feb. 15, 2017, by Jim Chih-Min Cheng et al., assigned to the present assignee and incorporated herein by reference.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4894138 | Gambert et al. | Jan 1990 | A |
| 4961834 | Kuhn et al. | Oct 1990 | A |
| 6376124 | Dodgson et al. | Apr 2002 | B1 |
| 8815065 | Yamamoto | Aug 2014 | B2 |
| 9989492 | Cheng et al. | Jun 2018 | B1 |
| 20050034987 | Zhou et al. | Feb 2005 | A1 |
| 20140311905 | Stetter | Oct 2014 | A1 |
| 20150226585 | Yang | Aug 2015 | A1 |
| 20150241375 | Merz et al. | Aug 2015 | A1 |
| 20150346138 | Allen et al. | Dec 2015 | A1 |
| 20160041114 | Neel et al. | Feb 2016 | A1 |
| 20160073930 | Stetter | Mar 2016 | A1 |
| 20160178565 | Chapples et al. | Jun 2016 | A1 |
| 20160349205 | Stetter et al. | Dec 2016 | A1 |
| 20170336343 | Bhat et al. | Nov 2017 | A1 |
| Number | Date | Country |
|---|---|---|
| 2007149471 | Dec 2007 | WO |
| Entry |
|---|
| EPO as ISA for PCT/US2018/018306, “International Search Report and Written Opinion”, 20 pages. |
| EPO as ISA PCT/US2017/033649, “International Search Report and Written Opinion”, dated Oct. 2, 2017, 25 pages. |
| Stetter et al., “Amperometric Gas Sensors—A Review”, Chemical Reviews, vol. 108, No. 2, Jan. 18, 2008, pp. 352-366, 15 pages. |
| SGX Sensortech, “Electrochemical Sensors Application Note 2 Design of Electronics for Electrochemical Gas Sensors”, pp. 1-5, England. |
| Dryden et al., “DStat: A Versatile, Open-Source Potentiostat for Electroanalysis and Integration”, Oct. 28, 2015, 11 pages, <http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0140349>. |
| Electronics Tutorials, “The Integrator Amplifier”, 6 pages, <http://www.electronics-tutorials.ws/opamp/ppamp_6.html>. |
| Unpublished U.S. Appl. No. 62/620,963, “Low Impedance Sensor for Low Density Material”, filed Jan. 23, 2018. |
| Unpublished U.S. Appl. No. 62/620,685, “Gas Diffusion Electrode and Membrane Electrode Assembly”, filed Jan. 23, 2018. |
| Number | Date | Country | |
|---|---|---|---|
| 20180231494 A1 | Aug 2018 | US |
| Number | Date | Country | |
|---|---|---|---|
| 62459597 | Feb 2017 | US |
