Reliable, fast-response, low-cost refrigerant leakage detection sensor

Abstract

Embodiments relate to sensors, sensor assemblies, and methods of making sensor coatings where the sensors comprise ACM or SAW based sensors with applied humidity tolerant coatings that target adsorption and desorption of selected refrigerant gas molecules via MOF coatings and particularly A2L and A3 refrigerants such as R32 and R290. In some embodiments the MOF coating include ZIF-67 or ZIF-8

Description

RELATED APPLICATIONS

[0001] This application claims benefit of U.S. Patent Application No. 63 / 710,968 filed Oct. 23, 2024. This referenced application is hereby incorporated herein by reference in its entirety including any appendices filed therewith and any other materials incorporated therein by reference.STATEMENT AS TO RIGHTS TO INVENTIONS MADE UNDER FEDERALLY-SPONSORED RESEARCH AND DEVELOPMENT

[0002] This invention was made with Government support under Contract DE-AC0576RL01830 awarded by the U.S. Department of Energy. The Government has certain rights in the invention.BACKGROUND OF THE INVENTIONField of the Invention

[0003] The invention generally relates to the field of refrigerant detection and more specifically to refrigerant detection using a QCM / MOF sensor or an SAW / MOF sensor, and even more particularly to sensors where the MOF comprises a targeted humidity tolerant MOF (HT-MOF) and wherein the gas being detected is an A2L or A3 refrigerant.Background Information

[0004] To lower the impact of refrigerant emissions and to comply with the AIM Act, HVAC equipment utilizing mildly flammable A2L refrigerants are being installed in buildings and a further transition to high flammability A3 ultra-low-GWP may be necessary. Refrigerant leak detection systems play a crucial role in safe transition to these new refrigerants.

[0005] Robust, reliable, and low-cost sensors are required to enable a safe refrigerant transition. If lower cost sensors are available, safety can be increased by being able to include more sensors in areas prone to refrigerant leakage. A limited number of sensors are available for A2L / A3 refrigerants that meet various safety standards such as UL, CSA, ASHRAE 15, etc. Safety standards of interest to sensor solutions include the following with only a few sensor solutions useable by manufacturers meeting all requirements of these safety standards:

[0006] 1) IEC 60335-2-40 Edition 6 (January 2018),

[0007] 2) UL / CSA 60335-2-40 (November 2019),

[0008] 3) ASHRAE Standard 15-2019 [4],

[0009] 4) ASHRAE proposed Standard 15.2P (Advisory Public Review), and

[0010] 5) JRA Standard 4068T: 2016R

[0011] Existing sensors have been used such as infrared, electrochemical, semiconductor, speed of sound, and Molecular Property Spectrometer technology. SAW based sensors with MOF coating have also been proposed

[0012] Additional information about prior refrigerant detection technology can be found, for example, in

[0013] 1) Reshniak, Viktor, Praveen Cheekatamarla, Vishaldeep Sharma, and Samuel Yana Motta. 2023. “A Review of Sensing Technologies for New, Low Global Warming Potential (GWP), Flammable Refrigerants” Energies 16, no. 18: 6499. https: / / doi.org / 10.3390 / en16186499

[0014] 2) Zheng, Jiao, Zang, Fiona, Yu, Cheng-Nien, Elbel, Stefan. 2021, Refrigerant Detector Characteristics for Use in HVACR Equipment, https: / / www.ahrinet.org / system / files / 2023-08 / AHRTI%209014%2001%20-%20Final%20Report%20-%20Final.pdf

[0015] 3) Wagner, Mark, and Rebecca Ferenchiak. 2017, Leak Detection of A2L Refrigerants in HVACR Equipment, https: / / www.ahrinet.org / system / files / 2023-08 / AHRTI_9009_Final_Report_0.pdf.

[0016] 4) U.S. Pat. No. 11,796,508, published Oct. 24, 2023, by Kunapuli et al., and entitled “Surface Acoustic Wave Sensor for Refrigerant Leakage Detection”, originally published as WO2021 / 041359 on Mar. 4, 2021. This referenced patent is incorporated herein by reference as if set forth in full herein.

[0017] A need remains for lower cost, robust, reliable, and standard meeting sensors for detecting refrigerants.SUMMARY OF THE INVENTION

[0018] It is an object of some embodiments of the invention is to provide humidity tolerant sensors for the detection of A2L and A3 refrigerants.

[0019] It is an object of some embodiments of the invention to provide QCM or SAW sensor capable of target selected refrigerants.

[0020] It is an object of some embodiments of the invention to provide a method for coating QCM or SAW sensors with selective MOF coatings.

[0021] Embodiments of this invention use QCM and an MOF material or SAW and an MOF material for A2L / A3 gas detection.

[0022] Some embodiments are directed to a coating technique for the MOF with polymer binder that shows stability and repeatability.

[0023] Some embodiments are configured so that sensing output is not negatively impacted by high humidity levels or changes in humidity (e.g. incorporate a humidity sensor and are calibrated).

[0024] Some embodiments are directed to QCM / SAW technologies having sensor response times of less than 10s of seconds and in some cases less than 10 seconds in combination with low manufacturing cost.

[0025] In some embodiments, the sensors may be integrated into HVAC equipment units using A2L or A3 gases.

[0026] In some embodiments, sensors are configured to provide detection of A2L or A3 refrigerant gases at different concentration levels. Sensors include Quartz Crystal Microbalance (QCM) technology or Surface Acoustic Wave (SAW) technology along with Metal Organic Framework materials (MOF). The QCM / SAW sensor is coated with an MOF material which adsorb / desorb the target gas. The adsorption amount of the target gas induces a change in operational parameters of the acoustic / vibration signal (e.g., resonance frequency, phase, dissipation) of the sensor. The parameter change is directly correlated to the actual concentration of the target gas present in the vicinity of the QCM sensor or SAW sensor. This sensing technology enables fast responses for refrigerant leak detection.

[0027] In some embodiments, a hydrophobic MOF coating is used that is hydrophobic, has a reversible water isotherm, and it tailored to target adsorption of a specific refrigerant gas of interest. In some embodiments a ZIF-67 MOF coating is placed on the QCM sensor or a SAW sensor and is used to detect an A2L refrigerant R32 (difluoromethane) or R454B (a blend of 68.9% difluoromethane (R-32) and 31.1% 2,3,3,3-tetrafluoropropene (R-1234yf)0. In some embodiments a ZIF-8 MOF coating is placed on the QCM sensor or a SAW sensor and is used to detect an A3 refrigerant R290 (propane).

[0028] In some embodiments, sensors of different embodiments of the invention may provide refrigerant gas concentration determinations based on mass loading using: (1) time of flight determinations, (2) frequency or frequency shift determinations, (3) phase shift determinations, (4) change of amplitude determinations, or (5) a combination of two or more of these specific determination methods.

[0029] In some embodiments, an additional coating may be placed over the MOF material.

[0030] In a first aspect of the invention a targeted sensor for detecting A2L or A3 refrigerants, comprising: (a) a QCM sensor or a SAW sensor, and (b) a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor.

[0031] Numerous variation of the first aspect exist and include, for example: (1) the first aspect wherein the QCM or the SAW sensor includes a QCM sensor; (2) The first aspect wherein the QCM or the SAW sensor includes a SAW sensor; (3) the first aspect or either of the first or second variations thereof wherein the MOF coating comprises a ZIF-67 coating; (4) the third variation of the first aspect wherein the sensor is configured to target detection of an A2L refrigerant; (5) the fourth variation of the first aspect wherein the A2L refrigerant comprises R32; (6) the sixth variation of the first aspect wherein the A2L refrigerant is R32; (7) the sixth variation of the first aspect wherein the A2L refrigerant is R454B; (8) The first aspect or either of the first or second variations thereof wherein the MOF coating comprises a ZIF-8 coating; (9) the eight variation of the first aspect wherein the sensor is configured to target detection of an A3 refrigerant; and (10) the ninth variation of the first aspect wherein the A3 refrigerant comprises R290.

[0032] In a second aspect of the invention a targeted sensor assembly for detecting A2L or A3 refrigerants, includes: (a) a QCM sensor or a SAW sensor; (b) a humidity tolerant MOF coating on the QCM sensor or the SAW sensor; (c) an oscillator triggering QCM sensor resonance or SAW sensor resonance; (d) a microprocessor or microcontroller for receiving information from the QCM sensor or from the SAW sensor, and using the information and / or outputting one or more of frequency, frequency shift, phase shift. amplitude shift, calibrated concentration of a detected targeted refrigerant, and / or other actionable or control information for enabling a safe operating environment around A2L or A3 refrigerants.

[0033] Numerous variations of the first aspect exist and include, for example: (1) the first aspect wherein the QCM sensor or the SAW sensor includes at least one QCM sensor; (2) the first variation of the second aspect wherein the at least one QCM comprises at least two QCM sensors with at least one QCM not including a humidity tolerant coating; (3) the second variation of the second aspect comprising at least one SAW sensor; (4) the third variation of the second aspect wherein the at least one SAW sensor comprises at least two SAW sensors with at least one SAW not including a humidity tolerant coating; (5) the second aspect or any of the first to fourth variations thereof wherein the MOF coating comprises a ZIF-67 coating; (6) the fifth variation of the second aspect wherein the sensor is configured to target detection of an A2L refrigerant; (7) the sixth variation of the second aspect wherein the A2L refrigerant comprises R32; (8) the seventh variation of the second aspect wherein the A2L refrigerant is R32; (9) the eighth variation of the second aspect wherein the A2L refrigerant is R454B; and (10) the second aspect or any of the first to fourth variations thereof wherein the MOF coating comprises a ZIF-8 coating.

[0034] Numerous additional variations of the second aspect of the invention exist, and include, for example: (11) the tenth variation of the second aspect wherein the sensor is configured to target detection of an A3 refrigerant; (12) the ninth variation of the second aspect wherein the A3 refrigerant comprises R290; (13) the second aspect or any of the first to twelfth variations thereof wherein the assembly further comprises at least one environmental sensor selected from the group consisting of: (A) one more temperatures sensors, (B) one or more pressure sensors, and (D) one or more humidity sensors, wherein the at least one environmental sensor provides an input to the microprocessor or microcontroller to allow inclusion or use of such information when providing output; (14) the second aspect or any of the first to thirteenth variations thereof wherein the one or more sensors and electrical components integrated onto a single PCB; (15) the second aspect or any of the first to fourteenth variations thereof wherein the sensor assembly is packaged to fit in a volume of less than 10 cubic inches; (16) the fifteenth variation of the second aspect wherein the sensor assembly is packaged to fit in a volume of less than 5 cubic inches; (17) the sixteenth variation of the second aspect wherein the sensor assembly is packaged to fit in a volume of less than 3 cubic inches; (18) the second aspect or any of the first to seventeenth variations thereof wherein a t90 sensing time is less than 30 seconds; (19) the eighteenth variation of the second aspect wherein the t90 sensing time is less than 10 seconds; and (20) the nineteenth variation of the second aspect wherein the t90 sensing time is less than 7 seconds.

[0035] Numerous further variations of the second aspect of the invention exist and include for example: (21) the twentieth variation of the second aspect wherein the t90 sensing time was no more than 4 seconds; (22) the second aspect or any of the first to twenty-first variations thereof wherein a t10 recovery time is less than 30 seconds; (23) the twenty-second variation of the second aspect wherein the t10 recovery time is less than 10 seconds; (24)) the twenty-third variation of the second aspect wherein the t10 recovery time is less than 5 seconds; (25) the twenty-fourth variation of the second aspect wherein the t10 recovery time is less than 3 seconds; and (26) the twenty-fifth variation of the second aspect wherein the t10 recovery time is no more than 1 seconds

[0036] In a third aspect of the invention a method of fabricating a targeted sensor for detecting A2L or A3 refrigerants, includes: (a) supplying a QCM sensor or a SAW sensor, and (b) applying a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor via spin coating.

[0037] In a fourth aspect of the invention a method of fabricating a targeted sensor for detecting A2L or A3 refrigerants, includes: (a) supplying a QCM sensor or a SAW sensor, and (b) applying a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor via spray coating.

[0038] In a fifth aspect of the invention a method of fabricating a targeted sensor for detecting A2L or A3 refrigerants, includes (a) supplying a QCM sensor or a SAW sensor, and (b) applying a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor via drop coating.

[0039] Numerous variations of the third to fifth aspects exist and include those noted with regard to the first and second aspects, mutatis mutandis, as well as numerous others that will be apparent form the teachings herein.

[0040] Other objects and advantages of various aspects and embodiments of the invention will be apparent to those of skill in the art upon review of the teachings herein. The various aspects and embodiments of the invention, set forth explicitly herein or otherwise ascertained from the teachings herein, may address any one of the above objects alone or in combination, or alternatively may address some other object of the invention ascertained from the teachings herein. It is not intended that any specific aspect or embodiment of the invention (that is explicitly set forth herein or that is ascertained from the teachings herein) necessarily address any of the objects set forth above let alone address all these objects simultaneously, but some aspects and embodiments may address more than one of these objects.BRIEF DESCRIPTION OF THE DRAWINGS

[0041] FIGS. 1A-1 and 1A-2 provide schematic cut side view of example stacked layers forming a QCM with an HT-MOF and a top view of each of those layers. respectively.

[0042] FIG. 1B provides a schematic representation of an example sensor that includes two quartz crystal resonators with one have an adsorption coating and the other not including such a coating.

[0043] FIG. 1C provides a schematic view of another example multi-detector sensor configuration including a reference sensor and two QCM sensors each with a different adsorption coating to allow detection of different gases of interest.

[0044] FIG. 2A provides a schematic illustration of a single QCM sensor assembly where the QCM is coated with the HT-MOF material for selective detection of a particular refrigerant or group of refrigerants while FIG. 2B provides a similar schematic but which incorporates an additional QCM and QCM oscillator.

[0045] FIGS. 3A-3C provide schematic illustrates similar to those of FIGS. 1A-1C but for a surface acoustic wave sensor or detector (SAW sensor) as opposed to a QCM sensor.

[0046] FIGS. 4A-4B provide sensor assembly configurations 410A and 410B, respectively, similar to 210A and 210B of FIGS. 2A and 2B, respectively but with the QCM sensors replaced by SAW sensors.

[0047] FIG. 5 provides a block diagram setting forth features of a targeted humidity tolerant metal organic framework coatings as well as examples of coating materials and refrigerants whose concentrations are detected.

[0048] FIG. 6 provides a block diagram settings forth a first process for creating targeted HT-MOF coatings

[0049] FIG. 7 provides a block diagram setting forth a second process for creating targeted HT-MOF coatings.

[0050] FIG. 8 provides a block diagram setting forth a third process for creating targeted HT-MOF coatings.

[0051] FIG. 9 sets forth a simplified process flow for using the sensors of some embodiments of the present invention.

[0052] FIG. 10 depicts real time response of QCM sensors with different sorbents coated there on under varying R32 refrigerant concentrations.

[0053] FIG. 11 shows plots of the QCM sensor with MIL-101 under dry and humid air conditions (˜60% RH) run for three cycles each with each cycling operating under varying R32 concentrations ranging from 0% to 4.36% v / v.

[0054] FIG. 12A provides a plot of frequency change vs time from a series of experiments using different R32 concentrations for a sorbent coating of carbonized Zn-MOF-74 under both dry and humid air conditions

[0055] FIG. 12B provides a plot of frequency change vs R32 refrigerant concentration for three set of tests with each involving seven different R32 concentration levels in dry air where the tests were performed on a first day, a seventh day, and a 14th day using a carbonized Zn-MOF-74 coating on a QCM sensor.

[0056] FIG. 13 provides a plot of frequency shift relative to humidity level where the frequency shift was caused by water adsorption in a hydrophobic MOF sensor.

[0057] FIG. 14A provides a plot of Frequency shift vs R32 concentration at different humidity levels for a QCM sensor with a hydrophobic MOF of ZIF-67 while FIG. 14B provides a table of data associated with the plot of FIG. 14A.

[0058] FIG. 15 provides a plot of frequency shift versus time for a series of tests performed with concentrations of R32 progressively oscillating between 0% and successively higher percents (up to 50% LFL) as time progresses and wherein three different plots are overlaid with each corresponding to a different humidity value (% RH).

[0059] FIG. 16A provides an image of peeling cross-hatch tape from a coated QCM sensor and particularly from a QCM sensor with a MIL-101 coating and an overlying parylene coating.

[0060] FIG. 16B provides a table of deposit MOF mass after coating and estimated remaining MOF mass after a tape test

[0061] FIG. 16C provides a table of resonance frequencies for two QCM sensors before and after coating with one being coated via a drop casting method and the other being coated via a spin casing method.

[0062] FIG. 17A is a table that correlates concentration in % LFL and concentration in (PPm,V / V) for an R32 refrigerant.

[0063] FIG. 17B provides plots of frequency shift in Hz against R32 Concentration in % LFL for three different % RH levels.

[0064] FIG. 18 provides a table setting forth estimated uncertainties in % LFL associated with uncertainties in humidity that accompany use of humidity sensors with finite accuracies for R32 concentration detection using humidity tolerant MOF coated QCM sensors.

[0065] FIG. 19 provides an annotated plot of frequency shift QCM sensor response for oscillating R32% LFL exposures and zero-air exposures where each successive R32 exposure had an increase in R32 concentration.

[0066] FIG. 20 provides a plot of R32 concentration values detected by a commercial R32 sensor in an experiment while a QCM / MOF sensor gathered frequency and frequency shift information associated with different concentrations of R32 gas during the experiment of FIG. 19.

[0067] FIGS. 21A-21F illustrate the distribution of frequency shift for R32 concentrations determined by a commercial sensor for variations in R32 concentration around different nominal values of 10, 20, 25, 30, 40, and 50% and with humidity varying from 10-60% RH where each figure focuses on the distribution around different nominal values.

[0068] FIG. 21G provides a table of means and standard deviations for the test data of FIG. 21A-21F based on the nominal R32 concentrations for each of FIGS. 21A-21F.

[0069] FIG. 21H provides a plot of frequency shift vs. R32 concentration when humidity ranged from 10-60% RH based on the average values from FIG. 21G along with a plot of the rest of the data from FIGS. 21A-21F

[0070] FIG. 21I provides a plot of frequency shift vs. R32 concentration when humidity ranged from 10-60% RH and linear fit using based on all of the data of FIGS. 21A-21F.

[0071] FIG. 22 provides a plot of frequency shift vs R32 concentration and vs RH data points for testing results for the spin-coated sensor with the mesh showing the calibration curve.

[0072] FIG. 23 provides a partial plot of over a frequency range of Y Hz to Y+10 Hz, where Y is an unspecified but relevant value for the experiment being run during a cycling of 11 periods of R32 gas presence at different concentrations separated by periods of humid zero-air presence during an example validation test for a spin coated sensor response at a fixed humidity of 36.3% RH such that frequency shift may be ascertained by subtracting a baseline frequency with no R32 present from (at the top of the plot) from a measured frequency obtained in the presence of R32 gas where the same six R32 concentrations are used from during each period.

[0073] FIG. 24 provides a partial plot of R32 concentration values obtained from a commercial sensor over 11 periods of repeated testing of six different R32 gas concentrations with each separated by a period of humid zero-air (i.e. no R32 presence) where each period of constant R32 value lasted one minute wherein the humidity during the entire test was held constant at 36.3% RH.

[0074] FIG. 25A-25C provide relative error amounts using linear regression based on Equation 11, at 6 different R32 concentrations at each of 18.4% RH (FIG. 25A), 36.9% RH (FIG. 25B), and 55.3% RH (FIG. 25C), while FIG. 25D provides summaries of the relative errors at 18.4, 36.9 and 55.3% RH, and a summary of relative errors for all humidity levels.

[0075] FIG. 25E provides a table of mean square error (MSE) and absolute values of the relative error summary at different concentrations for 3 humidity levels using the calculated R32 concentration from the linear regression Equation 11 and an R32 concentration from a commercial R32 gas sensor.

[0076] FIGS. 26A-26C provide relative error amounts using linear regression based on Equation 12, at 6 different R32 concentrations at each of 18.4% RH (FIG. 26A), 36.9% RH (FIG. 26B), and 55.3% RH (FIG. 26C), while FIG. 26D provides summaries of the relative errors at 18.4, 36.9 and 55.3% RH, and a summary of relative errors for all humidity levels.

[0077] FIG. 26E provides a table of mean square error (MSE) and absolute values of the relative error summary at different concentrations for 3 humidity levels using the calculated R32 concentration from the linear regression Equation 12 and an R32 concentration from a commercial R32 gas sensor.

[0078] FIGS. 27A-27C provide relative error amounts using linear regression based on Equation 13, at 6 different R32 concentrations at each of 18.4% RH (FIG. 27A), 36.9% RH (FIG. 27B), and 55.3% RH (FIG. 27C), while FIG. 27D provides summaries of the relative errors at 18.4, 36.9 and 55.3% RH, and a summary of relative errors for all humidity levels.

[0079] FIG. 27E provides a table of mean square error (MSE) and absolute values of the relative error summary at different concentrations for 3 humidity levels using the calculated R32 concentration from the linear regression Equation 13 and an R32 concentration from a commercial R32 gas sensor.

[0080] FIG. 28 provides a schematic representation of a Pierce-Gate Oscillator.

[0081] FIG. 29 provides images of an experimental set up used for testing the QCM / MOF sensor in combination with a frequency counter. a custom oscillator circuit board that included a Pierce-Gate oscillator circuit, and data logging software.

[0082] FIG. 30 provides two plots created using an experimental set up similar to that of FIG. 29 where the plots show frequency versus time as R32 gas levels change over one minute periods of fixed but varying R32 concentration separated by one minute periods without presence of R32 gas so that frequency changes can be seen with one plot showing frequency changes when detection is based on a QCM with no coating while the other shows changes when detection is based on a QCM sensor provided with an HKUST coating.

[0083] FIG. 31 provides an image of an example sensor assembly along with an exploded view of a QCM sensor (which may or may not include a gas selective coating such as an HT-MOF) and its housing.

[0084] FIGS. 32A-32G provide various views of an implementation of a single QCM sensor assembly according one embodiment of the invention, including an annotated isometric view of the circuit and sensor (FIG. 32A), a top view of a populated circuit board excluding mounted sensors (FIG. 32B), a top view of an unpopulated PCB showing traces and part identification markings (FIG. 32D), a bottom view of the unpopulated PCB showing traces and other features (FIG. 32E). a table correlating descriptions, values or part numbers, with PCB feature identifiers (FIG. 32F), and a schematic drawing of a plurality of individual circuit elements showing connections to other circuit elements (FIG. 32G).

[0085] FIGS. 33A-33G provide various views of an implementation of a dual QCM sensor assembly according one embodiment of the invention that includes a reference sensor and a selective gas sensor, with the FIGS including an annotated isometric view of the circuit and sensors (FIG. 33A), a top view of a populated circuit board excluding mounted sensors (FIG. 33B), a top view of an unpopulated PCB showing traces and part identification markings (FIG. 33D), a bottom view of the unpopulated PCB showing traces and other features (FIG. 33E). A table correlating descriptions, values or part numbers, with PCB feature identifiers (FIG. 33F), and a schematic drawing of a plurality of individual circuit elements showing connections to other circuit elements (FIG. 33G).DETAILED DESCRIPTION OF THE INVENTION

[0086] Various advantages and novel features of the present invention are described herein and will become more readily apparent to those skilled in this art from the following detailed description. In the preceding and following descriptions the preferred embodiment of the invention is shown by way of illustration of the best mode contemplated for carrying out the invention perhaps along with one or more variations. As will be realized, the invention is capable of modification in various respects without departing from the spirit of the invention as will be understood by those of skill in the art.Definitions and AbbreviationsQCM: Quartz Crystal Microbalance

[0088] SAW: Surface Acoustic Wave

[0089] ZIF-67: Zeolitic Imidazolate Framework-67

[0090] HT-MOF: Humidity tolerant metal organic framework crystalline porous material.

[0091] Targeted HT-MOF: A humidity tolerant MOF that is selected for its ability to target a particular refrigerant gas for adsorption and thus measurement.

[0092] Zero-air: Dry carrier gas (e.g. dry air) which may be mixed with pure R32 to obtain the different R32 gas concentrations used during testing. Aero-air may also be pass through a gas bubbler where it picks up moisture to become humid, where it is called humid zero-air. The composition of zero-air is: 20.9% v / v Oxygen, 79.1% v / v Nitrogen, <3 ppm H2O, <1 ppm THC (total hydrocarbon).GENERAL EMBODIMENTS

[0093] FIGS. 1A-1 and 1A-2 provide a schematic cut side view of example stacked layers forming a QCM with an HT-MOF and a top view of each of those layers respectively. In this example Layer 102B, i.e., the bottom layer, is a conductive electrode layer with circular central portion and a rectangular portion extending to the left where the left side of the rectangular portion may be used as a contact region when applying an electrical signal to or receiving an electrical signal from this layer. In an example configuration the circular part of 102B may be about 5 mm in diameter while a width of the rectangular portion may be about 3 mm. Layer 101 is a quartz crystal layer which provides the sensor with piezoelectric resonance oscillation when a properly oscillating drive voltage is applied. In an example configuration the nominal diameter of the crystal layer is 14 mm. The crystal layer is typically cut along a selected crystalline axis with its thickness dictating oscillation frequency which is typically around 330 μm for a standard 5 MHz crystal. The next layer is layer 102T or the top electrode layer which has a generally circular configuration which may include a tab region to the right which may extend beyond the quartz layer to allow electrical contact from below. In some examples the nominal diameter of layer 102T may be about 13 mm. some QCM sensors also include a final coating layer which may provide selective absorption of the particular gases who mass may change the oscillation of the frequency of the microbalance thus allow detection of gas concentration once a sensor has been calibrated. In some variations the coating may be limited to a selected portion of the underlying electrode area (e.g. to a circular central portion, to the entire lower electrode area, or even beyond the electrode area such that the coating contacts any extended quartz layer regions that extend beyond electrode 102T. In some embodiments, the above 4 layer stack forms the selective detection element of the sensor but requires a mount, an oscillating electrical input, output resonance frequency detection, and electrical signa processing to provide actionable detection information. In this example, the coating 103 is indicated to be a humidity tolerant (HT) metal organic framework (MOF) coating. In embodiments of the present invention the MOF coating is an HT-MOF that may be, for example, an ZIF-67 MOF. The coating can be tailored for selective adsorption of gas molecules (e.g. refrigerant gas) molecules and particularly A2L or A3 gas molecules at concentrations levels well below that of levels that could be dangerous). Also illustrated FIG. 1A-1 are two electrical connection locations, EC-B and EC-T, with one for contacting the 102B and the other for contact 102T where contact can be made from the bottom in both cases. In other embodiments, other contact and layer configurations may be used.

[0094] FIG. 1B provides a schematic representation of an example sensor that includes two quartz crystal resonators with one like that of FIGS. 1A-1 and 1A-2 but with the other lacking element 103 has it is provided as a reference sensor that provide frequency information without any adsorption coating. Such a reference sensor may be useful in isolating certain environmental changes that cause oscillation frequency shifts that have nothing to do with target gas presence.

[0095] FIG. 1C provides a schematic view of another example multi-detector sensor configuration including a reference sensor and two QCM sensors each with a different adsorption coating to allow detection of different gases of interest. In the FIG. one of the coating material is shown as 103-1 while the other is shown as 103-2.

[0096] FIG. 2A provides a schematic illustration of a single QCM sensor assembly 210A where the QCM is coated with the HT-MOF material for selective detection of a particular refrigerant or group of refrigerants. The assembly includes a The QCM / HT-MOF detector which is driven by an oscillator / QCM interface module IC which may be replaced by a plurality of discrete components in some embodiments. The Oscillator / QCM interface in-turn supplies resonant frequency information to a microprocessor or microcontroller which may provide information in a number of different forms depending on how it is programmed or hard coded. For example it may simply pass along raw frequency information to a separate processing circuit, it may pass along frequency shift or change information, it may include use of stored calibration information which may be used to pass on gas concentration information, or it may alternatively or additionally pass on warnings (e.g. contact relevant personnel or authorities, sound alarms, or the like) and / or it may automatically institute preventive measures (e.g. turn on venting fans, turn off gas supply lines, otherwise isolate gas into smaller volumes, and the like). In addition to receiving frequency information from the detector, the microprocess may receive additional information from other sensors that may form part of the assembly wherein the additional sensors may include one or more of a humidity sensor, a temperature sensor, and / or a pressure sensor which may be in turn used to provide more accurate output information. In some embodiments, multiple sensors may be integrated into a single device. In other embodiments, other sensors may be included while in still other embodiments, redundant sensors or processors may be included, in still further embodiments fail safe features and configurations may be included.

[0097] FIG. 2B provides a similar schematic illustration of a sensor assembly 210B similar to that of FIG. 2A but directed to a two QCM sensor assembly like that of 102-B where the second reference QCM and driving oscillator / interface circuit is included with its information also going to the microprocessor to provide enhanced processing and output.

[0098] FIGS. 3A-3C provide schematic illustrates similar to those of FIGS. 1A-1-1C but for a surface acoustic wave sensor or detector (SAW sensor) as opposed to a QCM sensor. As shown in FIG. 3A the saw detector 300A is formed by three primary layers: (1) a quartz piezoelectrical layer 301; (2) a patterned layer of metal traces 302T which provide a pattern for acoustic wave generation and a pattern for acoustic wave detection; and (3) a layer 303A of material that can selectively adsorb and release gases that are to be measured which change a response characteristic of the electrical circuit. In some embodiments of the present invention the coating 303 is in the form of an HT-MOF such as ZIF-67. More information about SAW detectors can be found in U.S. Pat. No. 11,796,508, which was referenced above, FIG. 3B provides a sensor group 300B that is similar to that of FIG. 1B and which includes two SAW detector elements with one including an HT-MOF while the other has no coating material and functions as a reference detector. FIG. 3C provides a sensor group 300C that is similar to that of FIG. 1C but all three sensors including SAW sensors as opposed to QCM sensors with one including no coating material, one including an HT-MOF material labeled as 303-1, and another including a different HT-MOF labeled as 303-2.

[0099] FIGS. 4A-4B provide sensor assembly configurations 410A and 410B, respectively, similar to 210A and 210B of FIGS. 2A and 2B, respectively but with the QCM sensors replaced by SAW sensors.

[0100] FIG. 5 provides a block diagram setting forth features of a targeted humidity tolerant metal organic framework coating and examples.

[0101] FIG. 6 provides a block diagram setting forth a first process for creating a HT-MOF coating involving spin coating of an HT-MOF slurry onto a QCM or SAW sensor surface to enable the surface to target adsorption of a selected refrigerant molecule.

[0102] FIG. 7 provides a block diagram setting forth a second process for creating a HT-MOF coating involving spray coating of an HT-MOF slurry onto a QCM or SAW sensor surface to enable the surface to target adsorption of a selected refrigerant molecule.

[0103] FIG. 8 provides a block diagram setting forth a third process for creating a HT-MOF coating involving drop coating of an HT-MOF slurry onto a QCM or SAW sensor surface to enable the surface to target adsorption of a selected refrigerant molecule.

[0104] FIG. 9 sets forth a method 921 of using sensors of some embodiments of the invention.EXPERIMENTS, IMPLEMENTATIONS, AND OTHER EMBODIMENTSOverview of Initial Sorbent Screening:

[0105] Five sorbents were located on a surface of an upper electrode of a Quartz Crystal Microbalance (QCM) sensor and were initially screened for their optimal response to R32 under ambient conditions. For all the tests performed, mass flow controllers (MFC) (Alicat Scientific, MCS-20SCCM-D / 5M) are used to measure the flow rates of the different gases. FIG. 10 shows response of the sorbents selected for study. These five were selected based on their use in previous experiments at PNNL and based on a literature review. Initial sensor response (frequency shift) was tested for each sorbent selected. These tests involved passing small pulses of R32 with a 0% to 4.36% v / v concentration in dry air over the sensor with the resulting frequency shift being recorded in real time. The response of the sensors to varying concentration of R32 were observed to be rapid and reversible. Based on these observations two of the Metal-Organic Framework (MOF) sorbents, MIL-101 and HKUST-1, were selected for further study due to their high signal to noise ratio and fast response to R32 gas.

[0106] Further testing of these sensors posed a different challenge, while the sensors showed sensitive, repeatable, and reversible signal change to R32, under humid conditions especially at high humid conditions, the signal from these sensors becomes damped and not reproducible on subsequent testing. FIG. 11 shows response of the QCM sensor loaded with MIL-101 under three consecutive dry condition runs and then underwent three humid condition runs where the humidity was set at approximately 60% relative humidity (RH) where each run underwent cycling with different R32 concentrations similar to that used in the experiment of FIG. 10. When humid air was introduced (˜60% RH) the signal from the same sensor became noisy. Similar observations were found for the other candidate sorbent, HKUST-1. These results suggest that water molecules compete with the adsorption of refrigerant molecules in the candidate sorbents.Humidity Mitigation Strategies

[0107] Initial attempts to mitigate the impact of moisture were focused on increasing the hydrophobicity of the sensor. This included coating the existing candidates with a hydrophobic polymer coating. Additional hydrophobic sorbents were also studied, including UiO-66, MCM-41 (mesoporous silica), carbonized Zn-MOF-74, and dealuminated zeolites. While these sorbents showed initial promise, the response of these sensors were not repeatable over time due to either unpredictable or large loadings of water molecules in the pores of these substrates. FIGS. 12A and 12B shows the response from one such attempt using a carbonized variant of Zn-MOF-74, where the MOF was thermally treated under inert conditions (N2 flow) to become more hydrophobic. While the hydrophobic sorbents initially tested gave a better response in humid conditions as shown in FIG. 12A When tested over time a decrease in their response was seen similar to that seen with the other sorbents that were tested even under dry testing conditions as shown in FIG. 12B. This observation suggests water adsorption within the pores of these materials accumulated over time, occupying active sites for R32 adsorption. This indicated that the pore opening which was not changed in the carbonization may be the reason for the irreversible water loading.

[0108] A hydrophobic MOF with small pore openings and reversible water isotherms is preferred because water adsorption is minimized and controlled to be the same under both the adsorption and desorption processes. A hydrophobic MOF with a reversible hysteresis free water isotherm was selected as the sensor material while following other selection criteria such as, for example, the material having hydrophobic character and having an appropriate pore size. In some embodiments the following selection criteria may apply:

[0109] 1. Hydrophobicity: the material should be hydrophobic toward water vapor.

[0110] 2. Water isotherm:

[0111] a. Material should have type II or III water isotherm defined by the International Union of Pure and Applied Chemistry (IUPAC) with low water adsorption.

[0112] b. No or small hysteresis for water adsorption and desorption.

[0113] 3. Pore size: Pore size should be slightly big enough to adsorb the target gas for fast adsorption and quick steady state sensor signal response. The pore size should promote fast adsorption and desorption of the target gas for gas sensing. An appropriate pore window in some cases may be close (0.9 to 1.1 X, where X is the size of the refrigerant molecule and has no pore cage larger than 2 nm. The pore window is important for selectivity while the pore cage is related to hysteresis

[0114] 4. Selectivity: The material should be selective to the target gas.

[0115] 5. Chemical and temperature stability: the material should exhibit stability in water and long-term stability when exposed to target refrigerant and moderate temperature changes. The material should have long term structural stability. Long term refers to a period of time comparable, greater, or substantially greater than anticipated sensor life.Test Results with Hydrophobic MOF ZIF-67

[0116] Water adsorption and gas adsorption isotherms were collected and studied for a series of sorbents. The MOFs with high surface area such as MIL-101 and HKUST-1, showed diminished responses over time due to moisture accumulation, and they did not sustain response to R32 under humid conditions. Based on this failure a hydrophobic MOF was selected for consideration as a sensor coating. ZIF-67 has chosen as a coating material as it showed relatively higher surface area, low water adsorption and appropriate pore size to provide a rapid and repeatable response to R32.

[0117] ZIF-67 has a water isotherm that is ideal for this application, it is considered hydrophobic, and its water adsorption is minimal and hysteresis free in moderate humid conditions. All test results described below were performed at a constant room temperature of 23° C. and a gas flowrate of 20 standard cubic centimeters per minute (sccm). Tests for water adsorption by the hydrophobic MOF sensor were performed. FIG. 13 shows the adjusted frequency shift result coming from such tests at different humidity levels (10, 20, 30, 40, 50 and 60% RH). The several days test results show a linear increase of the frequency shift as the humidity level increases. The linear fit curve enables compensation for the humidity response of the sensor during calibration.

[0118] FIG. 14 provides plots for sensor response in frequency shift (Hz) vs R32 concentration in terms of % of lower flammability level (LFL) for three different nominal humidity levels of 40% RH, 50% RH, and 60% RH and 6 different R32 concentrations of 10, 20, 25, 30, 40, and 50% LFL over several days of testing. Each point in the plot is the average of 3 runs at the same concentration of R32. The maximum standard deviation was 1.4 Hz which showed good repeatability. As shown in the figure, at a constant humidity, the frequency response of the sensor increased as the R32 concentration increased, showing almost perfect linearity.

[0119] In theory, as the concentration increases, the hydrophobic MOF adsorbs more R32 molecules, which increases its mass, consequently increasing the frequency shift of the resonance frequency of the sensor. As the humidity level increases, the frequency response of the sensor at a constant concentration of R32 decreases, indicating the competitive adsorption of R32 and water molecules. The test results of the sensor at constant humidity can be linearly fitted for sensor calibration, as shown in FIG. 14.

[0120] A validation of the results shown in FIG. 14A was performed at 43.31% RH humidity level. In this process three tests were performed with different concentrations of R32. To calculate the concentration of R32 using the sensor's frequency response obtained during testing, the linear fit curve equation of y=0.7109x (i.e. the linear curve fitting result for the plot of plot of 43.15% RH of FIG. 14) was used, where y is the sensor frequency shift and x the concentration of R32. A commercial sensor was used as a reference to benchmark the actual concentrations of R32. Test results for 3 tests are shown in FIG. 14B (i.e., for tests 1, 2, and 3 respectively). These results show that the targeted HT-MOF using ZIF-67 on a QCM sensor can detect the R32 concentration.

[0121] FIG. 15 shows an example of the time-based raw frequency shift signal of the sensor for different concentrations of R32 at different nominal humidity levels of 40, 50, and 60% RH. As previously explained, the sensor response increases with increasing R32 concentration at constant humidity and at a constant concentration, the response decreases with increasing humidity levels.

[0122] In summary, the sensor response with the hydrophobic MOF (ZIF-67) in different humidity conditions is linear, indicating an ability to easily compensate for humidity. Furthermore, the sensor response increases with increasing R32 concentration. The frequency response of the sensor in these experiments was relatively low for different concentrations of R32 at a maximum of 36 Hz at 50% LFL.Demonstration of Good Adhesion to Substrate

[0123] Sensing film with good adhesion to the substrate is important for long-term performance of the coated QCM sensors and for SAW (surface acoustic wave) sensors that may be used as an alternative to QCM sensors in some embodiments. One of the most common ways to test bonding between microfabricated layers is known as ‘the tape test’ which is an adaptation of ASTM D3359 “Standard Test Methods for Measuring Adhesion by Tape Test”. Cutting patterns of lines in the layer to be tested is often unnecessary when testing microfabricated devices. In an ASTM D3359 procedure, readily available tape is manually applied to the test layer and then rapidly removed without using significant equipment. However, this test has many limitations because parameters including the magnitude and duration of tape application pressure, waiting time between bonding and removing.

[0124] The test results for ZIF-67 were obtained on sensors without any Parylene coating which is sometimes used on some coated sensors. Polymer coatings such as Parylene can act as a protection layer and seemed to help reduce mass loss of the sensing film in the case of MIL-101.

[0125] Cross-hatch adhesion tape specifically for ASTM D3359-22 was used during the tape testing. Initially, the QOCM sensors were affixed to microscope slides. Then the adhesion tape was gently and firmly pressed onto the area coated with MOF. Finally, the tape was peeled off swiftly at the direction parallel to the QCM surface, as shown in FIG. 16A.

[0126] Instead of counting the missing crosscut area, we estimated the mass loss after the tape test by measuring the change of the resonance frequency of the QCM sensor before and after tape testing. This estimation can be obtained from the following equation which sets forth the relationship between the resonance frequency of the QCM sensor and the mass deposited on the sensor's facef=-2⁢f02A⁢ρq⁢μq⁢Δ⁢mEQ. 1where:

[0128] f0=Resonance frequency of the fundamental mode (Hz) which is 5 MHz in the sensor used

[0129] f=normalized frequency change (Hz)

[0130] Δm=Mass change (g)

[0131] A=Piezoelectrically active crystal area (Area between electrodes, A=1.539 cm2)

[0132] ρq=Density of quartz (ρq=2.648 g / cm3)

[0133] μq—Shear modulus of quartz for AT-cut crystal (μq=2.947E11 gcm−1s−2)The deposited mass of MOF on the surface of the QCM sensors after coating are calculated and summarized in the Table of FIG. 16B. The after coating resonance frequency was obtained at 10% RH. FIG. 16C provides a table of resonance frequencies for two QCM sensors before and after coating with one being coated via a drop casting method and the other being coated via a spin casing method. The deposited mass of MOF on the surface of QCM sensors after tape testing were calculated and are also summarized in the table of FIG. 16B along with the remaining mass portion as a percentage.

[0134] For the sensors only coated with MOF, the remaining portion is within the range of 51.11-69.9%. For the sensors with additional coating of Parylene, a greater portion of coating materials (including both MOF and parylene) was found to be remaining on the QCM surface. The test results for ZIF-67 were obtained on sensors without any Parylene coating. Polymer coatings such as Parylene acted as a protection layer and seems to help reduce mass loss of the sensing film in the case of MIL-101.QCM Sensors Data Analysis

[0135] Comparisons were made between ZIF-67 MOF coatings applied by two different method to a commercial QCM sensor from AWSensors. As with the experiments noted above, these experiments used an R32 target gas which is a target A2L gas for a sensing element of some embodiments of the invention. The fundamental frequency of the sensors are 5 MHz with one sensor coated with the MOF material via a drop casting technique and the other using a spin coating method. The sensors coated via these different methods showed different responses to different humidity levels.Relative Humidity Sensitivity Analysis for the Drop-Casted MOF Sensor.

[0136] A MOF sensor can be made using drop coating via a micropipette. For drop-casting process, 20 μL of MOF slurry was deposited on the QCM sensor and allowed for the solvent in the MOF slurry (ethanol but could also be methanol) to evaporate. After two minutes, the process was repeated with additional 20 μL and 10 μL MOF deposition with at least two-minute interval between each repeat for the earlier MOF layer to dry. The mass loading of the MOF material is estimated at 140 μg sensor testing, FIG. 17A provides a table correlating R32 concentration in % LFL to parts per million by volume (ppm, v / v). FIG. 17B provides plots of frequency shift in Hz against R32 Concentration in % LFL for three different % RH levels for a sensor with a drop cast ZIF-67 MOF coating. FIG. 17B is similar to that of FIG. 14A but with additional features shown. Plots like that of FIG. 17B may be used in a process for estimating the impact of uncertainties of relative humidities on the R32 concentration readings for a sensor.

[0137] For a humidity level of 43.15% RH (denoted as R1), the fitted equation for this relative humidity (RH) as shown in FIG. 17B is y=k1x=0.7109x. The corresponding angle with the positive x-axis is α1. For a humidity level of 52.34% RH (denoted as R2), the fitted equation is y=k2x=0.3955x as also shown in FIG. 17B. The corresponding angle with the positive x-axis is α2. The angle between these two lines is β, which can be calculated as:β=α2-α1=tan-1(k2)-tan-1(k1)=tan-1(0.3⁢9⁢5⁢5)-tan-1(0.7⁢1⁢0⁢9)=-3⁢5.4⁢0⁢9EQ. 2The unit angle per percentage humidity level can be obtained by:β(R2-R1)=-35.4095⁢2.3⁢4⁢%-4⁢3.1⁢5⁢%=-1.505⁢° / %EQ. 3For a certain humidity level R; within the range of R1 and R2, the angle with the positive x-axis at this RH level is:αi=α1+(Ri-R1)⁢β(R2-R1),EQ. 4The corresponding slope is ki=tan(αi), thus the relation between frequency shift and R32 concentration at humidity level Ri is:y=ki⁢x=tan⁡(αi)⁢x=tan⁡(αa+(Ri-R1)⁢β(R2-R1))⁢x.EQ. 5If the sensitivity of the humidity sensor is ε, the equations for RH level Ri−ε and Ri+ε are:y=tan⁡(αa+(Ri-ε-R1)⁢β(R2-R1))⁢x⁢ and⁢ y=tan⁡(αa+(Ri+ε-R1)⁢β(R2-R1))⁢x.EO. 6Given a certain frequency shift value F, the concentrations corresponding to RH levels Ri−θ, Ri, and Ri+θ are ci−1, ci, and ci+1 respectively, which can be calculated as:ci-1=Ftan⁡(αa+( RHi-θ- RH1)⁢β( RH1- RH2))EO. 7ci=Ftan⁡(αa+( RHi- RH1)⁢β( RH1- RH2))EQ. 8ci+1=Ftan⁡(αa+( RHi+θ- RH1)⁢β( RH1- RH2))EQ. 9The error at RH level of Ri is given as:-ϵ1=-<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[LeftBracketingBar]"< / annotation>< / semantics>ci-ci-1ci<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[RightBracketingBar]"< / annotation>< / semantics>⁢ and+ϵ2=+<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[LeftBracketingBar]"< / annotation>< / semantics>ci+1-cici<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[RightBracketingBar]"< / annotation>< / semantics>.EQ. 10Using the above method, R32 concentration estimates and maximum uncertainties of the R32 concentrations caused by the uncertainties of relative humidities can be estimated. The results for the current experiment and analysis are set forth in the table of FIG. 18. These uncertainties of the R32 concentration are obtained based on the relations between frequency shift (Hz) and R32 concentration (% LFL) at 43.15, 52.34 and 60.7% RH shown in FIG. 14A. Available example on-chip humidity sensors of certain accuracy are also listed in FIG. 18 along with details of unit price for bulk purchase (1000+ units), response time, and long-term drift.Product HDC3020DEFR is an integrated relative humidity and temperature sensor manufactured by Texas Instruments. Both products SHT45-AD1B-R2 and SHT40-AD1B-R2 also have integrated relative humidity and temperature sensors and are both from Sensirion AG. Products like these or other integrated sensors may may be integrated into a sensor assembly of some embodiments of the invention to provide necessary inputs for providing calibrated concentration levels of R32 or other refrigerants when a QCM or SAW sensor is used with a humidity tolerant MOF (HT-MOF) coating like ZIF-67 where “humidity tolerant” refers to a coating that has low water uptake (typically less than 5 wt. % at RH 50% and room temperature), has small pore openings (e.g. with a size of 0.3 to 0.5 nm, more preferably 0.35 to 0.45 nm, and has a reversible hysteresis free water adsorption-desorption isotherm such that capillary condensation in the pores is minimized and controlled to be the same under both the adsorption and desorption process.Due to the intrinsic dependencies on the ambient humidity, some commercial gas sensors were built with on-chip humidity sensors for self-calibration to get measurement accuracy across wide environmental range. Some of the available commercial gas sensor feature humidity compensation include, for example:1) SGP30 multi-gas (VOC and CO2eq) sensor, SGP 41 multi-gas (VOC and NOx) sensor from Sensirion AG,2) MPS™ refrigerant gas sensors (including R32, and R-454 sensors, etc.),3) AM4205H-LC-R32 Cubic R32 A2L Refrigerant Gas Sensor.None of these sensors are QCM or SAW sensors and more particularly none of these sensors are QCM or SAW sensors with humidity tolerant MOF coatings.Relative Humidity Sensitivity Analysis for the Spin-Coated MOF Sensor.Spin coating is a common technique for producing thin films of materials on various substrates, including sensors, and it's used in both scientific and industrial applications. An excess solution is poured onto the substrate which is rotated at a constant angular velocity which results in liquid flow radially outward, driven by centrifugal force. The excess liquid is quickly eliminated as droplets from the substrate perimeter and the film thickness becomes uniform during spin-off.Spin coating of MOF sensing films on QCM sensor were investigated. In one example of making such a film the following process was used. A 50 μl MOF particle dispersion using 2-10 g / L MOF with 0.2-1 g / L polymer binder was created in methanol which was then an amount of it was dropped to the center of a 5 MHz QCM substrate. The substrate was rotated at 500 rpm about its center for about 30 seconds. This process was repeatable until a desired mass loading of MOF was achieved on the substrate. The MOF-coated QCM sensor was air-dried before testing. The estimated mass loading of the ZIF-67 MOF material is 104 μg and the diameter of the coated substrate is about 14 mm.To measure the effect of humidity on the coated QCM sensor, multiple tests at a constant temperature of 23° C. and different humidity levels (10-60% RH) at 10% RH increments were performed. The tests consist of flowing different concentrations of R32 over the coated QCM sensor at constant relative humidity. R32 was introduced in successively increased concentrations (R32 plus air) for a period of time with each separated by a period of zero R32 concentration (i.e. humid zero-air). Each concentration of R32 is flowed over the sensor for one minute followed by one minute of humid zero-air flow. The following nominal concentration levels of R32 were used during the tests: 10% LFL, 20% LFL, 25% LFL, 30% LFL, 40% LFL, 50% LFL. The actual R32 concentrations were measured using a commercial R32 gas sensor (Nevada Nano, MPS™ A2L Refrigerant Gas Sensor) with a detection range of 5-100% LFL with an accuracy of ±3% LFL. Humidity in the tests was generated by flowing zero-air into liquid water in a gas washing bottle. Humid zero-air coming out of the gas washing bottle carries out water vapor into the QCM sensor flow cell. Humidity levels during the testing were varied by changing the humid zero-air gas flowrates. The volumetric flow rate of zero-air was {dot over (v)}1, humid zero-air was {dot over (v)}2, and dry R32 was {dot over (v)}3. The condition {dot over (v)}1+{dot over (v)}2+{dot over (v)}3=20 sccm was always maintained during all testing.The RH level was obtained from the built-in humidity sensor of the commercial R32 gas sensor, which was placed at the outlet of the flow cell holding the QCM sensors.The time series data of the frequency shift of the QCM coated sensor was recorded. The test results taken from multiple days were used for the sensor calibration.A computer algorithm was developed in MATLAB and was used to process the frequency shift time series data obtained from the experimental tests. The timeseries data was initially filtered using a moving average function with a moving average window size of 25 data samples. The algorithm calculates the maximum frequency shift values during the time the QCM sensor was exposed to different R32 concentrations. The maximum peak to peak frequency shift over a threshold value were saved and categorized for the corresponding R32 concentrations. FIG. 19 shows an annotated example of the algorithm output of the processed test data with peaks points in the frequency shift detected for increasing concentrations of R32. The MATLAB code is set forth in Appendix A attached to this application. During this test, the humidity was substantially constant and had a nominal value of 10% RH for the data shown.To get corresponding actual R32 concentration during the tests, a commercial R32 gas sensor's data is collected and saved. The commercial R32 gas sensor data is processed using another algorithm developed in MATLAB which is set forth in Appendix B. The commercial sensor data is initially filtered using a moving average function with a sample data window size of 2. Then, the significant peaks corresponding to each R32 concentration are detected and saved. The algorithm-detected concentration points are added to the corresponding frequency shift in the processed QCM / MOF sensor data. FIG. 20 shows an example of the output of the processed R32 concentration points obtained from the commercial R32 gas sensor data that measured R32 concentration while frequency change data of the QCM / MOF sensor was being recorded 859 data points for each R32 concentration and each different humidity level (10-60% RH) were collected from the testing of the spin coated QCM sensor.Calibration Curve without Humidity.An initial calibration evaluation process consisted of considering only the sensor's frequency shift with different R32 concentrations while excluding taking into consideration humidity level information. FIGS. 21A-21F illustrate the distribution of frequency shift for R32 concentrations determined by a commercial sensor for variations in R32 concentration around different nominal values of 10, 20, 25, 30, 40, and 50% LFL and with humidity varying from 10-60% RH, where FIG. 21A shows data associated with a nominal 10% LFL, FIG. 21B shows data associated with a nominal 20% LFL, 21C shows data associated with a nominal 25% LFL, FIG. 21D shows data associated with a nominal 30% LFL, 21E shows data associated with a nominal 40% LFL, FIG. 21F shows data associated with a nominal 50% LFL. FIG. 21G provides a table of means and standard deviations for the test data of FIG. 21A-21F based on the nominal R32 concentrations for each of FIGS. 21A-21F.The average values of FIG. 21G are plotted in FIG. 21H, alongside all the processed data clumped around their means. FIG. 21H also shows the results of a linear fit based on the average values which can be used to obtain a calibration curve for the sensor. The indicated r-square value of 0.9999 shown in FIG. 21H shows almost perfect linear fit for the average values. This indicates the sensor's frequency shift response is linear with increasing R32 concentrations. The initial linear fit equation is:y=0.1⁢0⁢7⁢xEQ. 11where y is the frequency shift response in Hz of the sensor and x is the R32 concentration in % LFL.Using a linear fit of all the testing data of FIG. 21H produces the distribution of and calibration curve of FIG. 21I. A calibration curve having an r-square value of 0.993 and mean square error value of 0.014, is obtained:y=-0.0⁢7⁢4+0.1⁢09⁢xEQ. 12Calibration Curve with Humidity as a ParameterConsidering the relative humidity values for each testing condition, a new calibration curve can be obtained for all testing results. The same testing data used in the initial calibration, previously discussed, is used here but with the addition of the relative humidity values. A total of 5154 data points for each parameter, notably R32 concentrations obtained from the commercial R32 gas sensor, the relative humidity values, and the frequency shift from the QCM sensor's response. Using the linear regression function in MATLAB, a calibration curve for the R32 concentration values with two-parameters as the RH values and sensor's frequency shift was obtained as:x=-1.1⁢3⁢2+9.1⁢2⁢5⁢y+0.0⁢53⁢zEQ. 13where z represents the RH value in % RH. The r-square value is 0.997, whereas the mean square error is 0.479. FIG. 22 shows the testing data for the new calibration with RH values, with the mesh grid showing the calibration curve.Sensor Response Validation Process for the Spin Coated QCM Sensor.Three sets of validation tests were performed, at constant temperature of 23° C. RH level is constant for each set of tests, within the range of 10-60% RH.For a specific set of tests, the volumetric flow rate of the humid zero-air gas {dot over (v)}2 was kept constant. During each cycle of the validation tests, R32 / zero-air mixture of six different concentrations were flowed over the target sensor intermittently for a duration of one minute, followed by a one-minute interval of gas with 0% LFL of R32 (i.e., {dot over (v)}3=0). This cycle was repeated sixty times. An example of the measured data for one set of validation testing at 36.3% RH is presented in FIG. 23. The peak frequency shifts were obtained from the algorithm developed in MATLAB (See Appendix A).The peak values of the commercial R32 gas sensor readings at different R32 concentrations were obtained from data processing and are shown in FIG. 24 with the peak points highlighted with light green dots. It is worth noting that each set of measurements involving 6 different R32 concentrations did not provide those concentrations in an order from smallest to largest, or vice-a-versa, but instead provided them in an order that mixed higher and lower values to provide improved results.Validation for the Calibration Curve without Humidity.For the spin coated QCM sensor, per the testing results' calibration curve without humidity discussed above, the relation between frequency shift y, and R32 concentration x is represented by the Equation 11. Thus, with known frequency shift y, the corresponding R32 concentration can be calculated as:x=y0.1⁢0⁢7.The relative error from the calculated concentrations obtained from the QCM sensor frequency shift and the measured concentration from commercial R32 gas sensor values is given as follows:Relative⁢ Error⁢ (%)=x-x0x0·100EQ. 14An error analysis of the linear fit Equation 11 at 18.4, 36.9 and 55.3% RH are summarized in box plots of FIGS. 25A-25D. With FIGS. 25A-25C respectfully providing errors for individual LFL bases for each of these three RH levels while FIG. 25D provides an overall summary for each RH level as well a single summary for all LFL and RH values combined. Average concentration values, MSE, average MSE, and maximum and minimum relative error values are provided in the Table of FIG. 25E. In FIGS. 25A, 25B, and 25C, each subplot or boxplot contains 6 boxes, each box corresponds to certain R32 concentration, which contains 60 points. In FIG. 25D, the first three boxes correspond to three different relative humidities, which contains 360 (60×6) data points each, the last box contains all the data points corresponding to all concentration and humidity levels, which contains 1080 (360×3) data points. As observed in FIG. 25D, most of the outliers of the points are located above 10%.With Equation 11 as shown in FIG. 21H, the minimum and maximum absolute relative errors for the validation tests at 18.4, 36.9 and 55.3% RH are, respectively,18.4% RH: 0.002% and 6.059%,36.9% RH: 0.005% and 10.099%, and

[0163] 55.3% RH: 0.015% and 14.930%To gauge the performance of the calculated concentration values to the commercial R32 gas sensor values Eq. 11 Was used, for the validation tests at each humidity with the mean square error (MSE) is given by:MSE=1N⁢∑i=1N(x-x0)2N is the number of tests, which equals to 60 for each humidity level.The calculated MSE values for the tests performed at each humidity level, and for all test data points are:18.4% RH: 2.02936.9% RH: 2.248

[0166] 55.3% RH: 2.685

[0167] All data: 2.321.

[0168] The low MSE values indicate that the calibration curve (equation 11) performed well in calculating the R32 concentrations from the spin coated QCM sensor frequency shift response. As shown in FIG. 25D, the relative error as well as the MSE values increase with increasing humidity level.

[0169] Similarly, using the linear fit curve given by Equation 12, an error analysis can be provided. FIG. 26A-26D depicts the results of such an analysis. While FIG. 26E provides a table of additional data. The illustrations of FIGS. 26A-26D and the data provided in the Table of 26E of analogous form to those of FIGS. 25A-25E.

[0170] With the linear fit of Equation 12, the minimum and maximum absolute relative errors for the validation tests at 18.4, 36.9 and 55.3% RH are shown as follow:

[0171] 18.4% RH: 0.015% and 4.762%,

[0172] 36.9% RH: 0.087% and 7.583%

[0173] 55.3% RH: 0.012% and 12.474%

[0174] The calculated MSE values for the tests performed at each humidity level, and for all test data points are:

[0175] 18.4% RH: 1.350

[0176] 36.9% RH: 1.957

[0177] 55.3% RH: 3.459

[0178] All data: 2.256.Validation for the Calibration Curve with Humidity as a Parameter.

[0179] By considering the factor of relative humidity level (% RH), a two-parameter regression equation, Equation 13, was proposed with a plot of a calibration curve being shown in FIG. 22. The error analysis corresponding to this updated two-parameter linear regression at 18.4, 36.9 and 55.3% RH are demonstrated in the box plots of FIG. 27A-27D and with summary data shown in FIG. 27E. The illustrations of FIGS. 27A-27D and the data provided in the Table of 27E of analogous form to those of FIGS. 25A-25E, and 26A-26E.

[0180] Using Equation 13, the minimum and maximum absolute relative errors for the validation tests at 18.4, 36.9 and 55.3% RH can be shown to be:

[0181] 18.4% RH: 0.012% and 9.205%,

[0182] 36.9% RH: 0.125% and 7.587%

[0183] 55.3% RH: 0.025% and 7.239%

[0184] The calculated MSE values for the tests performed at each humidity level, and for all test data points are:

[0185] 18.4% RH: 0.758

[0186] 36.9% RH: 1.644

[0187] 55.3% RH: 1.536

[0188] All data: 1.312.

[0189] With the calibration curve with humidity as a parameter, the overall maximum absolute relative error, especially maximum absolute relative error at higher RH levels, decreased. The MSE value declined, and the number of outliers were also observed to be reduced. The error analysis implies that the predicted concentration value from the two-parameter regression equation is closer to the real concentration, especially at higher RH levels. Thus, the accuracy of the calibration curve with humidity as a parameter is higher than the calibration curve without humidity.

[0190] In sum, compared to the drop cast sensor, the new QCM sensor using the spin coating technique is more tolerant to high humidity changes. It's important to note that the new spin coated sensor's outstanding performance under different humidity levels may be due to the uniformity of the coating on the QCM sensor as well as the low mass loading of the hydrophobic MOF material, which reduces the amount of water adsorbed by the MOF material thus reducing the competitive adsorption by water vapor and by target gas molecules.Initial Proof-of-Concept Design

[0191] Given that commercial QCM sensors (from AWSensors) with humidity tolerant MOF coatings added by the inventors can measure frequency shifts associated with R32 gas concentrations as demonstrated by the prior test results, the inventors decided to create their own oscillation circuitry for QCM sensing to verify frequency shifts on a custom circuit board. A classic Pierce-Gate oscillator was chosen for the initial effort. Such an oscillator is shown in as shown in FIG. 28 with X1 representing the QCM, U1 being an inverter, C1 and C2 being capacitors, Rf being a feedback resistor, and Rs being a resistor that isolates the output driver of the inverter U1 from the complex impedance formed by C2, C1, and X1. For the QCM to oscillate properly, the phase shift around the loop must be zero or an integer multiple of 2π (360°).

[0192] A test board for the Pierce-Gate oscillator circuit was created and is shown in an experimental set up in FIG. 29 with a left most image providing a wide shot view that shows the oscillator circuit including the QCM Sensor along with a frequency counter and an MPS sensor (i.e., a molecular property spectrometer) with a gas inlet and outlet which can provide an independent determination of gas concentration (e.g., R32 concentration) that flowed through the QCM / MOF sensor on the circuit board. The upper right image shows a close up view of the MPS, the circuit board including a QCM or QCM / MOF sensor in a holder that provides a gas flow passage with an inlet that receives a controlled mix of air, humidified air, and refrigerant to be used in validating and / or calibrating the QCM or QQCM / MOF sensor, an outlet that directs the mixture to the MPS sensor for independent characterization, and electrical connectors for supplying power to the circuit board and frequency information from the QCM or QCM / MOF sensor to the frequency counter. Data logging software was also provided with a screenshot shown in the lower right image.

[0193] In different experiments an empty QCM (i.e. without a coating) or a coated QCM were placed in the holder and tested using the experimental set up. FIG. 30 shows an example of the frequency changes at different R32 concentrations for an empty or blank sensor and for a HKUST coated sensor. As can be seen in the plots, the frequency of the empty QCM changes by less than 10 Hz from pure air to 25% LFL of R32, while the coated QCM exhibits more than a 300 Hz frequency change. To further enhance the oscillator circuitry and reduce costs, a quartz crystal oscillator chip was chosen for the benchtop sensor design. Unlike the Piece-Gate oscillator, which requires five components (1 inverter, 2 resistors, and 2 capacitors), the quartz crystal oscillator chip integrates resistors and capacitors into a single component. This chip supports a frequency range of up to 50 MHz, whereas the maximum frequency of the Piece-Gate oscillator is typically 20 MHz. In addition, in some embodiments, the circuit board may include one or more sensors for pressure, temperature, and / or humidity for environmental compensation or even a single integrated sensor that provides any two of these three or even all three of these. In other variations other sensors may be provided. In some embodiments, a reference channel (i.e. a QCM without a gas selective coating) may be included along with a gas detection channel (i.e. a QCM with a gas selective coating). The inclusion of the reference channel may be useful in providing compensation for frequency shifts under varying environmental conditions. In some embodiments, a full sensor assembly will also include a microcontroller or microprocessor is also included to directly read the frequency (eliminating the need for an external frequency counter} and possibly to provide other functions such as, for example, environmental compensation via data received from additional sensors, other calibration or testing functions, frequency shift calculation, gas concentration determination, and corrective action implementation. In some variations, a reference channel may be replaced by a second gas detection channel configured for detecting a different gas, while in still other variations, more than one gas detection sensor may be included along with a reference sensor.

[0194] An image of example sensor assembly is shown in FIG. 31 along with an exploded view of a QCM sensor (which may or may not include a gas selective coating such as an HT-MOF) and its housing. The example circuit includes two channels (e.g. a gas detection channel and a reference channel) along with a microprocessor. In the example shown, the two channels are provided with a flow inlet and a flow outlet for convenience in efficient experimentation, testing and validation. In other embodiments, where the assemblies may be included in an operational environment that is to be monitored the relevance of such housings or holders will dictate their inclusion or absenceSpecific Detailed Example Sensor Assembly Implementations

[0195] FIGS. 32A-32G provide various views of an implementation of a single QCM sensor assembly according one embodiment of the invention, including an annotated isometric view of the circuit and sensor (FIG. 32A), a top view of a populated circuit board excluding mounted sensors (FIG. 32B), a top view of an unpopulated PCB showing traces and part identification markings (FIG. 32D), a bottom view of the unpopulated PCB showing traces and other features (FIG. 32E). a table correlating descriptions, values or part numbers, with PCB feature identifiers (FIG. 32F), and a schematic drawing of a plurality of individual circuit elements showing connections to other circuit elements (FIG. 32G). In the example of FIG. 32A-32G the PCB has a width of 1 inch and a length of 2.5 inches with a component height of less than 1 inch and component width beyond that of the board being around 0.20 inches or less yielding a rectangular parallelepiped packaging size of around 3 to 5 cubic inches.

[0196] FIGS. 33A-33G provide various views of an implementation of a dual QCM sensor assembly according one embodiment of the invention that includes a reference sensor and a selective gas sensor, with the FIGS including an annotated isometric view of the circuit and sensors (FIG. 33A), a top view of a populated circuit board excluding mounted sensors (FIG. 33B), a top view of an unpopulated PCB showing traces and part identification markings (FIG. 33D), a bottom view of the unpopulated PCB showing traces and other features (FIG. 33E). a table correlating descriptions, values or part numbers, with PCB feature identifiers (FIG. 33F), and a schematic drawing of a plurality of individual circuit elements showing connections to other circuit elements (FIG. 33G). FIG. 33E has added yellow ovals that identify though holes, red circles that show a portion of the vias that connect the top and the bottom of the PCB and wherein the four corner holes surround by black mounting pads in the corners are for receiving mounting screws or pins. In the example of FIG. 33A-33G the PCB has a width of 1 inch and a length of 3.5 inches with a component height of less than 1 inch and component width beyond that of the board being around 0.20 inches or less yielding a rectangular parallelepiped packaging size of around 4 to 6 cubic inches.

[0197] The sensor and sensor assemblies of various embodiment of the invention may be used in any application where refrigerant concentration detection is needed. For example, the sensors may be located inside equipment that uses / contains refrigerant or in proximity to or within an environment that contains equipment such equipment or piping that. Examples: HVAC equipment, heat pump water heating equipment, ground source heat pump equipment, grocery refrigeration equipment, automotive HVAC equipment, machinery, and the like.ADDITIONAL REMARKS

[0198] Any materials referenced herein or in any appendix attached hereto are incorporated herein by reference as if set forth in full. To the extent that any definitions or other teachings set forth directly herein (i.e., not incorporated by reference) contract any teachings set forth in an appendix or in other material incorporated herein by reference the precedence given to the teachings or definitions are: (1) teachings set forth directly in the body of the application, then (2) teachings set forth in any appendix, and finally (3) teachings set forth in any incorporated material with material having more recent publication dates taking precedence over material having older publication dates.

[0199] It is intended that the aspects of the invention set forth specifically herein or otherwise ascertained from the present teachings represent independent invention descriptions which Applicant contemplates as full and complete, and that Applicant believes may be set forth as independent claims without need of importing additional limitations or elements from other embodiments or aspects set forth herein for interpretation or clarification. It is also understood that any variations of the aspects (as well as variations in any embodiments) set forth herein represent individual and separate features that may form separate claims or be individually or in any workable combination added to existing claims.

[0200] While various preferred embodiments of the invention are shown and described, it is to be distinctly understood that this invention is not limited thereto but may be variously embodied. From the foregoing description, it will be apparent that various changes may be made without departing from the spirit and scope of the invention.REFERENCE TO ATTACHED APPENDICES

[0201] Additional information about implementations and embodiments of the invention are set forth in the appendices filed herewith. These appendices are incorporated herein by reference:

[0202] Appendix A: Code for Processing Frequency Shift Time Series Data

[0203] Appendix B: Code for Processing Commercial R32 Gas Sensor Data

[0204] Appendix C: Unpublished Manuscription Setting Forth Additional Embodiment Information

[0205] Appendix D: Supplement Information for Appendix C

Examples

Embodiment Construction

[0086]Various advantages and novel features of the present invention are described herein and will become more readily apparent to those skilled in this art from the following detailed description. In the preceding and following descriptions the preferred embodiment of the invention is shown by way of illustration of the best mode contemplated for carrying out the invention perhaps along with one or more variations. As will be realized, the invention is capable of modification in various respects without departing from the spirit of the invention as will be understood by those of skill in the art.

Definitions and Abbreviations

QCM: Quartz Crystal Microbalance[0088]SAW: Surface Acoustic Wave[0089]ZIF-67: Zeolitic Imidazolate Framework-67[0090]HT-MOF: Humidity tolerant metal organic framework crystalline porous material.[0091]Targeted HT-MOF: A humidity tolerant MOF that is selected for its ability to target a particular refrigerant gas for adsorption and thus measurement.[0092]Zero-air...

RELATED APPLICATIONS

[0001] This application claims benefit of U.S. Patent Application No. 63 / 710,968 filed Oct. 23, 2024. This referenced application is hereby incorporated herein by reference in its entirety including any appendices filed therewith and any other materials incorporated therein by reference.STATEMENT AS TO RIGHTS TO INVENTIONS MADE UNDER FEDERALLY-SPONSORED RESEARCH AND DEVELOPMENT

[0002] This invention was made with Government support under Contract DE-AC0576RL01830 awarded by the U.S. Department of Energy. The Government has certain rights in the invention.BACKGROUND OF THE INVENTIONField of the Invention

[0003] The invention generally relates to the field of refrigerant detection and more specifically to refrigerant detection using a QCM / MOF sensor or an SAW / MOF sensor, and even more particularly to sensors where the MOF comprises a targeted humidity tolerant MOF (HT-MOF) and wherein the gas being detected is an A2L or A3 refrigerant.Background Information

[0004] To lower the impact of refrigerant emissions and to comply with the AIM Act, HVAC equipment utilizing mildly flammable A2L refrigerants are being installed in buildings and a further transition to high flammability A3 ultra-low-GWP may be necessary. Refrigerant leak detection systems play a crucial role in safe transition to these new refrigerants.

[0005] Robust, reliable, and low-cost sensors are required to enable a safe refrigerant transition. If lower cost sensors are available, safety can be increased by being able to include more sensors in areas prone to refrigerant leakage. A limited number of sensors are available for A2L / A3 refrigerants that meet various safety standards such as UL, CSA, ASHRAE 15, etc. Safety standards of interest to sensor solutions include the following with only a few sensor solutions useable by manufacturers meeting all requirements of these safety standards:

[0006] 1) IEC 60335-2-40 Edition 6 (January 2018),

[0007] 2) UL / CSA 60335-2-40 (November 2019),

[0008] 3) ASHRAE Standard 15-2019 [4],

[0009] 4) ASHRAE proposed Standard 15.2P (Advisory Public Review), and

[0010] 5) JRA Standard 4068T: 2016R

[0011] Existing sensors have been used such as infrared, electrochemical, semiconductor, speed of sound, and Molecular Property Spectrometer technology. SAW based sensors with MOF coating have also been proposed

[0012] Additional information about prior refrigerant detection technology can be found, for example, in

[0013] 1) Reshniak, Viktor, Praveen Cheekatamarla, Vishaldeep Sharma, and Samuel Yana Motta. 2023. “A Review of Sensing Technologies for New, Low Global Warming Potential (GWP), Flammable Refrigerants” Energies 16, no. 18: 6499. https: / / doi.org / 10.3390 / en16186499

[0014] 2) Zheng, Jiao, Zang, Fiona, Yu, Cheng-Nien, Elbel, Stefan. 2021, Refrigerant Detector Characteristics for Use in HVACR Equipment, https: / / www.ahrinet.org / system / files / 2023-08 / AHRTI%209014%2001%20-%20Final%20Report%20-%20Final.pdf

[0015] 3) Wagner, Mark, and Rebecca Ferenchiak. 2017, Leak Detection of A2L Refrigerants in HVACR Equipment, https: / / www.ahrinet.org / system / files / 2023-08 / AHRTI_9009_Final_Report_0.pdf.

[0016] 4) U.S. Pat. No. 11,796,508, published Oct. 24, 2023, by Kunapuli et al., and entitled “Surface Acoustic Wave Sensor for Refrigerant Leakage Detection”, originally published as WO2021 / 041359 on Mar. 4, 2021. This referenced patent is incorporated herein by reference as if set forth in full herein.

[0017] A need remains for lower cost, robust, reliable, and standard meeting sensors for detecting refrigerants.SUMMARY OF THE INVENTION

[0018] It is an object of some embodiments of the invention is to provide humidity tolerant sensors for the detection of A2L and A3 refrigerants.

[0019] It is an object of some embodiments of the invention to provide QCM or SAW sensor capable of target selected refrigerants.

[0020] It is an object of some embodiments of the invention to provide a method for coating QCM or SAW sensors with selective MOF coatings.

[0021] Embodiments of this invention use QCM and an MOF material or SAW and an MOF material for A2L / A3 gas detection.

[0022] Some embodiments are directed to a coating technique for the MOF with polymer binder that shows stability and repeatability.

[0023] Some embodiments are configured so that sensing output is not negatively impacted by high humidity levels or changes in humidity (e.g. incorporate a humidity sensor and are calibrated).

[0024] Some embodiments are directed to QCM / SAW technologies having sensor response times of less than 10s of seconds and in some cases less than 10 seconds in combination with low manufacturing cost.

[0025] In some embodiments, the sensors may be integrated into HVAC equipment units using A2L or A3 gases.

[0026] In some embodiments, sensors are configured to provide detection of A2L or A3 refrigerant gases at different concentration levels. Sensors include Quartz Crystal Microbalance (QCM) technology or Surface Acoustic Wave (SAW) technology along with Metal Organic Framework materials (MOF). The QCM / SAW sensor is coated with an MOF material which adsorb / desorb the target gas. The adsorption amount of the target gas induces a change in operational parameters of the acoustic / vibration signal (e.g., resonance frequency, phase, dissipation) of the sensor. The parameter change is directly correlated to the actual concentration of the target gas present in the vicinity of the QCM sensor or SAW sensor. This sensing technology enables fast responses for refrigerant leak detection.

[0027] In some embodiments, a hydrophobic MOF coating is used that is hydrophobic, has a reversible water isotherm, and it tailored to target adsorption of a specific refrigerant gas of interest. In some embodiments a ZIF-67 MOF coating is placed on the QCM sensor or a SAW sensor and is used to detect an A2L refrigerant R32 (difluoromethane) or R454B (a blend of 68.9% difluoromethane (R-32) and 31.1% 2,3,3,3-tetrafluoropropene (R-1234yf)0. In some embodiments a ZIF-8 MOF coating is placed on the QCM sensor or a SAW sensor and is used to detect an A3 refrigerant R290 (propane).

[0028] In some embodiments, sensors of different embodiments of the invention may provide refrigerant gas concentration determinations based on mass loading using: (1) time of flight determinations, (2) frequency or frequency shift determinations, (3) phase shift determinations, (4) change of amplitude determinations, or (5) a combination of two or more of these specific determination methods.

[0029] In some embodiments, an additional coating may be placed over the MOF material.

[0030] In a first aspect of the invention a targeted sensor for detecting A2L or A3 refrigerants, comprising: (a) a QCM sensor or a SAW sensor, and (b) a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor.

[0031] Numerous variation of the first aspect exist and include, for example: (1) the first aspect wherein the QCM or the SAW sensor includes a QCM sensor; (2) The first aspect wherein the QCM or the SAW sensor includes a SAW sensor; (3) the first aspect or either of the first or second variations thereof wherein the MOF coating comprises a ZIF-67 coating; (4) the third variation of the first aspect wherein the sensor is configured to target detection of an A2L refrigerant; (5) the fourth variation of the first aspect wherein the A2L refrigerant comprises R32; (6) the sixth variation of the first aspect wherein the A2L refrigerant is R32; (7) the sixth variation of the first aspect wherein the A2L refrigerant is R454B; (8) The first aspect or either of the first or second variations thereof wherein the MOF coating comprises a ZIF-8 coating; (9) the eight variation of the first aspect wherein the sensor is configured to target detection of an A3 refrigerant; and (10) the ninth variation of the first aspect wherein the A3 refrigerant comprises R290.

[0032] In a second aspect of the invention a targeted sensor assembly for detecting A2L or A3 refrigerants, includes: (a) a QCM sensor or a SAW sensor; (b) a humidity tolerant MOF coating on the QCM sensor or the SAW sensor; (c) an oscillator triggering QCM sensor resonance or SAW sensor resonance; (d) a microprocessor or microcontroller for receiving information from the QCM sensor or from the SAW sensor, and using the information and / or outputting one or more of frequency, frequency shift, phase shift. amplitude shift, calibrated concentration of a detected targeted refrigerant, and / or other actionable or control information for enabling a safe operating environment around A2L or A3 refrigerants.

[0033] Numerous variations of the first aspect exist and include, for example: (1) the first aspect wherein the QCM sensor or the SAW sensor includes at least one QCM sensor; (2) the first variation of the second aspect wherein the at least one QCM comprises at least two QCM sensors with at least one QCM not including a humidity tolerant coating; (3) the second variation of the second aspect comprising at least one SAW sensor; (4) the third variation of the second aspect wherein the at least one SAW sensor comprises at least two SAW sensors with at least one SAW not including a humidity tolerant coating; (5) the second aspect or any of the first to fourth variations thereof wherein the MOF coating comprises a ZIF-67 coating; (6) the fifth variation of the second aspect wherein the sensor is configured to target detection of an A2L refrigerant; (7) the sixth variation of the second aspect wherein the A2L refrigerant comprises R32; (8) the seventh variation of the second aspect wherein the A2L refrigerant is R32; (9) the eighth variation of the second aspect wherein the A2L refrigerant is R454B; and (10) the second aspect or any of the first to fourth variations thereof wherein the MOF coating comprises a ZIF-8 coating.

[0034] Numerous additional variations of the second aspect of the invention exist, and include, for example: (11) the tenth variation of the second aspect wherein the sensor is configured to target detection of an A3 refrigerant; (12) the ninth variation of the second aspect wherein the A3 refrigerant comprises R290; (13) the second aspect or any of the first to twelfth variations thereof wherein the assembly further comprises at least one environmental sensor selected from the group consisting of: (A) one more temperatures sensors, (B) one or more pressure sensors, and (D) one or more humidity sensors, wherein the at least one environmental sensor provides an input to the microprocessor or microcontroller to allow inclusion or use of such information when providing output; (14) the second aspect or any of the first to thirteenth variations thereof wherein the one or more sensors and electrical components integrated onto a single PCB; (15) the second aspect or any of the first to fourteenth variations thereof wherein the sensor assembly is packaged to fit in a volume of less than 10 cubic inches; (16) the fifteenth variation of the second aspect wherein the sensor assembly is packaged to fit in a volume of less than 5 cubic inches; (17) the sixteenth variation of the second aspect wherein the sensor assembly is packaged to fit in a volume of less than 3 cubic inches; (18) the second aspect or any of the first to seventeenth variations thereof wherein a t90 sensing time is less than 30 seconds; (19) the eighteenth variation of the second aspect wherein the t90 sensing time is less than 10 seconds; and (20) the nineteenth variation of the second aspect wherein the t90 sensing time is less than 7 seconds.

[0035] Numerous further variations of the second aspect of the invention exist and include for example: (21) the twentieth variation of the second aspect wherein the t90 sensing time was no more than 4 seconds; (22) the second aspect or any of the first to twenty-first variations thereof wherein a t10 recovery time is less than 30 seconds; (23) the twenty-second variation of the second aspect wherein the t10 recovery time is less than 10 seconds; (24)) the twenty-third variation of the second aspect wherein the t10 recovery time is less than 5 seconds; (25) the twenty-fourth variation of the second aspect wherein the t10 recovery time is less than 3 seconds; and (26) the twenty-fifth variation of the second aspect wherein the t10 recovery time is no more than 1 seconds

[0036] In a third aspect of the invention a method of fabricating a targeted sensor for detecting A2L or A3 refrigerants, includes: (a) supplying a QCM sensor or a SAW sensor, and (b) applying a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor via spin coating.

[0037] In a fourth aspect of the invention a method of fabricating a targeted sensor for detecting A2L or A3 refrigerants, includes: (a) supplying a QCM sensor or a SAW sensor, and (b) applying a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor via spray coating.

[0038] In a fifth aspect of the invention a method of fabricating a targeted sensor for detecting A2L or A3 refrigerants, includes (a) supplying a QCM sensor or a SAW sensor, and (b) applying a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor via drop coating.

[0039] Numerous variations of the third to fifth aspects exist and include those noted with regard to the first and second aspects, mutatis mutandis, as well as numerous others that will be apparent form the teachings herein.

[0040] Other objects and advantages of various aspects and embodiments of the invention will be apparent to those of skill in the art upon review of the teachings herein. The various aspects and embodiments of the invention, set forth explicitly herein or otherwise ascertained from the teachings herein, may address any one of the above objects alone or in combination, or alternatively may address some other object of the invention ascertained from the teachings herein. It is not intended that any specific aspect or embodiment of the invention (that is explicitly set forth herein or that is ascertained from the teachings herein) necessarily address any of the objects set forth above let alone address all these objects simultaneously, but some aspects and embodiments may address more than one of these objects.BRIEF DESCRIPTION OF THE DRAWINGS

[0041] FIGS. 1A-1 and 1A-2 provide schematic cut side view of example stacked layers forming a QCM with an HT-MOF and a top view of each of those layers. respectively.

[0042] FIG. 1B provides a schematic representation of an example sensor that includes two quartz crystal resonators with one have an adsorption coating and the other not including such a coating.

[0043] FIG. 1C provides a schematic view of another example multi-detector sensor configuration including a reference sensor and two QCM sensors each with a different adsorption coating to allow detection of different gases of interest.

[0044] FIG. 2A provides a schematic illustration of a single QCM sensor assembly where the QCM is coated with the HT-MOF material for selective detection of a particular refrigerant or group of refrigerants while FIG. 2B provides a similar schematic but which incorporates an additional QCM and QCM oscillator.

[0045] FIGS. 3A-3C provide schematic illustrates similar to those of FIGS. 1A-1C but for a surface acoustic wave sensor or detector (SAW sensor) as opposed to a QCM sensor.

[0046] FIGS. 4A-4B provide sensor assembly configurations 410A and 410B, respectively, similar to 210A and 210B of FIGS. 2A and 2B, respectively but with the QCM sensors replaced by SAW sensors.

[0047] FIG. 5 provides a block diagram setting forth features of a targeted humidity tolerant metal organic framework coatings as well as examples of coating materials and refrigerants whose concentrations are detected.

[0048] FIG. 6 provides a block diagram settings forth a first process for creating targeted HT-MOF coatings

[0049] FIG. 7 provides a block diagram setting forth a second process for creating targeted HT-MOF coatings.

[0050] FIG. 8 provides a block diagram setting forth a third process for creating targeted HT-MOF coatings.

[0051] FIG. 9 sets forth a simplified process flow for using the sensors of some embodiments of the present invention.

[0052] FIG. 10 depicts real time response of QCM sensors with different sorbents coated there on under varying R32 refrigerant concentrations.

[0053] FIG. 11 shows plots of the QCM sensor with MIL-101 under dry and humid air conditions (˜60% RH) run for three cycles each with each cycling operating under varying R32 concentrations ranging from 0% to 4.36% v / v.

[0054] FIG. 12A provides a plot of frequency change vs time from a series of experiments using different R32 concentrations for a sorbent coating of carbonized Zn-MOF-74 under both dry and humid air conditions

[0055] FIG. 12B provides a plot of frequency change vs R32 refrigerant concentration for three set of tests with each involving seven different R32 concentration levels in dry air where the tests were performed on a first day, a seventh day, and a 14th day using a carbonized Zn-MOF-74 coating on a QCM sensor.

[0056] FIG. 13 provides a plot of frequency shift relative to humidity level where the frequency shift was caused by water adsorption in a hydrophobic MOF sensor.

[0057] FIG. 14A provides a plot of Frequency shift vs R32 concentration at different humidity levels for a QCM sensor with a hydrophobic MOF of ZIF-67 while FIG. 14B provides a table of data associated with the plot of FIG. 14A.

[0058] FIG. 15 provides a plot of frequency shift versus time for a series of tests performed with concentrations of R32 progressively oscillating between 0% and successively higher percents (up to 50% LFL) as time progresses and wherein three different plots are overlaid with each corresponding to a different humidity value (% RH).

[0059] FIG. 16A provides an image of peeling cross-hatch tape from a coated QCM sensor and particularly from a QCM sensor with a MIL-101 coating and an overlying parylene coating.

[0060] FIG. 16B provides a table of deposit MOF mass after coating and estimated remaining MOF mass after a tape test

[0061] FIG. 16C provides a table of resonance frequencies for two QCM sensors before and after coating with one being coated via a drop casting method and the other being coated via a spin casing method.

[0062] FIG. 17A is a table that correlates concentration in % LFL and concentration in (PPm,V / V) for an R32 refrigerant.

[0063] FIG. 17B provides plots of frequency shift in Hz against R32 Concentration in % LFL for three different % RH levels.

[0064] FIG. 18 provides a table setting forth estimated uncertainties in % LFL associated with uncertainties in humidity that accompany use of humidity sensors with finite accuracies for R32 concentration detection using humidity tolerant MOF coated QCM sensors.

[0065] FIG. 19 provides an annotated plot of frequency shift QCM sensor response for oscillating R32% LFL exposures and zero-air exposures where each successive R32 exposure had an increase in R32 concentration.

[0066] FIG. 20 provides a plot of R32 concentration values detected by a commercial R32 sensor in an experiment while a QCM / MOF sensor gathered frequency and frequency shift information associated with different concentrations of R32 gas during the experiment of FIG. 19.

[0067] FIGS. 21A-21F illustrate the distribution of frequency shift for R32 concentrations determined by a commercial sensor for variations in R32 concentration around different nominal values of 10, 20, 25, 30, 40, and 50% and with humidity varying from 10-60% RH where each figure focuses on the distribution around different nominal values.

[0068] FIG. 21G provides a table of means and standard deviations for the test data of FIG. 21A-21F based on the nominal R32 concentrations for each of FIGS. 21A-21F.

[0069] FIG. 21H provides a plot of frequency shift vs. R32 concentration when humidity ranged from 10-60% RH based on the average values from FIG. 21G along with a plot of the rest of the data from FIGS. 21A-21F

[0070] FIG. 21I provides a plot of frequency shift vs. R32 concentration when humidity ranged from 10-60% RH and linear fit using based on all of the data of FIGS. 21A-21F.

[0071] FIG. 22 provides a plot of frequency shift vs R32 concentration and vs RH data points for testing results for the spin-coated sensor with the mesh showing the calibration curve.

[0072] FIG. 23 provides a partial plot of over a frequency range of Y Hz to Y+10 Hz, where Y is an unspecified but relevant value for the experiment being run during a cycling of 11 periods of R32 gas presence at different concentrations separated by periods of humid zero-air presence during an example validation test for a spin coated sensor response at a fixed humidity of 36.3% RH such that frequency shift may be ascertained by subtracting a baseline frequency with no R32 present from (at the top of the plot) from a measured frequency obtained in the presence of R32 gas where the same six R32 concentrations are used from during each period.

[0073] FIG. 24 provides a partial plot of R32 concentration values obtained from a commercial sensor over 11 periods of repeated testing of six different R32 gas concentrations with each separated by a period of humid zero-air (i.e. no R32 presence) where each period of constant R32 value lasted one minute wherein the humidity during the entire test was held constant at 36.3% RH.

[0074] FIG. 25A-25C provide relative error amounts using linear regression based on Equation 11, at 6 different R32 concentrations at each of 18.4% RH (FIG. 25A), 36.9% RH (FIG. 25B), and 55.3% RH (FIG. 25C), while FIG. 25D provides summaries of the relative errors at 18.4, 36.9 and 55.3% RH, and a summary of relative errors for all humidity levels.

[0075] FIG. 25E provides a table of mean square error (MSE) and absolute values of the relative error summary at different concentrations for 3 humidity levels using the calculated R32 concentration from the linear regression Equation 11 and an R32 concentration from a commercial R32 gas sensor.

[0076] FIGS. 26A-26C provide relative error amounts using linear regression based on Equation 12, at 6 different R32 concentrations at each of 18.4% RH (FIG. 26A), 36.9% RH (FIG. 26B), and 55.3% RH (FIG. 26C), while FIG. 26D provides summaries of the relative errors at 18.4, 36.9 and 55.3% RH, and a summary of relative errors for all humidity levels.

[0077] FIG. 26E provides a table of mean square error (MSE) and absolute values of the relative error summary at different concentrations for 3 humidity levels using the calculated R32 concentration from the linear regression Equation 12 and an R32 concentration from a commercial R32 gas sensor.

[0078] FIGS. 27A-27C provide relative error amounts using linear regression based on Equation 13, at 6 different R32 concentrations at each of 18.4% RH (FIG. 27A), 36.9% RH (FIG. 27B), and 55.3% RH (FIG. 27C), while FIG. 27D provides summaries of the relative errors at 18.4, 36.9 and 55.3% RH, and a summary of relative errors for all humidity levels.

[0079] FIG. 27E provides a table of mean square error (MSE) and absolute values of the relative error summary at different concentrations for 3 humidity levels using the calculated R32 concentration from the linear regression Equation 13 and an R32 concentration from a commercial R32 gas sensor.

[0080] FIG. 28 provides a schematic representation of a Pierce-Gate Oscillator.

[0081] FIG. 29 provides images of an experimental set up used for testing the QCM / MOF sensor in combination with a frequency counter. a custom oscillator circuit board that included a Pierce-Gate oscillator circuit, and data logging software.

[0082] FIG. 30 provides two plots created using an experimental set up similar to that of FIG. 29 where the plots show frequency versus time as R32 gas levels change over one minute periods of fixed but varying R32 concentration separated by one minute periods without presence of R32 gas so that frequency changes can be seen with one plot showing frequency changes when detection is based on a QCM with no coating while the other shows changes when detection is based on a QCM sensor provided with an HKUST coating.

[0083] FIG. 31 provides an image of an example sensor assembly along with an exploded view of a QCM sensor (which may or may not include a gas selective coating such as an HT-MOF) and its housing.

[0084] FIGS. 32A-32G provide various views of an implementation of a single QCM sensor assembly according one embodiment of the invention, including an annotated isometric view of the circuit and sensor (FIG. 32A), a top view of a populated circuit board excluding mounted sensors (FIG. 32B), a top view of an unpopulated PCB showing traces and part identification markings (FIG. 32D), a bottom view of the unpopulated PCB showing traces and other features (FIG. 32E). a table correlating descriptions, values or part numbers, with PCB feature identifiers (FIG. 32F), and a schematic drawing of a plurality of individual circuit elements showing connections to other circuit elements (FIG. 32G).

[0085] FIGS. 33A-33G provide various views of an implementation of a dual QCM sensor assembly according one embodiment of the invention that includes a reference sensor and a selective gas sensor, with the FIGS including an annotated isometric view of the circuit and sensors (FIG. 33A), a top view of a populated circuit board excluding mounted sensors (FIG. 33B), a top view of an unpopulated PCB showing traces and part identification markings (FIG. 33D), a bottom view of the unpopulated PCB showing traces and other features (FIG. 33E). A table correlating descriptions, values or part numbers, with PCB feature identifiers (FIG. 33F), and a schematic drawing of a plurality of individual circuit elements showing connections to other circuit elements (FIG. 33G).DETAILED DESCRIPTION OF THE INVENTION

[0086] Various advantages and novel features of the present invention are described herein and will become more readily apparent to those skilled in this art from the following detailed description. In the preceding and following descriptions the preferred embodiment of the invention is shown by way of illustration of the best mode contemplated for carrying out the invention perhaps along with one or more variations. As will be realized, the invention is capable of modification in various respects without departing from the spirit of the invention as will be understood by those of skill in the art.Definitions and AbbreviationsQCM: Quartz Crystal Microbalance

[0088] SAW: Surface Acoustic Wave

[0089] ZIF-67: Zeolitic Imidazolate Framework-67

[0090] HT-MOF: Humidity tolerant metal organic framework crystalline porous material.

[0091] Targeted HT-MOF: A humidity tolerant MOF that is selected for its ability to target a particular refrigerant gas for adsorption and thus measurement.

[0092] Zero-air: Dry carrier gas (e.g. dry air) which may be mixed with pure R32 to obtain the different R32 gas concentrations used during testing. Aero-air may also be pass through a gas bubbler where it picks up moisture to become humid, where it is called humid zero-air. The composition of zero-air is: 20.9% v / v Oxygen, 79.1% v / v Nitrogen, <3 ppm H2O, <1 ppm THC (total hydrocarbon).GENERAL EMBODIMENTS

[0093] FIGS. 1A-1 and 1A-2 provide a schematic cut side view of example stacked layers forming a QCM with an HT-MOF and a top view of each of those layers respectively. In this example Layer 102B, i.e., the bottom layer, is a conductive electrode layer with circular central portion and a rectangular portion extending to the left where the left side of the rectangular portion may be used as a contact region when applying an electrical signal to or receiving an electrical signal from this layer. In an example configuration the circular part of 102B may be about 5 mm in diameter while a width of the rectangular portion may be about 3 mm. Layer 101 is a quartz crystal layer which provides the sensor with piezoelectric resonance oscillation when a properly oscillating drive voltage is applied. In an example configuration the nominal diameter of the crystal layer is 14 mm. The crystal layer is typically cut along a selected crystalline axis with its thickness dictating oscillation frequency which is typically around 330 μm for a standard 5 MHz crystal. The next layer is layer 102T or the top electrode layer which has a generally circular configuration which may include a tab region to the right which may extend beyond the quartz layer to allow electrical contact from below. In some examples the nominal diameter of layer 102T may be about 13 mm. some QCM sensors also include a final coating layer which may provide selective absorption of the particular gases who mass may change the oscillation of the frequency of the microbalance thus allow detection of gas concentration once a sensor has been calibrated. In some variations the coating may be limited to a selected portion of the underlying electrode area (e.g. to a circular central portion, to the entire lower electrode area, or even beyond the electrode area such that the coating contacts any extended quartz layer regions that extend beyond electrode 102T. In some embodiments, the above 4 layer stack forms the selective detection element of the sensor but requires a mount, an oscillating electrical input, output resonance frequency detection, and electrical signa processing to provide actionable detection information. In this example, the coating 103 is indicated to be a humidity tolerant (HT) metal organic framework (MOF) coating. In embodiments of the present invention the MOF coating is an HT-MOF that may be, for example, an ZIF-67 MOF. The coating can be tailored for selective adsorption of gas molecules (e.g. refrigerant gas) molecules and particularly A2L or A3 gas molecules at concentrations levels well below that of levels that could be dangerous). Also illustrated FIG. 1A-1 are two electrical connection locations, EC-B and EC-T, with one for contacting the 102B and the other for contact 102T where contact can be made from the bottom in both cases. In other embodiments, other contact and layer configurations may be used.

[0094] FIG. 1B provides a schematic representation of an example sensor that includes two quartz crystal resonators with one like that of FIGS. 1A-1 and 1A-2 but with the other lacking element 103 has it is provided as a reference sensor that provide frequency information without any adsorption coating. Such a reference sensor may be useful in isolating certain environmental changes that cause oscillation frequency shifts that have nothing to do with target gas presence.

[0095] FIG. 1C provides a schematic view of another example multi-detector sensor configuration including a reference sensor and two QCM sensors each with a different adsorption coating to allow detection of different gases of interest. In the FIG. one of the coating material is shown as 103-1 while the other is shown as 103-2.

[0096] FIG. 2A provides a schematic illustration of a single QCM sensor assembly 210A where the QCM is coated with the HT-MOF material for selective detection of a particular refrigerant or group of refrigerants. The assembly includes a The QCM / HT-MOF detector which is driven by an oscillator / QCM interface module IC which may be replaced by a plurality of discrete components in some embodiments. The Oscillator / QCM interface in-turn supplies resonant frequency information to a microprocessor or microcontroller which may provide information in a number of different forms depending on how it is programmed or hard coded. For example it may simply pass along raw frequency information to a separate processing circuit, it may pass along frequency shift or change information, it may include use of stored calibration information which may be used to pass on gas concentration information, or it may alternatively or additionally pass on warnings (e.g. contact relevant personnel or authorities, sound alarms, or the like) and / or it may automatically institute preventive measures (e.g. turn on venting fans, turn off gas supply lines, otherwise isolate gas into smaller volumes, and the like). In addition to receiving frequency information from the detector, the microprocess may receive additional information from other sensors that may form part of the assembly wherein the additional sensors may include one or more of a humidity sensor, a temperature sensor, and / or a pressure sensor which may be in turn used to provide more accurate output information. In some embodiments, multiple sensors may be integrated into a single device. In other embodiments, other sensors may be included while in still other embodiments, redundant sensors or processors may be included, in still further embodiments fail safe features and configurations may be included.

[0097] FIG. 2B provides a similar schematic illustration of a sensor assembly 210B similar to that of FIG. 2A but directed to a two QCM sensor assembly like that of 102-B where the second reference QCM and driving oscillator / interface circuit is included with its information also going to the microprocessor to provide enhanced processing and output.

[0098] FIGS. 3A-3C provide schematic illustrates similar to those of FIGS. 1A-1-1C but for a surface acoustic wave sensor or detector (SAW sensor) as opposed to a QCM sensor. As shown in FIG. 3A the saw detector 300A is formed by three primary layers: (1) a quartz piezoelectrical layer 301; (2) a patterned layer of metal traces 302T which provide a pattern for acoustic wave generation and a pattern for acoustic wave detection; and (3) a layer 303A of material that can selectively adsorb and release gases that are to be measured which change a response characteristic of the electrical circuit. In some embodiments of the present invention the coating 303 is in the form of an HT-MOF such as ZIF-67. More information about SAW detectors can be found in U.S. Pat. No. 11,796,508, which was referenced above, FIG. 3B provides a sensor group 300B that is similar to that of FIG. 1B and which includes two SAW detector elements with one including an HT-MOF while the other has no coating material and functions as a reference detector. FIG. 3C provides a sensor group 300C that is similar to that of FIG. 1C but all three sensors including SAW sensors as opposed to QCM sensors with one including no coating material, one including an HT-MOF material labeled as 303-1, and another including a different HT-MOF labeled as 303-2.

[0099] FIGS. 4A-4B provide sensor assembly configurations 410A and 410B, respectively, similar to 210A and 210B of FIGS. 2A and 2B, respectively but with the QCM sensors replaced by SAW sensors.

[0100] FIG. 5 provides a block diagram setting forth features of a targeted humidity tolerant metal organic framework coating and examples.

[0101] FIG. 6 provides a block diagram setting forth a first process for creating a HT-MOF coating involving spin coating of an HT-MOF slurry onto a QCM or SAW sensor surface to enable the surface to target adsorption of a selected refrigerant molecule.

[0102] FIG. 7 provides a block diagram setting forth a second process for creating a HT-MOF coating involving spray coating of an HT-MOF slurry onto a QCM or SAW sensor surface to enable the surface to target adsorption of a selected refrigerant molecule.

[0103] FIG. 8 provides a block diagram setting forth a third process for creating a HT-MOF coating involving drop coating of an HT-MOF slurry onto a QCM or SAW sensor surface to enable the surface to target adsorption of a selected refrigerant molecule.

[0104] FIG. 9 sets forth a method 921 of using sensors of some embodiments of the invention.EXPERIMENTS, IMPLEMENTATIONS, AND OTHER EMBODIMENTSOverview of Initial Sorbent Screening:

[0105] Five sorbents were located on a surface of an upper electrode of a Quartz Crystal Microbalance (QCM) sensor and were initially screened for their optimal response to R32 under ambient conditions. For all the tests performed, mass flow controllers (MFC) (Alicat Scientific, MCS-20SCCM-D / 5M) are used to measure the flow rates of the different gases. FIG. 10 shows response of the sorbents selected for study. These five were selected based on their use in previous experiments at PNNL and based on a literature review. Initial sensor response (frequency shift) was tested for each sorbent selected. These tests involved passing small pulses of R32 with a 0% to 4.36% v / v concentration in dry air over the sensor with the resulting frequency shift being recorded in real time. The response of the sensors to varying concentration of R32 were observed to be rapid and reversible. Based on these observations two of the Metal-Organic Framework (MOF) sorbents, MIL-101 and HKUST-1, were selected for further study due to their high signal to noise ratio and fast response to R32 gas.

[0106] Further testing of these sensors posed a different challenge, while the sensors showed sensitive, repeatable, and reversible signal change to R32, under humid conditions especially at high humid conditions, the signal from these sensors becomes damped and not reproducible on subsequent testing. FIG. 11 shows response of the QCM sensor loaded with MIL-101 under three consecutive dry condition runs and then underwent three humid condition runs where the humidity was set at approximately 60% relative humidity (RH) where each run underwent cycling with different R32 concentrations similar to that used in the experiment of FIG. 10. When humid air was introduced (˜60% RH) the signal from the same sensor became noisy. Similar observations were found for the other candidate sorbent, HKUST-1. These results suggest that water molecules compete with the adsorption of refrigerant molecules in the candidate sorbents.Humidity Mitigation Strategies

[0107] Initial attempts to mitigate the impact of moisture were focused on increasing the hydrophobicity of the sensor. This included coating the existing candidates with a hydrophobic polymer coating. Additional hydrophobic sorbents were also studied, including UiO-66, MCM-41 (mesoporous silica), carbonized Zn-MOF-74, and dealuminated zeolites. While these sorbents showed initial promise, the response of these sensors were not repeatable over time due to either unpredictable or large loadings of water molecules in the pores of these substrates. FIGS. 12A and 12B shows the response from one such attempt using a carbonized variant of Zn-MOF-74, where the MOF was thermally treated under inert conditions (N2 flow) to become more hydrophobic. While the hydrophobic sorbents initially tested gave a better response in humid conditions as shown in FIG. 12A When tested over time a decrease in their response was seen similar to that seen with the other sorbents that were tested even under dry testing conditions as shown in FIG. 12B. This observation suggests water adsorption within the pores of these materials accumulated over time, occupying active sites for R32 adsorption. This indicated that the pore opening which was not changed in the carbonization may be the reason for the irreversible water loading.

[0108] A hydrophobic MOF with small pore openings and reversible water isotherms is preferred because water adsorption is minimized and controlled to be the same under both the adsorption and desorption processes. A hydrophobic MOF with a reversible hysteresis free water isotherm was selected as the sensor material while following other selection criteria such as, for example, the material having hydrophobic character and having an appropriate pore size. In some embodiments the following selection criteria may apply:

[0109] 1. Hydrophobicity: the material should be hydrophobic toward water vapor.

[0110] 2. Water isotherm:

[0111] a. Material should have type II or III water isotherm defined by the International Union of Pure and Applied Chemistry (IUPAC) with low water adsorption.

[0112] b. No or small hysteresis for water adsorption and desorption.

[0113] 3. Pore size: Pore size should be slightly big enough to adsorb the target gas for fast adsorption and quick steady state sensor signal response. The pore size should promote fast adsorption and desorption of the target gas for gas sensing. An appropriate pore window in some cases may be close (0.9 to 1.1 X, where X is the size of the refrigerant molecule and has no pore cage larger than 2 nm. The pore window is important for selectivity while the pore cage is related to hysteresis

[0114] 4. Selectivity: The material should be selective to the target gas.

[0115] 5. Chemical and temperature stability: the material should exhibit stability in water and long-term stability when exposed to target refrigerant and moderate temperature changes. The material should have long term structural stability. Long term refers to a period of time comparable, greater, or substantially greater than anticipated sensor life.Test Results with Hydrophobic MOF ZIF-67

[0116] Water adsorption and gas adsorption isotherms were collected and studied for a series of sorbents. The MOFs with high surface area such as MIL-101 and HKUST-1, showed diminished responses over time due to moisture accumulation, and they did not sustain response to R32 under humid conditions. Based on this failure a hydrophobic MOF was selected for consideration as a sensor coating. ZIF-67 has chosen as a coating material as it showed relatively higher surface area, low water adsorption and appropriate pore size to provide a rapid and repeatable response to R32.

[0117] ZIF-67 has a water isotherm that is ideal for this application, it is considered hydrophobic, and its water adsorption is minimal and hysteresis free in moderate humid conditions. All test results described below were performed at a constant room temperature of 23° C. and a gas flowrate of 20 standard cubic centimeters per minute (sccm). Tests for water adsorption by the hydrophobic MOF sensor were performed. FIG. 13 shows the adjusted frequency shift result coming from such tests at different humidity levels (10, 20, 30, 40, 50 and 60% RH). The several days test results show a linear increase of the frequency shift as the humidity level increases. The linear fit curve enables compensation for the humidity response of the sensor during calibration.

[0118] FIG. 14 provides plots for sensor response in frequency shift (Hz) vs R32 concentration in terms of % of lower flammability level (LFL) for three different nominal humidity levels of 40% RH, 50% RH, and 60% RH and 6 different R32 concentrations of 10, 20, 25, 30, 40, and 50% LFL over several days of testing. Each point in the plot is the average of 3 runs at the same concentration of R32. The maximum standard deviation was 1.4 Hz which showed good repeatability. As shown in the figure, at a constant humidity, the frequency response of the sensor increased as the R32 concentration increased, showing almost perfect linearity.

[0119] In theory, as the concentration increases, the hydrophobic MOF adsorbs more R32 molecules, which increases its mass, consequently increasing the frequency shift of the resonance frequency of the sensor. As the humidity level increases, the frequency response of the sensor at a constant concentration of R32 decreases, indicating the competitive adsorption of R32 and water molecules. The test results of the sensor at constant humidity can be linearly fitted for sensor calibration, as shown in FIG. 14.

[0120] A validation of the results shown in FIG. 14A was performed at 43.31% RH humidity level. In this process three tests were performed with different concentrations of R32. To calculate the concentration of R32 using the sensor's frequency response obtained during testing, the linear fit curve equation of y=0.7109x (i.e. the linear curve fitting result for the plot of plot of 43.15% RH of FIG. 14) was used, where y is the sensor frequency shift and x the concentration of R32. A commercial sensor was used as a reference to benchmark the actual concentrations of R32. Test results for 3 tests are shown in FIG. 14B (i.e., for tests 1, 2, and 3 respectively). These results show that the targeted HT-MOF using ZIF-67 on a QCM sensor can detect the R32 concentration.

[0121] FIG. 15 shows an example of the time-based raw frequency shift signal of the sensor for different concentrations of R32 at different nominal humidity levels of 40, 50, and 60% RH. As previously explained, the sensor response increases with increasing R32 concentration at constant humidity and at a constant concentration, the response decreases with increasing humidity levels.

[0122] In summary, the sensor response with the hydrophobic MOF (ZIF-67) in different humidity conditions is linear, indicating an ability to easily compensate for humidity. Furthermore, the sensor response increases with increasing R32 concentration. The frequency response of the sensor in these experiments was relatively low for different concentrations of R32 at a maximum of 36 Hz at 50% LFL.Demonstration of Good Adhesion to Substrate

[0123] Sensing film with good adhesion to the substrate is important for long-term performance of the coated QCM sensors and for SAW (surface acoustic wave) sensors that may be used as an alternative to QCM sensors in some embodiments. One of the most common ways to test bonding between microfabricated layers is known as ‘the tape test’ which is an adaptation of ASTM D3359 “Standard Test Methods for Measuring Adhesion by Tape Test”. Cutting patterns of lines in the layer to be tested is often unnecessary when testing microfabricated devices. In an ASTM D3359 procedure, readily available tape is manually applied to the test layer and then rapidly removed without using significant equipment. However, this test has many limitations because parameters including the magnitude and duration of tape application pressure, waiting time between bonding and removing.

[0124] The test results for ZIF-67 were obtained on sensors without any Parylene coating which is sometimes used on some coated sensors. Polymer coatings such as Parylene can act as a protection layer and seemed to help reduce mass loss of the sensing film in the case of MIL-101.

[0125] Cross-hatch adhesion tape specifically for ASTM D3359-22 was used during the tape testing. Initially, the QOCM sensors were affixed to microscope slides. Then the adhesion tape was gently and firmly pressed onto the area coated with MOF. Finally, the tape was peeled off swiftly at the direction parallel to the QCM surface, as shown in FIG. 16A.

[0126] Instead of counting the missing crosscut area, we estimated the mass loss after the tape test by measuring the change of the resonance frequency of the QCM sensor before and after tape testing. This estimation can be obtained from the following equation which sets forth the relationship between the resonance frequency of the QCM sensor and the mass deposited on the sensor's facef=-2⁢f02A⁢ρq⁢μq⁢Δ⁢mEQ. 1where:

[0128] f0=Resonance frequency of the fundamental mode (Hz) which is 5 MHz in the sensor used

[0129] f=normalized frequency change (Hz)

[0130] Δm=Mass change (g)

[0131] A=Piezoelectrically active crystal area (Area between electrodes, A=1.539 cm2)

[0132] ρq=Density of quartz (ρq=2.648 g / cm3)

[0133] μq—Shear modulus of quartz for AT-cut crystal (μq=2.947E11 gcm−1s−2)The deposited mass of MOF on the surface of the QCM sensors after coating are calculated and summarized in the Table of FIG. 16B. The after coating resonance frequency was obtained at 10% RH. FIG. 16C provides a table of resonance frequencies for two QCM sensors before and after coating with one being coated via a drop casting method and the other being coated via a spin casing method. The deposited mass of MOF on the surface of QCM sensors after tape testing were calculated and are also summarized in the table of FIG. 16B along with the remaining mass portion as a percentage.

[0134] For the sensors only coated with MOF, the remaining portion is within the range of 51.11-69.9%. For the sensors with additional coating of Parylene, a greater portion of coating materials (including both MOF and parylene) was found to be remaining on the QCM surface. The test results for ZIF-67 were obtained on sensors without any Parylene coating. Polymer coatings such as Parylene acted as a protection layer and seems to help reduce mass loss of the sensing film in the case of MIL-101.QCM Sensors Data Analysis

[0135] Comparisons were made between ZIF-67 MOF coatings applied by two different method to a commercial QCM sensor from AWSensors. As with the experiments noted above, these experiments used an R32 target gas which is a target A2L gas for a sensing element of some embodiments of the invention. The fundamental frequency of the sensors are 5 MHz with one sensor coated with the MOF material via a drop casting technique and the other using a spin coating method. The sensors coated via these different methods showed different responses to different humidity levels.Relative Humidity Sensitivity Analysis for the Drop-Casted MOF Sensor.

[0136] A MOF sensor can be made using drop coating via a micropipette. For drop-casting process, 20 μL of MOF slurry was deposited on the QCM sensor and allowed for the solvent in the MOF slurry (ethanol but could also be methanol) to evaporate. After two minutes, the process was repeated with additional 20 μL and 10 μL MOF deposition with at least two-minute interval between each repeat for the earlier MOF layer to dry. The mass loading of the MOF material is estimated at 140 μg sensor testing, FIG. 17A provides a table correlating R32 concentration in % LFL to parts per million by volume (ppm, v / v). FIG. 17B provides plots of frequency shift in Hz against R32 Concentration in % LFL for three different % RH levels for a sensor with a drop cast ZIF-67 MOF coating. FIG. 17B is similar to that of FIG. 14A but with additional features shown. Plots like that of FIG. 17B may be used in a process for estimating the impact of uncertainties of relative humidities on the R32 concentration readings for a sensor.

[0137] For a humidity level of 43.15% RH (denoted as R1), the fitted equation for this relative humidity (RH) as shown in FIG. 17B is y=k1x=0.7109x. The corresponding angle with the positive x-axis is α1. For a humidity level of 52.34% RH (denoted as R2), the fitted equation is y=k2x=0.3955x as also shown in FIG. 17B. The corresponding angle with the positive x-axis is α2. The angle between these two lines is β, which can be calculated as:β=α2-α1=tan-1(k2)-tan-1(k1)=tan-1(0.3⁢9⁢5⁢5)-tan-1(0.7⁢1⁢0⁢9)=-3⁢5.4⁢0⁢9EQ. 2The unit angle per percentage humidity level can be obtained by:β(R2-R1)=-35.4095⁢2.3⁢4⁢%-4⁢3.1⁢5⁢%=-1.505⁢° / %EQ. 3For a certain humidity level R; within the range of R1 and R2, the angle with the positive x-axis at this RH level is:αi=α1+(Ri-R1)⁢β(R2-R1),EQ. 4The corresponding slope is ki=tan(αi), thus the relation between frequency shift and R32 concentration at humidity level Ri is:y=ki⁢x=tan⁡(αi)⁢x=tan⁡(αa+(Ri-R1)⁢β(R2-R1))⁢x.EQ. 5If the sensitivity of the humidity sensor is ε, the equations for RH level Ri−ε and Ri+ε are:y=tan⁡(αa+(Ri-ε-R1)⁢β(R2-R1))⁢x⁢ and⁢ y=tan⁡(αa+(Ri+ε-R1)⁢β(R2-R1))⁢x.EO. 6Given a certain frequency shift value F, the concentrations corresponding to RH levels Ri−θ, Ri, and Ri+θ are ci−1, ci, and ci+1 respectively, which can be calculated as:ci-1=Ftan⁡(αa+( RHi-θ- RH1)⁢β( RH1- RH2))EO. 7ci=Ftan⁡(αa+( RHi- RH1)⁢β( RH1- RH2))EQ. 8ci+1=Ftan⁡(αa+( RHi+θ- RH1)⁢β( RH1- RH2))EQ. 9The error at RH level of Ri is given as:-ϵ1=-<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[LeftBracketingBar]"< / annotation>< / semantics>ci-ci-1ci<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[RightBracketingBar]"< / annotation>< / semantics>⁢ and+ϵ2=+<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[LeftBracketingBar]"< / annotation>< / semantics>ci+1-cici<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[RightBracketingBar]"< / annotation>< / semantics>.EQ. 10Using the above method, R32 concentration estimates and maximum uncertainties of the R32 concentrations caused by the uncertainties of relative humidities can be estimated. The results for the current experiment and analysis are set forth in the table of FIG. 18. These uncertainties of the R32 concentration are obtained based on the relations between frequency shift (Hz) and R32 concentration (% LFL) at 43.15, 52.34 and 60.7% RH shown in FIG. 14A. Available example on-chip humidity sensors of certain accuracy are also listed in FIG. 18 along with details of unit price for bulk purchase (1000+ units), response time, and long-term drift.Product HDC3020DEFR is an integrated relative humidity and temperature sensor manufactured by Texas Instruments. Both products SHT45-AD1B-R2 and SHT40-AD1B-R2 also have integrated relative humidity and temperature sensors and are both from Sensirion AG. Products like these or other integrated sensors may may be integrated into a sensor assembly of some embodiments of the invention to provide necessary inputs for providing calibrated concentration levels of R32 or other refrigerants when a QCM or SAW sensor is used with a humidity tolerant MOF (HT-MOF) coating like ZIF-67 where “humidity tolerant” refers to a coating that has low water uptake (typically less than 5 wt. % at RH 50% and room temperature), has small pore openings (e.g. with a size of 0.3 to 0.5 nm, more preferably 0.35 to 0.45 nm, and has a reversible hysteresis free water adsorption-desorption isotherm such that capillary condensation in the pores is minimized and controlled to be the same under both the adsorption and desorption process.Due to the intrinsic dependencies on the ambient humidity, some commercial gas sensors were built with on-chip humidity sensors for self-calibration to get measurement accuracy across wide environmental range. Some of the available commercial gas sensor feature humidity compensation include, for example:1) SGP30 multi-gas (VOC and CO2eq) sensor, SGP 41 multi-gas (VOC and NOx) sensor from Sensirion AG,2) MPS™ refrigerant gas sensors (including R32, and R-454 sensors, etc.),3) AM4205H-LC-R32 Cubic R32 A2L Refrigerant Gas Sensor.None of these sensors are QCM or SAW sensors and more particularly none of these sensors are QCM or SAW sensors with humidity tolerant MOF coatings.Relative Humidity Sensitivity Analysis for the Spin-Coated MOF Sensor.Spin coating is a common technique for producing thin films of materials on various substrates, including sensors, and it's used in both scientific and industrial applications. An excess solution is poured onto the substrate which is rotated at a constant angular velocity which results in liquid flow radially outward, driven by centrifugal force. The excess liquid is quickly eliminated as droplets from the substrate perimeter and the film thickness becomes uniform during spin-off.Spin coating of MOF sensing films on QCM sensor were investigated. In one example of making such a film the following process was used. A 50 μl MOF particle dispersion using 2-10 g / L MOF with 0.2-1 g / L polymer binder was created in methanol which was then an amount of it was dropped to the center of a 5 MHz QCM substrate. The substrate was rotated at 500 rpm about its center for about 30 seconds. This process was repeatable until a desired mass loading of MOF was achieved on the substrate. The MOF-coated QCM sensor was air-dried before testing. The estimated mass loading of the ZIF-67 MOF material is 104 μg and the diameter of the coated substrate is about 14 mm.To measure the effect of humidity on the coated QCM sensor, multiple tests at a constant temperature of 23° C. and different humidity levels (10-60% RH) at 10% RH increments were performed. The tests consist of flowing different concentrations of R32 over the coated QCM sensor at constant relative humidity. R32 was introduced in successively increased concentrations (R32 plus air) for a period of time with each separated by a period of zero R32 concentration (i.e. humid zero-air). Each concentration of R32 is flowed over the sensor for one minute followed by one minute of humid zero-air flow. The following nominal concentration levels of R32 were used during the tests: 10% LFL, 20% LFL, 25% LFL, 30% LFL, 40% LFL, 50% LFL. The actual R32 concentrations were measured using a commercial R32 gas sensor (Nevada Nano, MPS™ A2L Refrigerant Gas Sensor) with a detection range of 5-100% LFL with an accuracy of ±3% LFL. Humidity in the tests was generated by flowing zero-air into liquid water in a gas washing bottle. Humid zero-air coming out of the gas washing bottle carries out water vapor into the QCM sensor flow cell. Humidity levels during the testing were varied by changing the humid zero-air gas flowrates. The volumetric flow rate of zero-air was {dot over (v)}1, humid zero-air was {dot over (v)}2, and dry R32 was {dot over (v)}3. The condition {dot over (v)}1+{dot over (v)}2+{dot over (v)}3=20 sccm was always maintained during all testing.The RH level was obtained from the built-in humidity sensor of the commercial R32 gas sensor, which was placed at the outlet of the flow cell holding the QCM sensors.The time series data of the frequency shift of the QCM coated sensor was recorded. The test results taken from multiple days were used for the sensor calibration.A computer algorithm was developed in MATLAB and was used to process the frequency shift time series data obtained from the experimental tests. The timeseries data was initially filtered using a moving average function with a moving average window size of 25 data samples. The algorithm calculates the maximum frequency shift values during the time the QCM sensor was exposed to different R32 concentrations. The maximum peak to peak frequency shift over a threshold value were saved and categorized for the corresponding R32 concentrations. FIG. 19 shows an annotated example of the algorithm output of the processed test data with peaks points in the frequency shift detected for increasing concentrations of R32. The MATLAB code is set forth in Appendix A attached to this application. During this test, the humidity was substantially constant and had a nominal value of 10% RH for the data shown.To get corresponding actual R32 concentration during the tests, a commercial R32 gas sensor's data is collected and saved. The commercial R32 gas sensor data is processed using another algorithm developed in MATLAB which is set forth in Appendix B. The commercial sensor data is initially filtered using a moving average function with a sample data window size of 2. Then, the significant peaks corresponding to each R32 concentration are detected and saved. The algorithm-detected concentration points are added to the corresponding frequency shift in the processed QCM / MOF sensor data. FIG. 20 shows an example of the output of the processed R32 concentration points obtained from the commercial R32 gas sensor data that measured R32 concentration while frequency change data of the QCM / MOF sensor was being recorded 859 data points for each R32 concentration and each different humidity level (10-60% RH) were collected from the testing of the spin coated QCM sensor.Calibration Curve without Humidity.An initial calibration evaluation process consisted of considering only the sensor's frequency shift with different R32 concentrations while excluding taking into consideration humidity level information. FIGS. 21A-21F illustrate the distribution of frequency shift for R32 concentrations determined by a commercial sensor for variations in R32 concentration around different nominal values of 10, 20, 25, 30, 40, and 50% LFL and with humidity varying from 10-60% RH, where FIG. 21A shows data associated with a nominal 10% LFL, FIG. 21B shows data associated with a nominal 20% LFL, 21C shows data associated with a nominal 25% LFL, FIG. 21D shows data associated with a nominal 30% LFL, 21E shows data associated with a nominal 40% LFL, FIG. 21F shows data associated with a nominal 50% LFL. FIG. 21G provides a table of means and standard deviations for the test data of FIG. 21A-21F based on the nominal R32 concentrations for each of FIGS. 21A-21F.The average values of FIG. 21G are plotted in FIG. 21H, alongside all the processed data clumped around their means. FIG. 21H also shows the results of a linear fit based on the average values which can be used to obtain a calibration curve for the sensor. The indicated r-square value of 0.9999 shown in FIG. 21H shows almost perfect linear fit for the average values. This indicates the sensor's frequency shift response is linear with increasing R32 concentrations. The initial linear fit equation is:y=0.1⁢0⁢7⁢xEQ. 11where y is the frequency shift response in Hz of the sensor and x is the R32 concentration in % LFL.Using a linear fit of all the testing data of FIG. 21H produces the distribution of and calibration curve of FIG. 21I. A calibration curve having an r-square value of 0.993 and mean square error value of 0.014, is obtained:y=-0.0⁢7⁢4+0.1⁢09⁢xEQ. 12Calibration Curve with Humidity as a ParameterConsidering the relative humidity values for each testing condition, a new calibration curve can be obtained for all testing results. The same testing data used in the initial calibration, previously discussed, is used here but with the addition of the relative humidity values. A total of 5154 data points for each parameter, notably R32 concentrations obtained from the commercial R32 gas sensor, the relative humidity values, and the frequency shift from the QCM sensor's response. Using the linear regression function in MATLAB, a calibration curve for the R32 concentration values with two-parameters as the RH values and sensor's frequency shift was obtained as:x=-1.1⁢3⁢2+9.1⁢2⁢5⁢y+0.0⁢53⁢zEQ. 13where z represents the RH value in % RH. The r-square value is 0.997, whereas the mean square error is 0.479. FIG. 22 shows the testing data for the new calibration with RH values, with the mesh grid showing the calibration curve.Sensor Response Validation Process for the Spin Coated QCM Sensor.Three sets of validation tests were performed, at constant temperature of 23° C. RH level is constant for each set of tests, within the range of 10-60% RH.For a specific set of tests, the volumetric flow rate of the humid zero-air gas {dot over (v)}2 was kept constant. During each cycle of the validation tests, R32 / zero-air mixture of six different concentrations were flowed over the target sensor intermittently for a duration of one minute, followed by a one-minute interval of gas with 0% LFL of R32 (i.e., {dot over (v)}3=0). This cycle was repeated sixty times. An example of the measured data for one set of validation testing at 36.3% RH is presented in FIG. 23. The peak frequency shifts were obtained from the algorithm developed in MATLAB (See Appendix A).The peak values of the commercial R32 gas sensor readings at different R32 concentrations were obtained from data processing and are shown in FIG. 24 with the peak points highlighted with light green dots. It is worth noting that each set of measurements involving 6 different R32 concentrations did not provide those concentrations in an order from smallest to largest, or vice-a-versa, but instead provided them in an order that mixed higher and lower values to provide improved results.Validation for the Calibration Curve without Humidity.For the spin coated QCM sensor, per the testing results' calibration curve without humidity discussed above, the relation between frequency shift y, and R32 concentration x is represented by the Equation 11. Thus, with known frequency shift y, the corresponding R32 concentration can be calculated as:x=y0.1⁢0⁢7.The relative error from the calculated concentrations obtained from the QCM sensor frequency shift and the measured concentration from commercial R32 gas sensor values is given as follows:Relative⁢ Error⁢ (%)=x-x0x0·100EQ. 14An error analysis of the linear fit Equation 11 at 18.4, 36.9 and 55.3% RH are summarized in box plots of FIGS. 25A-25D. With FIGS. 25A-25C respectfully providing errors for individual LFL bases for each of these three RH levels while FIG. 25D provides an overall summary for each RH level as well a single summary for all LFL and RH values combined. Average concentration values, MSE, average MSE, and maximum and minimum relative error values are provided in the Table of FIG. 25E. In FIGS. 25A, 25B, and 25C, each subplot or boxplot contains 6 boxes, each box corresponds to certain R32 concentration, which contains 60 points. In FIG. 25D, the first three boxes correspond to three different relative humidities, which contains 360 (60×6) data points each, the last box contains all the data points corresponding to all concentration and humidity levels, which contains 1080 (360×3) data points. As observed in FIG. 25D, most of the outliers of the points are located above 10%.With Equation 11 as shown in FIG. 21H, the minimum and maximum absolute relative errors for the validation tests at 18.4, 36.9 and 55.3% RH are, respectively,18.4% RH: 0.002% and 6.059%,36.9% RH: 0.005% and 10.099%, and

[0163] 55.3% RH: 0.015% and 14.930%To gauge the performance of the calculated concentration values to the commercial R32 gas sensor values Eq. 11 Was used, for the validation tests at each humidity with the mean square error (MSE) is given by:MSE=1N⁢∑i=1N(x-x0)2N is the number of tests, which equals to 60 for each humidity level.The calculated MSE values for the tests performed at each humidity level, and for all test data points are:18.4% RH: 2.02936.9% RH: 2.248

[0166] 55.3% RH: 2.685

[0167] All data: 2.321.

[0168] The low MSE values indicate that the calibration curve (equation 11) performed well in calculating the R32 concentrations from the spin coated QCM sensor frequency shift response. As shown in FIG. 25D, the relative error as well as the MSE values increase with increasing humidity level.

[0169] Similarly, using the linear fit curve given by Equation 12, an error analysis can be provided. FIG. 26A-26D depicts the results of such an analysis. While FIG. 26E provides a table of additional data. The illustrations of FIGS. 26A-26D and the data provided in the Table of 26E of analogous form to those of FIGS. 25A-25E.

[0170] With the linear fit of Equation 12, the minimum and maximum absolute relative errors for the validation tests at 18.4, 36.9 and 55.3% RH are shown as follow:

[0171] 18.4% RH: 0.015% and 4.762%,

[0172] 36.9% RH: 0.087% and 7.583%

[0173] 55.3% RH: 0.012% and 12.474%

[0174] The calculated MSE values for the tests performed at each humidity level, and for all test data points are:

[0175] 18.4% RH: 1.350

[0176] 36.9% RH: 1.957

[0177] 55.3% RH: 3.459

[0178] All data: 2.256.Validation for the Calibration Curve with Humidity as a Parameter.

[0179] By considering the factor of relative humidity level (% RH), a two-parameter regression equation, Equation 13, was proposed with a plot of a calibration curve being shown in FIG. 22. The error analysis corresponding to this updated two-parameter linear regression at 18.4, 36.9 and 55.3% RH are demonstrated in the box plots of FIG. 27A-27D and with summary data shown in FIG. 27E. The illustrations of FIGS. 27A-27D and the data provided in the Table of 27E of analogous form to those of FIGS. 25A-25E, and 26A-26E.

[0180] Using Equation 13, the minimum and maximum absolute relative errors for the validation tests at 18.4, 36.9 and 55.3% RH can be shown to be:

[0181] 18.4% RH: 0.012% and 9.205%,

[0182] 36.9% RH: 0.125% and 7.587%

[0183] 55.3% RH: 0.025% and 7.239%

[0184] The calculated MSE values for the tests performed at each humidity level, and for all test data points are:

[0185] 18.4% RH: 0.758

[0186] 36.9% RH: 1.644

[0187] 55.3% RH: 1.536

[0188] All data: 1.312.

[0189] With the calibration curve with humidity as a parameter, the overall maximum absolute relative error, especially maximum absolute relative error at higher RH levels, decreased. The MSE value declined, and the number of outliers were also observed to be reduced. The error analysis implies that the predicted concentration value from the two-parameter regression equation is closer to the real concentration, especially at higher RH levels. Thus, the accuracy of the calibration curve with humidity as a parameter is higher than the calibration curve without humidity.

[0190] In sum, compared to the drop cast sensor, the new QCM sensor using the spin coating technique is more tolerant to high humidity changes. It's important to note that the new spin coated sensor's outstanding performance under different humidity levels may be due to the uniformity of the coating on the QCM sensor as well as the low mass loading of the hydrophobic MOF material, which reduces the amount of water adsorbed by the MOF material thus reducing the competitive adsorption by water vapor and by target gas molecules.Initial Proof-of-Concept Design

[0191] Given that commercial QCM sensors (from AWSensors) with humidity tolerant MOF coatings added by the inventors can measure frequency shifts associated with R32 gas concentrations as demonstrated by the prior test results, the inventors decided to create their own oscillation circuitry for QCM sensing to verify frequency shifts on a custom circuit board. A classic Pierce-Gate oscillator was chosen for the initial effort. Such an oscillator is shown in as shown in FIG. 28 with X1 representing the QCM, U1 being an inverter, C1 and C2 being capacitors, Rf being a feedback resistor, and Rs being a resistor that isolates the output driver of the inverter U1 from the complex impedance formed by C2, C1, and X1. For the QCM to oscillate properly, the phase shift around the loop must be zero or an integer multiple of 2π (360°).

[0192] A test board for the Pierce-Gate oscillator circuit was created and is shown in an experimental set up in FIG. 29 with a left most image providing a wide shot view that shows the oscillator circuit including the QCM Sensor along with a frequency counter and an MPS sensor (i.e., a molecular property spectrometer) with a gas inlet and outlet which can provide an independent determination of gas concentration (e.g., R32 concentration) that flowed through the QCM / MOF sensor on the circuit board. The upper right image shows a close up view of the MPS, the circuit board including a QCM or QCM / MOF sensor in a holder that provides a gas flow passage with an inlet that receives a controlled mix of air, humidified air, and refrigerant to be used in validating and / or calibrating the QCM or QQCM / MOF sensor, an outlet that directs the mixture to the MPS sensor for independent characterization, and electrical connectors for supplying power to the circuit board and frequency information from the QCM or QCM / MOF sensor to the frequency counter. Data logging software was also provided with a screenshot shown in the lower right image.

[0193] In different experiments an empty QCM (i.e. without a coating) or a coated QCM were placed in the holder and tested using the experimental set up. FIG. 30 shows an example of the frequency changes at different R32 concentrations for an empty or blank sensor and for a HKUST coated sensor. As can be seen in the plots, the frequency of the empty QCM changes by less than 10 Hz from pure air to 25% LFL of R32, while the coated QCM exhibits more than a 300 Hz frequency change. To further enhance the oscillator circuitry and reduce costs, a quartz crystal oscillator chip was chosen for the benchtop sensor design. Unlike the Piece-Gate oscillator, which requires five components (1 inverter, 2 resistors, and 2 capacitors), the quartz crystal oscillator chip integrates resistors and capacitors into a single component. This chip supports a frequency range of up to 50 MHz, whereas the maximum frequency of the Piece-Gate oscillator is typically 20 MHz. In addition, in some embodiments, the circuit board may include one or more sensors for pressure, temperature, and / or humidity for environmental compensation or even a single integrated sensor that provides any two of these three or even all three of these. In other variations other sensors may be provided. In some embodiments, a reference channel (i.e. a QCM without a gas selective coating) may be included along with a gas detection channel (i.e. a QCM with a gas selective coating). The inclusion of the reference channel may be useful in providing compensation for frequency shifts under varying environmental conditions. In some embodiments, a full sensor assembly will also include a microcontroller or microprocessor is also included to directly read the frequency (eliminating the need for an external frequency counter} and possibly to provide other functions such as, for example, environmental compensation via data received from additional sensors, other calibration or testing functions, frequency shift calculation, gas concentration determination, and corrective action implementation. In some variations, a reference channel may be replaced by a second gas detection channel configured for detecting a different gas, while in still other variations, more than one gas detection sensor may be included along with a reference sensor.

[0194] An image of example sensor assembly is shown in FIG. 31 along with an exploded view of a QCM sensor (which may or may not include a gas selective coating such as an HT-MOF) and its housing. The example circuit includes two channels (e.g. a gas detection channel and a reference channel) along with a microprocessor. In the example shown, the two channels are provided with a flow inlet and a flow outlet for convenience in efficient experimentation, testing and validation. In other embodiments, where the assemblies may be included in an operational environment that is to be monitored the relevance of such housings or holders will dictate their inclusion or absenceSpecific Detailed Example Sensor Assembly Implementations

[0195] FIGS. 32A-32G provide various views of an implementation of a single QCM sensor assembly according one embodiment of the invention, including an annotated isometric view of the circuit and sensor (FIG. 32A), a top view of a populated circuit board excluding mounted sensors (FIG. 32B), a top view of an unpopulated PCB showing traces and part identification markings (FIG. 32D), a bottom view of the unpopulated PCB showing traces and other features (FIG. 32E). a table correlating descriptions, values or part numbers, with PCB feature identifiers (FIG. 32F), and a schematic drawing of a plurality of individual circuit elements showing connections to other circuit elements (FIG. 32G). In the example of FIG. 32A-32G the PCB has a width of 1 inch and a length of 2.5 inches with a component height of less than 1 inch and component width beyond that of the board being around 0.20 inches or less yielding a rectangular parallelepiped packaging size of around 3 to 5 cubic inches.

[0196] FIGS. 33A-33G provide various views of an implementation of a dual QCM sensor assembly according one embodiment of the invention that includes a reference sensor and a selective gas sensor, with the FIGS including an annotated isometric view of the circuit and sensors (FIG. 33A), a top view of a populated circuit board excluding mounted sensors (FIG. 33B), a top view of an unpopulated PCB showing traces and part identification markings (FIG. 33D), a bottom view of the unpopulated PCB showing traces and other features (FIG. 33E). a table correlating descriptions, values or part numbers, with PCB feature identifiers (FIG. 33F), and a schematic drawing of a plurality of individual circuit elements showing connections to other circuit elements (FIG. 33G). FIG. 33E has added yellow ovals that identify though holes, red circles that show a portion of the vias that connect the top and the bottom of the PCB and wherein the four corner holes surround by black mounting pads in the corners are for receiving mounting screws or pins. In the example of FIG. 33A-33G the PCB has a width of 1 inch and a length of 3.5 inches with a component height of less than 1 inch and component width beyond that of the board being around 0.20 inches or less yielding a rectangular parallelepiped packaging size of around 4 to 6 cubic inches.

[0197] The sensor and sensor assemblies of various embodiment of the invention may be used in any application where refrigerant concentration detection is needed. For example, the sensors may be located inside equipment that uses / contains refrigerant or in proximity to or within an environment that contains equipment such equipment or piping that. Examples: HVAC equipment, heat pump water heating equipment, ground source heat pump equipment, grocery refrigeration equipment, automotive HVAC equipment, machinery, and the like.ADDITIONAL REMARKS

[0198] Any materials referenced herein or in any appendix attached hereto are incorporated herein by reference as if set forth in full. To the extent that any definitions or other teachings set forth directly herein (i.e., not incorporated by reference) contract any teachings set forth in an appendix or in other material incorporated herein by reference the precedence given to the teachings or definitions are: (1) teachings set forth directly in the body of the application, then (2) teachings set forth in any appendix, and finally (3) teachings set forth in any incorporated material with material having more recent publication dates taking precedence over material having older publication dates.

[0199] It is intended that the aspects of the invention set forth specifically herein or otherwise ascertained from the present teachings represent independent invention descriptions which Applicant contemplates as full and complete, and that Applicant believes may be set forth as independent claims without need of importing additional limitations or elements from other embodiments or aspects set forth herein for interpretation or clarification. It is also understood that any variations of the aspects (as well as variations in any embodiments) set forth herein represent individual and separate features that may form separate claims or be individually or in any workable combination added to existing claims.

[0200] While various preferred embodiments of the invention are shown and described, it is to be distinctly understood that this invention is not limited thereto but may be variously embodied. From the foregoing description, it will be apparent that various changes may be made without departing from the spirit and scope of the invention.REFERENCE TO ATTACHED APPENDICES

[0201] Additional information about implementations and embodiments of the invention are set forth in the appendices filed herewith. These appendices are incorporated herein by reference:

[0202] Appendix A: Code for Processing Frequency Shift Time Series Data

[0203] Appendix B: Code for Processing Commercial R32 Gas Sensor Data

[0204] Appendix C: Unpublished Manuscription Setting Forth Additional Embodiment Information

[0205] Appendix D: Supplement Information for Appendix C

Examples

Embodiment Construction

[0086]Various advantages and novel features of the present invention are described herein and will become more readily apparent to those skilled in this art from the following detailed description. In the preceding and following descriptions the preferred embodiment of the invention is shown by way of illustration of the best mode contemplated for carrying out the invention perhaps along with one or more variations. As will be realized, the invention is capable of modification in various respects without departing from the spirit of the invention as will be understood by those of skill in the art.

Definitions and Abbreviations

QCM: Quartz Crystal Microbalance[0088]SAW: Surface Acoustic Wave[0089]ZIF-67: Zeolitic Imidazolate Framework-67[0090]HT-MOF: Humidity tolerant metal organic framework crystalline porous material.[0091]Targeted HT-MOF: A humidity tolerant MOF that is selected for its ability to target a particular refrigerant gas for adsorption and thus measurement.[0092]Zero-air...

Claims

1. A targeted sensor for detecting A2L or A3 refrigerants, comprising:a) a QCM sensor or a SAW sensor, andb) a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor.

2. The targeted sensor of claim 1 comprising a QCM sensor.

3. The targeted sensor of claim 1 comprising a SAW sensor.

4. The targeted sensor of claim 1 wherein the MOF coating comprises a ZIF-67 coating.

5. The targeted sensor of claim 4 wherein the refrigerant comprises R32.

6. The targeted sensor of claim 1 wherein the MOF coating comprises a ZIF-8 coating.

7. The targeted sensor of claim 6 wherein the refrigerant comprises R290.

8. A targeted sensor assembly for detecting A2L or A3 refrigerants, comprising:a) a QCM sensor or a SAW sensor;b) a humidity tolerant MOF coating on the QCM sensor or the SAW sensor;c) an oscillator triggering QCM sensor resonance or SAW sensor resonance;d) a microprocessor or microcontroller for receiving information from the QCM sensor or from the SAW sensor and using the information and / or outputting one or more of frequency, frequency shift, phase shift. amplitude shift, calibrated concentration of a detected targeted refrigerant, and / or other actionable or control information for enabling a safe operating environment around A2L or A3 refrigerants.

9. The targeted sensor assembly of claim 8 comprising at least one QCM sensor.

10. The targeted sensor assembly of claim 9 wherein the at least one QCM comprises at least two QCM sensors with at least one QCM not including a humidity tolerant coating.

11. The targeted sensor assembly of claim 8 comprising at least one SAW sensor.

12. The targeted sensor assembly of claim 11 wherein the at least one SAW sensor comprises at least two SAW sensors with at least one SAW not including a humidity tolerant coating.

13. The targeted sensor assembly of claim 8 wherein the MOF coating comprises a ZIF-67 coating.

14. The targeted sensor assembly of claim 8 wherein the MOF coating comprises a ZIF-8 coating.

15. The targeted sensor assembly of claim 14 wherein the A3 refrigerant comprises R290.

16. The targeted sensor assembly of claim 8 wherein the assembly further comprises at least one environmental sensor selected from the group consisting of: (1) one more temperatures sensors, (2) one or more pressure sensors, and (3) one or more humidity sensors, wherein the at least one environmental sensor provides an input to the microprocessor or microcontroller to allow inclusion or use of such information when providing output.

17. The targeted sensor assembly of claim 8 wherein a t90 sensing time is less than 30 seconds.

18. The targeted sensor assembly of claim 17 wherein the t90 sensing time is less than 10 seconds19. The targeted sensor assembly of claim 8 wherein a t10 recovery time is less than 30 seconds20. A method of fabricating a targeted sensor for detecting A2L or A3 refrigerants, comprising:a) supplying a QCM sensor or a SAW sensor, andb) applying a targeted humidity tolerant MOF coating on a surface of the QCM Sensor or the SAW sensor via spin coating.

Images