Sensor fabrication and testing system

Abstract

A sensor system for fabricating optical fiber sensors comprises a positioning assembly, a gas microenvironment module for generating controlled gas environments, an optical interrogator for measuring optical path length of a sensing element, and a data acquisition and control unit configured to perform fabricating, calibration, and testing operations according to corresponding methods.

Description

Docket No. 3510410.019901; 25MST024SENSOR FABRICATION AND TESTING SYSTEMSTATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

[0001] This disclosure was made with Government support under grant W911NF2120274 awarded by the Defense Finance and Accounting Service. The Government has certain rights.FIELD

[0002] The present disclosure relates to manufacturing and testing systems, and more particularly to an integrated system for fabricating, calibrating, and testing optical fiber sensors.BACKGROUND

[0003] Gas sensing technologies have broad applications. Optical fiber-based sensors offer advantages including remote sensing capability, immunity to electromagnetic interference, and suitability for deployment in harsh environments. Metal-organic frameworks (MOFs) represent a class of porous materials characterized by high surface area and tunable chemical functionality, making them useful for gas adsorption and detection applications.

[0004] Fabrication of optical fiber sensors incorporating MOF materials has traditionally involved multiple discrete steps, including separate MOF synthesis, purification, and attachment of the MOF material to an optical fiber. Calibration and testing of gas sensors similarly involve separate procedures that are typically performed using external gas chambers or testing apparatuses distinct from the fabrication equipment.SUMMARY

[0005] According to an aspect of the present disclosure, a sensor system for fabricating an optical fiber sensor from an optical fiber is provided. The sensor system comprises a positioning assembly configured to receive and hold the optical fiber and to move the optical fiber along at least two dimensions. The sensor system further comprises a gas microenvironment module including a plurality of gas ports configured to fluidly couple to one or more gas reservoirs, the gas microenvironment module being controllable to dispense one or more gases from the plurality of gas ports to generate a microenvironment in proximity to the gas microenvironment module. The sensor system further comprises an optical interrogator configured to optically couple to the optical fiber at a proximal end thereof and to measure an optical path length of a sensing element formed at an end face of the optical fiber at a distal1CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024end thereof. The sensor system further comprises a data acquisition and control unit operably coupled to the positioning assembly, the gas microenvironment module, and the optical interrogator. The data acquisition and control unit comprises a processor and memory, the processor being communicatively coupled to the memory. The memory includes instructions which, when executed by the processor, cause the processor to control the sensor system to perform one or more fabricating operations on the optical fiber to form the optical fiber sensor, control the sensor system to perform one or more calibration operations on the optical fiber sensor, and control the sensor system to perform one or more testing operations on the optical fiber sensor.

[0006] According to another aspect of the present disclosure, a method for identifying gas analytes is provided. The method comprises fabricating an optical fiber sensor having a sensing element at an end face of an optical fiber. The method further comprises generating, using a gas microenvironment module, a plurality of gas microenvironments, each gas microenvironment including at least one target gas analyte at a defined concentration. The method further comprises sequentially exposing the optical fiber sensor to the plurality of gas microenvironments by positioning the optical fiber sensor within each gas microenvironment. The method further comprises recording, using an optical interrogator, a sensor response curve for each exposure, each sensor response curve capturing a time-dependent response of the optical fiber sensor during adsorption of the at least one target gas analyte. The method further comprises training a machine learning model using the recorded sensor response curves to recognize patterns in the sensor response curves corresponding to the at least one target gas analyte for enabling the machine learning model to identify a concentration of the at least one target gas analyte present at the sensing element based on the time-dependent response of the sensing element thereto. The method further comprises exposing the optical fiber sensor to an unknown gas environment. The method further comprises recording, using the optical interrogator, a sensor response curve from the unknown gas environment. The method further comprises applying the trained machine learning model to the sensor response curve from the unknown gas environment to identify a concentration of the at least one gas analyte in the unknown gas environment.

[0007] According to another aspect of the present disclosure, a method for identifying gas analytes using a sensor array is provided. The method comprises fabricating a plurality of optical fiber sensors, each optical fiber sensor having a sensing element at an end face of an optical fiber. The method further comprises generating, using a gas microenvironment module, a gas microenvironment having a defined gas composition including at least one target gas analyte at a defined concentration. The method further comprises exposing the plurality of optical fiber sensors to the gas microenvironment such that each of the optical fiber sensing elements adsorb the at least one target gas analyte. The method further comprises recording, using an optical interrogator, a sensor response curve from each of the plurality of optical fiber 2CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024sensors, wherein each sensor response curve captures a time-dependent response of the respective sensing element during adsorption of the at least one target gas analyte. The method further comprises repeating said generating, exposing, and recording for a plurality of gas compositions to generate a training dataset comprising a plurality of collective response patterns from the plurality of optical fiber sensors to the plurality of gas compositions. The method further comprises training a machine learning model using the collective response patterns to recognize patterns in the collective response patterns corresponding to concentrations of the at least one target gas analyte. The method further comprises exposing the plurality of optical fiber sensors to an unknown gas environment. The method further comprises recording, using the optical interrogator, a sensor response curve from each of the plurality of optical fiber sensors in the unknown gas environment. The method further comprises applying the trained machine learning model to the sensor response curves from the unknown gas environment to identify a concentration of the at least one target gas analyte in the unknown gas environment based on the collective response patterns from the plurality of optical fiber sensors.

[0008] According to yet another aspect of the present disclosure, a method for manufacturing optical fiber sensors is provided. The method comprises fabricating a plurality of optical fiber sensors, each optical fiber sensor being fabricated by forming a sensing element at an end face thereof by dipping the optical fiber into a precursor solution to form a microdroplet at the end face and evaporating the microdroplet to form a composite layer at the end face. The method further comprises recording, using an optical interrogator, optical path length data of each sensing element of the plurality of optical fiber sensors during fabrication. The method further comprises recording performance data of each of the plurality of optical fiber sensors following fabrication thereof. The method further comprises training a machine learning model, using the optical path length data and the performance data, to identify patterns in the optical path length data that correspond to the performance data. The method further comprises fabricating a subsequent optical fiber sensor while recording, using the optical interrogator, subsequent optical path length data therefrom. The method further comprises applying the machine learning model to the subsequent optical path length data to predict performance characteristics of the subsequent optical fiber sensor.

[0009] Other objects and features of the present disclosure will be in part apparent and in part pointed out hereinafter.BRIEF DESCRIPTION OF THE DRAWINGS

[0010] FIG. 1 is an elevation of a sensor system according to an embodiment of the present disclosure;3CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024

[0011] FIG. 2 is a front perspective of a gas microenvironment module according to an embodiment of the present disclosure;

[0012] FIG. 3 is a front perspective of another gas microenvironment module according to another embodiment of the present disclosure;

[0013] FIG. 4 shows a bare optical fiber before being dipped into a precursor solution as part of an operation of an optical fiber sensor fabrication method according to an embodiment;

[0014] FIG. 5 is similar to FIG. 4, but shows a microdroplet at an end face of the optical fiber after being dipped into and withdrawn from the precursor solution;

[0015] FIG. 6 is similar to FIG. 5, but shows the distal end of the optical fiber and the microdroplet positioned in a controlled microenvironment;

[0016] FIG. 7 is similar to FIG. 6 but shows the microdroplet partially evaporated;

[0017] FIG. 8 is similar to FIG. 7, but shows the microdroplet fully evaporated;

[0018] FIG. 9 is a block diagram of the sensor system;

[0019] FIG. 10 is a flow chart generally corresponding to a method of using the sensor system according to an embodiment;

[0020] FIG. 11 is a flow chart generally corresponding to an optical fiber fabrication method according to an embodiment;

[0021] FIG. 12 is a flow chart generally corresponding to a method of identifying one or more gas analytes according to an embodiment;

[0022] FIG. 13 is a flow chart generally corresponding to a method of selecting optical fiber sensors according to an embodiment;

[0023] FIG. 14 is an optical path length vs time graph of an example;

[0024] FIG. 15 is a sensor response vs ethanol concentration graph of an example; and

[0025] FIG. 16 is a wavelength vs time graph and an optical path length vs time graph for an example.4CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024DETAILED DESCRIPTION

[0026] Throughout the present disclosure, the term "optical fiber sensor" is used to refer to an optical fiber with a sensing element (e.g., a sensor head). Thus, when an optical fiber has a sensing element disposed thereon, it will be referred to as an optical fiber sensor.

[0027] Traditional fabrication of optical fiber sensors incorporating metal-organic framework materials involves multiple discrete steps requiring manual intervention and separate equipment. Calibration and testing of such sensors are typically performed using external gas chambers distinct from the fabrication equipment, and the transfer of sensors between systems introduces handling steps and potential sources of variability. These manual processes are labor-intensive, time-consuming, and impractical for high-throughput screening of sensor candidates. Furthermore, training artificial intelligence models for gas analysis requires extensive datasets that would be prohibitively slow and expensive to generate through manual methods.

[0028] The present disclosure addresses these limitations via a sensor system that integrates fabrication, calibration, and testing of optical fiber sensors into a continuous sequence without requiring separate systems or manual intervention. The disclosure describes novel methods of fabricating optical fiber sensors using a pendant micro-droplet evaporation technique within controlled gas microenvironments. The disclosure further describes the use of artificial intelligence (e.g., machine learning) in optimizing fabrication processes by correlating real-time fabrication data with sensor performance characteristics, as well as the use of machine learning in calibration and testing operations for identifying gas analytes within complex gas mixtures.

[0029] Referring to FIGS. 1 and 9, an exemplary embodiment of the sensor system in accordance with the present disclosure is generally indicated at reference number 100. The sensor system 100 broadly comprises a positioning assembly 114, a base 104, a gas microenvironment module 156, an optical interrogator 176, and a data acquisition and control unit (DACU) 178. The positioning assembly 114 is configured to receive and hold an optical fiber 132 and to move the optical fiber 132 along multiple dimensions relative to the base 104 and the gas microenvironment module 156. A vial 108 is configured to hold a precursor solution 110 from which a sensing element 144 is formed at an end face of the optical fiber 132. The gas microenvironment module 156 includes a plurality of gas ports 160 configured to dispense one or more gases to generate controlled microenvironments 168 in proximity thereto.5CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024The optical interrogator 176 is configured to optically couple to the optical fiber 132 and to measure an optical path length at the sensing element 144. The data acquisition and control unit 178 is operably coupled to the positioning assembly 114, the gas microenvironment module 156, and the optical interrogator 176 to coordinate operations of the sensor system 100. As will be explained in further detail below, the sensor system 100 enables autonomous fabrication of optical fiber sensors through a pendant micro-droplet evaporation technique, followed by calibration and testing of the fabricated optical fiber sensors within controlled gas microenvironments, thereby integrating sensor fabrication, calibration, and testing into a continuous sequence without requiring separate systems or manual intervention.

[0030] Referring to FIG. 1, the base 104 is disposed within an enclosure 102 and supports a vial heater 106. The vial heater 106 is configured to receive and heat a vial 108 containing the precursor solution 110. The vial heater 106 is configured to heat the vial 108 to a temperature in an inclusive range of from about ambient temperature to about 200°C. In one embodiment, the base 104 itself is temperature controlled and is configured to heat the vial 108 directly. Coolers (e.g., Peltier coolers) may also be placed on the base 104 for cooling the vial 108 containing the precursor solution 110.

[0031] A stirrer 112 is positioned in proximity to the vial 108 and is configured to mix the contents of the vial to maintain uniform composition of the precursor solution 110. In the illustrated embodiment, the stirrer 112 is a propeller-type stirrer. In another embodiment, the stirrer is a magnetic stir bar-type stirrer. In a further embodiment, the stirrer is a sonicator-type stirrer.

[0032] The positioning assembly 114 is configured to receive and hold the optical fiber 132 to move the optical fiber along at least two dimensions. In the illustrated embodiment, the positioning assembly 114 is configured to move the optical fiber 132 along three dimensions comprising an X-dimension for side-to-side positioning (broadly, an X-Axis 126), a Y-dimension for vertical positioning (broadly, a Y-Axis 128), and a Z-dimension for back-and-forth positioning (broadly, a Z-Axis 130). According to an embodiment, positioning assembly 114 comprises a gantry 116 having a gantry beam 118 supported by a gantry rail 119. The gantry beam 118 provides horizontal movement along the X-Axis 126, while the gantry rail 119 provides vertical movement along the Y-Axis 128. Another gantry rail oriented substantially parallel to the Z-Axis 130 provides movement therealong. An arm 120 extends from the gantry beam 118 and terminates at a fiber holder 122 configured to receive and hold the optical fiber 132.6CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024

[0033] Precise positioning and rapid movement capabilities of the positioning assembly 114 enable accurate placement of the optical fiber 132 within the precursor solution 110 and the gas microenvironments 168, as well as efficient transitions between fabrication, calibration, and testing operations as will be described in greater detail below. In the illustrated embodiment, the positioning assembly 114 has a positioning resolution of 10 micrometers or finer in each of the three dimensions (e.g., 5 micrometers or finer in each of the three dimensions). The positioning assembly 114 allows movement speeds in both the X and Y dimensions in an inclusive range of from about 1 mm / s to about 500 mm / s. These parameters may be varied in alternative embodiments depending on the application requirements.

[0034] The optical fiber 132 has a proximal end 138 and a distal end 140 with an optical fiber end face 142. According to aspects of the present disclosure, a sensing element 144 is formed at the optical fiber end face 142 at the distal end 140 of the optical fiber 132 from a microdroplet 146 at the optical fiber end face 142. When acquired by the optical fiber holder 122, the distal end 140 of optical fiber 132 is in proximity to the holder, such that the position of the distal end of the optical fiber may be finely controlled by the positioning assembly 114.

[0035] The gas microenvironment module 156 includes a plurality of gas ports 160 configured to fluidly couple to one or more gas reservoirs 170. The gas microenvironment module 156 is controllable to dispense one or more gases from the plurality of gas ports 160 to generate a microenvironment 168 in proximity to the gas microenvironment module. As used herein, a "microenvironment" refers to a controlled gas environment (e.g., having controlled temperature, composition, flow, etc.). The microenvironment is generally defined by a gas 164 (e.g., a gas cone) or multiple gases emanating from the gas ports 160. As further shown at FIG.1, the gas microenvironment module 156 includes a gas port plate 158 with the plurality of gas ports 160. A plurality of gas tubes 162 connect the gas ports 160 to the gas reservoirs 170 through gas valves 172; each gas port of the plurality of gas ports is connected to the one or more gas reservoirs 170 through a gas tube 162. In an alternative embodiment, multiple ports are connected to a shared gas reservoir to create a larger microenvironment of a single gas.

[0036] In one embodiment, the gas tubes 162 are constructed from stainless steel material, which provides durability and chemical resistance. In another embodiment, the gas tubes 162 are constructed from copper material. In a further embodiment, the gas tubes 162 are constructed from PEEK plastic material, which offers chemical inertness and flexibility for routing within the sensor system 100. In yet another embodiment, the gas tubes 162 are polyimide capillary tubes with small diameters for more precise delivery of gas to smaller7CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024microenvironment regions. The material or materials forming the gas tubes 162 may be chosen for the specific application requirements and may differ from one tube to another.

[0037] The gas ports 160 are configured to dispense gases 164 in combination to form mixed gas microenvironments 166 or individually to form individual (e.g., homogeneous) microenvironments 168 in proximity to the gas microenvironment module 156. The gas port plate 158 (broadly, a gas port structure) may be fabricated from plastic, metal, or other suitable materials. The gas ports 160 are configured to generate specific geometric flow patterns such as conical or cylindrical gas jets of adjustable size. The gas ports 160 have uniform or varied diameters depending on the desired gas flow characteristics. The direction of each gas port's 160 axis is precisely controlled to influence gas mixing dynamics. In an exemplary embodiment, internal fins are included within the gas ports 160 to introduce additional turbulence or controlled mixing for improving gas homogeneity. In a further embodiment, high-voltage electrical means are used to create micro-plasmas at the gas outputs for in-situ gas modifications including cleaning, etching, melting, or annealing of the optical fiber 132 surface.

[0038] The sensor system 100 further comprises a gas control system for regulating gas flow from the gas reservoirs 170 to the gas ports 160. The gas control system includes one or more electronically actuated valves (broadly, the gas valves 172) operably coupled to the DACU 178 (e.g., a processor 180 of the DACU). In the illustrated embodiment, the valves 172 are flow rate valves controllable by the DACU 178 to regulate gas flow rates. For example, the valves are controllable by the DACU to output a constant flow of target analyte gas (e.g., 100 ml per minute) to the gas ports 160.

[0039] Gas tube temperature controllers 174 are provided to regulate the temperature of gas flowing through the gas tubes 162. The gas tube temperature controllers 174 comprise heaters (e.g., resistive heaters) configured to increase a temperature of gas flowing through the tubes 162 and / or coolers (e.g., Peltier coolers) configured to decrease a temperature of gas flowing through the tubes. The gas microenvironment module 156 may include further heating and / or cooling means to provide additional control over the temperature of each gas 164 dispensed therefrom.

[0040] The optical interrogator 176 is optically coupled to the proximal end 138 of the optical fiber 132 and is configured to measure an optical path length of the sensing element 144 as it is formed from a microdroplet 146 at the optical fiber end face 142 at the distal end8CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024140 of the optical fiber 132. In the illustrated embodiment, the optical interrogator 176 and sensing element 144 form an extrinsic Fabry-Perot interferometer system. The function and utility of this system will be described in greater detail below.

[0041] The DACU 178 is operably coupled to the positioning assembly 114, the gas microenvironment module 156, the stirrer 112, the vial heater 106, the gas valves 172, and the optical interrogator 176. The DACU 178 comprises the processor 180 and a memory 182. The processor 180 is communicatively coupled to the memory 182. The memory 182 includes instructions which, when executed by the processor 180, cause the processor 180 to control the sensor system 100 to perform fabricating, calibration, and testing operations on the optical fiber 132. A user interface 184 is provided for user interaction with the sensor system 100.

[0042] Referring to FIG. 2, the gas microenvironment module 156 is shown in greater detail. The gas microenvironment module 156 is configured to hold the gas tubes 162 in a structured arrangement for formation of gas flow fields. The gas port plate 158 securely holds the gas tubes 162 at the ports 160, allowing for the controlled release of selected gas streams. In the illustrated embodiment, the gas port plate 158 is a grid-patterned gas port plate (e.g., wherein the plurality of gas ports 160 are arranged in a grid). In alternative embodiments, the plurality of gas ports 160 are arranged in a linear arrangement, a honeycomb arrangement, a triangular arrangement, or other arrangements. When gas flows through the gas tubes 162 and exits through the gas ports 160, cones of gas 164 are formed in proximity to their respective gas ports. As depicted in FIG. 2, two cones of gas 164 are shown intersecting (e.g., populating a shared space). At their intersection, the mixed gas microenvironment 166 is formed. At a mixed gas microenvironment 166, two or more gases from adjacent gas ports 160 combine to create a binary or multi-component gas mixture.

[0043] Referring now to FIG. 3, an alternative embodiment of a gas microenvironment module in accordance with the present disclosure is generally indicated at reference number 156'. The gas microenvironment module 156' broadly comprises a gas port post 159, a plurality of gas ports 160', and a plurality of gas tubes 162. The gas microenvironment module 156' serves the same function as the gas microenvironment module 156 of FIG. 2 but with a more compact linear configuration having fewer ports.

[0044] Referring now to FIGS. 4-8, an optical fiber sensor fabrication method in accordance with the present disclosure is illustrated. Broadly, the fabrication method comprises9CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024a dip operation and a gas exposure operation performed in sequence to form the optical fiber sensor from the optical fiber 132.

[0045] Referring to FIGS. 4 and 5, the optical fiber 132 is shown being dipped into the precursor solution 110 (as held by the vial 108). The optical fiber 132 includes an optical fiber core 134 and an optical fiber cladding 136 surrounding the optical fiber core. The distal end of optical fiber 140 terminates at an optical fiber end face 142.

[0046] In the illustrated embodiment, the precursor solution 110 is a metal-organic-compound precursor solution comprising HKUST-1 precursor. The HKUST-1 precursor solution uses dimethyl sulfoxide (DMSO) as a solvent and contains copper(II) nitrate trihydrate as an inorganic metal cluster and 1,3,5-benzenetricarboxylic acid (BTC), also referred to as trimesic acid, as an organic linker. In alternative embodiments, other MOF precursors are employed to form alternative MOF sensors.

[0047] The dip operation includes controlling the positioning assembly 114 to move the optical fiber 132 along the vertical dimension (broadly, the Y-Axis 128) to dip the optical fiber into the precursor solution 110. In the illustrated embodiment, the data acquisition and control unit 178 controls the gantry 116 to lower the optical fiber 132 along the Y-Axis 128 until the optical fiber end face 142 contacts and enters the precursor solution 110 within the vial 108. A velocity of the dip process is optimized to prevent detachment of a droplet from the optical fiber end face 142 and to prevent precursor residues on the sides of the optical fiber 132. In one embodiment, the velocity is 5 mm / s. A height parameter of the dip process from a meniscus of the precursor solution 110 is optimized to avoid a vapor-liquid equilibrium region such that subsequent evaporation of the droplet is not affected. In one embodiment, the height is 25 mm.

[0048] At FIG. 5, the optical fiber 132 is shown after being dipped into and withdrawn from the precursor solution 110. Upon withdrawal of the optical fiber 132 from the precursor solution 110, a microdroplet 146 (pendant droplet) forms at the optical fiber end face 142. The microdroplet 146 has a microdroplet surface 149 at an outer boundary thereof. The microdroplet 146 has a microdroplet depth 148 corresponding to a vertical dimension of the microdroplet. The size of the microdroplet 146 is influenced by factors including the withdrawal velocity and properties of the precursor solution 110 (e.g., viscosity). The positioning assembly 114 controls the withdrawal speed along the Y-Axis 128 to regulate the microdroplet 146 size and ensure uniform deposition across multiple sensor fabrication cycles.10CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024

[0049] During this and other operations, the pendant microdroplet 146 and / or resulting sensing element 144 function as an extrinsic Fabry-Perot interferometer (EFPI) cavity. When the optical fiber 132 is dipped into and withdrawn from the precursor solution 110, the microdroplet 146 establishes two reflective interfaces: a first interface at the boundary between the optical fiber end face 142 and the liquid microdroplet 146, and a second interface at the boundary between the liquid microdroplet and the surrounding air (i.e, at a microdroplet surface 149). These two interfaces function as partially reflective surfaces that form a low-finesse EFPI cavity, wherein the microdroplet 146 acts as a thin film cavity between the reflectors. Light from the optical interrogator 176 travels through the optical fiber core 134 and partially reflects at the fiber-liquid interface (first reflector), while the remaining light continues through the microdroplet 146 and reflects at the liquid-air interface (second reflector). The reflected light from both interfaces travels back through the optical fiber 132 and interferes, producing an interference pattern (interferogram) that is detected by the optical interrogator 176.

[0050] The optical path length of this EFPI cavity is determined by both the physical thickness of the microdroplet 146 and the refractive index of the liquid precursor solution 110. As the microdroplet 146 evaporates and a composite layer 150 forms, changes in both the physical cavity length and the refractive index of the material cause shifts in the interferogram. During nucleation, an increase in solute concentration may cause an increase in the refractive index, resulting in a red shift of the interferogram. As the solvent continues to evaporate and crystallization progresses, the physical cavity length decreases, resulting in a blue shift of the interferogram. These shifts are monitored in real time by the optical interrogator 176 to track the fabrication process. After complete evaporation of the solvent, the composite layer 150 remains at the optical fiber end face 142, and the EFPI cavity transitions from a liquid-based cavity to a solid-based cavity. In this configuration, the sensing element 144 interfaces with the surrounding gas environment, and subsequent adsorption of gas analytes into the composite layer 150 causes changes in the refractive index and / or physical thickness of the composite layer, which are detected as shifts in the interferogram by the optical interrogator 176.

[0051] Referring now to FIG. 6, a gas exposure operation is shown wherein the distal end of optical fiber 140 (e.g., the end face 142) and the microdroplet 146 are positioned in a controlled microenvironment 168. The gas exposure operation includes generating, using the gas microenvironment module 156, a gaseous microenvironment 168 in proximity to the gas microenvironment module and positioning, using the positioning assembly 114, the distal end region 140 of the optical fiber 132 in the gaseous microenvironment. The gas 164 generally11CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024defines the microenvironment 168 that surrounds the distal end 140 of the optical fiber 132 and the microdroplet 146. The gas valves 172 control dispensing of gas from the gas reservoirs 170 through the gas tubes 162 to the gas ports 160. In one embodiment, the gaseous microenvironment 168 comprises dry nitrogen gas to minimize water incorporation into a composite layer during evaporation of the solvent. In another embodiment, the gaseous microenvironment 168 comprises a controlled humidity environment containing a known and / or measured quantity of water vapor to influence a crystal hydration level in the composite layer. In yet another embodiment, the gaseous microenvironment comprises an alternative carrier gas. Suitable alternative carrier gases include argon or carbon dioxide.

[0052] During fabrication and following creation of the microdroplet 146 at the optical fiber end face 142, the optical fiber 132 is generally held at rest. This comprises holding the positioning assembly 114 steady for a rest period of a duration sufficient to permit the precursor solution 110 to form a composite layer 150 of the sensing element 144. The composite layer 150 is formed via evaporation of the solvent portion of the precursor solution 110 and crystallization of the precursor portion of the precursor solution 110. In the illustrated embodiment, the gas exposure operation is performed concurrently with the rest operation such that evaporation of the solvent portion of the precursor solution 110 occurs within the gaseous microenvironment. The duration of the rest period may be in an inclusive range of from about 20 seconds to about 100 seconds. In one embodiment, the duration is about 30 seconds.

[0053] Testing indicates crystal layer formation is approximately 24 seconds under nitrogen assistance (when exposed to a nitrogen microenvironment) and 90 seconds under ambient conditions (standard atmospheric conditions). The precursor solution temperature is maintained in an inclusive range of from about 24°C to about 80°C for crystal layer formation. In one embodiment, HKUST-1 synthesis is performed in an inclusive range from about room temperature (about 20°C) to 57°C for large mass deposition and high crystal quality. Heating the precursor solution 110 to approximately 150°C facilitates synthesis of larger crystals approximately (e.g., about 17.5 micrometers in length).

[0054] Referring to FIG. 7, the microdroplet 146 is shown partially evaporated and beginning to form the composite layer 150. As evaporation progresses, the composite layer 150 begins to form at the optical fiber end face 142. The evaporation of the microdroplet 146 causes a rapid increase in a surface-to-volume ratio thereof. The rapid increase in the surface-to-volume ratio results in supersaturation, which drives the droplet 146 to shift from a disordered state at an air-liquid boundary to a partially ordered state. Evaporation triggers nucleation of12CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024the microdroplet 146. As the solvent (e.g., DMSO) in the microdroplet 146 continues to evaporate, a parallel process of crystallization and polymerization is triggered. In one embodiment, a significant portion (e.g., around 8-21%) of a total mass of the composite layer 150 comprises polymeric BTC species. The polymeric BTC species act as an adhesive that secures MOF crystallites to the optical fiber 132.

[0055] Referring now to FIG. 8, the microdroplet 146 is fully evaporated, leaving a fully formed composite layer 150 defining the sensing element 144. The thickness of the composite layer 150 is largely controlled via parameters of fabrication. In general, the thickness of the composite layer 150 is in an inclusive range of about 3 micrometers to 20 micrometers. The composite layer 150 shown at FIG. 8 is an active sensing element ready for subsequent calibration and testing operations (e.g., within the controlled gas microenvironment).

[0056] The composite layer 150 comprises a crystal-polymer composite layer containing HKUST-1 crystallites as a primary sensing component. In an alternative embodiment, ZIF-8 MOF is used as a super hydrophobic material for sensor fabrication to resist water adsorption. In other embodiments, other MOF materials may be used to similar effect.

[0057] Referring to FIG. 9, a block diagram of the sensor system 100 illustrates the command and network architecture thereof. The data acquisition and control unit (DACU) 178 is the control hub for the sensor system 100. The DACU 178 includes the processor 180 communicatively coupled to a memory 182. The memory 182 stores instructions which, when executed by the processor 180, cause the processor 180 to control operations of the sensor system 100.

[0058] The DACU 178 is operably coupled to a transmitter 186 for communication with external systems and devices 192. The transmitter 186 enables bidirectional data exchange between the sensor system 100 and remote locations. In one embodiment, the transmitter 186 provides wireless connectivity such as Bluetooth or WiFi, or wired connectivity such as USB, to enable direct connection to external devices.

[0059] The sensor system 100 is configured to communicate with a cloud 188 (e.g., internet) via the transmitter 186. The cloud 188 includes a training module 191 configured for training machine learning models and a cloud storage module 190 configured for storing information, data, and trained machine learning models. The cloud 188 is accessible by remote devices 192 such that the sensor system 100 is controllable remotely. In one embodiment, the cloud storage module 190 stores sensor response data collected during fabrication, calibration,13CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024and testing operations, and the training module 191 processes the stored data to train machine learning models for gas analyte identification and sensor fabrication / selection. These processes will be described in greater detail below with reference to FIGS. 10-13.

[0060] With continued reference to FIG. 9, the DACU 178 controls and monitors various components of the sensor system 100. The DACU 178 is operably coupled to motors 124 that drive the gantry 116 to position the fiber holder 122 along multiple dimensions. The DACU 178 controls the stirrer 112 for mixing contents of the vial 108 and controls the vial heater 106 for heating the precursor solution. The DACU 178 is further operably coupled to the gas ports 160 of the gas microenvironment module 156 (e.g., to control gas mixing / dispensing), the gas valves 172 for regulating gas flow from the gas reservoirs 170, and the gas tube temperature controllers 174 for controlling the temperature of gas flowing through the gas tubes 162.

[0061] The DACU 178 maintains two-way connections with the optical interrogator 176, the user interface 184, and the transmitter 186. The two-way connection with the optical interrogator 176 enables the DACU 178 to control functions of the optical interrogator and to receive optical path length data therefrom during sensor fabrication, calibration, and testing. The two-way connection with the user interface 184 enables the DACU 178 to receive direct commands from a user and to display information to the user. The two-way connection with the transmitter 186 enables the DACU 178 to send and receive data and commands to and from the cloud 188 and connected devices 192.

[0062] In some embodiments, the network architecture of the sensor system 100 is varied. In one embodiment, the training module 191 and the cloud storage module 190 are implemented locally within the DACU 178, such as within the memory 182, such that all operations including machine learning model training, data storage, and sensor control are performed without requiring an external network connection. In this configuration, the sensor system 100 operates as a self-contained unit capable of autonomous fabrication, calibration, testing, and machine learning operations. In an alternative embodiment, the DACU 178 comprises a microchip connected to the transmitter 186, wherein all functions of the sensor system 100 are controlled and recorded via the cloud 188. In this configuration, the cloud 188 performs data processing, machine learning model training, and storage functions, while the DACU 178 serves primarily as an interface between the cloud 188 and the physical components of the sensor system 100. Those of skill in the art will recognize that various intermediate configurations are within the scope of the present disclosure, wherein some functions are14CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024performed locally by the DACU 178 and other functions are performed remotely via the cloud 188.

[0063] The DACU 178 is configured to autonomously perform (e.g., to perform in the absence of user assistance, to perform automatically) the one or more fabricating operations, the one or more calibration operations, and the one or more testing operations in a continuous sequence. The processor 180 executes instructions stored in the memory 182 to coordinate the positioning assembly 114, gas microenvironment module 156, optical interrogator 176, and other components.

[0064] In some embodiments, the sensor system 100 includes additional components for enabling high-throughput autonomous operation. A cache of fibers may be provided to supply sequential or concurrent fabrication operations. The cache may comprise a carousel, magazine, or rack configured to hold a plurality of optical fibers in indexed positions from which the positioning assembly 114 can retrieve individual fibers. A repository for storing formed optical fiber sensors after fabrication is further provided. The repository may comprise indexed storage positions or compartments configured to receive and retain fabricated optical fiber sensors following completion of fabrication, calibration, and testing operations. Means for automatically pairing or coupling the optical fibers to the optical interrogator 176 are also provided, such as an automated coupling mechanism for aligning and connecting the proximal end 138 of each optical fiber 132 to the optical interrogator 176 without manual intervention. These components enable many sensors to be formed concurrently or sequentially without user intervention.

[0065] Referring to FIG. 10, a method 1000 of using the sensor system 100 in accordance with the present disclosure is generally indicated. The method 1000 broadly comprises a fabrication operation 1002 for fabricating a sensor, a calibration operation 1004 for calibrating the sensor, and a testing operation 1006 for testing the sensor. The method 1000 represents a continuous sequence of operations performed by the sensor system 100 for preparing an optical fiber sensor for use in gas detection applications.

[0066] Referring briefly to FIG. 11, the fabrication operation 1002 is shown individually as fabrication method 1100 generally corresponding to the illustrated operations of FIGS. 4-8. The method 1100 broadly comprises operation 1102 for preparing the optical fiber (e.g., cleaving to produce a flat, clean end face), operation 1104 for preparing the precursor solution (e.g., dissolving precursor components in a solvent), operation 1106 for dipping the optical15CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024fiber into the precursor solution, operation 1108 for withdrawing the optical fiber from the precursor solution such that a pendant microdroplet forms at the end face, and operation 1110 for positioning the optical fiber in a gas microenvironment 168 where the microdroplet undergoes evaporation to form the composite layer 150 at the optical fiber end face 142, thereby forming the sensing element 144. Throughout the method 1100, the optical interrogator 176 monitors the optical path length of the forming sensing element in real time, providing data indicative of the fabrication progress and the final thickness of the sensing element.

[0067] Referring again to FIG. 10, following the fabrication operation 1002, the method 1000 proceeds to the calibration operation 1004, which establishes response characteristics of the optical fiber sensor for one or more target gas analytes. The calibration operation includes an activation operation, an adsorption operation, and a desorption operation. At the activation operation, the gas microenvironment module 156 generates a purge gas microenvironment (e.g., dry air or nitrogen gas), and the positioning assembly 114 positions the sensing element 144 therein to purge previously adsorbed molecules and establish a baseline sensor output. At the adsorption operation, the gas microenvironment module 156 generates an analyte gas microenvironment containing a target gas analyte at a defined concentration, and the positioning assembly 114 positions the sensing element 144 therein. The optical interrogator 176 records sensor response data as the target gas analyte adsorbs into the sensing element 144 until adsorption equilibrium is reached. At the desorption operation, the positioning assembly 114 repositions the optical fiber sensor within the purge gas microenvironment to desorb the target gas analyte and return the sensor response to baseline. In one embodiment, purging with alternative gases such as 1,1 -difluoroethane is performed to produce distinct desorption time profiles for identifying specific adsorbed gas analytes.

[0068] The adsorption and desorption operations are repeated for a plurality of concentrations of the target gas analyte to generate a calibration curve representing sensor response as a function of target gas analyte concentration. In one embodiment, the plurality of concentrations are incrementally varied according to a calibration scheme comprising concentration increments in an inclusive range of from about 1 ppb to about 1 ppm. The processor 180 determines a sensitivity parameter from the slope of the calibration curve.

[0069] The calibration operation 1004 further comprises a cross-sensitivity calibration operation wherein the gas microenvironment module 156 generates a cross-sensitivity gas microenvironment (e.g., the mixed gas microenvironment 166 shown at FIG. 2) containing the target gas analyte at an analyte concentration and an interferent gas (e.g., water vapor) at an16CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024interferent concentration. The optical interrogator 176 records sensor response data as both gases adsorb into the sensing element 144, and the processor 180 generates a cross-sensitivity calibration curve to quantify the effect of the interferent gas on detection of the target gas analyte. The calibration operations may be repeated for a plurality of target gas analytes to characterize the sensor response to each target gas analyte.

[0070] In some embodiments, the sensor system 100 is configured to coat the composite layer 150 with a polymer to improve durability, chemical resistance, and / or selectivity. Hydrophobic polymer coatings may be applied to reduce moisture interference, and functionalized polymer layers may be used to introduce selective binding sites for specific gas molecules.

[0071] Following calibration, the method 1000 proceeds to the testing operation 1006, which evaluates performance of the optical fiber sensor under realistic operating conditions. The testing operation 1006 comprises generating, using the gas microenvironment module 156, a test gas microenvironment containing a predefined gas mixture comprising a target gas analyte and at least one background gas (e.g., dry air, nitrogen, or water vapor at a controlled humidity level). The positioning assembly 114 positions the sensing element 144 within the test gas microenvironment, and the optical interrogator 176 records sensor response data during exposure to the predefined gas mixture. The positioning assembly 114 then repositions the optical fiber sensor within a purge gas microenvironment to flush the sensor and restore the baseline sensor output.

[0072] In one embodiment, the testing operations are repeated in a repeated cycle under identical conditions to assess repeatability, accuracy, and measurement precision of the optical fiber sensor. Repeatability is assessed by comparing sensor response data across multiple cycles, accuracy is assessed by comparing recorded sensor response data to expected values based on known concentrations, and measurement precision is assessed by evaluating variability in sensor response data across the repeated cycles.

[0073] The sensor system 100 is configured to autonomously perform the fabricating operation 1002, the calibrating operation 1004, and the testing operation 1006 in a continuous sequence without manual intervention between operations. All sensor response data recorded during execution of the method 1000 is stored by the data acquisition and control unit 178 (e.g., at memory 182, at cloud storage module 190) for subsequent analysis and for training machine learning models to improve sensor manufacturing, sensor selectivity, and / or sensor accuracy.17CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024

[0074] Referring to FIG. 12, a method 1200 for identifying gas analytes using a machine learning model is generally indicated. Single sensors may lack sufficient selectivity in complex gas mixtures containing interfering molecules. Similar to the biological olfactory system wherein distinct but overlapping responses from multiple receptors enable pattern recognition, sensor arrays employ multiple sensors where each sensor responds differently to the same gas mixture, and the collective responses create unique patterns. Machine learning algorithms analyze these patterns to identify specific gases within complex mixtures. The selection of fabricated sensors for the sensor array enables the generation of collective response patterns from a plurality of optical fiber sensors, which provides enhanced pattern recognition capabilities compared to a single sensor. In general, the method 1200 comprises sequentially exposing optical fiber sensors to gas microenvironments, recording sensor response curves, training a machine learning model based on the response curves, exposing the optical fiber sensors to an unknown gas environment, and applying the trained machine learning model to identify gas analytes in the unknown gas environment.

[0075] The method 1200 begins with operation 1202 at which fabricated sensors are selected for a sensor array. The selection of fabricated sensors for the sensor array enables the generation of collective response patterns from a plurality of optical fiber sensors, which provides enhanced pattern recognition capabilities compared to a single sensor.

[0076] With continued reference to FIG. 12, following operation 1202 the method 1200 proceeds to operation 1204 at which the sensor array is exposed to simulated environments. With respect to the illustrated embodiment of FIG. 1, the gas microenvironment module 156 generates a plurality of gas microenvironments, each gas microenvironment 168 including at least one target gas analyte at a defined concentration. The plurality of gas microenvironments 168 includes varying concentrations of the at least one target gas analyte. It may include a plurality of different target gas analytes, and / or one or more interferent gases (e.g., for characterization of cross-sensitivity effects). The positioning assembly 114 sequentially positions the optical fiber sensors within each gas microenvironment 168 such that each of the optical fiber sensing elements 144 adsorb the at least one target gas analyte.

[0077] At operation 1206, sensor response data is recorded during exposure to the simulated environments. Thus, operations 1204 and 1206 are performed concurrently (e.g., simultaneously). The optical interrogator 176 records a sensor response curve for each exposure. Each sensor response curve captures a time-dependent response of the optical fiber sensor during adsorption of the at least one target gas analyte. The sensor response curve18CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024comprises an optical path length of the sensing element 144 as a function of time. The timedependent response captures the dynamic adsorption behavior of the target gas analyte into the sensing element 144, including the rate of adsorption and the equilibrium optical path length value.

[0078] In one embodiment, the method 1200 includes purging the sensing element 144 with a purge gas (e.g., nitrogen) between sequential exposures to the plurality of gas microenvironments to remove adsorbed molecules from the sensing element and returns the sensor response to a baseline output before the next exposure.

[0079] In another embodiment, the method 1200 includes recording a desorption transient curve during purging. The desorption transient curve captures the time-dependent response of the sensing element 144 as the target gas analyte desorbs from the metal-organic framework film. Different analytes desorb at different rates under varying purge conditions, and the desorption transient curves provide additional information for distinguishing gas species beyond direct adsorption measurements.

[0080] At operation 1208, a machine learning model is trained using the recorded sensor response data. The machine learning model is trained to recognize patterns in the sensor response curves corresponding to the at least one target gas analyte; that is, the training enables the machine learning model to identify a concentration of the at least one target gas analyte present at the sensing element 144 based on the time-dependent response of the sensing element thereto. When a plurality of optical fiber sensors are employed as a sensor array, the generating, exposing, and recording operations are repeated for a plurality of gas compositions to generate a training dataset comprising a plurality of collective response patterns from the plurality of optical fiber sensors to the plurality of gas compositions. The machine learning model is trained using the collective response patterns to recognize patterns in the collective response patterns corresponding to concentrations of the at least one target gas analyte. In one embodiment, the machine learning model is trained using the desorption transient curves in addition to the adsorption response curves. The machine learning model may be trained to identify many target analytes simultaneously.

[0081] Any of the calibrating operations or testing operations described above with respect to the method 1000 of FIG. 10 or the method 1100 of FIG. 11 may be included in the method 1200. For example, the activation, adsorption, desorption, and cross-sensitivity calibration operations described with respect to the calibration operation 1004 may be performed as part19CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024of operations 1204 and 1206 to generate the training dataset. Similarly, the testing operations described with respect to the testing operation 1006 may be performed as part of operations 1210 and 1212 when exposing the sensor array to the unknown gas environment.

[0082] In various embodiments, the machine learning model comprises a principle component analysis (PCA) model, a support vector machine (SVM), or an artificial neural networks (ANNs).

[0083] The method 1200 continues to operation 1210 at which the sensor array is exposed to an unknown gas environment. Each of the optical fiber sensing elements 144 adsorb gas analytes present in the unknown gas environment.

[0084] At operation 1212, sensor response data is recorded from the unknown gas environment. This generally comprises recording a sensor response curve from each of the plurality of optical fiber sensors in the unknown gas environment.

[0085] At operation 1214, one or more analytes in the unknown environment are identified via the trained machine learning model. The trained machine learning model is applied to the sensor response curves from the unknown gas environment to identify a concentration of the at least one gas analyte in the unknown gas environment based on the collective response patterns from the plurality of optical fiber sensors. The collective response patterns from the sensor array provide enhanced discrimination between gas analytes compared to a single sensor, as each sensor in the array responds differently to the same gas mixture, creating patterns that the machine learning model recognizes.

[0086] Referring to FIG. 13, a method 1300 for manufacturing optical fiber sensors and selecting optical fiber sensors for inclusion in a sensor array is generally indicated. In general, the method 1300 enables identification of patterns in fabrication data that correspond to sensor performance, allowing selection of sensors meeting predefined performance criteria for the sensor array. The method 1300 broadly comprises fabricating a plurality of optical fiber sensors, recording sensor response data during fabrication, recording sensor performance data after fabrication, training a machine learning model, fabricating subsequent sensors, and screening the subsequent sensors using the trained machine learning model.

[0087] The method 1300 begins with operation 1302. At operation 1302, a plurality of optical fiber sensors are fabricated using the dip method. As described previously with respect to FIGS. 4-8, each optical fiber sensor is fabricated by forming a sensing element at an end20CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024face thereof by dipping the optical fiber into a precursor solution to form a microdroplet at the end face and evaporating the microdroplet to form a composite layer at the end face.

[0088] At operation 1304 optical path length data is recorded, using an optical interrogator, for each sensing element of the plurality of optical fiber sensors during fabrication. Monitoring the optical path length of each sensing element during fabrication comprises monitoring the optical path length during evaporation of the microdroplet. The optical interrogator 176 records the optical path length data in real time as the microdroplet evaporates and the composite layer forms at the optical fiber end face 142. The optical path length data captures the time-dependent evolution of the sensing element during the fabrication process, including changes resulting from solvent evaporation and crystallization of the precursor portion of the precursor solution.

[0089] At operation 1306, performance data is recorded for each of the plurality of optical fiber sensors following fabrication thereof. Performance characteristics may include sensitivity to a target gas analyte, response time, recovery time, long-term stability, or other characteristics. In one embodiment, the performance data is collected by exposing each fabricated optical fiber sensor to one or more target gas analytes at defined concentrations and recording the sensor response using the optical interrogator 176. The sensitivity to a target gas analyte is determined from the slope of a calibration curve representing sensor response as a function of target gas analyte concentration. The response time is determined from the time required for the sensor to reach adsorption equilibrium upon exposure to the target gas analyte. The recovery time is determined from the time required for the sensor to return to a baseline sensor output following purging with a purge gas. Long-term stability is determined from repeated measurements over an extended period.

[0090] At operation 1308, a machine learning model (e.g., PCA, SVM, ANN, CNN, etc.) is trained using the optical path length data and the performance data to identify patterns in the optical path length data that correspond to the performance data. The training process correlates features in the optical path length data recorded during fabrication with the performance characteristics measured after fabrication. In this manner, the machine learning model learns to recognize patterns in the fabrication data that are predictive of sensor performance.

[0091] At operation 1310, subsequent optical fiber sensors are fabricated while recording, using the optical interrogator 176, subsequent optical path length data therefrom. The subsequent optical fiber sensor is fabricated using the same dip method as described with21CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024respect to the operation 1302. During fabrication of the subsequent optical fiber sensors, the optical interrogator 176 records the subsequent optical path length data in real time.

[0092] At operation 1312, the machine learning model is applied to the subsequent optical path length data to predict performance characteristics of the subsequently fabricated optical fiber sensors. Operation 1312 comprises screening the subsequent optical fiber sensors to determine whether the predicted performance characteristics meet predefined performance criteria.

[0093] In one embodiment, optical fiber sensors are selected for inclusion in a sensor array based on the predicted performance characteristics. For example, optical fiber sensors having predicted performance characteristics that meet or exceed the predefined performance criteria are selected for the sensor array, while optical fiber sensors having predicted performance characteristics that do not meet the predefined performance criteria are rejected. The predefined performance criteria are established based on the intended application of the sensor array and the desired sensitivity, response time, recovery time, and long-term stability for detecting one or more target gas analytes.

[0094] The trained model may also be used for general screening purposes beyond sensor array selection to save time and resources by allowing the sensor system 100 to identify undesirable characteristics before expending gas and other resources on calibrating and testing. For example, if a composite layer is formed incorrectly, this can be identified during or shortly after fabrication based on the optical path length data, allowing the sensor to be rejected before proceeding to calibration and testing operations.

[0095] Together, methods 1000, 1100, 1200, and 1300 enable the sensor system 100 to autonomously produce, characterize, and deploy optical fiber sensors while leveraging machine learning for both sensor selection and gas analyte identification. Method 1000 generally describes the framework for fabrication, calibration, and testing in a continuous sequence. Method 1100 more specifically details the fabrication process (e.g., the dip method). Method 1200 facilitates gas analyte identification via machine learning and sensor arrays, wherein collective response patterns from multiple optical fiber sensors are analyzed to distinguish target analytes for complex gas mixtures. Method 1300 streamlines screening and selection of sensors based on fabrication data correlated with performance characteristics for enabling the sensor system 100 to predict sensor performance from optical path length data recorded during fabrication.22CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024EXAMPLES

[0096] The following examples demonstrate the fabrication and gas sensing performance of optical fiber sensors produced using the pendant micro-droplet evaporation technique (dip method) described herein. These examples illustrate real-time monitoring of sensing element formation, calibration of sensor response to ethanol, and sensor behavior during humidity and ethanol exposure cycles. The data presented in these examples were obtained from implementations of the sensor system and validates the principles underlying the integrated fabrication, calibration, and testing approach of the present disclosure.

[0097] Referring to FIG. 14, a graph 1400 depicts optical path length in micrometers on the vertical axis versus time in seconds on the horizontal axis during sensor fabrication under ambient conditions. FIG. 14 provides a detailed representation of real-time data collected during fabrication of the sensing element 144. The optical path length measurements are recorded using an extrinsic Fabry -Perot interferometer system, tracking the evolution of the sensing element 144 in real time as the pendant microdroplet 146 at the optical fiber end face 142 evaporates. The graph 1400 displays two overlapping traces representing an acquired signal and a filtered signal. The optical path length initially reflects the vertical diameter of the pendant microdroplet 146, which typically starts at around 50 micrometers when the optical fiber 132 is withdrawn from the precursor solution 110. As the solvent evaporates, the microdroplet 146 shrinks, causing a gradual reduction in the optical path length. Changes in the refractive index of the microdroplet 146, resulting from increasing solute concentration and the nucleation of MOF crystals, also contribute to variations in the recorded optical path length. The optical path length remains relatively stable with minor fluctuations until approximately 70 seconds, after which a rapid decrease occurs. The optical path length drops from approximately 55 micrometers to approximately 7 micrometers between approximately 70 seconds and 100 seconds, indicating the evaporation of the microdroplet 146 and formation of the composite layer 150 at the optical fiber end face 142. The raw data (acquired signal) may contain noise due to environmental factors such as minor vibrations, temperature variations, or fluctuations in the interrogator system. The filtered signal removes this noise, providing a smoothed representation of the acquired signal while preserving the overall trend of the optical path length change during the fabrication process.

[0098] With continued reference to FIG. 14, as the evaporation proceeds, the rate of change in optical path length slows, indicating the transition from a liquid-phase microdroplet to a solid-phase MOF film. The plateau in the optical path length signifies the completion of23CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024the sensing element 144 formation, marking the point at which the composite layer 150 has fully crystallized and adhered to the optical fiber end face 142. The final thickness of the MOF layer can be inferred from the final optical path length measurement. The initial plateau region of the graph 1400 corresponds to the period during which the microdroplet 146 is present at the optical fiber end face 142 and solvent evaporation is occurring at a rate that maintains the microdroplet 146 in a relatively stable configuration. The rapid decrease in optical path length corresponds to the final stages of solvent evaporation and the formation of the composite layer 150. This real-time data is valuable for Al training and optimization of the fabrication process as described above. For example, a machine learning model (e.g., that described with respect to FIG. 13) can be trained to identify anomalies in the evaporation profile that may indicate defects in film formation, allowing for automatic adjustments in precursor concentration, drying conditions, or droplet size to correct these issues in future fabrication cycles.

[0099] Referring to FIG. 15, a graph 1500 depicts sensor response to ethanol. Calibration establishes the sensor's response characteristics for a given target gas analyte. In FIG. 15, a calibration process is demonstrated for ethanol as the target gas analyte, though the same general approach applies to a wide range of gases. The sensor data was recorded autonomously, wherein the sensor under test was transitioned between a pure nitrogen microenvironment and a gas microenvironment containing ethanol in nitrogen at varying concentrations. The process began with the sensor exposed to a 130 ppm ethanol -in-nitrogen environment until equilibrium was reached, followed by a return to the nitrogen microenvironment to allow for ethanol desorption. This sequence was repeated for 208 ppm and 260 ppm concentrations.

[0100] The graph 1500 plots sensor response in micrometers on the vertical axis against ethanol concentration in nitrogen in parts per million on the horizontal axis. The graph 1500 includes data points at 130 ppm, 208 ppm, and 260 ppm ethanol concentrations, with corresponding sensor responses of approximately 7.3 micrometers, 10.7 micrometers, and 12.5 micrometers respectively. The recorded sensor response (optical path length) is plotted against the ethanol concentrations, forming the calibration curve for ethanol. A dashed linear trend line extends through the data points, indicating a linear relationship between ethanol concentration and sensor response. The graph 1500 indicates an ethanol sensitivity of 36.4 nm / ppm. An inset within the graph 1500 displays an ethanol adsorption isotherm, plotting equilibrium uptake in weight percent on the vertical axis against relative pressure on the horizontal axis.

[0101] With continued reference to FIG. 15, the linear relationship between ethanol concentration and sensor response demonstrated in the graph 1500 indicates that the optical24CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024fiber sensor fabricated using the dip method described with reference to FIGS. 4-8 exhibits predictable and reproducible sensing characteristics. The sensitivity of 36.4 nm / ppm provides a quantitative measure of the sensor response per unit concentration of ethanol, enabling calibration of the optical fiber sensor for ethanol detection applications.

[0102] Referring to FIG. 16, a graph 1600 A depicts wavelength in nanometers versus time in seconds and a graph 1600B depicts optical path length in micrometers versus time in seconds. FIG. 16 illustrates the formation and response behavior of a thin crystalline film layer (approximately 6 micrometers) initially drawn as a micro-droplet from a precursor solution. This layer, as grown from a micro-droplet of precursor solution on the fiber tip under ambient conditions, is maintained at room temperature, purged with dry nitrogen gas, and subjected to multiple humidity-purge cycles. Each cycle consists of exposure to relative humidity until equilibrium is reached, followed by a nitrogen purge until an empty MOF equilibrium is reached. The sensor is interrogated using an optical fiber interrogator functioning as an extrinsic Fabry -Perot interferometer. MOFs inherently do not selectively detect a single gas molecule; instead, any molecule capable of diffusing into the MOF's cages through its windows will contribute to the sensor response. After the humidity adsorption cycles, a single ethanol-purge cycle is performed.

[0103] The graph 1600 A illustrates sensor response during transitions between water vapor exposure and nitrogen purge cycles, labeled as H2O to N2 Purge for relative humidity, followed by exposure to 260 ppm ethanol transitioning to nitrogen purge. The graph 1600 A shows wavelength values in an inclusive range of from 1540 nm to 1580 nm. The graph 1600B corresponds to the same time period and shows optical path length measurements at a relative humidity of 55% and a concentration of 17579 ppm. The graph 1600B displays optical path length values in an inclusive range of from 6.36 micrometers to 6.46 micrometers.

[0104] The graphs 1600 A and 1600B demonstrate the response of the optical fiber sensor to repeated humidity-purge cycles followed by ethanol exposure. The nitrogen purge removes all adsorbed molecules, effectively "activating" the MOF by clearing its cavities, ensuring accurate subsequent adsorption measurements. The data shown in the graphs 1600A and 1600B illustrates the sensor equilibrium behavior and desorption characteristics under varying nitrogen purge flow rates. The wavelength shifts shown in the graph 1600A correspond to changes in the optical path length of the sensing element 144 as water vapor and ethanol molecules adsorb into and desorb from the composite layer 150. The desorption process varies25CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024due to fluctuations in nitrogen purge flow rates. The desorption slope is dependent on the nitrogen purge flow rate and the adsorbed target gas analyte, providing additional information for Al-based (e.g., machine learning based) sensor output analysis. Purging with alternative gases, such as 1,1 -difluoroethane, produces distinct desorption time profiles, offering an additional differentiation metric for identifying specific adsorbed gas analytes. Thus, purge gas composition and flow rate are tunable variables that generate desorption transient curves, which can serve as input for Al model training to improve sensor selectivity. Since different analytes desorb at different rates under varying purge conditions, this process enhances the ability to distinguish gas species beyond direct adsorption measurements.

[0105] While the systems and methods above have been described and disclosed in certain terms and have disclosed certain embodiments or modifications, persons skilled in the art who have acquainted themselves with the disclosure, will appreciate that it is not necessarily limited by such terms, nor to the specific embodiments and modification disclosed herein. Thus, a wide variety of alternatives, suggested by the teachings herein, can be practiced without departing from the spirit of the disclosure, and rights to such alternatives are particularly reserved and considered within the scope of the disclosure.

[0106] Embodiments of the present disclosure may comprise a special purpose computer including a variety of computer hardware, as described in greater detail herein.

[0107] For purposes of illustration, programs and other executable program components may be shown as discrete blocks. It is recognized, however, that such programs and components reside at various times in different storage components of a computing device, and are executed by a data processor(s) of the device.

[0108] Although described in connection with an example computing system environment, embodiments of the aspects of the disclosure are operational with other special purpose computing system environments or configurations. The computing system environment is not intended to suggest any limitation as to the scope of use or functionality of any aspect of the disclosure. Moreover, the computing system environment should not be interpreted as having any dependency or requirement relating to any one or combination of components illustrated in the example operating environment. Examples of computing systems, environments, and / or configurations that may be suitable for use with aspects of the disclosure include, but are not limited to, personal computers, server computers, hand-held or laptop devices, multiprocessor systems, microprocessor-based systems, set top boxes,26CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024programmable consumer electronics, mobile telephones, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.

[0109] Embodiments of the aspects of the present disclosure may be described in the general context of data and / or processor-executable instructions, such as program modules, stored in memory, i.e., one or more tangible, non-transitory storage media, and executed by one or more processors or other devices. Generally, program modules include, but are not limited to, routines, programs, objects, components, and data structures that perform particular tasks or implement particular abstract data types. Aspects of the present disclosure may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote storage media including memory storage devices.

[0110] In operation, processors, computers and / or servers may execute the processorexecutable instructions (e.g., software, firmware, and / or hardware) such as those illustrated herein to implement aspects of the disclosure.

[0111] Embodiments may be implemented with processor-executable instructions. The processor-executable instructions may be organized into one or more processor-executable components or modules on a tangible processor readable storage medium. Also, embodiments may be implemented with any number and organization of such components or modules. For example, aspects of the present disclosure are not limited to the specific processor-executable instructions or the specific components or modules illustrated in the figures and described herein. Other embodiments may include different processor-executable instructions or components having more or less functionality than illustrated and described herein.

[0112] The order of execution or performance of the operations in accordance with aspects of the present disclosure illustrated and described herein is not essential, unless otherwise specified. That is, the operations may be performed in any order, unless otherwise specified, and embodiments may include additional or fewer operations than those disclosed herein. For example, it is contemplated that executing or performing a particular operation before, contemporaneously with, or after another operation is within the scope of the disclosure.

[0113] When introducing elements of the disclosure or embodiments thereof, the articles "a," "an," "the," and "said" are intended to mean that there are one or more of the elements. The27CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024terms "comprising," "including," and "having" are intended to be inclusive and mean that there may be additional elements other than the listed elements.

[0114] Not all of the depicted components illustrated or described may be required. In addition, some implementations and embodiments may include additional components. Variations in the arrangement and type of the components may be made without departing from the spirit or scope of the claims as set forth herein. Additional, different or fewer components may be provided and components may be combined. Alternatively, or in addition, a component may be implemented by several components.

[0115] The above description illustrates embodiments by way of example and not by way of limitation. This description enables one skilled in the art to make and use aspects of the disclosure, and describes several embodiments, adaptations, variations, alternatives and uses of the aspects of the disclosure, including what is presently believed to be the best mode of carrying out the aspects of the disclosure. Additionally, it is to be understood that the aspects of the disclosure are not limited in application to the details of construction and the arrangement of components set forth in the following description or illustrated in the drawings. The aspects of the disclosure are capable of other embodiments and of being practiced or carried out in various ways. Also, it will be understood that the phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting.

[0116] It will be apparent that modifications and variations are possible without departing from the scope of the disclosure defined in the appended claims. As various changes could be made in the above constructions and methods without departing from the scope of the disclosure, it is intended that all matter contained in the above description and shown in the accompanying drawings shall be interpreted as illustrative and not in a limiting sense.

[0117] In view of the above, it will be seen that several advantages of the aspects of the disclosure are achieved and other advantageous results attained.

[0118] The Abstract and Summary are provided to help the reader quickly ascertain the nature of the technical disclosure. They are submitted with the understanding that they will not be used to interpret or limit the scope or meaning of the claims. The Summary is provided to introduce a selection of concepts in simplified form that are further described in the Detailed Description. The Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the claimed subject matter.28CORE / 3510410.019901 / 238905605.1

Claims

Docket No. 3510410.019901; 25MST024WHAT IS CLAIMED IS:

1. A sensor system for fabricating an optical fiber sensor from an optical fiber, the sensor system comprising:a positioning assembly configured to receive and hold the optical fiber and to move the optical fiber along at least two dimensions;a gas microenvironment module including a plurality of gas ports configured to fluidly couple to one or more gas reservoirs, the gas microenvironment module being controllable to dispense one or more gases from the plurality of gas ports to generate a microenvironment in proximity to the gas microenvironment module;an optical interrogator configured to optically couple to the optical fiber at a proximal end thereof and to measure an optical path length of a sensing element formed at an end face of the optical fiber at a distal end thereof; anda data acquisition and control unit operably coupled to: the positioning assembly; the gas microenvironment module; and the optical interrogator, the data acquisition and control unit comprising a processor and memory, the processor being communicatively coupled to the memory, the memory including instructions which, when executed by the processor, cause the processor to:control the sensor system to perform one or more fabricating operations on the optical fiber to form the optical fiber sensor;control the sensor system to perform one or more calibration operations on the optical fiber sensor; andcontrol the sensor system to perform one or more testing operations on the optical fiber sensor.

2. The sensor system set forth in claim 1, wherein the at least two dimensions comprise a vertical dimension, and wherein the one or more fabricating operations comprises a dip operation, the dip operation including controlling the positioning assembly to move the optical fiber along the vertical dimension to dip the optical fiber into a precursor solution for forming a microdroplet of the precursor solution at the end face of the optical fiber.

3. The sensor system set forth in claim 2, wherein the precursor solution is a metal-organic-compound precursor solution.29CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST0244. The sensor system set forth in claims 2 or 3, wherein the precursor solution comprises HKUST-1 precursor.

5. The sensor system set forth in any one of claims 3-4, wherein the one or more fabricating operations comprise a rest operation, the rest operation comprising holding the positioning assembly steady for a rest period of a duration sufficient to permit the precursor solution to form a composite layer of the sensing element, the composite layer being formed via evaporation of a solvent portion of the precursor solution and crystallization of a precursor portion of the precursor solution.

6. The sensor system set forth in claim 5, wherein the duration is in the range of 20 seconds to 100 seconds.

7. The sensor system set forth in claims 5 or 6, wherein the duration is about 30 seconds.

8. The sensor system set forth in any one of claims 5-7, wherein the one or more fabricating operations comprise a gas exposure operation, the gas exposure operation including:generating, using the gas microenvironment module, a gaseous microenvironment in proximity to the gas microenvironment module; andpositioning, using the positioning assembly, the distal end region of the optical fiber in the gaseous microenvironment.

9. The sensor system set forth in claim 8, wherein the gas exposure operation is performed concurrently with the rest operation such that evaporation of the solvent portion of the precursor solution occurs within the gaseous microenvironment.

10. The sensor system set forth in claim 9, wherein the gaseous microenvironment comprises dry nitrogen gas to minimize water incorporation into the composite layer during evaporation of the solvent.30CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST02411. The sensor system set forth in claim 9, wherein the gaseous microenvironment comprises a controlled humidity environment containing water vapor to influence a crystal hydration level in the composite layer.

12. The sensor system set forth in claim 9, wherein the gaseous microenvironment comprises an alternative carrier gas.

13. The sensor system set forth in claim 12, wherein the alternative carrier gas is selected from argon or carbon dioxide.

14. The sensor system set forth in any one of claims 9-12, further comprising a gas control system including one or more electronically actuated valves operably coupled to the processor, wherein the one or more electronically actuated valves are controllable by the processor to regulate gas flow from the one or more gas reservoirs to the plurality of gas ports.

15. The sensor system set forth in claim 14, wherein the electronically actuated valves are flow rate valves.

16. The sensor system set forth in any one of claims 1-15, wherein the one or more calibration operations comprise an activation operation including:generating, using the gas microenvironment module, a purge gas microenvironment in proximity to the gas microenvironment module, the purge gas microenvironment including a purge gas; andpositioning, using the positioning assembly, the sensing element within the purge gas microenvironment to purge the sensing element for establishing a baseline sensor output.

17. The sensor system set forth in claim 16, wherein the purge gas comprises dry air or nitrogen gas.31CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST02418. The sensor system set forth in claims 16 or 17, wherein the one or more calibration operations further comprise:an adsorption operation including:generating, using the gas microenvironment module, an analyte gas microenvironment in proximity to the gas microenvironment module, the analyte gas microenvironment including a target gas analyte at a first concentration;positioning, using the positioning assembly, the sensing element at the analyte gas microenvironment; andrecording, using the optical interrogator, sensor response data of the optical fiber sensor as the target gas analyte adsorbs into the sensing element until the sensor response data indicates an adsorption equilibrium has been reached at the sensing element; anda desorption operation including:repositioning, using the positioning assembly, the optical fiber sensor within the purge gas microenvironment to desorb the target gas analyte from the sensing element and return the sensor response to the baseline sensor output.

19. The sensor system set forth in claim 18, wherein the one or more calibration operations further comprise repeating the adsorption operation and the desorption operation for a plurality of concentrations of the target gas analyte to generate a calibration curve representing sensor response as a function of target gas analyte concentration.

20. The sensor system set forth in claim 19, wherein the plurality of concentrations are incrementally varied according to a calibration scheme.

21. The sensor system set forth in claim 20, wherein the calibration scheme comprises concentration increments in an inclusive range of from about Ippb to about 1 ppm.

22. The sensor system set forth in any one of claims 19-21, wherein the processor is further configured to determine a sensitivity parameter from the calibration curve.32CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST02423. The sensor system set forth in any one of claims 18-22, wherein the one or more calibration operations further comprise a cross-sensitivity calibration operation including: generating, using the gas microenvironment module, a cross-sensitivity gas microenvironment in proximity to the gas microenvironment module, the cross-sensitivity gas microenvironment including the target gas analyte at an analyte concentration and an interferent gas at an interferent concentration;positioning, using the positioning assembly, the sensing element at the cross-sensitivity gas microenvironment;recording, using the optical interrogator, sensor response data of the optical fiber sensor as the target gas analyte and the interferent gas adsorb into the sensing element; and generating a cross-sensitivity calibration curve to quantify an effect of the interferent gas on detection of the target gas analyte.

24. The sensor system set forth in claim 23, wherein the interferent gas comprises water vapor.

25. The sensor system set forth in any one of claims 18-24, wherein the one or more calibration operations are repeated for a plurality of target gas analytes to characterize the sensor response to each target gas analyte.

26. The sensor system set forth in claim 1, wherein the one or more testing operations comprise:generating, using the gas microenvironment module, a test gas microenvironment in proximity to the gas microenvironment module, the test gas microenvironment including a predefined gas mixture comprising a target gas analyte and at least one background gas; positioning, using the positioning assembly, the sensing element at the test gas microenvironment;recording, using the optical interrogator, sensor response data of the optical fiber sensor during exposure to the predefined gas mixture; andrepositioning, using the positioning assembly, the optical fiber sensor within a purge gas microenvironment to flush the optical fiber sensor with a purge gas and restore a baseline sensor output.33CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST02427. The sensor system set forth in claim 26, wherein the one or more testing operations further comprise repeating the generating, positioning, recording, and repositioning in a repeated cycle under identical conditions to assess at least one of repeatability, accuracy, and measurement precision of the optical fiber sensor.

28. The sensor system set forth in any one of claims 26-27, wherein the at least one background gas comprises at least one of dry air, nitrogen, and water vapor at a controlled humidity level.

29. The sensor system set forth in claim 1, wherein each gas port of the plurality of gas ports is connected to the one or more gas reservoirs through a gas tube.

30. The sensor system set forth in claim 29, wherein the gas microenvironment module comprises a grid-patterned gas port plate configured to hold gas tubes in a structured arrangement for formation of gas flow fields.

31. The sensor system set forth in claim 30, wherein the plurality of gas ports are arranged in one of a linear arrangement, a honeycomb arrangement, or a triangular arrangement.

32. The sensor system set forth in any one of claims 29-31, further comprising at least one of:heaters configured to increase a temperature of gas flowing through the gas tubes; and coolers configured to decrease a temperature of gas flowing through the gas tubes.

33. The sensor system set forth in claim 1, further comprising a base configured to support a vial containing a precursor, wherein the base is temperature controlled.34CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST02434. The sensor system set forth in claim 33, wherein the base is configured to heat the vial to a temperature in an inclusive range of from about ambient temperature to about 200°C.

35. The sensor system set forth in claim 1, wherein the positioning assembly is configured to move the optical fiber along three dimensions.

36. The sensor system set forth in claim 35, wherein the three dimensions comprise an X-dimension for side-to-side positioning, a Y-dimension for vertical positioning, and a Z-dimension for back-and-forth positioning.

37. The sensor system set forth in claim 36, wherein the positioning assembly has a positioning resolution of 10 micrometers or finer in each of the three dimensions.

38. The sensor system set forth in any one of claims 35-37, wherein the positioning assembly has a positioning resolution of 5 micrometers or finer in each of the three dimensions.

39. The sensor system set forth in claim 1, wherein the optical interrogator comprises an extrinsic Fabry-Perot interferometer system configured to provide real-time monitoring of the sensing element during formation thereof.

40. The sensor system set forth in claim 1, wherein the data acquisition and control unit is configured to autonomously perform the one or more fabricating operations, the one or more calibration operations, and the one or more testing operations in a continuous sequence.

41. A method of identifying gas analytes, the method comprising:fabricating an optical fiber sensor having a sensing element at an end face of an optical fiber;generating, using a gas microenvironment module, a plurality of gas microenvironments, each gas microenvironment including at least one target gas analyte at a defined concentration;35CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024sequentially exposing the optical fiber sensor to the plurality of gas microenvironments by positioning the optical fiber sensor within each gas microenvironment;recording, using an optical interrogator, a sensor response curve for each exposure, each sensor response curve capturing a time-dependent response of the optical fiber sensor during adsorption of the at least one target gas analyte;training a machine learning model using the recorded sensor response curves to recognize patterns in the sensor response curves corresponding to the at least one target gas analyte for identifying a concentration of the at least one target gas analyte present at the sensing element;exposing the optical fiber sensor to an unknown gas environment;recording, using the optical interrogator, a sensor response curve from the unknown gas environment; andapplying the trained machine learning model to the sensor response curve from the unknown gas environment to identify a concentration of the least one gas analyte in the unknown gas environment.

42. The method set forth in claim 41, wherein the plurality of gas microenvironments includes varying concentrations of the at least one target gas analyte.

43. The method set forth in claims 41 or 42, wherein the plurality of gas microenvironments includes a plurality of different target gas analytes.

44. The method set forth in any one of claims 41-43, wherein each sensor response curve comprises an optical path length of the sensing element as a function of time.

45. The method set forth in any one of claims 41-44, wherein the machine learning model comprises at least one of principal component analysis (PCA), support vector machines (SVM), or artificial neural networks (ANNs).

46. The method set forth in any one of claims 41-45, further comprising purging the sensing element with a purge gas between sequential exposures to the plurality of gas microenvironments.36CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST02447. The method set forth in claim 46, further comprising recording a desorption transient curve during purging and training the machine learning model using the desorption transient curves.

48. The method set forth in claims 46 or 47, wherein the purge gas comprises nitrogen gas.

49. A method for identifying gas analytes using a sensor array, the method comprising:fabricating a plurality of optical fiber sensors, each optical fiber sensor having a sensing element at an end face of an optical fiber;generating, using a gas microenvironment module, a gas microenvironment having a defined gas composition including at least one target gas analyte at a defined concentration;exposing the plurality of optical fiber sensors to the gas microenvironment such that each of the optical fiber sensing elements adsorb the at least one target gas analyte;recording, using an optical interrogator, a sensor response curve from each of the plurality of optical fiber sensors, wherein each sensor response curve captures a time-dependent response of the respective sensing element during adsorption of the at least one target gas analyte;repeating said generating, exposing, and recording for a plurality of gas compositions to generate a training dataset comprising a plurality of collective response patterns from the plurality of optical fiber sensors to the plurality of gas compositions;training a machine learning model using the collective response patterns to recognize patterns in the collective response patterns corresponding to concentrations of the at least one target gas analyte;exposing the plurality of optical fiber sensors to an unknown gas environment; recording, using the optical interrogator, a sensor response curve from each of the plurality of optical fiber sensors in the unknown gas environment; andapplying the trained machine learning model to the sensor response curves from the unknown gas environment to identify a concentration of the at least one target gas analyte in37CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST024the unknown gas environment based on the collective response patterns from the plurality of optical fiber sensors.

50. The method set forth in claim 49, wherein each sensing element comprises a composite layer.

51. The method set forth in claim 50, wherein the composite layer comprises HKUST-1.

52. The method set forth in any one of claims 49-51, wherein each sensor response curve captures an optical path length of the respective sensing element as a function of time.

53. The method set forth in any one of claims 49-52, wherein the plurality of gas compositions includes varying concentrations of the at least one target gas analyte.

54. The method set forth in any one of claims 49-53, wherein the plurality of gas compositions includes a plurality of different target gas analytes.

55. The method set forth in any one of claims 49-54, wherein the plurality of gas compositions includes at least one interferent gas.

56. The method set forth in any one of claims 49-55, wherein the machine learning model comprises at least one of principal component analysis (PCA), support vector machines (SVM), or artificial neural networks (ANNs).

57. The method set forth in any one of claims 49-56, further comprising purging each sensing element with a purge gas between exposures to the plurality of gas compositions.38CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST02458. The method set forth in claim 57, further comprising recording a desorption transient curve from each of the plurality of optical fiber sensors during purging, and training the machine learning model using the desorption transient curves.

59. The method set forth in claims 57 or 58, wherein the purge gas comprises nitrogen gas.

60. A method for manufacturing optical fiber sensors, the method comprising: fabricating a plurality of optical fiber sensors, each optical fiber sensor being fabricated by forming a sensing element at an end face thereof by dipping the optical fiber into a precursor solution to form a microdroplet at the end face and evaporating the microdroplet to form a composite layer defining the sensing element at the end face;recording, using an optical interrogator, optical path length data of each sensing element of the plurality of optical fiber sensors during fabrication;recording performance data of each of the plurality of optical fiber sensors following fabrication thereof;training a machine learning model, using the optical path length data and the performance data, to identify patterns in the optical path length data that correspond to the performance data;fabricating a subsequent optical fiber sensor while recording, using the optical interrogator, subsequent optical path length data therefrom; andapplying the machine learning model to the subsequent optical path length data to predict performance characteristics of the subsequent optical fiber sensor.

61. The method set forth in claim 60, wherein recording the optical path length of each sensing element during fabrication comprises monitoring the optical path length during evaporation of the microdroplet.

62. The method set forth in claims 60 or 61 , wherein the performance characteristics comprise at least one of sensitivity to a target gas analyte, response time, recovery time, or long-term stability.39CORE / 3510410.019901 / 238905605.1Docket No. 3510410.019901; 25MST02463. The method set forth in any one of claims 60-62, wherein the machine learning model comprises at least one of principal component analysis (PCA), support vector machines (SVM), or artificial neural networks (ANNs).

64. The method set forth in any one of claims 60-63, wherein the composite layer comprises metal-organic framework crystallites.

65. The method set forth in claim 64, wherein the metal-organic framework crystallites comprise HKUST-1.

66. The method set forth in any one of claims 60-65, wherein the precursor solution is a metal-organic framework precursor solution.

67. The method set forth in any one of claims 60-66, further comprising selecting, based on the predicted performance characteristics, optical fiber sensors for inclusion in a sensor array.

68. The method set forth in claim 67, wherein said selecting comprises screening the subsequent optical fiber sensor to determine whether the predicted performance characteristics meet predefined performance criteria for the sensor array.40CORE / 3510410.019901 / 238905605.1

Images