Plasmonic coffee-ring pattern diagnostic devices and methods of making and using them

Abstract

The invention provides devices and methods that allow detection of analytes of interest present in a sample at concentrations as low as 3 pg / mL. The methods involve allowing a drop of sample containing or not containing an analyte of interest to dry on a nanofibrous or mesofibrous hydrophilic membrane treated to render it intermediate between hydrophilic and hydrophobic and overlapping the area of the dried sample drop with a drop containing gold nanoshells functionalized to detect the analyte of interest. Patterns of the interaction between the functionalized gold nanoshells and the analyte of interest (if present) allow detection of the analyte and, in machine learning-assisted embodiments, quantifying the analyte present in the sample.

Description

PATENT APPLICATIONPLASMONIC COFFEE-RING PATTERN DIAGNOSTIC DEVICES AND METHODS OF MAKING AND USING THEMInventor(s): Kamyar BEHROUZI, a resident of Albany, CALiwei LIN, a resident of Moraga, CAApplicant: The Regents of the University of CaliforniaEntity: SmallCanady + Lortz, LLP3435 Wilshire Blvd., Suite 1400Los Angeles, CA 90010Tel: (415) 440-3100PLASMONIC COFFEE-RING PATTERN DIAGNOSTIC DEVICES AND METHODS OF MAKING AND USING THEMCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to, and the benefit of, U.S. Provisional Patent Application No. 63 / 625,582, filed January 26, 2024, the contents of which are hereby incorporated by reference.STATEMENT OF FEDERAL FUNDING

[0002] Not applicable.BACKGROUND OF THE INVENTION

[0003] The majority of diseases can potentially be prevented or diagnosed earlier by providing affordable and user-friendly biosensing platforms. Access to uncomplicated biosensors holds the potential to elevate global health quality, leading to significant cost savings and the preservation of numerous lives [1]—

[0010] . While considerable efforts are being made to develop straightforward and cost-effective biosensors for various diseases, such as SARS-CoV-2

[0011] —

[0016] , cancer

[0017] —

[0022] , and bloodstream infections

[0023] —

[0027] , the lateral flow immunoassay (“LFIA”) stands out as the most effective platform to date, considering factors like sensitivity, simplicity, and the necessary instrumentation

[0028] —

[0031] ,

[0004] Nevertheless, when compared to the gold standard, enzyme-linked immunosorbent assays, (“ELISA”), currently available LFIAs fall short in sensitivity for the early diagnosis of critical diseases such as sepsis

[0031] ,

[0032] , Sepsis, defined as an overreaction of the immune system to an infection, affects some 50 million people annually causing about 19% of all global death

[0033] , Early detection and effective treatment of diseases such as sepsis relies significantly on the sensitivity and response time of the biosensor. Current hospital practices involve using ELISA, which takes between 2 to 4 hours

[0034] , However, in sepsis, starting antibiotic administration just one hour earlier can increase the chance of survival by 7.6%

[0035] , and delayed diagnostics can exacerbate the situation

[0036] ,

[0037] , Thus, reducing the time to diagnosis becomes crucial

[0033] , The problem is further exacerbated by delayed hospitalvisits due to the unavailability of a simple test

[0034] ,

[0035] , The concentration of procalcitonin (“PCT”), a key sepsis biomarker

[0038] , can reach around 50-500 pg / ml three hours after the start of infection once symptoms emerge

[0039] , Most cases reach hospitals, however, after 12 hours, when the concentration of PCT exceeds 10 ng / ml

[0039] , Conventional visual LFIA, which operates on the level of ng / ml

[0029] ,

[0039] , cannot reliably detect PCT at the pg / ml levels required for early detection of sepsis

[0039] ,

[0005] To address the sensitivity issue, researchers have been exploring next generation LFIAs by incorporating fluorescent labeling and imaging to improve their detection limits

[0040] -

[0042] , For example, combining LFIA with activated fluorescent nanoparticles has shown a significant improvement in the sensitivity for biomarkers like SARS-CoV-2 Nucleocapsid protein (N-Protein), with a limit of detection (LOD) of 212 pg / ml

[0041] , but requires special instrumentation. Research works on biosensors detectable by the naked eye or a smartphone image are also progressing, such as using an aggregation-induced color change plasmonic biosensor

[0043] and smartphone nanoparticles on-chip sensor

[0044] , These studies reveal a trade-off between simplicity and sensitivity. The former provides easy sample mixing and visual changes in solution color within 5 minutes with limited sensitivity (around 300 ng / ml for N-Protein SARS-CoV-2 detection via the naked eye). The latter achieves better sensitivity (around 250 copies / ml) for various viral infections, including Zika and HBV (Hepatitis B Virus), but requires reactions on a microfluidic device and post-processing images

[0044] ,

[0006] In general, the sensitivity of a biosensor can be enhanced by pre-concentrating biomarkers for improved signal-to-noise ratio or, equivalently, a lower false-positive rate. For example, polymerase chain reaction (“PCR”) effectively achieves the pre-concentration for DNA sensing

[0045] ,

[0046] , but the detections of proteins and small molecules do not have such a process. The indirect use of PCR to amplify the fluorescence signal resulting from DNA- attached antibodies by specific proteins, known as quantitative immune-PCR

[0047] , represents a successful protein pre-concentration method. Another method, known as the single molecule array (“SiMoA”) captures proteins with magnetic particles to amplify fluorescence signals and achieve a low limits of detection (LODs) of 5 fg / ml for prostate-specific antigen as compared to 100 pg / ml with ELISA

[0048] , However, these pre-concentration methods require expensive instruments and reagents.

[0007] It would be desirable to have additional articles and methods that allow rapid detection of analytes diagnostic for diseases of interest with high sensitivity and low cost. Surprisingly, the present invention fulfills these and other needs.BRIEF SUMMARY OF THE INVENTION

[0008] In a first group of embodiments, the invention provides methods for indicating whether a sample contains a soluble analyte of interest. The methods comprising the following steps, in the following order: (a) obtaining a nanofibrous or mesofibrous membrane disposed on a solid support, (b) depositing a drop of an aqueous solution containing said sample on a first position on said nanofibrous or mesofibrous membrane, (c) allowing said drop of said sample to evaporate, thereby creating an area where said drop of said sample evaporated, said area having an interior portion and an edge around at least some of said interior portion, which edge bears residual material from said drop of said sample, (d) depositing at a second position on said nanofibrous or mesofibrous membrane, a drop containing gold nanoshells, which gold nanoshells have been functionalized to bind to said analyte of interest on said nanofibrous or mesofibrous membrane, which second position is disposed on said nanofibrous or mesofibrous membrane to allow (1) some of said drop containing said gold nanoshells to overlap with said edge bearing said residual material from said drop of said sample, thereby creating a first zone, and (2) some of said drop containing said gold nanoshells to overlap with some of said interior portion of said area, thereby creating a second zone, (e) allowing said gold nanoshells to interact with any of said analyte of interest present in said first detection zone and with themselves, thereby forming a first pattern in said first zone and a second pattern in said second zone, thereby forming a combination of said first pattern and said second pattern, whereby said combination of said first pattern and said second pattern indicates whether said analyte of interest is present in said sample or is not present in said sample. In some embodiments, a smooth pattern in said first zone indicates that said analyte is present in said sample. In some embodiments, a lumpy pattern in said first zone indicates that said analyte is not present in said sample. In some embodiments, the nanofibrous or mesofibrous membrane is intermediate between being hydrophilic and being hydrophobic and is wettable. In some embodiments, the nanofibrous or mesofibrous membrane has been rendered intermediate between being hydrophilic and being hydrophobic and is wettable, by being heated to 80 °C ±15 °C for a period of time and then cooled. In some embodiments, the nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, of cellulose nitrite, or of nylon. In some embodiments, the nanofibrous or mesofibrous membrane is of polytetrafluoroethylene. In some embodiments, the nanofibrous or mesofibrous membrane is nanofibrous. In some embodiments, the sample drop has a volume of 3-10 pL. In some embodiments, the drop containing said goldnanoshells has a volume of 1-5 pL. In some embodiments, the gold nanoshells are functionalized to bind said analyte of interest by being attached to an antibody that specifically binds said analyte of interest. In some embodiments, the analyte of interest is a protein. In some embodiments, the protein is a protein of a virus, a protein of a bacterium, a biomarker for a cancer, or a biomarker for a disease or a disease state. In some embodiments, step (c) is performed at a temperature of 80 °C ± 10 °C. In some embodiments, step (c) is performed at a temperature of 80 °C ± 5 °C. In some embodiments, determining whether there is a smooth pattern in the first zone is made by visual inspection. In some embodiments, determining whether or not said smooth pattern exists in the first zone is made by a device programmed to identify and compare a pattern in said first zone to a dataset of smooth patterns made by one or more analytes of interest bound by gold nanoshells functionalized to bind said one or more analytes of interest. In some embodiments, an amount of the one or more analytes of interest is quantified by determining an intensity level along a central line through said edge around at least some of said interior portion. In some embodiments, the dataset of smooth patterns made by one or more analytes of interest bound by gold nanoshells functionalized to bind said one or more analytes of interest is derived from training a convolutional neural network (CNN) using a plurality of test samples of known concentrations of the one or more analytes of interest bound by gold nanoshells and determining whether there is the smooth pattern in the first zone comprises processing an image including the first zone by the CNN to yield a positive result or a negative result for the smooth pattern. In some embodiments, the image yielding the positive result for the smooth pattern is further processed by a generator guided by a discriminator providing a conditional generative adversarial network (C-GAN) to segment the first zone to identify a central line through said edge around at least some of said interior portion and an intensity level along the central line is derived to quantify an amount of the one or more analytes of interest. In some embodiments, a length of the central line is normalized for all processed test images.

[0009] In a further group of embodiments, the invention provides methods for making a device for determining whether a sample contains an analyte of interest. The methods comprise (a) providing a solid support having a lower surface and an upper surface, (b) providing a thermally stable film with a lower surface and an adhesive upper surface, (c) adhering said upper surface of said solid support to said lower surface of said thermally stable film, (d) layering a hydrophilic nanofibrous or mesofibrous membrane onto adhesive uppersurface of said thermally stable film, thereby creating an assembled device, and (e) heating said assembled device to 80 °C ± 15 °C for a time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane to intermediate between hydrophilic and hydrophobic, and then allowing said nanofibrous or mesofibrous membrane it to cool to a temperature below 80 °C, thereby forming said device for determining whether a sample contains an analyte of interest. In some embodiments, the adhesive upper surface of said thermally stable film comprises silicon nanoparticles. In some embodiments, the method further comprises step (d)’ between steps (d) and (e): pressing said hydrophilic nanofibrous or mesofibrous membrane onto said adhesive upper surface of said thermally stable film after said hydrophilic nanofibrous or mesofibrous membrane has been layered onto said adhesive upper surface. In some embodiments, the hydrophilic nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, cellulose nitrite, or nylon. In some embodiments, the hydrophilic nanofibrous or mesofibrous membrane is of hydrophilic polytetrafluoroethylene. In some embodiments, the nanofibrous or mesofibrous membrane of hydrophilic polytetrafluoroethylene is nanofibrous. In some embodiments, the thermally stable film is of polyimide. In some embodiments, the heating of step (e) is to a temperature of 80 °C ± 10 °C. In some embodiments, the heating of step (e) is to a temperature of 80 °C ± 5 °C. In some embodiments, the heating of step (e) is to a temperature of 80 °C. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 2 minutes and 240 hours. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 5 minutes and 48 hours. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 10 minutes and 12 hours. In some embodiments, the said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 10 minutes and 4 hours. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 15 minutes and 2 hours. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 15 minutes and 1 hour. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is 30 minutes ± 10 minutes. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is 30 minutes ± 5 minutes. In someembodiments, the solid support is of glass. In some embodiments, the solid support is of a plastic that does not melt or soften when heated to 100 °C. In some embodiments, the plastic that does not melt or soften when heated to 100 °C is a thermosetting polymer or a thermoplastic plastic. In some embodiments, the thermosetting polymer or thermoplastic is a polyether ether ketone, polytetrafluoroethylene, polybenzimidazole, or polydicyclopentadiene. In some embodiments, the solid support is of metal. In some embodiments, the metal is stainless steel.

[0010] In yet a further group of embodiments, the invention provides devices for determining whether a sample contains an analyte of interest. The devices comprise, in the following order: (a) a solid support having an upper surface having an area, (b) a thermally stable film having (1) a lower surface adhered to said upper surface of said solid support, and (2) an upper surface, and (c) a nanofibrous or mesofibrous membrane adhered to said upper surface of said thermally stable film. In some embodiments, the thermally stable film is adhered to said upper surface of said solid support by an adhesive. In some embodiments, the thermally stable film is adhered to said upper surface of said solid support by double-sided tape. In some embodiments, the nanofibrous or mesofibrous membrane is adhered to said upper surface of said thermally stable film film by an adhesive. In some embodiments, the nanofibrous or mesofibrous membrane is adhered to said upper surface of said thermally stable film by double-sided tape. In some embodiments, the nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, cellulose nitrite, or nylon. In some embodiments, the nanofibrous or mesofibrous membrane is of polytetrafluoroethylene. In some embodiments, the nanofibrous or mesofibrous membrane is nanofibrous. In some embodiments, the thermally stable film is of polyimide. In some embodiments, the solid support is of glass. In some embodiments, the solid support is of a plastic that does not melt or soften when heated to 100 °C. In some embodiments, the plastic that does not melt or soften when heated to 100 °C is a thermosetting polymer or a thermoplastic plastic. In some embodiments, the thermosetting polymer or thermoplastic is a polyether ether ketone, polytetrafluoroethylene, polybenzimidazole, or polydicyclopentadiene. In some embodiments, the solid support is of metal. In some embodiments, the metal is stainless steel. In some embodiments, the nanofibrous or mesofibrous membrane is surrounded by a liquid-impermeable barrier. In some embodiments, the device further comprises at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample. In some embodiments, the device further comprises at least one mark, a physical guide, or both, showing where on saidmembrane to place a plasmonic drop. In some embodiments, (1) said solid support is a thermosetting polymer or a thermoplastic, (2) said thermally stable film is polyimide and is adhered to said upper surface of said solid support by an adhesive, (3) said nanofibrous or mesofibrous membrane is nanofibrous, is made of polytetrafluoroethylene, and is surrounded by a liquid-impermeable barrier, and (4) comprises at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample. In some embodiments, the device further comprises (5) at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop. In some embodiments, (1) said solid support is a thermosetting polymer or a thermoplastic, (2) said thermally stable film is polyimide and is adhered to said upper surface of said solid support by a double-sided tape, (3) said nanofibrous or mesofibrous membrane is nanofibrous, is made of polytetrafluoroethylene, and is surrounded by a liquid-impermeable barrier, and (4) comprises at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample. In some embodiments, the device further comprises (5) at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop.

[0011] In still a further group of embodiments, the invention provides methods for making a device for determining whether a sample contains an analyte of interest. The methods comprise (a) providing a solid support having a lower surface and an upper surface, (b) providing a thermally stable film with a lower surface and an adhesive upper surface, (c) adhering said upper surface of said solid support to said lower surface of said thermally stable film, and(d) layering a nanofibrous or mesofibrous membrane that has been treated to be wettable but to be intermediate between hydrophilic and hydrophobic, onto adhesive upper surface of said thermally stable film, thereby forming said device for determining whether a sample contains an analyte of interest. In some embodiments, the adhesive upper surface of said thermally stable film comprises silicon nanoparticles. In some embodiments, the method further comprises step (e): pressing said hydrophilic nanofibrous or mesofibrous membrane onto said adhesive upper surface of said thermally stable film after said nanofibrous or mesofibrous membrane has been layered onto said adhesive upper surface. In some embodiments, the nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, cellulose nitrite, or nylon. In some embodiments, the nanofibrous or mesofibrous membrane is of polytetrafluoroethylene. In some embodiments, the nanofibrous or mesofibrous polytetrafluoroethylene is nanofibrous. In some embodiments, the thermally stable film is of polyimide. In some embodiments, the treatment of said nanofibrous or mesofibrousmembrane is by heating said nanofibrous or mesofibrous membrane to 80 °C ± 15 °C for a time sufficient to change a hydrophilic nanofibrous or mesofibrous membrane from hydrophilic to intermediate between hydrophilic and hydrophobic or to hydrophobic, but wettable. In some embodiments, the heating is to 80 °C ± 10 °C. In some embodiments, the heating is to 80 °C ± 5 °C. In some embodiments, the heating is to 80 °C ± 2 °C. In some embodiments, the time period sufficient to change a hydrophilic nanofibrous or mesofibrous membrane to intermediate between hydrophilic and hydrophobic is between 2 minutes and 240 hours. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 10 minutes and 24 hours. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 10 minutes and 4 hours. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 15 minutes and 1 hour. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is 30 minutes ± 10 minutes. In some embodiments, the time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is 30 minutes ± 5 minutes. In some embodiments, the treatment of said nanofibrous or mesofibrous membrane is by chemical vapor deposition, by plasma treatment, by grafting hydrophobic polymers onto the nanofibers or mesofibers of the membrane, by spin-coating or dip coating, by a sol-gel coating, by depositing one or more self-assembled monolayers of hydrophobic molecules on said nanofibrous or mesofibrous membrane, by deposition of hydrophobic nanoparticles on said nanofibrous or mesofibrous membrane, by thermal or ultraviolet crosslinking of hydrophobic monomers on said nanofibrous or mesofibrous membrane, by depositing alternating layers of oppositely-charged hydrophobic polymers or polyelectrolytes on said nanofibrous or mesofibrous membrane, or by in situ polymerization of hydrophobic polymers on said nanofibrous or mesofibrous membrane. In some embodiments, the solid support is of glass. In some embodiments, the solid support is of a plastic that does not melt or soften when heated to 100 °C. In some embodiments, the plastic that does not melt or soften when heated to 100 °C is a thermosetting polymer or a thermoplastic plastic. In some embodiments, the thermosetting polymer or thermoplastic is a polyether ether ketone, polytetrafluoroethylene, polybenzimidazole, or polydicyclopentadiene. In some embodiments, the solid support is of metal. In some embodiments, the metal is of stainless steel.

[0012] In another group of embodiments, the invention provides devices for determining whether a sample contains an analyte of interest, said device comprising, in the following order: (a) a solid support having an upper surface having an area, (b) a thermally stable film having (1) a lower surface adhered to said upper surface of said solid support, and (2) an upper surface, and (c) a wettable nanofibrous or mesofibrous membrane that is intermediate between being hydrophilic and being hydrophobic, which wettable nanofibrous or mesofibrous membrane is disposed on said upper surface of said thermally stable film. In some embodiments, the thermally stable film is adhered to said upper surface of said solid support by an adhesive. In some embodiments, the thermally stable film is adhered to said upper surface of said solid support by double-sided tape. In some embodiments, the nanofibrous or mesofibrous membrane is adhered to said upper surface of said thermally stable film film by an adhesive. In some embodiments, the nanofibrous or mesofibrous membrane is adhered to said upper surface of said thermally stable film by double-sided tape. In some embodiments, the nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, cellulose nitrite, or nylon. In some embodiments, the nanofibrous or mesofibrous membrane is of polytetrafluoroethylene. In some embodiments, the nanofibrous or mesofibrous membrane is nanofibrous. In some embodiments, the thermally stable film is of polyimide. In some embodiments, the solid support is of glass. In some embodiments, the solid support is of a plastic that does not melt or soften when heated to 100 °C. In some embodiments, the plastic that does not melt or soften when heated to 100 °C is a thermosetting polymer or a thermoplastic plastic. In some embodiments, the thermosetting polymer or thermoplastic is a polyether ether ketone, polytetrafluoroethylene, polybenzimidazole, or polydicyclopentadiene In some embodiments, the solid support is of metal. In some embodiments, the nanofibrous or mesofibrous membrane is surrounded by a liquid-impermeable barrier. In some embodiments, the device further comprises at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample. In some embodiments, the device further comprises at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop. In some embodiments, the device further comprises a cover over said membrane. In some embodiments, the solid support is a thermosetting polymer or a thermoplastic, (2) said thermally stable film is polyimide and is adhered to said upper surface of said solid support by an adhesive, (3) said nanofibrous or mesofibrous membrane is nanofibrous, is made of polytetrafluoroethylene, and is surrounded by a liquid-impermeable barrier, and (4) comprises at least one mark, a physical guide, or both, showing where on said membrane toplace a drop of a sample. In some embodiments, the device further comprises (5) at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop. In some embodiments, the device further comprises a removable or slidable cover over said membrane. In some embodiments, (1) said solid support is a thermosetting polymer or a thermoplastic, (2) said thermally stable film is polyimide and is adhered to said upper surface of said solid support by a double-sided tape, (3) said nanofibrous or mesofibrous membrane is nanofibrous, is made of polytetrafluoroethylene, and is surrounded by a liquid- impermeable barrier, and (4) comprises at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample. In some embodiments, the device further comprises (5) at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop. In some embodiments, the device further comprises a removable or slidable cover over said membrane.BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figs. 1A-E. Fig. 1A. Fig. 1A is a cartoon depicting sequential evaporation of a sample drop which may or may not contain an analyte of interest (labeled “Analyte Drop Evaporation”) and a drop containing plasmonic nanaparticles deposited on opposite sides of a nanofibrous membrane, followed by a graphic depiction of imaging of the nanofibrous membrane and image analysis, with two squares at the bottom showing potential results of evaporating the two drops: the square on the left has just a “lumpy” pattern within the two circles, showing a negative result, and the square on the right showing both a smooth pattern and a lumpy pattern, showing a positive result. Fig. IB. Fig. IB is a cartoon showing how evaporation-induced flow within the sessile drops pushes nanoparticles towards the boundary of the drop as it evaporates, pre-concentrating the nanoparticles, and forming a “coffee-ring.” The first droplet evaporation deposits any analyte present in the sample drop on the membrane, while the second drop, containing plasmonic nanoparticles functionalized with an antibody that specifically binds the analyte of interest, permits visualizing the analyte (if present in the sample) through antibody-antigen interaction. Dark circles: N-protein. Lighter circles: antibody-coated gold nanoshells (“GNSh”). Fig. 1C. Fig. 1C presents, on the left, a large diagram showing an asymmetric plasmonic pattern consisting of two regions, referred to as “specific” and “non-specific” zones. Inside the common area between the coffee-ring created by the evaporation of the drop of the sample and the coffee-ring created by the evaporation of the drop containing the plasmonic nanoparticles, antibody conjugated GNShs are specifically interacting with already-coated proteins through antibody-antigeninteractions, generating dispersed, 2D-like pattern. Outside the specific zone of the plasmonic coffee-ring enclosed region, GNShs do not have a sufficient number of antigens to bind to and engage in self-interactions, leading to the generation of large 3D aggregates. This accumulation of large 3D aggregates is referrred to as a non-specific pattern. On the right side of the large drawing in Fig. 1C are two smaller cartoons. The top one of the two cartoons shows the small particles formed by the specific interaction of the protein of interest and antibodies on the GNShs. The bottom one of the two cartoons depicts larger, 3D aggregates formed by self-interactions of GNShs that have not bound the protein of interest. Fig. ID. Fig. ID shows the first step of the evaporation process on the nanofibrous membrane to illustrate how it affects the deposition and pattern of particles on the membrane. Fig. ID presents, on the left, a drawing, and on the right, a photograph, showing the “coffee-ring” formed by evaporation of a drop that may or may not contain an analyte of interest (such drops are sometimes referred to herein as “sample drops” or “sample droplets”). Fig. IE. Fig. IE presents, on the left, a drawing, and on the right, a photograph, showing the area of overlap of the “coffee-rings” formed by evaporation of (a) a drop or droplet that may or may not contain an analyte of interest and (b) a drop or droplet contain gold nanoshells that have been functionalized to specifically bind an analyte of interest (a “plasmonic drop” or “plasmonic droplet”). In this case, the sample drop has been placed to the right of center of the membrane (as shown in Fig. ID) and then allowed to evaporate, following which the plasmonic droplet was placed to the left of the center of the membrane and allowed to evaporate, with some of the plasmonic droplet flowing over the area in which the sample drop evaporated. The area in which the two droplets overlapped is labeled in Fig. IE as the “overlapping zone.” Fig. IE shows a smooth pattern in the overlapping zone and a rougher, “lumpy” pattern in the area in which the plasmonic drop dried, but in which it did not overlap with the sample drop.

[0014] Figs. 2A-2F. Fig. 2A. Fig. 2A is a depiction of the spreading that occurs when a sample or plasmonic nanoparticles drop is placed on a nanofibrous membrane and evaporates on the membrane. The drop spreads through and above the membrane (as shown by the arrows), pushing particles in the drop towards the outer boundary of the detection zone and non-specifically collecting them. Fig. 2B. Fig. 2B depicts how evaporation induces a flow inside the droplet that pushes particles in the drop towards the edge of the drop, thereby forming the so-called “coffee-ring” in fixed contact radius evaporation. Fig. 2C. Fig. 2C is a depiction of fixed contact angle evaporation. When the contact angle of the residual dropdecreases below a certain value (less than 5°), it starts to shrink down the contact radius while keeping the contact angle fixed. This step brings the rest of the particles towards the center of the residual drop and makes the stain-like pattern. Fig. 2D. Fig. 2D is a depiction of backward evaporation through the membrane. The rest of the solution inside the membrane evaporates at the end, and since the evaporation rate increases with the effective surface area, it starts to evaporate from the outer ring towards the center. As the majority of the particles have already settled in or on the membrane, this step has minimum effect on the particle deposition. Fig. 2E. Fig. 2E is presented in parts (i) and (ii). Fig. 2E(i) presents a confocal microscopy image of the deposition of fluorescent dyes during the evaporation of an exemplar sample drop. A high concentration of dyes is visible at the spreading boundary (outer ring) and the sample drop coffee-ring (inner ring). The area between outer and inner ring is depleted of particles, provides room to push these two rings towards each other and enhance pre-concentration factor, while letting small chemical compounds, such as salt molecules pass though the membrane and minimize crystallization at or near the coffee-ring. Fig. 2E(ii). Fig. 2E(ii) is a graph showing the profile of the central line passing across the rings. The higher gray index values correspond to higher concentrations. X axis: normalized axial distance. Y axis: normalized gray index value. The numbers along the top of the graph and vertical dashed lines correspond to the numbered positions in the microscopy image. Fig. 2F. Fig. 2F is presented in parts (i) and (ii). Fig. 2F(i) presents a diagram of the drop deposition pattern of a sample drop and a plasmonic drop. The grayscale image at the top shows different areas on the detection zone, including residual sample drop, sample spreading zone, GNShs residual drop, and GNShs spreading zone. By looking closely at the asymmetric pattern generated inside the area enclosed by the GNShs residual drop, one can see the clear pattern difference between specific and non-specific zones. Since the right side of the plasmonic residual drop coffee-ring is closer to the coffee-ring of the sample residual drop, the boundary is more intense than the left side, showing a non-uniformity in the coffee-ring pattern. Fig. 2F(ii) is a graph showing the profile of the central line passing through the plasmonic residual drop coffee-ring. The observed qualitative non-uniformity in the coffeering intensity and the gradient in the specific pattern is clear in the profile, providing information for further processing. Note, a higher gray index value means less particle has accumulated. The presented image is the detection zone of N-Protein at 1000 ng / ml concentration. X and Y axes are as described for Fig. 2E(ii).

[0015] Figs. 3A-E. Fig. 3A. Fig. 3 A depicts a work flow for making an embodiment of a device of the present invention. On the bottom left. Fig. 3 A shows, from the bottom up, a glass slide providing a solid support, an adhesive layer containing silicon nanoparticles, and a hydrophilic nanofibrous PTFE membrane, layered on top of one another. On the top right, a sample drop is depicted having been deposited on the nanofibrous membrane. The device is on a hot plate heating the device to 80 °C, causing the sample drop to evaporate, as depicted by the arrows facing outbound from the curved dome of the drop. Fig. 3B. Fig. 3B presents three fluorescent images from three separate experiments examining how heating exemplar membranes to different temperatures changes how droplets spread on the membranes. The images have been vertically stacked for ready comparison, with horizontal lines denoting where the bottom of one image ends and the top of the image below it begins. In each experiment, droplets were placed on a membrane heated to a selected temperature, with the images stacked in order, with the membrane heated to the lowest temperature of the three at the top and the membrane heated to the highest temperature of the three at the bottom, as shown by the arrow to the right side of the figure. As the temperature of the membrane increases, the coffee-ring of the sample drop (“CR”, the two vertical, dashed lines on the right-hand side of Fig. 3B) shrinks and moves away from the spreading line (“SL”, the vertical, dashed line on the left side of Fig. 3B). At too high a temperature, crystalline objects form at the coffee-ring, preventing GNShs in the plasmonic drop from accessing biomarkers in the sample drop (z.e., the SL and the CR lines coalesce at the same position). Fig. 3C. Fig. 3C presents a graph with an inset drawing. The graph shows the spreading rate (dashed line and circles) and the maximum spreading distance (solid line and diamonds) as the temperature at which the assay was performed using the exemplar devices was increased from 20 °C to 160 °C. The inset drawing is a representation of how liquid in a sample drop behaves when the drop is placed on a membrane. A portion of the liquid beads to form a drop sitting on the surface of the membrane (the portion of a sample drop sitting on top of the membrane is referred to herein as a “residual drop”) and is shown in the inset as a dark gray circle. Some of the liquid, however, moves into and through the membrane, and extends beyond the residual drop. The liquid that has moved beyond the edge of the residual drop is shown in the inset in light gray. The furthest distance (in mm) that the liquid has moved beyond the edge of the residual drop is referred to in Fig. 3C as the ARmax. X axis: temperature. Y axis: ARmax, in mm. The graph shows that increasing the temperature at which the assays were run above 80 °C decreases membrane wettability significantly, leading to less droplet spread through the membrane. At and above 120 °C, the droplet did not spreadthrough the membrane, presumably because the membrane is no longer porous. Fig. 3D. Fig. 3D is a graph showing the evaporation time (circles) and the contact angle (diamond shapes) of assays run at temperatures from 20 °C to 160 °C. The graph shows that increasing the temperature above 80 °C increases the contact angle significantly. Evaporation time increases with the treatment temperature, it is believed because more solution stays inside the residual drop (the portion of the drop remaining as the rest of the drop evaporates), leaving less area available for evaporation. Fig. 3E. Fig. 3E shows images showing the asymmetric patterns generated at various concentrations of an exemplar analyte, PSA. As seen in the images, the pattern in the area of overlap becomes less intense as the PSA concentration decreases from 1000 ng / ml to 10 pg / ml, which is the limit of detection (“LOD”) in this assay. Below the LOD, the specific and non-specific patterns are indistinguishable from the control sample (buffer solution without any analyte). The size of the common area is user-dependent and is not used for biosensing, however, the pattern intensity and gradient of the pattern, and higher order statistics, together with coffee-ring intensities are all features that can be used for discriminating between positive and negative results and for quantifying the concentration of analyte present in the sample.

[0016] Figs. 4A-D. Fig. 4A. Fig. 4A shows different aggregation patterns of GNShs inside and outside of the zone in which the sample drop and the plasmonic drop overlap, induced by specific versus non-specific interactions. The insets show scanning electron microscopy images of the area close to the sample residual drop coffee-ring. Plasmonic nanoparticles inside the overlapping zone generate dispersed 2D-like pattern due to their interaction with already coated proteins and making small aggregates. Outside the zone of specific interaction with analyte in the sample, GNShs interact with themselves because of insufficient or lack of deposited proteins, and form large aggregates. The images are of an exemplar analyte (the N- Protein of SARS-CoV-2) in the area shown, at a concentration of 1000 ng / ml. Fig. 4B. Fig. 4B compares the electric field distribution on the resonance frequency of a gold nanoparticle (GNP) having a diameter of 40 nm and of a gold-coated nanoshell (GNSh) having a diameter of 150 nm. The larger size and higher localization factor at the metal / dielectric boundary makes GNShs better candidates than GNPs to generate visible patterns, due to their higher optical absorption cross section. Fig. 4C. Fig. 4C shows a simplified periodic array analysis of GNSh aggregates. On resonance, field has been strongly localized in area between the nanoparticles. Fig. 4D. Fig. 5D is a graph of normalized absorption per atom (in a.u.) versus wavelength, showing the optical absorption spectrum per particle. Different spectra are seenfor specific aggregates (those binding analyte) versus non-specific aggregates (those between GNPs or between GNShs). The graph shows that GNShs provide higher effective absorption than do GNPs.

[0017] Figs. 5A-J. Fig. 5A. Fig. 5 A shows a CNN network, with VGG-16 architecture for detection zone classification. The network takes the grayscale image of the asymmetric pattern shown on the left and outputs the probability of a positive result compared to a negative result. Fig. 5B. Fig. 5B is a depiction of the generator and prediction portion of a C- GAN network for automatic specific zone segmentation. The generative and discriminator networks (shown in Fig. 5C) are training together to balance their performance and making the generative algorithm strong enough to provide manually segmented images. The network gets detection zone patterns as inputs and segments out the specific zone together with plasmonic residual drop coffee-ring area. Moreover, the generator removes the artifacts and noisy features, facilitating subsequent concentration quantification step. Note that this network is designed to make the sensing and quantification fully automatic. Fig. 5C. Fig. 5C shows the discriminator function and fine tuning aspects of a C-GAN network for automatic specific zone segmentation. Fig. 5D. Fig. 5D shows the comparison of the actual image to the prediction in a C-GAN network for automatic specific zone segmentation. Fig. 5E. Fig. 5E presents graphs showing the CNN output probabilities for four exemplar proteins of interest, N-protein, PCT, CEA, and PSA. The screening performance of the coffee-ring biosensor for the N-Protein, PCT, CEA and PSA proteins with the standard LOD of 50 pg / ml, 100 pg / ml, 750 pg / ml, and 10 pg / ml, respectively, compares to the probabilistic LOD of ~50 pg / ml, ~30 pg / ml, -650 pg / ml, and -3 pg / ml, respectively. Fig. 5F. Fig. 5F shows the FC regression network for concentration prediction from crossline profiles. The network gets crosslines around the central line as input and extracts features to build a non-linear relationship between the critical features (e.g., specific zone intensity, gradient and higher order statistics, and coffee-ring intensities) and sample concentration. Changing the concentration clearly varies these features, providing information for machine learning algorithms. The predicted overlapping area by C-GAN network provides necessary information to define the central line. Fig. 5G Fig. 5G presents graphs showing the predicted concentrations versus actual concentrations for four different biomarkers, N-Protein, PCT, CEA and PSA. The results show excellent predictions by the FC network for a point-of-care biosensor (less than half of order of magnitude error for a majority of the data). The variations at each point are induced by uncertainties in performing the tests and the proteinquality degradation over time. Fig. 5H. Fig. 5H shows a protocol for using exemplar coffeering biosensors to analyze human derived samples. Samples of pooled human saliva are taken by medical swabs, then mixed with samples spiked with N-Protein by inserting the swab into a solution spiked with N-protein. Sample drops of the resulting mixtures were applied at different concentrations to exemplar coffee-ring biosensors. Fig. 51. Fig. 51 is a graph of the results of studies following the protocol set forth in Fig. 5H. The screening performance of the coffee-ring biosensor for N-Protein detection in human saliva, comparing the standard LOD of 100 pg / ml and probabilistic LOD of 50 pg / ml, showing more than two orders of magnitude better sensitivity than the equivalent LFIA test. Fig. 5J. Fig. 5J is a graph of predicted concentrations (Y axis) versus actual concentrations (X axis) for N-Protein in human saliva. The results show that the predictions were very similar to the actual results.

[0018] Fig. 6, photos (i)-(iv). Fig. 6, photo (i). Fig. 6, photo (i) shows a membrane on which a sample drop containing PSA (the analyte of interest in this assay) at a concentration of 1000 ng / ml has been assayed with GNShs functionalized to bind PSA. The arrow coming from the smooth pattern shows specific detection of the PSA (the analyte of interest in this assay). Fig. 6, photo (ii). Fig. 6, photo (ii), shows the same assay but in which the sample drop also contained carcinoembryonic antigen (CEA) at a concentration of 1000 ng / ml. The overlap area again shows a smooth pattern which, as denoted by the arrow coming from the smooth pattern, shows specific detection of the PSA. Fig. 6, photo (iii). Fig. 6, photo (iii) shows a membrane on which a sample drop containing phosphate buffered saline (“PBS”), and CEA at a concentration of 1000 ng / ml has been tested with a plasmonic drop containing GNShs functionalized to bind PSA, but not CEA. There is no smooth pattern, indicating that the detection of a smooth pattern was specific for the detection of PSA (the analyte of interest in this assay) by the GNShs. Fig. 6, photo (iv). Fig. 6, photo (iv), shows a membrane on which a sample drop containing PBS, but no analyte, has been tested with a plasmonic drop containing GNShs functionalized to bind PSA. There is no smooth pattern in the detection zone, indicating that PSA, the analyte of interest, was not detected.

[0019] Figs. 7A-D. Fig. 7A. Fig. 7A shows an exemplar embodiment of an at-home multiplex coffee-ring biosensor kit. Fig. 7B. Fig. 7B is a diagram of a film heater made of patterned copper wires on a poliimide (“PI”) substrate, designed to heat the biosensor kit to 80°C at 12 V applied voltage. Fig. 7C. Fig. 7C is a diagram of a microtube to facilitate handling microliter volumes of the sample solution and the plasmonic solution. Exemplar tubes were cut into estimated lengths to provide sample and plasmonic drops of 5 pl and 2 pl,respectively, providing accurate yet simple control over the required volumes. The equation in the Figure allows choosing microtubes that will hold the volume desired when the user tilts the microtube and introduces it into a solution to be sampled or a solution containing the plasmonic drops. A 3D-printed flexible bulb attached to a pipette tip can be used to hold the microtube, enabling a droplet of the desired size to be released from the microtube when the bulb is squeezed. Fig. 7D. Fig. 7D is a diagram showing in the inset a 3D-printed guide piece for precise control over the placement of the sample droplet and the plasmonic droplet, ensuring a consistent gap between them for higher repeatability. Open windows in the guide piece allow accurate positioning of the guide piece relative to the substrate.

[0020] Figs. 8A-D. Fig. 8A. Fig. 8A presents a photograph and a diagram. The photograph shows the asymmetric coffee-ring pattern formed testing for the N-protein of SARS-CoV-2 at a 100 ng / ml concentration when the drops are applied at a gap distance of 1.66 mm. The “Overlapping Region” diagram shows the area in which the two drops overlap when applied at that gap distance. Fig. 8B. Fig. 8B presents a photograph and a diagram. The photograph shows the asymmetric coffee-ring pattern formed testing for the N-protein of SARS-CoV-2 at a 100 ng / ml concentration when the drops are applied at a gap distance of 0.60 mm. The “Overlapping Region” diagram shows the area in which the two drops overlap when applied at that gap distance. A comparison with Fig. 8A shows that the Overlapping Region is much larger when the gap distance is smaller. Fig. 8C. Fig. 8C is a graph of the central crossline gray index pattern for the photograph of the coffee-ring pattern formed by the drops applied at the 1.66 mm gap distance. Fig. 8D. Fig. 8D is a graph of the central crossline gray index pattern for the photograph of the coffee-ring pattern formed by the drops applied at the 0.60 mm gap distance. A comparison of Fig. 8C and Fig. 8D shows distinct characteristics due to the difference in gap distance. The variations in the crossline patterns highlight the importance of maintaining consistent droplet placement locations and gap distances for repeatable results.DETAILED DESCRIPTIONIntroduction

[0021] As noted in the Background, the lack of inexpensive, rapid methods of detecting biomarkers for various conditions with high sensitivity slows the diagnosis of diseases, with adverse consequences for patients. In 2021, the present inventors disclosed methods anddevices at the 2021 21st International Conference on Solid-State Sensors, Actuators and Microsystems (Transducers) | 978-1-6654-1267-4 / 21, pages 381-384; https: / / ieeexplore.ieee.org / document / 9495602, that were capable of detecting analytes with a limit of detection of 5 ng / mL, similar to that of commonly-used lateral flow immunoassays (“LFIAs”). The methods and devices discussed in the 2021 disclosure are sometimes referred to herein as the “2021 methods” and the “2021 devices.”

[0022] Surprisingly, improvements to the 2021 devices and methods have been discovered that increase the sensitivity of the devices and methods by 500 fold, lowering the limit of detection from the 5 nanogram per mL sensitivity of the 2021 versions to as low as 3 picograms per mL. This remarkable increase in sensitivity (over two orders of magnitude higher than conventional LFIAs) provides inexpensive, rapid, and sensitive methods for detecting biomarkers that are far more sensitive than the art-standard LFIAs, and provides inexpensive articles of manufacture to perform the methods with a sensitivity never before achieved by a simple biosensor that can be used both in non-medical settings, such as by the public at home, in addition to settings such as drugstore pharmacies, urgent care centers, and emergency rooms. The discussion below describes the methods developed in 2021, and then the improvements that together increase the sensitivity by 500 times over the 2021 embodiments.

[0023] In 2021, the present inventors disclosed devices and methods exploiting the properties of the natural mechanism by which an evaporating sessile drop creates a pattern sometimes referred to as the “coffee-ring effect.” The coffee-ring effect occurs because conductive and other forces within an evaporating drop tend to impel or otherwise drive substances within the drop to the edge of the drop, concentrating them along the edge to form a so-called “coffee ring” where the edge of the drop was before it evaporated away. When the drop is of a sample containing an analyte of interest, the coffee-ring effect concentrates the analyte of interest at the edge of the drop as it evaporates. The 2021 inventive devices and methods exploited this tendency by placing a drop containing (or not containing) an analyte of interest on a thin but porous nanomembrane. (A drop being tested to see if it does or does not contain an analyte of interest is sometimes referred to herein as “the sample drop”. The terms “drop” and “droplet” are generally used interchangably herein unless otherwise noted or required by context.) The sample drop was allowed to evaporate, which drove to the edge of the drop any particles and analytes in the drop other than water, thereby concentrating the particles and analytes compared to their average distribution in the sample drop prior to the sample drophaving been evaporated on the nanomembrane surface. A second drop, this one containing 40 nanometer gold nanoparticles (“GNPs”) functionalized by covalently attaching to the GNPs antibodies which specifically bind the analyte of interest if they come into contact with it, was then deposited on the area on which the sample drop dried on the nanomembrane, creating a plasmonic resonance pattern (the second drop, which is functionalized to bind to an analyte of interest, is sometimes to herein as the “plasmonic drop ”) The 2021 results showed that this method allowed detection of an exemplar analyte down to a limit of detection of 5 nanograms per mL.

[0024] While the 2021 methods and devices had a sensitivity about equivalent to art-standard LFIAs, there were also problems that made the 2021 methods and devices less than ideal. Notably, they were not readily adapable for use by members of the public to perform tests in their homes, something of particular concern in 2021 as the Covid- 19 pandemic made at- home testing for SARS-CoV-2 a public health priority.

[0025] The plasmonic resonance pattern produced by the 2021 methods showing whether the analyte was or was not present was “dirty.” As noted, the coffee-ring effect drives any analyte in the sample drop to the edge of the sample drop as the sample drop dries. When the plasmonic drop is placed over the dried sample drop spreads and dries, the GNPs contact the coffee-ring containing the analyte over a limited area, creating patterns that can obscure whether the result was positive or negative.

[0026] The present disclosure improves the ability to obtain positive or negative results using the coffee-drop effect by at least two orders of magntitude compared to the methods and devices disclosed in 2021. The present disclosure changes the previous results and methods in several ways. Together, these changes to the previously disclosed methods and devices allow the detection of exemplar analytes down to a limit of detection as low as 3 picograms per mL, an increase in sensitivity of over 500 times compared to the 5 nanogram limit of detection provided by the 2021 methods.

[0027] First, the inventive methods change the approach used in the 2021 methods. Instead of placing the sample drop on the membrane, allowing it to dry, and then placing the plasmonic drop where the sample drop was placed, as in the 2021 methods, in the inventive methods, the sample drop is placed on the membrane off to one side of the membrane, allowed to dry, and the plasmonic drop is then added off to the other side of the membrane, but close enough to the first drop so that, as the plasmonic drop spreads, it spreads over theedge of the “coffee-ring” left by now-dried sample drop. Studies found that this offset of the two drops creates two patterns, which together make it significantly easier to detect if the analyte of interest (the one specifically bound by the antibody or other capture moiety attached to the gold nanoparticles) is present: (a) a smooth area of small aggregates where the plasmonic drop overlapped with the dried sample drop, and plasmonic particles interacted with the analyte of interest, if it is present in the sample, and (b) a rougher pattern, in areas in which the plasmonic drop overlapped with the dried sample drop and there was no analyte in the sample drop which the nanoparticles in the plasmonic drop were functionalized to bind, in which case the functionalized nanoparticles interact with each other rather than the (absent) analyte of interest. The inventive devices exploit this improvement by having the user place the sample drop and the plasmonic drop on the membrane at separate positions that are offset from one another. Conveniently, the device has markings to denote where the user should place the sample drop, the plasmonic drop, or both, such that when the user places the drops on the membrane, the desired overlap of the two drops results. In some embodiment, the device may have a guide piece disposed over some or all of the top of the device with holes (“windows”) through which the drops can be placed on the membrane, to guide the user as to where to place the sample drop, the plasmonic drop, or both, on the membrane. The windows can be marked to indicate which drop is to go in which window. The guide piece can be attached to the device or can be positioned over a device that does not have an integral guide piece at the time of use.

[0028] Second, the 2021 methods and devices used 40 nanometer gold nanoparticles. Studies to improve the patterns provided by those methods revealed that switching to nanoshells of silica coated with a thin layer of gold (the exemplar nanoshells are 140 nm silica shells coated with a layer of gold, for a total size of 150 nm), provided a stronger signal due not just to the larger size of the shells, but also due to a difference in the plasmonic resonance and optical properties of a 40 nm gold nanoparticle and gold disposed over the surface of a 140 nm sphere. The use of the gold nanoshells (sometimes referred to herein as “GNShs”) resulted in a darker pattern that makes it easier to determine if the results of an assay conducted following the inventive methods are positive or negative.

[0029] A third improvement resulted from an accidental finding. The nanofibrous membranes used in the 2021 methods and devices are hydrophilic. The membranes are held to a thermally stable film by an adhesive typically disposed on the thermally stable film. A prototype device was accidentally left on a hot plate and, after it was found, was tested. Thethermally -treated device proved to provide much better readability of the results than did non-thermally treated devices. Investigation to uncover the basis for this improvement revealed that heating the membrane in the presence of the adhesive had changed the membrane’s properties from hydrophilic to somewhere between hydrophilic and hydrophobic. Without wishing to be bound by theory, it is believed that this intermediate property is responsible for the improved readability of the inventive devices. (Thermal treatment of membranes and other methods for changing the hydrophilicity of the membrane to a degree where liquid can still pass through it but the contact angle of the residual droplet is close to the hydrophobic value are discussed further below).

[0030] Together, these improvements make the methods and devices of the present invention surprisingly more sensitive than the 2021 methods and devices. Some embodiments of these surprisingly more sensitive devices are suitable for use by pharmacy staff and staff in medical offices. In some embodiments, these surprisingly more sensitive devices are suitable for use by members of the public wishing to use an at-home test to determine whether they have a particular condition or have contracted a particular disease. In preferred embodiments, the condition or disease is one which can be detected or diagnosed by the presence of a protein or other analyte of interest that can be dissolved in a water droplet.

[0031] To help clarify embodiments of the inventive methods and devices, certain features will now be discussed. An analyte the GNShs have been functionalized to detect is sometimes referred to herein as an “analyte of interest.” Many of the analytes of interest are proteins and such a protein will sometimes be referred to more specifically as a “protein of interest.” (The term “analyte of interest” can, however, refer to a protein unless otherwise indicated in the text or required by context.)

[0032] As noted above, the sample drop is placed on the surface of the nanofibrous membrane offset to one side, while the plasmonic drop is placed on the surface offset from where the sample drop was placed, but close enough so that a portion of the plasmonic drop can overlap with the area of the coffee-ring from the sample drop. The interaction of the functionalized GNShs with the coffee-ring from the sample drop in the area in which the two drops overlap generates patterns that differ depending on whether the analyte of interest (a) was not present in the sample drop, (b) was present in the sample drop in high concentration, or (c) was present in the sample drop at a level close to the limit of detection of the device.

[0033] In some important embodiments, the devices provide results that can be read by eye. This is true both in cases in which the concentration of the analyte in the sample drop is well above the limit of detection of the device, and in cases in which the analyte is not present in the sample drop. An analyte present in an amount an order of magnitude lower than that detectable by a lateral flow assay will still be present in an amount an order of magnitude above the limit of detection of the inventive devices, and thus easily detectable by eye using the inventive devices. Thus, the inventive methods and devices are expected to provide home users with more sensitive ability to detect, for example, whether a member of their household has contracted covid- 19 or another condition whose presence can be detected by an analytes present in a bodily fluid than is available from currently available LIFA-based home test kits. Bodily fluids that can be used for detection of analytes include, for example, saliva, nasal secretions, urine, and droplets of blood.

[0034] In embodiments in which the analyte is present close to the limit of detection of the device, the presence or absence of the analyte, and the concentration, the asymmetric plasmonic patterns can be further processed using deep neural networks for concentration quantification. In some embodiments, these analyses can be performed on an app downloaded onto a smartphone. In some embodiments, the analyses can be performed on a program installed on a laptop or a desktop computer.

[0035] Exemplar embodiments of the inventive devices were tested on multiple exemplar proteins for the ability to detect them as analytes of interest, including the N-protein of SARS-CoV-2, the sepsis biomarker procalcitonin (sometimes referred to as “PCT”), the prostate cancer biomarker prostate specific antigen (sometimes referred to as “PSA”), and the cancer marker carcinoembryonic antigen (sometimes referred to as “CEA”). Each of these proteins of interest was detected by the exemplar devices.

[0036] While the coffee-ring effect has sometimes been employed in previous protein detection methods using fluorescent imaging or magnetic particle trapping, the signals generated by these sensors have been challenging to analyze or have required complex instruments for signal capture

[0049] —

[0055] .

[0037] Surprisingly, the asymmetric plasmonic patterns created by concentrating analytes in the coffee-ring left by an evaporated drop of a sample on a thin, nanofibrous membrane, and overlapping them with plasmonic gold nanoshells functionalized to bind to the analytes, generates specific visual signals.

[0038] In some embodiments, inventive devices combining these features provide binary (yes-or-no) results that the user can identify by eye. These embodiments should be particularly useful in providing home test kits to determine if an individual has contracted a disease, such as covid-19. Home test kits based on embodiments of the invention are expected to produce more sensitive and more reliable detection of covid-19 infection than is provided by current LIFA-based covid- 19 tests. The unique combination of the coffee-ring effect pre-concentration mechanism and the introduced asymmetric plasmonic patterns enhances the 2021 coffee-ring biosensing methods, which had a detection sensitivity similar to that of LFIA, by two-to-three orders of magnitude

[0056] ,

[0039] Further surprisingly, in some embodiments, the asymmetric plasmonic patterns can be detected and analyzed by scanning the patterns with readily available camera technology connected to a coded processor with access to a database trained on data of such patterns. Such computer-assisted embodiments permit detecting even lower concentrations of the analytes than can be detected by the unassisted eye and, in some embodiments allow quantitating the amount of analyte present using deep neural network analyses on a smartphone image. It is expected that such analyses can be conducted on servers accessed over the internet (commonly referred to as “the cloud”), or on a user’s smartphone.

[0040] In some embodiments, the inventive devices may be intended to detect if a single, particular analyte of interest is present or absent in the sample. In these embodiments, the gold nanoshells deposited on the surface in the plasmonic drop are all functionalized so that they will selectively bind to that analyte of interest. In some embodiments, the inventive devices may be intended to detect whether two or more selected analytes of interest are present in the sample or if none of the two or more analytes of interest are present. In these embodiments, some of the gold nanoshells deposited on the surface are functionalized to bind the first analyte of interest, others are functionalized to bind to the second analyte of interest, and so on, for each of the analytes whose presence the device is intended to detect.

[0041] In preferred embodiments, the analyte or analytes of interest are proteins. In some embodiments, at least one protein of interest is N-protein of SARS-CoV-2. In some embodiments, at least one protein of interest is procalcitonin. In some embodiments, at least one protein of interest is prostate specific antigen. In some embodiments, at least one protein of interest is carcinoembryonic antigen. While covid- 19 is one example of a disease for which home test kits are now in common use, it is anticipated that embodiments of the inventive devices and methods can be used to detect any disease or condition which can bedetected by the presence of soluble proteins or other soluble biomarkers and for which an antibody or other chemical moiety which specifically binds the soluble protein or soluble biomarker can be attached to a gold nanoshell. Kits intended for home use will typically be intended to detect the soluble protein or other soluble biomarker in a bodily fluid that is accessible to persons without medical training, such as the nasal secretions swabbed in home kits testing for covid, saliva, tears, or urine. Such kits can also be provided with an alcohol wipe to sterilize skin, and a small lancet to allow the user to obtain and test a drop of blood. Kits for use in medical offices, urgent care centers, and the like will, of course, already have the items needed to obtain blood drops for testing.Membranes

[0042] The devices of the invention provide a solid surface having a mesofibrous or a nanofibrous membrane on top. Nanofibers are fibers with diameters in the nanometer range, which can be used to form highly porous meshes

[0088] , while mesofibers are fibers with diameters in the micron range. In preferred embodiments, the membranes are nanofibrous.

[0043] In typical embodiments, the amount of liquid in the sample drop and the plasmonic drop is quite small, typically less than 10 pL (larger volumes can be used, but usually require more time to evaporate and thus increase the time before a result is available). Given the small volumes of fluids, the assay can be conducted in a small area. For convenience of handling, devices intended for use by consumers will typically be sized larger than the area in which the assay itself is conducted. In some embodiments, the devices are sized about the size of an over-the-counter CO VID-19 test kit.

[0044] Preferably, the membrane is slightly porous and is wettable. In the context of this disclosure, a “wettable” membrane is one that allows liquid to pass through. In some embodiments, the membrane is hydrophilic, which permits part of the sample drop to move through the membrane, but it is preferably intermediate between hydrophilic and hydrophobic. The porosity and thickness of the membrane is chosen so that most of the drop remains of the surface of the membrane in contact with air, permitting water in the drop to evaporate; use of a porous membrane is preferred as it increases relative surface area of the drop available to evaporate.

[0045] In preferred embodiments, the porous membrane is a nanofibrous membrane. The very small interstices between the nanofibers of such membranes provides a high surface to volume ratio and more surface for particles to attach. Nanofibous membranes are particularlyuseful, as the fibers retain particles as the coffee-ring forms. Water penetrates the small pores, but at a small flow rate due to the small pore size. This leaves a large residual drop (the portion of the original drop that remains on top of the membrane rather than penetrating into the pores). The drop evaporates from both the residue drop on top of the membrane, and from the membrane.

[0046] The membranes used in the studies were circular, with a diameter of 4.5 mm. In some embodiments, the top of the membrane or the whole of the membrane, is surrounded, in whole or in part, by a water-impermeable substance to define an area in which the liquid of any drops placed on the membrane will be confined. While membranes 3-7 mm are preferred, and preferably 4-6 mm, for use with sample drops of 5-8 pl, larger membranes can be used if the impermeable substance confines the drops within an area of the stated area.Alternatively, a larger membrane can be used with a larger sample drop, such as 10 pl or 15 pl. As noted above, however, larger drops will take longer to dry.

[0047] The membranes used in studies underlying the disclosure used membranes from Direct Detect® Assay-free Cards (Millipore Sigma, Burlington, MA, item DDAC00010-GR), which come with hydrophilic polytetrafluoroethylene filters surrounded by a hydrophobic barrier. (Polytetrafluoroethylene is often abbreviated as “PTFE.” PTFE membranes for use in the inventive methods and devices herein refers to hydrophilic PTFE, the hydrophilicity of which is preferably modified before use, as discussed herein.)

[0048] In an accidental discovery made in the course of studies underlying the present disclosure, a prototype device bearing a hydrophilic PTFE membrane was left on a hot plate. An assay using the device with the membrane that has been heated revealed it provided much better readability of results of an embodiment of the inventive assays than did membranes which had not been heated. The membrane was resting on a thermally stable film which contained silicon nanoparticles. Without wishing to be bound by theory, it is believed that silicon nanoparticles from the adhesive migrated into the hydrophilic membrane, changing its properties to be intermediate between hydrophilic and hydrophobic, and that it is that change in properties that improved the ability to determine the results of the assay. In preferred embodiments, the membrane is on a film bearing silicon nanoparticles before the membrane is heated. Temperatures for heating the devices are discussed in a later section. In some embodiments, the membrane is heated once the device has been assembled. In some embodiments in which the hydrophilicity of the membrane is to be altered by heating, themembrane is first heated on an adhesive containing silicon nanoparticles and then used to assemble into a device.

[0049] Membranes with any particular combination of porosity and thickness can be readily tested for their suitability to serve as a surface for use in different embodiments of the inventive devices by taking equally sized pieces of the membrane or surface being evaluated for suitability (the “test membrane or surface”) and of a nanofibrous PTFE membrane, placing the membranes on identical solid surfaces, and running assays in parallel using equal aliquots of a sample drop containing an exemplar protein of interest in a concentration of 50 pg / mL to 1000 pg / mL (that is, a concentration expected to be readily detectable by eye using a membrane in the inventive methods) and following the methods set forth in the Examples. Test membranes or surfaces that provide a visual pattern in the area of overlap of the sample drop and of the plasmonic drop that is approximately equal to that of the visual pattern in the area of overlap of the sample drop and of the plasmonic drop using the PTFE membrane, as judged subjectively by an unbiased observer are considered suitable for use in the inventive devices and methods, while test membranes or surfaces that provide a visual pattern in the area of overlap of the sample drop and of the plasmonic drop that are hard to distinguish compared to the visual pattern in the area of overlap of the sample drop and of the plasmonic drop using the PTFE membrane, as judged subjectively by the same unbiased observer are considered not to be suitable for use in the inventive devices and methods.

[0050] It is expected that a variety of materials can form suitable surfaces for making the inventive devices. Such materials include cellulose nitrite, nylon, and PTFE. Conveniently, thin membranes of PTFE are commercially available and is made of nanofibers. Further, it is transparent to infrared light. Transparency to infrared light allows the use of spectrographic methods for practitioners wishing to use a second method to confirm the presence of high amounts of protein. PTFE is a preferred allowing it to be used with sprat a preferred material for making some embodiments of the devices provided by the present disclosure.Methods Other Than Heating for Changing the Hydrophilicity and Wettability of the Membrane

[0051] As noted in the preceding section, it was found that heating an exemplar hydrophilic nanofibrous membrane in the presence of an adhesive containing silicone nanoparticles caused silicone nanoparticles to migrate into the membrane, changing it to a more hydrophobic surface, while having good wettability, as graphically depicted in Fig. 3 A. Fig.3C shows how the ARmax changes as a function of the membrane temperature. In preferred embodiments, the hydrophobicity and wettability of the membrane is tuned to result in an ARmax between and including 0.4 mm and 0.6 mm.

[0052] In preferred embodiments, it is contemplated that the size of the membrane will be selected so that it will be approximately the size through which the volume of the sample drop will spread. The membrane is preferably surrounded by an impermeable barrier so that the contents of the sample drop are retained within an area in which the plasmonic drop can interact with the sample drop. The membranes used in the exemplar devices were 4.5 mm squares. The results as to spreading rate can be normalized to the size of the detection zone

[0053] The hydrophilicity and wettability of the membranes of the exemplar devices were changed by thermal treatment with the membranes (which were originally hydrophilic as obtained from the manufacturer) resting on an adhesive containing silicone nanoparticles, which migrated into the membrane and made it more hydrophobic, as graphically depicted in Fig. 3 A. Thermal treatment is not, however, the only way to change the properties of the membrane to increase the readability of assays using embodiments of the inventive devices. It is believed that changing the membrane from being hydrophilic to being hydrophobic, but having a wettable surface will similarly improve sensitivity in detecting analytes of interest.

[0054] A number of surface modification techniques are known that can be employed to alter the surface energy of a membrane while maintaining some level of wettability. By carefully selecting the technique and materials, it is expected that the practitioner can achieve the desired balance between hydrophobicity and wettability, ensuring the membrane remains functional for use in different embodiments of the inventive devices. As noted above, the hydrophobicity and wettability of the membrane is preferably tuned to result in an ARmax between and including 0.4 mm and 0.6 mm.

[0055] A brief summary of a number of techniques that can be used is set forth below. As each of the techniques is well known in the art, it is expected that persons of skill in material science need only the explanation that these methods can be used to tune the hydrophobicity and wettability of the membrane to result in an ARmax between and including 0.4 mm and 0.6 mm to allow them to use these art-recognized techniques to change the properties of membranes that are provided as hydrophilic membranes to membranes that have a hydrophobic surface but that are wettable. Measuring and changing wettability of surfaces is well known in the art, as exemplified by, e.g., reference

[0087] ,

[0056] In some embodiments, chemical vapor deposition, or “CVD,” can be used to confer the desired characteristics to the membranes. In these embodiments, a thin hydrophobic coating is deposited on the membrane. In some of these embodiments, the hydrophobic coating is of a silane-based compound, such as trimethylchlorosilane or a perfluorosilane. This technique allows precise control of surface chemistry, introducing hydrophobic functional groups (e.g., -CH3, -CF3) while maintaining wettability due to nanostructures or surface roughness.

[0057] Alternatively, in some embodiments, the membrane can be subjected to plasma treatment, in which plasma polymerization is used to deposit a hydrophobic coating, such as a fluorocarbon-based polymer, on the membrane. Plasma-induced etching or modification can create a dual-scale roughness that provides a hydrophobic yet wettable surface.

[0058] Further, in some embodiments, hydrophobic polymers are chemically grafted to the membrane. For example, the membrane can be subjected to graft polymerization with hydrophobic polymers such as polydimethylsiloxane or fluoropolymers. UV-induced polymerization or initiator-based surface grafting can be used to ensure uniform hydrophobicity.

[0059] Yet another technique known in the art to alter the hydrophilicity of the membrane to provide a hydrophobic, but wettable membrane is the application of a thin layer of hydrophobic but water-permeable polymer, such as poly(vinylidene fluoride). Controlled coating of the thickness can retain the underlying membrane’s wettability.

[0060] Another method known in the art is the so-called sol-gel process. Sol-gel refers to a process in which nanoparticles in a liquid (a sol) form a continuous three-dimensional network extending throughout the liquid (a gel). With regard to membranes for use in embodiments of the inventive devices, the sol-gel process can be used to coat the surface of the membrane with a sol-gel layer containing hydrophobic moieties like fluorosilanes. This method can provide hydrophobicity while ensuring fine control over the porosity and wettability.

[0061] Yet another method known in the art involves the use of self-assembled monolayers, or “SAMs.” In this method, self-assembled monolayers of hydrophobic molecules (e.g., long-chain alkanethiols or silanes) are deposited on the membrane. SAMs can create hydrophobic surfaces with nanometric precision, retaining wettability depending on the chosen molecular structure.

[0062] Still another method is nanostructuring. Using techniques like laser ablation, etching, or nanoparticle deposition, mi cro / nano- scale roughness can be introduced in combination with low-surface-energy coatings. The surface structure promotes a Cassie-Baxter state, where water partially wets the surface.

[0063] Thermal or UV crosslinking can also be used to adjust the hydrophilicity or hydrophobicity of the membranes to have the degree of hydrophobicity and wettability desired by the practitioner. In this methods, hydrophobic monomers or oligomers are crosslinked on the membrane surface using heat or UV light. Examples include crosslinked polyethylene glycol derivatives with hydrophobic end groups.

[0064] Similarly, “Layer-by-Layer” (“LbL”) assembly can be used. In these embodiments, alternating layers of oppositely charged hydrophobic polymers or polyelectrolytes are deposited onto the membrane. Hydrophobicity and wettability can be tuned by adjusting the number and type of layers.

[0065] In Situ Polymerization is performed by initiating hydrophobic polymer formation directly on the surface, such as by polymerizing fluorinated monomers or acrylates. This method allows hydrophobicity and wettability to be tuned by adjusting the number and type of layers.Solid Surfaces as Supports for the Membrane

[0066] The membrane is typically positioned on a solid, impermeable surface, such as a plastic or glass rectangle, circle, or square. As explained above, the size of the overall device is larger than the small area in which the test itself is conducted (for ease of reference, the area in which the sample drop and the plasmonic drop are placed and in which they can then interact is sometimes referred to herein as the “test area”). To keep the concentration of the analyte and of the functionalized nanoshells high in the test area, in some embodiments, an impermeable barrier is positioned around the test area to keep the fluid of the sample drop, the plasmonic drop, or both, from diffusing outside of the test area. Hydrophobic or other impermeable barriers suited for use as such barriers are known, as exemplified by the hydrophobic barrier around the PTFE membranes on Millipore Sigma Direct Detect® Assay- free Cards. In some embodiments, the barrier can be a silicon-based ink. See, e.g., Rajendra, el al., Analyst, 2014, 139(24):6361-6365; doi: 10.1039 / C4AN01626B.

[0067] In other embodiments, the membrane placed on the solid support is is smaller than the solid support on which it is placed and because the volume of fluid comprising the sampledrop placed on the membrane is so low, the surface tension of the fluid prevents it from flowing from the membrane onto the solid support.

[0068] Exemplar devices used in studies underlying the present disclosure were handmade from the bottom up. A workflow for making the examplar devices is shown in Fig. 3 A. Starting with a glass slide (which was the solid support in these exemplar devices), double sided tape was placed on the slide. A one-sided polyimide (“PI”) film was then placed on the double-sided tape with the adhesive side of the PI film up, and a thin hydrophilic PTFE membrane was then laid over the PI film. As explained in the Examples, PI film has good thermal stability and provides a robust substrate for the membrane during thermal treatment. Other polymeric films that have an adhesive and that are resistant to temperatures up to 100 °C are also expected to be suitable. For example, polyethylene terephthalate (“PET”) film bearing an adhesive is expected to work.

[0069] Studies underlying the present disclosure found that heating the device improved the wettability and contact angle of the membrane. Exemplar devices were made by pressing the membrane firmly onto the PI film to provide good attachment of the membrane to the PI film, and then heating the device for 30 minutes at 80 °C (heating of the devices is sometimes referred to herein as “thermal treatment.) Through this process, silicon nanoparticles from the adhesive penetrate the membrane, attaching to its nanofibers, increasing their size, and thus controlling the porosity of the membrane. As shown in Fig. 3C, heating the device changes these characteristics. For the combination of a device comprising a PTFE membrane over a PI film, studies showed that temperatures between 60 °C and 100 °C can be used, with temperatures between 70 °C and 90 °C being preferred, and 80 °C ± 5 °C, 80 °C ± 3 °C, 80 °C ± 2 °C, and 80 °C ± 1 °C being more preferred, with each stated range of temperature being preferred to any stated before it. While 80 °C ± 1 °C is a particularly preferred temperature for a device bearing a PTFE membrane on a PI film, it is anticipated that devices can be made using other hydrophilic nanomembranes (such as cellulose nitrite or nylon) over a PI film, or with a hydrophilic membrane of PTFE, cellulose nitrite or nylon, over a film other than PI. In some preferred embodiments, an adhesive containing silicon nanoparticles is used to adhere the membrane to the thermally stable film.

[0070] A person of skill wishing to make a device from any particular combination of membrane and film can readily do so using the method explained in the descriptions of Fig. 3C and Fig. 3D and determine which temperatures provide the best wettability and contact angle for the membrane. Whichever combination of membrane and film is chosen, andwhatever temperature is selected by the practitioner, the device then is heated to the chosen temperature for 5 to 120 minutes, for 10 to 90 minutes, for 15 to 60 minutes, for 20 to 40 minutes, or for 30 ± 5 minutes, with each successive range of time being preferred to any previously stated range of time.

[0071] The membrane may further have a water-impermeable material disposed around some or all of the membrane to keep the drop on the surface of the membrane within its confines. Suitable water-impermeable materials are known in the art. As the inventive devices are made using a thermal treatment, typically 30 minutes at 80 °C, and preferably heated again to 80 °C in advance of the sample drop being placed on the membrane, the impermeable material selected should be one that can be heated to at least 80 °C without melting or losing integrity as a water-impermeable barrier.Sample Drops, Plasmonic Drops, and Placement of the Drops on the Membrane

[0072] In some embodiments, an inventive device can be used by the end-user to determine whether they, or a family member, are suffering from a condition which can be diagnosed from a bodily fluid, such as saliva (human saliva was used to verify the detection of analytes in some studies underlying the disclosure herein). For saliva, a drop of the subject’s saliva as the sample drop can be placed directly on the device. In cases in which the bodily fluid may be too viscous to be applied as a drop, a small amount of a saline solution, such as phosphate buffered saline, can be added to the sample to make it easier to place on the device. To avoid diluting the concentration of analyte below the limit of detection of the inventive device, the amount of saline solution added is preferably no larger than the volume of the sample to be diluted. In other embodiments, the bodily fluid may be nasal secretions (such as used in the lateral flow immunoassay kits provided for CO VID-19 testing).

[0073] In urgent care, pharmacy, or hospital settings, the bodily fluid may be blood plasma. It is believed that the presence of red blood cells (“RBCs”) in the blood will likely block the ability of the plasma to diffuse through the porous membrane. Blood is preferably centrifuged to drive the RBC to the bottom of the centrifuge tube. The plasma, now free of RBC can be removed from the centrifuge tube by conventional means, such as pipetting, and a sample drop placed on one of the inventive devices for testing.

[0074] Initial studies underlying the present disclosure were performed using pipettes to place sample droplets on the membrane of exemplar devices. Pipettes are, of course, not usually found in household settings and providing them in a kit intended for home testingwould unnecessarily raise the price of the kit. One convenient and inexpensive option for picking up a sample of bodily fluid and placing it on the membrane of a device of the invention is to use a microtube. The length of the microtube is chosen based on how much sample needs to be provided to the membrane. The microtube can be provided with a bulb at a first end of the microtube that can be squeezed to force the sample out the second end of the microtube onto the membrane of the device when the second end has been placed over a position on the membrane designated to receive it.

[0075] Lateral flow immunoassay devices typically require the use of a sample with a volume of 40 pL. The devices of the present invention can use sample sizes that are almost an order of magnitude lower. In studies underlying the present disclosure, detection of analytes of interest were successful using sample drops with a volume of only 5 pL, while the plasmonic drop had a volume of only 2 pL. Thus, embodiments of the present invention allow sensitive detection of analytes with much smaller sample volumes than needed for typical lateral flow immunoassays. This allows using smaller amounts of materials to produce the inventive devices and smaller amounts of reagents to conduct them compared to lateral flow immunoassays.

[0076] The volume of the drop is limited to some extent by the size of what may be termed the detection zone of the membrane. The exemplar devices used in the studies reported in the Examples had membranes that were 4.5 mm in diameter, and a 10 pL drop would fill the entire detection zone, leaving no room for a plasmonic drop to be deposited on the membrane in a position that would provide an asymmetric pattern (the second drop cannot be placed on the membrane on top of the sample drop and cause an asymmetric pattern to form). The sample drop is preferably as large as possible without ruining the ability of the drops to form an asymmetric pattern. It will be appreciated that, as a liquid, the drop will spread into and along the surface of the membrane even as it dries, and the area encompassed by the residual drop by the time it is fully evaporated will have a larger radius than that of the drop as it was first placed on the membrane.

[0077] An equation that can be used to determine the volume of a sample drop for any particular radius of a detection zone is:Equation 1 :where Vs is the sample drop, Vo is the volume of empty space within the membrane available to be filled by liquid and Rd is the radius of the detection zone. The equation is derived by assuming that the drop forms a hemisphere when first placed on the membrane and it covers half the membrane. R = R il, with Rsbeing the residual drop radius. Since the hemisphere assumption is not the case and the residual drop has a larger radius than the original drop as it was placed on the membrane, the term Rs= Rail is a good approximation to provide an asymmetric pattern.

[0078] The designed value for the plasmonic droplet is 2 pL for a detection zone of 4.5 mm. This value is based on the concentration of the gold nanoshells (GNShs) and the surface tension of the plasmonic droplet. The concentration of the GNShs used in the plasmonic drops preferably remains the same, both for stability and for limiting the accumulation of particles. Equation 1 can still be used, however, because the plasmonic drop has lower surface tension than the sample drop and will expand more for each pL than will a sample drop. So, instead of Rs = Rail, Rd / 3 can be used.

[0079] As noted above, the plasmonic drops comprise gold nanoshells. Gold nanoshells suitable for use in the inventive devices and methods are commercially available. The gold nanoshells are functionalized to bind to the analyte of interest. In some embodiments, the functionalization is by attaching to the gold nanoshells antibodies (“Abs”) that specifically bind the analyte of interest. Methods of functionalizing gold nanoshells, including attaching antibodies to them by conventional chemistry, are known. An exemplar method to attach antibodies to the gold nanoshells is set forth in the studies reported in the Examples. Studies underlying the present invention used silica shells having a diameter of 140 nm coated with gold, resulting in a nanoparticle having a total diameter of 150 nm. The electric field distributions of a 40 nm diameter gold nanoparticle (GNP) and a 150 nm diameter GNSh are compared at the resonance frequency in Fig. 4B. The large size and high localization factor at the metal / dielectric boundary makes GNShs a better candidate for pattern visualization.

[0080] The GNShs used in the studies underlying this disclosure are widely available, easy to synthesize, and robust plasmonic nanoparticles, making them particularly convenient embodiments for use in at-home biosensing kits. However, it is expected that other plasmonic nanoparticles, with different shapes, sizes, and compositions, can be used make embodiments of the inventive devices biosensors. Aspects of plasmonic silver and gold nanoparticles, andthe effect of changes in shape and structure on their functionality is set forth in e.g., reference

[0070] ,

[0081] The simplified periodic array analysis of GNShs aggregates shows the electric field has been strongly localized in areas between the nanoparticles in Fig. 4C. Based on the simplified model, the optical absorption spectrum reveals two broadband peaks around 420 nm (in violet color range) and 660 nm (in red color range) for the dispersed 2D-like GNShs aggregates, compared to 550 nm (in green color range) for the GNPs. The 3D aggregates have a more uniform absorption across the visible spectrum as shown in Fig. 4D. Furthermore, GNShs have been compared with conventional 40 nm GNPs to show that GNShs exhibit higher optical absorption compared to those of the 40 nm GNPs, which is consistent with findings in the literature

[0073] , As the wavelengths at which GNShs absorb light increases as particle size increases, GNShs larger than the 150 nm in diameter GNShs used in the studies reported herein so long as the light that they absorb is still in the visible range.

[0082] The sample drop is placed on the membrane first and allowed to evaporate. Forces within the drying drop change the distribution of analytes within the drop compared to their original distribution: analytes within the drop concentrate along the edge of the drop, while a correspondingly lower concentration of the analytes remain within the interior of drop. This concentration of analytes at the edge of the drying drop results in what is sometimes referred to as a “coffee-ring” along what was the edge of where the drop dried.

[0083] The plasmonic drop is then placed on the membrane, offset from, or opposite to, where the sample drop was placed, but close enough so that, as the plasmonic drop spreads on the membrane, there is an area of overlap between the edge of the plasmonic drop and the coffee-ring left behind by the sample drop, but an area in which they do not. Comparing the difference in patterns between the area of overlap and the area in which there is no overlap between the area occupied by the now-dried sample drop and the plasmonic drop helps the reviewer discern whether the sample contained the analyte of interest or did not contain the analyte of interest. As noted below, where functionalized GNShs bind to analyte, a smooth pattern is formed, while areas in which the analyte of interest is not present, the Abs tend to interact and create a lumpy pattern that makes it easier to distinguish the smooth pattern caused by binding to the analyte of interest.

[0084] In some embodiments, the device bearing the membrane is heated and the sample and plasmonic drops are then sequentially placed on the membrane. The heat both increases the evaporation rate, allowing the assay to be conducted in less time, and also helps cause more of the analyte to accumulate at the boundary of the drying drop, enhancing the sensitivity of detection of the analyte. The device can be heated to a temperature from 60 °C to 95 °C, from 70 °C to 90 °C, from 75 °C to 85 °C, or 80 °C ± 4 °C, 80 °C ± 3 °C, 80 °C ± 2 °C, 80 °C ± 1 °C, or 80 °C, with each stated temperature or range being preferred to any stated before it. In some embodiments, the drops are allowed to evaporate at room temperature, which removes the need for a means to heat the device, but reduces the device’s sensitivity.

[0085] As persons of skill will appreciate, the positions at which the two drops (the now- dried sample drop and the newly placed on the membrane plasmonic drop) will overlap will depend in part on the size of the drops - a 10 pl sample drop will spread further across the membrane than will a 5 pl sample drop. The plasmonic drop should not be placed directly over where the sample drop dried, or so close to it that all or virtually all of the plasmonic drop overlaps with where the sample drop dried. (For convenience of reference, the area where the sample drop dried and the later-applied plasmonic drop intersect is sometimes referred to herein as an overlap of the two drops, even though the sample drop has already dried before the plasmonic drop is deposited on the membrane.) Positions at which to place first the sample drop and then the plasmonic drop so that there is an area in which they overlap and areas in which they do not are readily determined for any particular combination of drop sizes and membranes of different pore sizes and thickness and is easily within the skill of even untrained members of the public, let alone that of persons of skill. Placing the sample drop on a device of the present invention, followed by the plasmonic drop, to create patterns allowing a readout of the device and then determining whether the patterns created by the two drops indicates that the analyte is present (a “positive result”) or not present (a “negative result”) is sometimes referred to herein as “conducting an assay” or “performing an assay.”

[0086] Conveniently, marks are provided on the inventive devices to indicate to the user where to place the sample drop and where to place the plasmonic drop so that the drops overlap in a manner that will facilitate determining whether the analyte of interest is present in the sample at a concentration above the limit of detection of the device. As the layers (including the membrane) over the solid support are thin, the markings can be on the glass slide or other solid support and seen by the user. Alternatively, they can be on a layerbetween the solid support and the membrane, they can be on the membrane, or they can be on a portion of the device extending beyond the membrane, but adjacent to it. The markings may be, for example, numbers (such as “1” and “2”, indicating to the user where the first (sample) drop is placed and where the second (plasmonic) drop is placed), letters (such as “S” for “sample” and “I” for indicator), or symbols, such as plus signs or Xs. The markings can be in ink. Inks used for lateral flow immunoassays are expected to be suitable for use in the inventive devices.Covers

[0087] The inventive devices can have a cover to protect the membrane and keep it clean before use. The cover can, for example, be a piece that can slide along the length of the device to expose the membrane for use, can be on a hinge to be swiveled up when the device is to be used, or can be detachable so that it can be removed from the device when an assay using the device is to be performed.Detection of the Analyte of Interest at High and at Low Concentrations

[0088] As noted above, the inventive devices and methods allow detection of analytes of interest at much lower concentrations than those that can be detected using lateral flow immunoassays, with limits of detection (“LOD”) as low as 10 pg / ml. As depicted in Fig. 1c, concentrations well above the LOD of the inventive devices can typically be detected by eye by observing the smooth pattern in the area of overlap compared to the grainy appearance of the area in which the two drops do not overlap.

[0089] When the analyte of interest is present at a concentration close to the LOD of the inventive devices, it can be detected with the assistance of a computer. First, a dataset is created by performing a series of assays using known concentrations of the particular analyte of interest. An algrorithm finds the central line of the patterns in the photograph for each known concentration, and records the pattern in the central line in the dataset, correlated to the concentration of the analyte.

[0090] For a test sample, an assay is performed, and a photograph is taken of the device after any patterns caused by the addition of the plasmonic drop have been created. Conveniently, the photograph is taken with the camera of a smartphone, which can then upload the photograph to a program. (If the photograph is not taken with a smartphone, it is provided to the computer by another means.) The computer is programmed to find the central line in thephotograph from the test sample and to compare it to the information in the dataset on which it was trained. Studies underlying the invention and reported in the Examples show that this computer assisted technique allows determination of whether an analyte is present or not at concentrations of analyte at which a reviewer cannot make a determination by eye.

[0091] Further, since the algorithm is trained on different concentrations of the analyte of interest, it can determine the concentration of the analyte of interest in the sample. Practitioners wishing to determine the concentration of the analyte of interest in the sample can use the computer-assisted method even for samples in which the presence of the presence of the analyte of interest in the sample can be determined by eye.EXAMPLESExample 1

[0092] In some embodiments, the asymmetric pattern formed in the area of overlap between where the sample drop evaporated and the plasmonic drop extends allows detection of a positive reaction by eye. This can be more readily understood by referring to Fig. 1c, in which the two circles formed by the sample drop (indicated by the circle drawn with a solid line) and the plasmonic drop (indicated by the circle drawn with a dashed line), respectively, have an area of overlap which is labeled as being characterized by having small aggregates formed by the interaction of the analyte of interest with gold nanoshells functionalized with an antibody that specifically binds the analyte.

[0093] The volume of the sample drop can be readily adjusted to provide the best pattern clarity given the particular membrane used and the particular analyte to be detected. The volumes of the drop, however, is typically quite small: the sample drop is typically 3-7 pl, more typically 5 pl ± 0.5 pl, and preferably 5 pl. The volume of the drop of the plasmonic nanoparticles is similarly small, and is preferably 2 pl ± 0.5 pl. The maximum usable volume of the drop will depend on the size of the membrane. For the exemplar membranes used in the studies reported here, which have a diameter of 4.5mm, the maximum drop size is expected to be 4 pl. It is expected that the user can readily empirically determine the maximum size drop that can be used with a membrane of any particular size and thickness by running assays as described herein.

[0094] Without wishing to be bound by theory, it is believed that the underlying mechanism of the coffee-ring biosensor concerns two phenomena. First, the local concentration of theprotein of interest in the drop increases through the coffee-ring process at the position at which the sample drop was placed. Second, the plasmonic drop then allows visualization of the deposited proteins through specific interactions, creating a dispersed 2D-like plasmonic pattern (see Figs. IB and 1C). Plasmonic nanoparticles outside the specific region interact with each other due to the lack of specific proteins, forming large 3D aggregates that contribute to an asymmetric pattern which itself contributes to specific detection of the protein or proteins of interest (Fig.1C). Although the plasmonic drop undergoes the same evaporation process, its lower surface tension causes it to spread more, removing the nonspecific signal generated by an excess of GNShs through pushing them towards the hydrophobic boundary of the detection zone. Moreover, the coffee-ring effect also applies brings more GNShs in contact with specific proteins on the membrane, increasing the reaction rate and enhancing the visual signal.

[0095] The evaporation process on the thin nanofibrous membrane for both the sample drop and for the plasmonic nanoparticle drop contains four major steps: spreading, fixed contact radius evaporation, fixed contact angle evaporation, and backward evaporation. These steps, which are graphically depicted in Figs. 2A-D, will now be discussed. First, once a droplet is placed on the membrane, it spreads towards the hydrophobic barrier at the boundary of the detection zone in two different formats. Part of the fluid passes through the membrane while a second part, the residual drop, moves along the top until equilibrium is reached between capillary force, surface tension, and gravity effects (see, Fig. 2A). During the spreading process, a small portion of particles in the drop (e.g., proteins in the sample drop and GNShs in the plasmonic nanoparticle drop) aggregates non-specifically at the boundary of the spreading circle, while the majority of particles stay within the residual drop, merging toward the inner coffee-ring through evaporation-induced flow. This process pre-concentrates the sample concentration during the fixed contact radius evaporation step, as depicted in Fig. 2B. The last two steps have less effect on the pre-concentration process. During the fixed contact angle evaporation, the residual drop contact radius decreases while keeping the contact angle fixed (less than 5°), leaving the rest of the particles in the central part of the detection zone, as depicted in Fig. 2C. Fourth, the remaining fluid inside the membrane evaporates from the outer ring towards the center, as the larger surface area has a higher evaporation rate (Fig. 2D). Sessile drop evaporation on a porous membrane is explained in more detail in

[0052] , Note that having a thin membrane minimizes the fluid volume within the membrane, which in turn minimizes the number of particles lost to the non-specific boundary. However, by tuning thesurface properties of the membrane, we can further reduce this number and amplify the preconcentration.

[0096] In some embodiments, the invention provides devices exploiting asymmetric plasmonic patterns to allow the detection of a positive or negative result by visual inspection. In some cases, however, the concentration of the protein of interest is too low, or too near the limit of detection, to distinguish between different patterns and draw a conclusion using the unaided eye. Moreover, it may be desirable for the user to automate the detection process. In some embodiments, the pattern is captured using a smartphone and processed using a deep machine learning algorithm (see Fig. lA). In some embodiments, the invention provides a diagnostic method based on a dataset of asymmetric patterns developed for diverse biomarkers and control samples. A dataset can be developed and used train a Convolutional Neural Network (“CNN”) with VGG-16 architecture (see Fig. 5A and reference

[0074] ). To test for a specific biomarker, one evaporates the sample drop, then the plasmonic drop, and then takes a photo of the zone in which the sample droplet and the plasmonic droplet overlap (the zone in which the two droplets overlap is sometimes referred to herein as the “detection zone” or the “overlapping zone”), conveniently by using a smartphone. The image is then provided to a machine learning algorithm, which processes the images, and extracts key features, which a Fully Connected (“FC”) network then uses to classify the test outcome, as graphically depicted in Fig. 5A.

[0097] A Conditional Generative Adversarial Network (“C-GAN”) is used to extract information from the overlapping zone for further processing and sample concentration quantifications. In an exemplar embodiment, the generative network is designed based on the U-Net network, originally created for the segmentation of biological samples

[0075] , as graphically depicted in Fig. 5B. To train the generator algorithm, a dataset was prepared of overlapping zone images as input and manually segmented specific zones as labels, performing the training process in tandem with the discriminator network, itself a CNN network, as depicted in Fig. 5C. Training these two networks together enabled the training of a C-GAN model capable of extracting the specific zone from detection zone images that originally had noisy patterns not relevant to the sensing process (refer to Fig. 5D). It should be noted that while a U-Net model alone worked for the dataset used, the literature suggests that GAN networks are robust in the presence of noise

[0076] , This is useful for coffee-ring biosensor outputs, as they can include noisy patterns that may decrease the performance of asingle U-Net network. The presented C-GAN network was trained using the Pix2Pix method

[0077] ,

[0098] In Fig. 5D, the designed network has the detection zone patterns as inputs and segments out the specific zone as the output of C-GAN. Next, the extracted specific zone is fed into the image processing unit (“IPU”) to extract the critical points of the overlapping zone, called central points, (shown as golden spheres in Fig. 5D) for crosslines extraction, which is being used in the next step for the concentration quantification. The generator removes the artifacts and noisy features to facilitate the subsequent concentration quantification step. This network makes the sensing and quantification fully automatic.

[0099] A CNN network with fine-tuned VGG-16 architecture was used to analyze images from plasmonic biosensing images of N-Protein, PCT, CEA, and PSA. Fig. 5E shows results using a 0.5 probability as the boundary between positive and negative results, the screening performance of the coffee-ring biosensor for each of the described biomarkers. Specifically, a minimum LOD of about 3 pg / ml was achieved for PSA, which is about 30 times better than a comparable ELISA test

[0048] , Notably, the observed LODs for N-Protein and PCT were well below the relevant concentrations in the corresponding diseases

[0039] ,

[0078] -

[0080] , indicating that the inventive devices can be used for early SARS-CoV-2 and sepsis diagnosis. While the observed LOD for CEA and PSA was also within relevant ranges, early cancer diagnosis would be expected to require the detection of additional biomarkers, such as circulating tumor DNAs (ctDNAs) by highly sensitive genetic diagnostics such as PCR

[0081] ,

[0100] The observed variations in LODs among different proteins are believed to be due to the affinity of the antibody-antigen interactions and the quality of the proteins received from vendors. Apart from having low LOD values, a biosensor should respond specifically to the designed biomarker. To test the specificity of exemplars of the inventive devices, CEA and PSA proteins were mixed, as these two proteins are related to cancer diagnosis and may be present in samples simultaneously. PSA-specific plasmonic droplets were applied to various sample droplets, including (1) a protein-free buffer, (2) a sample that was spiked with highly concentrated (1000 ng / ml) CEA, (3) a sample that was spiked with PSA, and (4) a sample that was spiked with PSA plus a high concentration (1000 ng / ml) of CEA. The results showed that specific patterns were found only in the case of samples containing PSA or PSA plus CEA, indicating that the exemplar devices detected the proteins with high specificity.

[0101] Each image of a nanofibrous membrane on which a sample droplet and a plasmonic droplet has been placed contains information in the overlapping and non-overlapping zones of the residual droplets. An FC network can be trained based on the gray index values of the lines connecting both sides of the plasmonic residual droplet and passing through its center (Fig. 5F). This automatic algorithm can use the outputs of the C-GAN network to find the central crosslines without interfering with noisy irregular patterns outside the plasmonic coffee-ring region. By applying the trained network to four different biomarkers, the concentration quantification performance can be evaluated. Results showed less than half an order of magnitude deviations from expected values for most points based on a simple and affordable diagnostic method (Fig. 5G). Moreover, results show that analyzing images from exemplars of the inventive devices allowed detecting biomarkers over a concentration range of 4-5 orders of magnitude. To assist in achieving repeatable results, a 3D-printed mount was attached to the substrate, allowing even untrained users to precisely control the droplet placement locations and the gap distances between them. This setup ensured that consistent asymmetric plasmonic patterns were produced in both the calibration and test sets.

[0102] Lastly, to demonstrate the capability of the coffee-ring biosensor in detecting biomarkers within complex biological samples, N-protein spiked samples were mixed with pooled human saliva (see Figs. 5H-5J). The results indicated that the coffee-ring biosensor retained its sensitivity and successfully detected N-protein within the biological solution, showing no loss in performance. Additionally, the results sensitivity of detection over two orders of magnitude greater than that of the equivalent LFIA.Example 2

[0103] This Example presents the results of a study comparing an exemplar biosensor embodiment of the present invention against other biosensors.

[0104] Table 1 presents a comparison of the detection of an exemplar biomarker (the N- Protein of SARS-CoV-2) by various biosensors, including an exemplar embodiment of the present invention.Table 1.Example 3

[0105] This Example illustrates the particle distribution during the droplet evaporation process.

[0106] Fluorescent dyes were used to represent proteins in the sample drop, as shown in Fig.2E. In agreement with our observation of evaporation steps, two bright lines appear — one corresponding to the spreading boundary (non-specific aggregation of nanoparticles) and the other showing the coffee-ring of the now-evaporated sample residual drop as the inner circle (see Fig. 2E(i)). (A “residual drop” is the portion of the sample drop that remains as a droplet on the membrane, in contrast to the portion that moves into and through the membrane).

[0107] Between these two rings, there is a particle-depleted zone that clearly separates the spreading boundary from the coffee-ring, showing that by minimizing drop spreading, one can concentrate particles in the drop at a single ring, achieving a higher local concentration ofproteins than was originally present in the drop as a whole (Fig. 2E(i)). The gray index profile across the central line passing over outer and inner rings (Fig. 2E(ii)) shows high contrast at the coffee-ring, representing that the majority of particles within the residual drop accumulated there due to evaporation-induced flow toward the boundary of the residual drop. It should be noted that some biomarkers (e.g., proteins) can be trapped within the forest of micro-crystalline residues formed by different salts existing inside the biological solution, such as potassium and sodium (see, Fig. 2E(i), inset 5. Note: the insets are numbered as shown in the large photograph from which the insets are taken). Spreading helps push these small salt molecules towards the outer boundary, reducing biomarker trapping inside the inner circle.

[0108] To generate the asymmetric plasm onic pattern, the plasm onic droplet is placed on the side of the membrane away from where the sample drop was placed (Fig. 2F(i) shows the relative location of each zone). In the common area between the now-evaporated sample drop and the current plasm onic residual drop, GNShs interact with proteins already covering the surface of the membrane, resulting in specific aggregation, as shown in Fig. 2F(i), inset 2. Outside the common zone, the protein concentration is very small, so GNShs self-aggregation is stronger than surface adsorption, leading to large non-specific aggregation and dark dots in inset 4 of Fig. 2F(i). As mentioned earlier, a similar coffee-ring effect happens to the plasm onic drop, increasing the local concentration of GNShs at the boundary of the plasmonic residual drop (see insets 1 and 5 of Fig. 2F(i)). Note that the coffee-ring intensity on the outer side (inset 5, Fig. 2F(i)) is stronger than the inner part (inset 1, Fig. 2F(i)), as the majority of GNShs are attracted by previously coated proteins on the membrane in the common area, and less GNSh is available to aggregate at the inner coffee-ring. Each of the explained regions is clearly depicted in the central crossline profile of the detection zone (see Fig. 2F(ii)). Since the protein concentration drops from the coffee-ring of the sample drop towards the center of the sample residual drop, there is a gradient in the gray index intensity in the common area (the concentration drops from point 3 towards point 1, as shown in Fig. 2F(i). The transition between specific aggregation and non-specific aggregation is clearly shown at boundary 3 (coffee-ring of sample residual drop), moving from a continuous pattern to a noisy and granular one (see, Fig. 2F(ii)). The captured image profile contains information about the concentration of the sample drop, which can be used for further processing to quantify the concentration of the sample.

[0109] The impact of humidity on the coffee-ring pattern was also examined. Previous studies have shown that humidity can significantly affect droplet evaporation patterns on non- porous and non-heated substrates, particularly for smaller droplets (smaller than 100 pm)

[0058] , However, notable variations were not observed in studies conducted in the course of the current work, even under extreme conditions.

[0110] More information about the theory of particle deposition pattern during sessile drop evaporation is set forth in

[0057] ,

[0059] -

[0061] .Example 4

[0111] This Example discusses how a nanofibrous membrane can be tuned to adjust the movement of a drop through the membrane.

[0112] Enhancing particle accumulation at the coffee-ring boundary involves parameters including substrate material, solution evaporation rate, droplet viscosity, and droplet volume. As previously discussed, during the drop spreading step, nanoparticles such as proteins migrate toward the hydrophobic barrier, leading to non-specific aggregation and consequent loss of biomarkers, resulting in an increase in the lowest detectable concentration. To manipulate the spreading rate and retain the majority of the solution at the residual drop, the size of nanofibers within the thin hydrophilic membrane was adjusted through thermal treatment of the supporting substrate, as shown in Fig. 3 A. The detection zone in the exemplar device shown comprises five materials stacked together: glass, double-sided tape, polyimide (“PI”) film, adhesive, and a hydrophilic polytetrafluoroethylene (“PTFE”) nanofibrous membrane. By applying thermal flux to the substrate, silicon nanoparticles leave the adhesive layer, penetrate the nanofibrous membrane, and cover the hydrophilic PTFE nanofibers, resulting in an increase in nanofiber size (Fig. 3 A). Notably, PI film is temperature stable

[0062] , which provides a robust substrate for the membrane during the thermal treatment process, and the nanofibrous hydrophilic PTFE nanopores allow large protein clusters to bind while permitting the passage of small molecule compounds, forming crystalline objects at the outer boundary. Furthermore, PTFE is transparent in the infrared (IR) spectrum

[0063] , which enables additional optical characterization of the samples if desired. ). Membranes suitable for use in embodiments of the inventive devices preferably are very thin (z.e., <10 pm), have high protein absorptivity, have controlled wettability, and have a high surface-to-volume ratio. In some embodiments, the membranes are nanofibrous.Information about selecting nanofibrous membranes suitable for use in biosensing applications is available in, for example, references

[0064] and

[0065] ,

[0113] Thermal treatment provides a convenient means by which to control the spreading rate. Increasing the applied temperature over a critical range can effectively block the nanoscale pores inside the membrane, preventing the solution from spreading (see, Fig. 3B). By obstructing the spreading step and containing the entire solution within the residual drop, salt molecules concentrate at the coffee-ring position, forming large crystals that impede access to biomarkers. Consequently, studies underlying this disclosure explored a range of applied temperatures from 20 °C to 160 °C, monitoring both spreading and evaporation times, as shown in Figs. 3C and 3D. Increasing the applied temperature to 80 °C resulted in a gradual reduction in the spreading rate and the distance between the spreading boundary and the residual drop boundary, with a significant drop observed above 80 °C up to 120 °C. Beyond 120 °C, all nanopores are blocked, preventing the solution from spreading through the membrane (Fig. 3C). Additionally, a comparison of evaporation time and the contact angle of the residual drop at different temperatures revealed that the contact angle jumps after 80 °C, indicating increased membrane hydrophobicity at higher temperatures (see Fig. 3D). Higher thermal treatment temperatures also extended the evaporation time due to increased confinement of the solution inside the residual drop compared to cases with higher wettability (Fig. 3D).

[0114] Based on the study conducted on the influence of thermal treatment temperature on particle accumulation (Figs. 3 A-D), 80 °C was selected to provide uniform conditions for further studies. Although temperature-sweeping experiments indicated better performance at 100 °C, the decision to work with lower temperatures was driven by the fact that we decided to conduct the sample evaporation process at 80 °C (plasmonic evaporation was conducted at room temperature), ensuring faster evaporation and minimizing the response time of our biosensor, reaching around 10 min from the beginning of the sample evaporation step. Therefore, the membrane undergoes another step of thermal treatment during the sample droplet evaporation, which adds more to the size of nanofibers. It is important to note that while higher temperatures below the boiling point of the buffer solution, phosphate-buffered saline (PBS), can further minimize evaporation time, such temperatures may lead to protein structure denaturation

[0058] , affecting the affinity of the protein of interest to the antibodies on the GNShs intended to bind the protein of interest. Additionally, heated substrate evaporation below certain temperatures

[0059] enhances the coffee-ring effect). For clarity, it is noted thatthe inventive coffee-ring biosensors work without applying heat during sample evaporation. Heat was used in some of the studies underlying the present disclosure to reduce response time and to accumulate more particles at the coffee-ring.Example 5

[0115] This Example discusses coffee-ring patterns and biosensing.

[0116] Utilizing the process derived over the course of a series of experiments, we applied our novel biosensing method to a range of proteins, including the N-protein of SARS-CoV-2, the sepsis biomarker PCT, and the cancer biomarkers PSA and CEA (Fig. 3E). Starting from high concentrations of the biomarkers (e.g., 1000 ng / ml) we gradually diluted the samples until the pattern became indistinguishable from the control solution of pure buffer (that is, with no biomarker). As depicted in Fig. 3E, the asymmetric deposition of the sample and the plasmonic droplets resulted in a non-symmetric pattern in the detection zone. With decreasing sample concentration, the intensity of the common area decreased, indicating fewer GNShs specifically attaching to the antigens previously coated on the membrane. At concentrations below the limit of detection (“LOD”), the pattern inside and outside the common area became indistinguishable, as the sample concentration was too low to bring sufficient antigens in and around the coffee-ring of the sample residual drop. Consequently, GNShs attracted each other more strongly than to the substrate, even in the common area and close to the coffee-ring, forming large aggregates in the entire regions surrounded by the coffee-ring of the plasmonic residual drop (see Fig. 3E). As persons of skill will appreciate, the size of the common area is user-dependent and therefore not consistent, but the intensity of the specific pattern inside the common area is sufficient for distinguishing positive versus negative results. In the example shown in Fig. 3E, the LOD for the protein of interest was 10 pg / ml, 5 orders of magnitude lower than the starting concentration, and hundreds of times better than the LOD for a lateral flow assay.

[0117] While signals can be detected by the naked eye, additional signals within each image not visible to the naked eye, such as the relative intensities of the coffee-rings in the detection zone, can be captured by a smartphone, followed by further processing using computer vision methods, including deep learning image processing, to extract more information for sample quantification. That aspect will be discussed in a later section.Example 6

[0118] This Example discusses plasmonic asymmetric patterns.

[0119] One of the key mechanisms for achieving a LOD as low as 10 pg / ml is the asymmetric pattern generated by plasmonic gold nanoshells (“GNShs”). As discussed earlier, in the presence of biomarkers acting as antigens on the membrane, bioconjugated GNShs are attracted to them, forming dispersed 2D aggregates that result in smooth patterns, as can be seen in Fig. 5a. However, in areas lacking sufficient antigens, GNShs exhibit stronger interactions with themselves rather than the substrate, forming large 3D aggregates characterized by a noisy “dark dots” pattern, also seen in Fig. 4A, clearly visible in scanning electron microscope (“SEM”) images at the coffee-ring of the plasmonic residual drop. The inherent difference between these two patterns serves as a distinct indicator of a positive sample, achievable by breaking the symmetry of the deposited droplets

[0120] It is believed that the fundamental reason behind this difference in patterns lies in the role of nanoplasmonic aggregate shape, size, and orientation in light-matter interaction

[0041] , As a full numerical analysis of the nanoplasmonic aggregates requires substantial computational resources, which is not necessarily in our case, we simplified our model to a periodic array of GNShs with different density to represent the role of the aggregates in the diverse optical response of the plasmonic patterns (see Figs. 4B and 4C). Background information concerning the role of geometrical properties in the optical response of nanoplasmonic aggregates are discussed in

[0041] ,

[0060] ,

[0061] ,

[0121] Based on our model, when GNShs form dispersed 2D-like aggregates, specifically in the presence of antigens, the optical absorption spectrum reveals two broadband peaks around 420 nm (in the violet color range) and 660 nm (in the red color range), compared to 550 nm (in the green color range). However, in 3D aggregates, we observed a more uniform absorption across the visible spectrum (see, Fig. 4D). Furthermore, we compared GNShs with conventional 40 nm gold nanoparticles (GNPs) to underscore the importance of transitioning from GNPs to GNShs. Our study demonstrates that GNShs exhibit higher optical absorption compared to the 40 nm GNPs (Fig. 4D), which aligns with the literature

[0062] ,Example 7

[0122] This Example sets forth information about the use of neural networks for protein sensing and quantitation.

[0123] As discussed earlier in the Detailed Description, asymmetric plasmonic patterns generated at high biomarker concentrations are visible to the naked eye. However, at lowerconcentrations, especially near the LOD, it can be difficult to distinguish between different patterns by eye to determine if the biosensor is providing a positive or a negative response to the sample.

[0124] Use of the methods described herein to various proteins over multiple tests resulted in the development of an extensive dataset of asymmetric patterns for diverse biomarkers and control samples. The dataset was used to train a Convolutional Neural Network (CNN) to automate the detection process, as graphically depicted in Fig. 5A. To test for a specific biomarker, one first evaporates the sample and plasmonic droplets, and then takes a photo of the detection zone. Conveniently, the photo can be taken using a smartphone. A machine learning algorithm processes the images, extracting key features, and a Fully Connected (FC) network classifies the test outcome, as graphically depicted in Fig. 5 A.

[0125] As noted above, each image contains more information than just pattern differences between specific and non-specific zones, primarily in the coffee-ring of plasmonic residual drops. As we intentionally placed droplets asymmetrically on the membrane, the relationship between sample concentration and the relative intensities of the opposite sides of the coffeering is no longer linear compared to symmetric placement of the droplets. To extract this nonlinear relationship, we trained an FC network based on the gray index values of the lines connecting both sides of the plasmonic residual drop and passing through its center, as graphically depicted in Fig. 5F. The fully automatic algorithm can use the output of the C- GAN network to find the central crosslines without interfering with noisy irregular patterns outside the plasmonic coffee-ring area. By applying the trained network to the four different biomarkers, we evaluated the concentration quantification performance of our coffee-ring biosensor, revealing less than half an order of magnitude deviations from expected values for the majority of points, as depicted in Fig. 5G, which is excellent performance for a simple and affordable diagnostic method Moreover, our data shows that the biosensor is capable of detecting biomarkers over a very large range, 4 to 5 orders of magnitudes (that is, 1000 ng / ml to 10 pg / ml).Example 8

[0126] This Example sets forth the materials and methods used in the studies discussed in the preceding Examples.Materials

[0127] GNShs with an average diameter of 150 nm (exhibiting a variation in size of 15% or less) were obtained from nanoComposix (San Diego, CA). These nanoshells consist of a silica core with an average diameter of 130 nm and gold nanoshells of ~ 10 nm. SARS-CoV- 2 N-Protein, PCT, CEA, and PSA proteins, as well as polyclonal antibodies, were obtained from Sino Biological US Inc. (Wayne, PA). Antibody purification was carried out using Amicon® Ultra 0.5 ml filters (MilliporeSigma, Burlington, MA). Hydrophilic PTFE nanofibrous membranes were sourced from MilliporeSigma. All reagents were procured from nanoComposix (San Diego, CA) unless otherwise stated.GNShs Functionalization

[0128] We introduced 1 ml of a solution containing 5mM potassium phosphate and 0.5% PEG with a molecular weight of 20 kDa to the dried NHS GNShs. This addition served to activate the NHS-Esters for subsequent antibody conjugation. Following this, the solution of GNShs was combined with purified antibodies at a specified concentration and allowed to incubate for one hour. After the incubation period, 5 pl of a 5% (w / v) hydroxylamine solution was introduced to block any unbound NHS-Esters, and the mixture was allowed to interact for an additional 10 minutes. To eliminate excess antibodies in the solution, the covalently conjugated GNShs were centrifuged three times at 3.8k RCF (relative centrifugal field). Subsequently, all non-bonded antibodies were removed, and the conjugated GNShs were supplemented with the volume of a solution containing 0.1X PBS, 0.5% BSA, 0.5% Tween 20, and 0.05% sodium azide necessary to achieve a final concentration of 5 pg / ml for the antibody conjugate.Substrate Fabrication and Thermal Treatment Process

[0129] Exemplar devices for the studies discussed in the Examples were made by hand. The device (sometimes referred to herein as a “substrate”) was made by stacking a glass slide on a surface, placing double-sided tape and PI film with adhesive on the glass slide, and adding a thin film of nanofibrous hydrophilic PTFE membrane on top. After all the layers were stacked, the middle of the membrane was pressed (conveniently with the thumb, when being hand-assembled) for 10 seconds, so that the whole of the membrane contacted the adhesive film underneath it. To control the wettability and surface properties of the membrane, the substrate was heated. In the initial studies, the assembled device was placed on a hot plate at a fixed temperature for 30 minutes. After 30 minutes, the device was cooled by applying nitrogen gas to the glass side of the substrate for about one minute.

[0130] To characterize the wettability of the thermally treated membrane, a droplet was placed on a membrane and the membrane placed it on a hot plate at 80 °C. The evaporation process was captured with a smartphone camera. The recorded video was processed to extract information about the spreading rate, residual drop radius and evaporation time. To characterize the surface properties of the membrane, we captured an image of buffer solution sessile droplet and extracted the contact angle by curve-fitting (refer to Figs. 23(a) and (b)). To structure of the nanofibers was examined under a Philips XL30 scanning electron microscope (“SEM”) (FEI Co., Hillsboro, OR).

[0131] As hot plates may not be readily available to persons outside of a laboratory setting, later studies were conducted using a wire resistance heater and heating the device to 80 °C. In some embodiments, the device is heated to 80 °C ±10 °C. In some embodiments, the device is heated to 80 °C ± 7.5 °C. In some embodiments, the device is heated to 80 °C ± 5 °C. In some embodiments, the device is heated to 80 °C ± 2.5 °C. In some embodiments, the device is heated to 80 °C ±1 °C. In some embodiments, the device is heated for 30 minutes ±10 minutes. In some embodiments, the device is heated for 30 minutes ± 7.5 minutes. In some embodiments, the device is heated for 30 minutes ± 5 minutes. In some embodiments, the device is heated for 30 minutes ± 2.5 minutes.Sample Droplet Evaporation Analysis

[0132] To analyze the evaporation process, the entire evaporation process was recorded using a smartphone fixed above the sample droplet. Recorded videos were then analyzed to characterize each step in the sessile drop evaporation process. To get particle deposition patterns, buffer solution was mixed with Alexa Fluor™ 488 NHS-Ester (ThermoFisher Scientific), at Ipg / ml concentration and confocal microscopy was used to capture fluorescence images after the droplet evaporation.Biosensing Process

[0133] As depicted in Figs. 1 A-E, the biosensing process includes successive evaporation of a drop which may or may not contain an analyte of interest (sometimes referred to herein as the “sample drop” or “drop containing a sample”), and a drop containing gold nanoshells (“GNShs”) which have been functionalized to specifically bind to the analyte of interest (the GNShs may also be functionalized to specifically bind to more than one analyte of interest). Conveniently, the GNShs are functionalized by attaching to them antibodies that specifically bind the analyte or analytes of interest.

[0134] To perform sensing using exemplar embodiments of the inventive coffee-ring biosensors, sample stock solutions were prepared by diluting stock protein solution from 1000 ng / ml to 1 pg / ml, by successive one-tenth dilution processes. The dilution process accuracy at high concentrations was checked using Millipore Direct-Detect and the correction factor was applied to prepared samples (concentrations multiplied by factor of 3). For each test, the hot plate was heated to 80 °C. The biosensor performance proved not to be very sensitive to temperature variations ±10 °C from 80 °C.

[0135] After achieving the chosen temperature on each membrane, we applied 5 pl of sample drop on one side of the membrane and waited for it to evaporate completely, typically around 1 minute. The device was then cooled by applying nitrogen gas on the glass side of the substrate. Next, we applied 2 pl of GNShs drop on the opposite side and waited about 9 minutes for it to fully evaporate.

[0136] The optical image was captured using a smartphone, an iPhone XR or an iPhone 14 (Apple, Inc., Cupertino, CA), using basic settings, from about 15 cm above. For better visualization at higher resolutions, a standard light microscope capable of low magnification can be used, such as an AmScope trinocular 5X-10X-15X-30X stereo microscope (AmScope.com).Biosensing in Human Samples

[0137] The ability of the inventive coffee-ring biosensors to detect analytes in biological samples, pooled human saliva from Innovative Research was studied. Medical swabs were immersed in the saliva sample following standard procedures, and then mixed with N-Protein SARS-CoV-2 spiked buffer. The rest of the procedure followed the same steps as the in-PBS tests. For a comparison between the performance of the coffee-ring biosensor and conventional LFIA assays, we applied the equivalent sample volume plus the running buffer to commercially available Flowflex® COVID-19 antigen LIFA home test kits (ACON Laboratories, Inc., San Diego, CA) at various concentrations. The background-corrected ratio of the test line to the control line of the LIFA was used to assess the performance of the LFIA tests.At-Home Coffee-Ring Biosensing Kit

[0138] To enable untrained users to perform tests at home without needing special tools like a hot plate or precision pipettes, a testing kit was designed that includes glass microtubes, 3D-printed components, and a film heater. In an embodiment, a battery-powered film heater was attached to the bottom of the glass substrate, allowing the substrate to be heated to a predetermined temperature. In an exemplar embodiment, a film heater was used that heated the substrate to 80°C when 12 volts is applied to the heater’s electric circuit. To handle accurately specific volumes of sample and of plasmonic solution, precisely cut glass microtubes were used. A solution contacting the mouth of such microtubes undergoes capillary force that automatically draws a sample of the solution in, and the length of the tube is selected such that capillary force will draw into the microtube the desired volume. A 3D- printed flexible bulb, attached to a pipette tip, was assembled with the microtube, enabling accurate yet simple sample handling, and the squeezable bulb providing a means for extruding droplet(s) from the microtube. Additionally, a 3D-printed mount was placed on top of the substrate to guide accurate droplet placement onto the detection zone of the membrane, while maintaining consistent spacing between droplets. Thus, in this embodiment, the 3-D printed mount allowed for high repeatability across different tests.

[0139] The most sensitive material, the plasmonic solution, remains stable for about six months at 4°C. In some embodiments, the plasmonic solution can be dried on a membrane and redissolved in reagent solvent for use, thus avoiding the need for storate at 4°C. Devices incorporating pre-dried plasmonic drops can therefore be provided to users in kits similar to current conventional LFIA covid test kits.Numerical Modeling

[0140] Electromagnetic field distribution and optical absorption spectrum of specific and non-specific aggregated GNPs and GNShs was modeled by the finite element method (“FEM”), using COMSOL Multiphysics software (V.6.1) (COMSOL, Inc., Burlington, MA). To simplify the numerical modeling, periodic arrays of GNP or GNSh placed on top of the hydrophilic PTFE substrate with an effective refractive index of 1.37 (considering the membrane porosity) and the background refractive index set to 1 as air. The Floquet boundary condition was used to represent periodicity in solving the Maxwell’s equation and perfectly matched layer (PML) on top and bottom boundaries were used. The box height was larger than half of the maximum wavelength in each material and the mesh size was optimized for dielectric and plasmonic materials. The Lorentz -Drude model was chosen to represent plasmonic material (gold)

[0082] , To model specific and non-specific aggregates, the periodicity of the unit cell was manipulated in a way that the density of particles representsthe specificity of the aggregates: tightly pack array for the non-specific form (large aggregates), and dispersed nanoparticles for the specific form (2D-like dispersed pattern). The optical absorption was normalized by the number of nanoparticles in the same 2D area.Deep Neural Network Training

[0141] All machine learning algorithms were implemented using the TensorFlow library in Python. For CNN-based detection of positive versus negative samples, cropped grayscale images of the detection zone were used as inputs for training the network with positive and negative class labels. Experiments were carried out at varying concentrations for four different proteins and repeated multiple times, resulting in a dataset of approximately 100 unique images. This dataset was expanded using augmentation techniques such as rotation to create a sufficient amount of data for training the deep neural network. However, even with augmentation, to build a robust model, a dataset on the scale of thousands of images is typically required, which is difficult, costly, and time-consuming to generate experimentally. To address this challenge, transfer learning was employed by fine-tuning a pre-trained VGG- 16 network, initially trained on the ImageNet dataset containing around 12 million images

[0074] , Both VGG-16 and ResNet-50 were tested; it was observed that VGG-16 reaches lower loss values (binary cross-entropy) and higher accuracies at the same iteration number. The model was modified by adjusting the last layer from 1000 nodes to 256 and replacing the softmax function with a sigmoid function, since our task was binary classification. The model outputs a probability between 0 and 1, with values below 0.5 considered negative and those above 0.5 considered positive.

[0142] For the C-GAN based segmentation of the specific zone, the Pix2Pix algorithm was used, with grayscale detection zone images used as input and manually extracted specific zones used as the labels. Generative and discriminator networks were trained in tandem. The training was stopped when the generated specific zone and the ground truth are similar enough, since loss values are not a reliable source of action in training generative networks

[0084] , All images are normalized prior to training, and their sizes are 256 by 256.

[0143] The algorithm to quantify concentration value uses extracted lines to cross the plasmonic residual droplet coffee-ring around the center to increase the training data. The simple FC network is used to replace the non-linear relationship between the hidden information in each image and the sample concentration. Since the crossline sizes are not the same for all samples, we fixed the size and attached constant 1 to the extended part,resampling white region (255 in grayscale, 1 in normalized scale) outside the coffee-ring region. The random sampling and flipping augmentation are important for getting reliable predictions from the network. The networks training process, test, and evaluation were done on Google Colab. Additionally, the performance of the model was evaluated under various conditions, including camera tilt angle, lighting variations, and different levels of imaging sensor noise. Our analysis revealed minimal variation, highlighting the robustness of the trained model. The tilt angle variation was restored by scaling back the detection zone into the circular shape.

[0144] Those skilled in the art will appreciate that implementation of the described computer processing of the test sample images including deep learning comprising training and processing using a convolutional neural network (CNN) and / or a conditional generative adversarial network (C-GAN) can be performed by any known suitable computing device and software. In one notable example embodiment as previously described, a smart phone application can be conveniently employed to implement both capture (using the built-in camera) and processing of a given test sample image. In this case, the application can be installed including appropriately pre-trained CNN and / or C-GAN ready for processing test samples in the field.Example 9

[0145] This Example discusses the results set forth in the preceding Examples.

[0146] The studies discussed above show the creation of a novel, simple, rapid (about 10 minutes of response time), and affordable biosensing method with high sensitivity and specificity, utilizing a natural pre-concentration process, coffee-ring, and asymmetric plasmonic patterns. The devices can be used by individuals without prior training. In some embodiments, the user can receive an analysis of the sample informed by deep learning algorithms.

[0147] We combined passively controlled evaporation-induced flow within evaporating sessile drops of sample with the high optical absorption of GNShs to accumulate biomarkers to the boundaries and to visualize them. By thermally treating the nanofibrous membrane, we improved particle flow inside the evaporating droplets. We demonstrated that our biosensor is general and can be applied to a wide range of biomarkers, including the N-Protein of SARS-CoV-2, the sepsis marker PCT, the cancer marker CEA and PSA, with an LOD as low as 10 pg / ml, which is 500 times lower than the LOD of conventional LFIA and comparable toELISA tests. Furthermore, we illustrated that the introduced biosensor is highly specific by combining non-specific and specific biomarkers in a single sample drop. Different deep learning networks were used to automate the sensing process and enhance the concentration quantification capabilities, showing a sensitive range of 4 to 5 orders of magnitudes with high accuracy.

[0148] In conclusion, the newly developed biosensor opens a new way to revolutionize current widely accepted LFIA platforms and provides a simple and affordable diagnostic method sensitive enough to replace expensive and complex ELISA-based methods.REFERENCES[1] E. Fu, P. Yager, P. N. Floriano, N. Christodoulides, and J. T. McDevitt, “Perspective on Diagnostics for Global Health,” IEEE Pulse, vol. 2, no. 6, pp. 40-50, 2011, doi:10.1109 / MPUL.2011.942766.[2] T. L. G. Health, “Essential diagnostics: mind the gap,” Lancet Glob. Heal., vol. 9, no. 11, p. el474, Nov. 2021, doi: 10.1016 / S2214-109X(21)00467-8.[3] H. Yadav, D. Shah, S. Sayed, S. Horton, and L. F. Schroeder, “Availability of essential diagnostics in ten low-income and middle-income countries: results from national health facility surveys,” Lancet Glob. Heal., vol. 9, no. 11, pp. el553-el560, Nov. 2021, doi: 10.1016 / S2214- 109X(21 )00442-3.[4] N. Engel et al., “Addressing the challenges of diagnostics demand and supply: insights from an online global health discussion platform,” BMJ Glob. Heal., vol. 1, no. 4, p. e000132, Dec. 2016, doi: 10.1136 / bmjgh-2016-000132.[5] D. Melander, S. Sivarasu, I. Yekinni, C. Yunzhang, and A. Erdman, “Diagnostics as the Key to Advances in Global Health: Proposed Methods for Making Reliable Diagnostics Widely Available,” J. Med. Device., vol. 14, no. 1, Feb. 2020, doi: 10.1115 / 1.4046046.[6] R. Komatsu et al., “Lives saved by Global Fund-supported HIV / AIDS, tuberculosis and malaria programs: estimation approach and results between 2003 and end-2007,” BMC Infect. Dis., vol. 10, no. 1, p. 109, 2010, doi: 10.1186 / 1471-2334-10-109.[7] D. Low-Beer, R. Komatsu, and O. Kunii, “Saving Lives in Health: Global Estimates and Country Measurement,” PLOS Med., vol. 10, no. 10, p. el001523, Oct. 2013, [Online], Available: https: / / doi.org / 10.1371 / journal.pmed.1001523.[8] D. R. Hogan, G. A. Stevens, A. R. Hosseinpoor, and T. Boerma, “Monitoring universal health coverage within the Sustainable Development Goals: development and baseline data for an index of essential health services,” Lancet Glob. Heal., vol. 6, no. 2, pp. el52-el68, Feb. 2018, doi: 10.1016 / S2214-109X(17)30472-2.[1] E. Fu, P. Yager, P. N. Floriano, N. Christodoulides, and J. T. McDevitt, “Perspective on Diagnostics for Global Health,” IEEE Pulse, vol. 2, no. 6, pp. 40-50, 2011, doi:10.1109 / MPUL.2011.942766.[2] T. L. G. Health, “Essential diagnostics: mind the gap,” Lancet Glob. Heal., vol. 9, no. 11, p. el474, Nov. 2021, doi: 10.1016 / S2214-109X(21)00467-8.[3] H. Yadav, D. Shah, S. Sayed, S. Horton, and L. F. Schroeder, “Availability of essential diagnostics in ten low-income and middle-income countries: results from national health facility surveys,” Lancet Glob. Heal., vol. 9, no. 11, pp. el553-el560, Nov. 2021, doi: 10.1016 / S2214- 109X(21 )00442-3.[4] N. Engel et al., “Addressing the challenges of diagnostics demand and supply: insights from an online global health discussion platform,” BMJ Glob. Heal., vol. 1, no. 4, p. e000132, Dec. 2016, doi: 10.1136 / bmjgh-2016-000132.[5] D. Melander, S. Sivarasu, I. Yekinni, C. Yunzhang, and A. Erdman, “Diagnostics as the Key to Advances in Global Health: Proposed Methods for Making Reliable Diagnostics Widely Available,” J. Med. Device., vol. 14, no. 1, Feb. 2020, doi: 10.1115 / 1.4046046.[6] R. Komatsu et al., “Lives saved by Global Fund-supported HIV / AIDS, tuberculosis and malaria programs: estimation approach and results between 2003 and end-2007,” BMC Infect. Dis., vol. 10, no. 1, p. 109, 2010, doi: 10.1186 / 1471-2334-10-109.[7] D. Low-Beer, R. Komatsu, and O. Kunii, “Saving Lives in Health: Global Estimates and Country Measurement,” PLOS Med., vol. 10, no. 10, p. el001523, Oct. 2013, doi:10.1371 / j ournal .pmed.1001523.[8] D. R. Hogan, G. A. Stevens, A. R. Hosseinpoor, and T. Boerma, “Monitoring universal health coverage within the Sustainable Development Goals: development and baseline data for an index of essential health services,” Lancet Glob. Heal., vol. 6, no. 2, pp. e!52-e!68, Feb. 2018, doi: 10.1016 / S2214-109X(17)30472-2.[9] W. R. de Araujo, H. Lukas, M. D. T. Torres, W. Gao, and C. de la Fuente-Nunez, “Low-Cost Biosensor Technologies for Rapid Detection of COVID-19 and Future Pandemics,” ACS Nano, vol. 18, no. 3, pp. 1757-1777, Jan. 2024, doi: 10.1021 / acsnano.3c01629.

[0010] G. Karuppaiah, A. Vashist, M. Nair, M. Veerapandian, and P. Manickam, “Emerging trends in point-of-care biosensing strategies for molecular architectures and antibodies of SARS-CoV-2,” Biosens. Bioelectron. X, vol. 13, p. 100324, 2023, doi: https: / / doi.Org / 10.1016 / j.biosx.2023.100324.

[0011] T. Chaibun et al., “Rapid electrochemical detection of coronavirus SARS-CoV-2,” Nat. Commun, vol. 12, no. 1, p. 802, 2021, doi: 10.1038 / s41467-021-21121-7.

[0012] D. K. Ban et al., “Rapid self-test of unprocessed viruses of SARS-CoV-2 and its variants in saliva by portable wireless graphene biosensor,” Proc. Natl. Acad. Sci., vol. 119, no. 28, p. e2206521119, Jul. 2022, doi: 10.1073 / pnas.2206521119.

[0013] L. F. de Lima, A. L. Ferreira, M. D. T. Torres, W. R. de Araujo, and C. de la Fuente- Nunez, “Minute-scale detection of SARS-CoV-2 using a low-cost biosensor composed of pencil graphite electrodes,” Proc. Natl. Acad. Sci., vol. 118, no. 30, p. e2106724118, Jul. 2021, doi: 10.1073 / pnas.2106724118.

[0014] B. D. Kevadiya et al., “Diagnostics for SARS-CoV-2 infections,” Nat. Mater., vol. 20, no. 5, pp. 593-605, 2021, doi: 10.1038 / s41563-020-00906-z.

[0015] G. Seo et al., “Rapid Detection of COVID-19 Causative Virus (SARS-CoV-2) in Human Nasopharyngeal Swab Specimens Using Field-Effect Transistor-Based Biosensor,” ACS Nano, vol. 14, no. 4, pp. 5135-5142, Apr. 2020, doi: 10.1021 / acsnano.0c02823.

[0016] J. Hu et al., “Rapid genetic screening with high quality factor metasurfaces,” Nat. Commun., vol. 14, no. 1, p. 4486, 2023, doi: 10.1038 / s41467-023-39721-w.

[0017] R. C. Fitzgerald, A. C. Antoniou, L. Fruk, and N. Rosenfeld, “The future of early cancer detection,” Nat. Med., vol. 28, no. 4, pp. 666-677, 2022, doi: 10.1038 / s41591-022- 01746-x.

[0018] V. Fedyuk et al., “Multiplexed, single-molecule, epigenetic analysis of plasma- isolated nucleosomes for cancer diagnostics,” Nat. Biotechnol., vol. 41, no. 2, pp. 212-221, 2023, doi: 10.1038 / s41587-022-01447-3.

[0019] H. Shin et al., “Single test-based diagnosis of multiple cancer types using Exosome- SERS-AI for early stage cancers,” Nat. Commun., vol. 14, no. 1, p. 1644, 2023, doi: 10.1038 / s41467-023-37403-l.

[0020] L. Hao et al., “CRISPR-Cas-amplified urinary biomarkers for multiplexed and portable cancer diagnostics,” Nat. Nanotechnol., vol. 18, no. 7, pp. 798-807, 2023, doi: 10.1038 / s41565-023-01372-9.

[0021] J. D. Cohen et al., “Detection and localization of surgically resectable cancers with a multi-analyte blood test,” Science (80-. )., vol. 359, no. 6378, pp. 926-930, Feb. 2018, doi: 10.1126 / science.aar3247.

[0022] X. Chen et al., “Non-invasive early detection of cancer four years before conventional diagnosis using a blood test,” Nat. Commun., vol. 11, no. 1, p. 3475, 2020, doi:10.1038 / s41467-020- 17316-z.

[0023] T. Rydzak et al., “Metabolic preference assay for rapid diagnosis of bloodstream infections,” Nat. Commun., vol. 13, no. 1, p. 2332, 2022, doi: 10.1038 / s41467-022-30048-6.

[0024] K. L. Kalantar et al., “Integrated host-microbe plasma metagenomics for sepsis diagnosis in a prospective cohort of critically ill adults,” Nat. Microbiol., vol. 7, no. 11, pp. 1805-1816, 2022, doi: 10.1038 / s41564-022-01237-2.

[0025] F. Ellett et al., “Diagnosis of sepsis from a drop of blood by measurement of spontaneous neutrophil motility in a microfluidic assay,” Nat. Biomed. Eng., vol. 2, no. 4, pp. 207-214, 2018, doi: 10.1038 / s41551-018-0208-z.

[0026] S. Shang, G. Chen, Y. Wu, L. Du, and Z. Zhao, “Rapid Diagnosis of Bacterial Sepsis with PCR Amplification and Microarray Hybridization in 16S rRNA Gene,” Pediatr. Res., vol. 58, no. 1, pp. 143-148, 2005, doi: 10.1203 / 01. PDR.0000169580.64191.8B.

[0027] U. Hassan et al., “A point-of-care microfluidic biochip for quantification of CD64 expression from whole blood for sepsis stratification,” Nat. Commun., vol. 8, no. 1, p. 15949, 2017, doi: 10.1038 / ncomms 15949.

[0028] J. Budd et al., “Lateral flow test engineering and lessons learned from COVID-19,” Nat. Rev. Bioeng., vol. 1, no. 1, pp. 13-31, 2023, doi: 10.1038 / s44222-022-00007-3.

[0029] L. Shen et al., “A rapid lateral flow immunoassay strip for detection of SARS-CoV-2 antigen using latex microspheres,” J. Clin. Lab. Anal., vol. 35, no. 12, pp. 1-7, 2021, doi: 10.1002 / jcla.2409L

[0030] S. Bayoumy et al., “Glycovariant-based lateral flow immunoassay to detect ovarian cancer-associated serum CA125,” Commun. Biol., vol. 3, no. 1, p. 460, 2020, doi:10.1038 / s42003 -020-01191 -x.

[0031] Z. Bradley and N. Bhalla, “Point-of-care diagnostics for sepsis using clinical biomarkers and microfluidic technology,” Biosens. Bioelectron., vol. 227, p. 115181, 2023, doi: 10.1016 / j .bios.2023.115181.

[0032] M. M. Serrano et al., “Comparison of commercial lateral flow immunoassays and ELISA for SARS-CoV-2 antibody detection,” J. Clin. Virol., vol. 129, p. 104529, 2020, doi: 10.1016 / j .jcv.2020.104529.

[0033] M. Faisal et al., “Computer-aided National Early Warning Score to predict the risk of sepsis following emergency medical admission to hospital: a model development and external validation study,” Can. Med. Assoc. J., vol. 191, no. 14, p. E382 LP-E389, Apr. 2019, doi: 10.1503 / cmaj.181418.

[0034] K. L. Singampalli, J. Li, and P. B. Lillehoj, “Rapid magneto-enzyme-linked immunosorbent assay for ultrasensitive protein detection,” Anal. Chim. Acta, vol. 1225, p. 340246, 2022, doi: 10.1016 / j. aca.2022.340246.

[0035] A. Kumar et al., “Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock,” Crit. Care Med., vol.34, no. 6, 2006, doi: 10.1097 / 01. CCM.0000217961.75225.E9.

[0036] C. W. Seymour et al., “Delays From First Medical Contact to Antibiotic Administration for Sepsis,” Crit. Care Med., vol. 45, no. 5, 2017, doi:10.1097 / CCM.0000000000002264.

[0037] H. K. Neilson et al., “Diagnostic Delays in Sepsis: Lessons Learned From a Retrospective Study of Canadian Medico-Legal Claims,” Crit. Care Explor., vol. 5, no. 2, 2023, doi: 10.1097 / CCE.0000000000000841.

[0038] A. L. Vijayan et al., “Procalcitonin: a promising diagnostic marker for sepsis and antibiotic therapy,” J. Intensive Care, vol. 5, no. 1, p. 51, 2017, doi: 10.1186 / s40560-017- 0246-8.

[0039] X. E. Cao et al., “A diagnostic platform for rapid, simultaneous quantification of procalcitonin and C-reactive protein in human serum,” eBioMedicine, vol. 76, Feb. 2022, doi: 10.1016 / j.ebiom.2022.103867.

[0040] S. J. Park et al., “PCR-like performance of rapid test with permselective tunable nanotrap,” Nat. Commun., vol. 14, no. 1, p. 1520, 2023, doi: 10.1038 / s41467-023-37018-6.

[0041] R. Gupta et al., “Ultrasensitive lateral-flow assays via plasm onically active antibody- conjugated fluorescent nanoparticles,” Nat. Biomed. Eng., vol. 7, no. 12, pp. 1556-1570, 2023, doi: 10.1038 / s41551-022-01001-1.

[0042] D. Wang et al., “Rapid lateral flow immunoassay for the fluorescence detection of SARS-CoV-2 RNA,” Nat. Biomed. Eng., vol. 4, no. 12, pp. 1150-1158, 2020, doi: 10.1038 / s41551 -020-00655-z.

[0043] K. Behrouzi and L. Lin, “Gold nanoparticle based plasmonic sensing for the detection of SARS-CoV-2 nucleocapsid proteins,” Biosens. Bioelectron., vol. 195, p. 113669, 2022, doi: https: / / doi.Org / 10.1016 / j.bios.2021.113669.

[0044] M. S. Draz et al., “Virus detection using nanoparticles and deep neural network- enabled smartphone system,” Sci. Adv., vol. 6, no. 51, p. eabd5354, Dec. 2023, doi: 10.1126 / sciadv.abd5354.

[0045] K. Kleppe, E. Ohtsuka, R. Kleppe, I. Molineux, and H. G. Khorana, “Studies on polynucleotides: XCVI. Repair replication of short synthetic DNA’s as catalyzed by DNA polymerases,” J. Mol. Biol., vol. 56, no. 2, pp. 341-361, 1971, doi: 10.1016 / 0022- 2836(71)90469-4.

[0046] J. D. Kaunitz, “The Discovery of PCR: ProCuRement of Divine Power,” Dig. Dis. Sci., vol. 60, no. 8, pp. 2230-2231, 2015, doi: 10.1007 / sl0620-015-3747-0.

[0047] C. M. Niemeyer, M. Adler, and R. Wacker, “Detecting antigens by quantitative immuno-PCR,” Nat. Protoc., vol. 2, no. 8, pp. 1918-1930, 2007, doi: 10.1038 / nprot.2007.267.

[0048] S. M. Schubert et al., “Ultra-sensitive protein detection via Single Molecule Arrays towards early stage cancer monitoring,” Sci. Rep., vol. 5, pp. 1-8, 2015, doi:10.1038 / srep 11034.

[0049] J. Yang, Y. Zhang, L. Wu, and Y. Lu, “A coffee-ring effect-based paper sensor chip for the determination of beta-lactoglobulin in foods via a smartphone,” Sensors Actuators B Chem., vol. 374, no. September 2022, p. 132807, 2023, doi: 10.1016 / j.snb.2022.132807.

[0050] Y. Li et al., “A portable visual coffee ring based on carbon dot sensitized lanthanide complex coordination to detect bisphenol A in water,” RSC Adv., vol. 12, no. 12, pp. 7306- 7312, 2022, doi: 10.1039 / d2ra00039c.

[0051] A. S. Rourke-Funderburg, A. B. Walter, B. Carroll, A. Mahadevan-Jansen, and A. K. Locke, “Development of a Low-Cost Paper-Based Platform for Coffee Ring-Assisted SERS,” ACS Omega, vol. 8, no. 37, pp. 33745-33754, 2023, doi: 10.1021 / acsomega.3c03690.

[0052] H. Zargartalebi et al., “Capillary-Assisted Molecular Pendulum Bioanalysis,” J. Am. Chem. Soc., vol. 144, no. 40, pp. 18338-18349, 2022, doi: 10.1021 / jacs.2c06192.

[0053] Y. Hong et al., “Label-free diagnosis for colorectal cancer through coffee ring- assisted surface-enhanced Raman spectroscopy on blood serum,” J. Biophotonics, vol. 13, no.4, p. e201960176, Apr. 2020, doi: https: / / doi.org / 10.1002 / jbio.201960176.

[0054] T. Parsain, A. Tripathi, and A. Tiwari, “Detection of milk adulteration using coffee ring effect and convolutional neural network,” Food Addit. Contam. Part A, vol. 41, no. 7, pp. 730-741, Jul. 2024, doi: 10.1080 / 19440049.2024.2358518.

[0055] X. Li, A. R. Sanderson, S. S. Allen, and R. H. Lahr, “Tap water fingerprinting using a convolutional neural network built from images of the coffee-ring effect,” Analyst, vol. 145, no. 4, pp. 1511-1523, 2020, doi: 10.1039 / C9AN01624D.

[0056] K. Behrouzi and L. Lin, “Double-coffee ring nanoplasmonic effects with convolutional neural learning for Sars-Cov-2 detection,” in 2021 21st InternationalConference on Solid-State Sensors, Actuators and Microsystems (Transducers), 2021, pp. 381-384.

[0057] S. K. Wilson and H.-M. D’ Ambrosio, “Evaporation of Sessile Droplets,” Annu. Rev. Fluid Meeh., vol. 55, no. 1, pp. 481-509, Jan. 2023, doi: 10.1146 / annurev-fluid-031822- 013213.

[0058] X. Shen, C.-M. Ho, and T.-S. Wong, “Minimal Size of Coffee Ring Structure,” J. Phys. Chem. B, vol. 114, no. 16, pp. 5269-5274, Apr. 2010, doi: 10.1021 / jp912190v.

[0059] A. Kumar, D. Sen, A. Das, and J. Bahadur, “Pattern of an Evaporated Colloidal Droplet on a Porous Membrane Dictated by Competitive Processes of Flow and Absorption,” Langmuir, vol. 38, no. 23, pp. 7121-7128, Jun. 2022, doi: 10.1021 / acs.langmuir. lc03442.

[0060] M. Pack, H. Hu, D.-O. Kim, X. Yang, and Y. Sun, “Colloidal Drop Deposition on Porous Substrates: Competition among Particle Motion, Evaporation, and Infiltration,” Langmuir, vol. 31, no. 29, pp. 7953-7961, Jul. 2015, doi: 10.1021 / acs.langmuir.5b01846.

[0061] A. W. Wray, P. S. Wray, B. R. Duffy, and S. K. Wilson, “Contact-line deposits from multiple evaporating droplets,” Phys. Rev. Fluids, vol. 6, no. 7, p. 73604, Jul. 2021, doi: 10.1103 / PhysRevFluids.6.073604.

[0062] P. Ma et al., “A review on high temperature resistant polyimide films: Heterocyclic structures and nanocomposites,” Compos. Commun., vol. 16, pp. 84-93, 2019, doi: 10.1016 / j.coco.2019.08.011.

[0063] J. Piwowarczyk, R. J^drzejewski, D. Moszynski, K. Kwiatkowski, A. Niemczyk, and J. Baranowska, “XPS and FTIR Studies of Polytetrafluoroethylene Thin Films Obtained by Physical Methods.,” Polymers (Basel)., vol. 11, no. 10, Oct. 2019, doi: 10.3390 / polyml 1101629.

[0064] L. A. Mercante, A. Pavinatto, T. S. Pereira, F. L. Migliorini, D. M. dos Santos, and D. S. Correa, “Nanofibers interfaces for biosensing: Design and applications,” Sensors and Actuators Reports, vol. 3, p. 100048, 2021, doi: https: / / doi.Org / 10.1016 / j.snr.2021.100048.

[0065] X. Liu et al., “Plasmonic Coupling of Au Nanoclusters on a Flexible MXene / Graphene Oxide Fiber for Ultrasensitive SERS Sensing,” ACS Sensors, vol. 8, no. 3, pp. 1287-1298, Mar. 2023, doi: 10.1021 / acssensors.2c02808.

[0066] D. Marti, J. Torras, O. Bertran, P. Turon, and C. Aleman, “Temperature effect on the SARS-CoV-2: A molecular dynamics study of the spike homotrimeric glycoprotein,” Comput. Struct. Biotechnol. J., vol. 19, pp. 1848-1862, 2021, doi: 10.1016 / j.csbj.2021.03.037.

[0067] R. Sliz, J. Czajkowski, and T. Fabritius, “Taming the Coffee Ring Effect: Enhanced Thermal Control as a Method for Thin-Film Nanopatteming,” Langmuir, vol. 36, no. 32, pp. 9562-9570, Aug. 2020, doi: 10.102 l / acs.langmuir.0c01560.

[0068] C. Bravin and V. Amendola, “Plasmonic Absorption in Antigen-Induced Aggregated Gold Nanoparticles: Toward a Figure of Merit for Optical Nanosensors,” ACS Appl. Nano Mater., vol. 5, no. 1, pp. 578-586, Jan. 2022, doi: 10.1021 / acsanm. lc03380.

[0069] A. Aubry, D. Y. Lei, S. A. Maier, and J. B. Pendry, “Interaction between Plasmonic Nanoparticles Revisited with Transformation Optics,” Phys. Rev. Lett., vol. 105, no. 23, p. 233901, Nov. 2010, doi: 10.1103 / PhysRevLett.105.233901.

[0070] Y. Hang, A. Wang, and N. Wu, “Plasmonic silver and gold nanoparticles: shape- and structure-modulated plasmonic functionality for point-of-caring sensing, bio-imaging and medical therapy,” Chem. Soc. Rev., vol. 53, no. 6, pp. 2932-2971, 2024, doi: 10.1039 / D3CS00793F.

[0071] Q. Wu, G. Chen, S. Qiu, S. Feng, and D. Lin, “A target-triggered and self-calibration aptasensor based on SERS for precise detection of a prostate cancer biomarker in human blood,” Nanoscale, vol. 13, no. 16, pp. 7574-7582, 2021, doi: 10.1039 / D1NR00480H.

[0072] H.-E. Lee et al., “Amino-acid- and peptide-directed synthesis of chiral plasmonic gold nanoparticles,” Nature, vol. 556, no. 7701, pp. 360-365, 2018, doi: 10.1038 / s41586-018- 0034-1.

[0073] P. Tuersun and X. Han, “Optical absorption analysis and optimization of gold nanoshells,” Appl. Opt., vol. 52, no. 6, pp. 1325-1329, 2013, doi: 10.1364 / A0.52.001325.

[0074] K. Simonyan and A. Zisserman, “Very Deep Convolutional Networks for Large-Scale Image Recognition,” arXiv, 2014.

[0075] O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation,” in Medical Image Computing and Computer- Assisted Intervention - MICC Al, 2015, pp. 234-241, doi: 10.48550 / arXiv.1505.04597.

[0076] L. D. Tran, S. M. Nguyen, and M. Arai, “GAN-Based Noise Model for Denoising Real Images,” Leet. Notes Comput. Sci. (including Subser. Leet. Notes Artif. Intell. Leet. Notes Bioinformatics), vol. 12625 LNCS, pp. 560-572, 2021, doi: 10.1007 / 978-3-030- 69538-5_34.

[0077] P. Isola, J.-Y. Zhu, T. Zhou, and A. A. Efros, “Image-to-image translation with conditional adversarial networks,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 1125-1134, doi: 10.1109 / CVPR.2017.632.

[0078] M. R. Dixon et al., “Carcinoembryonic Antigen and Albumin Predict Survival in Patients With Advanced Colon and Rectal Cancer,” Arch. Surg., vol. 138, no. 9, pp. 962-966, Sep. 2003, doi: 10.1001 / archsurg.138.9.962.

[0079] Y. Zhou, L. Zhang, Y.-H. Xie, and J. Wu, “Advancements in detection of SARS- CoV-2 infection for confronting COVID-19 pandemics,” Lab. Investig., vol. 102, no. 1, pp. 4-13, 2022, doi: 10.1038 / s41374-021-00663-w.

[0080] I. M. Thompson et al., “Prevalence of Prostate Cancer among Men with a Prostate- Specific Antigen Level <4.0 ng per Milliliter,” N. Engl. J. Med., vol. 350, no. 22, pp. 2239- 2246, May 2004, doi: 10.1056 / NEJMoa031918.

[0081] Y. Li et al., “Current status of ctDNA in precision oncology for hepatocellular carcinoma,” J. Exp. Clin. Cancer Res., vol. 40, no. 1, p. 140, 2021, doi: 10.1186 / sl3046-021- 01940-8.

[0082] A. D. Rakic, A. B. Djurisic, J. M. Elazar, and M. L. Majewski, “Optical properties of metallic films for vertical-cavity optoelectronic devices,” Appl. Opt., vol. 37, no. 22, pp. 5271-5283, 1998, doi: 10.1364 / A0.37.005271.

[0083] W. Wang, K. Chen, X. Ma, and J. Guo, “Artificial intelligence reinforced upconversion nanoparticle-based lateral flow assay via transfer learning,” Fundam. Res., vol.3, no. 4, pp. 544-556, 2023, doi: https: / / doi.Org / 10.1016 / j.fmre.2022.03.025.

[0084] A. Borji, “Pros and cons of GAN evaluation measures,” Comput. Vis. Image Underst., vol. 179, no. February 2018, pp. 41-65, 2019, doi: 10.1016 / j.cviu.2018.10.009.

[0085] H. Cho et al., “Electrochemical Impedance-Based Biosensors for the Label-Free Detection of the Nucleocapsid Protein from SARS-CoV-2,” ACS Sensors, vol. 7, no. 6, pp. 1676-1684, 2022, doi: 10.1021 / acssensors.2c00317.

[0086] D. R. P. Nicollete et al., “Enhancing a SARS-CoV-2 nucleocapsid antigen test sensitivity with cost efficient strategy through a cotton intermembrane insertion,” Sci. Rep., vol. 13, no. 1, pp. 1-13, 2023, doi: 10.1038 / s41598-023-31641 -5.

[0087] M. Edalatpour, L. Liu, A.M. Jacobi, K.F. Eid, A.D. Sommers, Applied Energy, vol. 222, pp. 967-992, 2018, doi: 10.1016 / j.apenergy.2018.03.178.

[0088] Kenry and C. T. Lim, Progress Polymer Sci., vol. 70, pp. 1-17, 2017, doi : 10.1016 / j .progpoly sci .2017.03.002.

[0149] It is understood that the examples and embodiments described herein are for illustrative purposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to be included within the spirit and purview of this application and scope of the appended claims. All publications, patents, and patent applications cited herein are hereby incorporated by reference in their entirety for all purposes.

Claims

CLAIMS1. A method of indicating whether a sample contains a soluble analyte of interest, said method comprising the following steps, in the following order:(a) obtaining a nanofibrous or mesofibrous membrane disposed on a solid support,(b) depositing a drop of an aqueous solution containing said sample on a first position on said nanofibrous or mesofibrous membrane,(c) allowing said drop of said sample to evaporate, thereby creating an area where said drop of said sample evaporated, said area having an interior portion and an edge around at least some of said interior portion, which edge bears residual material from said drop of said sample,(d) depositing at a second position on said nanofibrous or mesofibrous membrane, a drop containing gold nanoshells, which gold nanoshells have been functionalized to bind to said analyte of interest on said nanofibrous or mesofibrous membrane, which second position is disposed on said nanofibrous or mesofibrous membrane to allow (1) some of said drop containing said gold nanoshells to overlap with said edge bearing said residual material from said drop of said sample, thereby creating a first zone, and (2) some of said drop containing said gold nanoshells to overlap with some of said interior portion of said area, thereby creating a second zone,(e) allowing said gold nanoshells to interact with any of said analyte of interest present in said first detection zone and with themselves, thereby forming a first pattern in said first zone and a second pattern in said second zone, thereby forming a combination of said first pattern and said second pattern, whereby said combination of said first pattern and said second pattern indicates whether said analyte of interest is present in said sample or is not present in said sample.

2. The method of claim 1, wherein a smooth pattern in said first zone indicates that said analyte is present in said sample.

3. The method of claim 1, wherein a lumpy pattern in said first zone indicates that said analyte is not present in said sample.

4. The method of claim 1, wherein said nanofibrous or mesofibrous membrane is intermediate between being hydrophilic and being hydrophobic and is wettable.

5. The method of claim 4, wherein said nanofibrous or mesofibrous membrane has been rendered intermediate between being hydrophilic and being hydrophobic and is wettable, by being heated to 80 °C ±15 °C for a period of time and then cooled.

6. The method of claim 5, wherein said nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, of cellulose nitrite, or of nylon.

7. The method of claim 4, wherein said nanofibrous or mesofibrous membrane is of poly tetrafluoroethyl ene .

8. The method of claim 5, wherein said nanofibrous or mesofibrous membrane is of polytetrafluoroethylene9. The method of claim 1, wherein said nanofibrous or mesofibrous membrane is nanofibrous.

10. The method of claim 1, wherein said sample drop has a volume of 3-10 pL.

11. The method of claim 1, wherein said drop containing said gold nanoshells has a volume of 1-5 pL.

12. The method of claim 1, wherein said gold nanoshells are functionalized to bind said analyte of interest by being attached to an antibody that specifically binds said analyte of interest.

13. The method of claim 1, wherein said analyte of interest is a protein.

14. The method of claim 13, wherein said protein is a protein of a virus, a protein of a bacterium, a biomarker for a cancer, or a biomarker for a disease or a disease state.

15. The method of claim 1, wherein step (c) is performed at a temperature of 80 °C ± 10 °C.

16. The method of claim 1, wherein step (c) is performed at a temperature of 80 °C ±5 °C.

17. The method of claim 1, wherein said determining whether there is a smooth pattern in the first zone is made by visual inspection.

18. The method of claim 1, wherein said determining whether or not said smooth pattern exists in the first zone is made by a device programmed to identify and compare a pattern in said first zone to a dataset of smooth patterns made by one or more analytes of interest bound by gold nanoshells functionalized to bind said one or more analytes of interest.

19. The method of claim 18, further wherein an amount of the one or more analytes of interest is quantified by determining an intensity level along a central line through said edge around at least some of said interior portion.

20. The method of claim 18, wherein the dataset of smooth patterns made by one or more analytes of interest bound by gold nanoshells functionalized to bind said one or more analytes of interest is derived from training a convolutional neural network (CNN) using a plurality of test samples of known concentrations of the one or more analytes of interest bound by gold nanoshells and determining whether there is the smooth pattern in the first zone comprises processing an image including the first zone by the CNN to yield a positive result or a negative result for the smooth pattern.

21. The method of claim 20, further wherein the image yielding the positive result for the smooth pattern is further processed by a generator guided by a discriminator providing a conditional generative adversarial network (C-GAN) to segment the first zone to identify a central line through said edge around at least some of said interior portion and an intensity level along the central line is derived to quantify an amount of the one or more analytes of interest.

22. The method of claim 21, wherein a length of the central line is normalized for all processed test images.

23. A method for making a device for determining whether a sample contains an analyte of interest, said method comprising(a) providing a solid support having a lower surface and an upper surface,(b) providing a thermally stable film with a lower surface and an adhesive upper surface,(c) adhering said upper surface of said solid support to said lower surface of said thermally stable film,(d) layering a hydrophilic nanofibrous or mesofibrous membrane onto adhesive upper surface of said thermally stable film, thereby creating an assembled device, and(e) heating said assembled device to 80 °C ± 15 °C for a time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane to intermediate between hydrophilic and hydrophobic, and then allowing said nanofibrous or mesofibrous membrane it to cool to a temperature below 80 °C, thereby forming said device for determining whether a sample contains an analyte of interest.

24. The method of claim 23, wherein said adhesive upper surface of said thermally stable film comprises silicon nanoparticles.

25. The method of claim 23, further comprising step (d)’ between steps (d) and (e): pressing said hydrophilic nanofibrous or mesofibrous membrane onto said adhesive upper surface of said thermally stable film after said hydrophilic nanofibrous or mesofibrous membrane has been layered onto said adhesive upper surface.

26. The method of claim 23, wherein said hydrophilic nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, cellulose nitrite, or nylon.

27. The method of claim 26, wherein said hydrophilic nanofibrous or mesofibrous membrane is of hydrophilic polytetrafluoroethylene.

28. The method of claim 27, wherein said nanofibrous or mesofibrous membrane of hydrophilic polytetrafluoroethylene is nanofibrous.

29. The method of claim 23, wherein said thermally stable film is of polyimide.

30. The method of claim 23, wherein said heating of step (e) is to a temperature of 80 °C ± 10 °C.

31. The method of claim 23, wherein said heating of step (e) is to a temperature of 80 °C ± 5 °C.

32. The method of claim 23, wherein said heating of step (e) is to a temperature of 80 °C.

33. The method of claim 23, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 2 minutes and 240 hours.

34. The method of claim 33, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 5 minutes and 48 hours.

35. The method of claim 33, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 10 minutes and 12 hours.

36. The method of claim 33, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 10 minutes and 4 hours.

37. The method of claim 33, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 15 minutes and 2 hours.

38. The method of claim 33, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 15 minutes and 1 hour.

39. The method of claim 33, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is 30 minutes ± 10 minutes.

40. The method of claim 33, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is 30 minutes ± 5 minutes.

41. The method of claim 23, wherein said solid support is of glass.

42. The method of claim 23, wherein said solid support is of a plastic that does not melt or soften when heated to 100 °C.

43. The method of claim 42, wherein said plastic that does not melt or soften when heated to 100 °C is a thermosetting polymer or a thermoplastic plastic.

44. The method of claim 43, wherein said thermosetting polymer or thermoplastic is a polyether ether ketone, polytetrafluoroethylene, polybenzimidazole, or polydicyclopentadiene.

45. The method of claim 23, wherein said solid support is of metal.

46. The method of claim 45, wherein said metal is of stainless steel.

47. A device for determining whether a sample contains an analyte of interest, said device comprising, in said order:(a) a solid support having an upper surface having an area,(b) a thermally stable film having (1) a lower surface adhered to said upper surface of said solid support, and (2) an upper surface, and(c) a nanofibrous or mesofibrous membrane adhered to said upper surface of said thermally stable film.

48. The device of claim 47, wherein said thermally stable film is adhered to said upper surface of said solid support by an adhesive.

49. The device of claim 47, wherein said thermally stable film is adhered to said upper surface of said solid support by double-sided tape.

50. The device of claim 47, wherein said nanofibrous or mesofibrous membrane is adhered to said upper surface of said thermally stable film film by an adhesive.

51. The device of claim 47, wherein said nanofibrous or mesofibrous membrane is adhered to said upper surface of said thermally stable film by double-sided tape.

52. The device of claim 47, wherein said nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, cellulose nitrite, or nylon.

53. The device of claim 47, wherein said nanofibrous or mesofibrous membrane is of poly tetrafluoroethyl ene .

54. The device of claim 47, wherein said nanofibrous or mesofibrous membrane is nanofibrous.

55. The device of claim 47, wherein said thermally stable film is of polyimide.

56. The device of claim 47, wherein said solid support is of glass.

57. The method of claim 47, wherein said solid support is of a plastic that does not melt or soften when heated to 100 °C.

58. The method of claim 57, wherein said plastic that does not melt or soften when heated to 100 °C is a thermosetting polymer or a thermoplastic plastic.

59. The method of claim 58, wherein said thermosetting polymer or thermoplastic is a polyether ether ketone, polytetrafluoroethylene, polybenzimidazole, or polydicyclopentadiene.

60. The method of claim 47, wherein said solid support is of metal.

61. The method of claim 60, wherein said metal is stainless steel.

62. The device of claim 47, wherein said nanofibrous or mesofibrous membrane is surrounded by a liquid-impermeable barrier.

63. The device of claim 47, further comprising at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample.

64. The device of claim 47, further comprising at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop.

65. The device of claim 47, wherein (1) said solid support is a thermosetting polymer or a thermoplastic, (2) said thermally stable film is polyimide and is adhered to said upper surface of said solid support by an adhesive, (3) said nanofibrous or mesofibrous membrane is nanofibrous, is made of polytetrafluoroethylene, and is surrounded by a liquid-impermeable barrier, and (4) comprises at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample.

66. The device of claim 65, further comprising (5) at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop.

67. The device of claim 47, wherein (1) said solid support is a thermosetting polymer or a thermoplastic, (2) said thermally stable film is polyimide and is adhered to said upper surface of said solid support by a double-sided tape, (3) said nanofibrous or mesofibrous membrane is nanofibrous, is made of polytetrafluoroethylene, and is surrounded by a liquid-impermeable barrier, and (4) comprises at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample.

68. The device of claim 67, further comprising (5) at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop.

69. A method for making a device for determining whether a sample contains an analyte of interest, said method comprising(a) providing a solid support having a lower surface and an upper surface,(b) providing a thermally stable film with a lower surface and an adhesive upper surface,(c) adhering said upper surface of said solid support to said lower surface of said thermally stable film, and(d) layering a nanofibrous or mesofibrous membrane that has been treated to be wettable but to be intermediate between hydrophilic and hydrophobic, onto adhesive upper surface of said thermally stable film, thereby forming said device for determining whether a sample contains an analyte of interest.

70. The method of claim 69, wherein said adhesive upper surface of said thermally stable film comprises silicon nanoparticles.

71. The method of claim 69, further comprising step (e): pressing said hydrophilic nanofibrous or mesofibrous membrane onto said adhesive upper surface of said thermally stable film after said nanofibrous or mesofibrous membrane has been layered onto said adhesive upper surface.

72. The method of claim 69, wherein said nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, cellulose nitrite, or nylon.

73. The method of claim 72, wherein said nanofibrous or mesofibrous membrane is of poly tetrafluoroethyl ene .

74. The method of claim 73, wherein said nanofibrous or mesofibrous polytetrafluoroethylene is nanofibrous.

75. The method of claim 69, wherein said thermally stable film is of polyimide.

76. The method of claim 69, wherein said treatment of said nanofibrous or mesofibrous membrane is by heating said nanofibrous or mesofibrous membrane to 80 °C ± 15 °C for a time sufficient to change a hydrophilic nanofibrous or mesofibrous membrane from hydrophilic to intermediate between hydrophilic and hydrophobic or to hydrophobic, but wettable.

77. The method of claim 76, wherein said heating is to 80 °C ± 10 °C.

78. The method of claim 76, wherein said heating is to 80 °C ± 5 °C.

79. The method of claim 76, wherein said heating is to 80 °C ± 2 °C.

80. The method of claim 76, wherein said time period sufficient to change a hydrophilic nanofibrous or mesofibrous membrane to intermediate between hydrophilic and hydrophobic is between 2 minutes and 240 hours.

81. The method of claim 76, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 10 minutes and 24 hours.

82. The method of claim 76, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 10 minutes and 4 hours.

83. The method of claim 76, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is between 15 minutes and 1 hour.

84. The method of claim 76, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is 30 minutes ± 10 minutes.

85. The method of claim 76, wherein said time period sufficient to modify said hydrophilic property of said hydrophilic nanofibrous or mesofibrous membrane is 30 minutes ± 5 minutes.

86. The method of claim 69, wherein said treatment of said nanofibrous or mesofibrous membrane is by chemical vapor deposition, by plasma treatment, by grafting hydrophobic polymers onto the nanofibers or mesofibers of the membrane, by spin-coating or dip coating,by a sol-gel coating, by depositing one or more self-assembled monolayers of hydrophobic molecules on said nanofibrous or mesofibrous membrane, by deposition of hydrophobic nanoparticles on said nanofibrous or mesofibrous membrane, by thermal or ultraviolet crosslinking of hydrophobic monomers on said nanofibrous or mesofibrous membrane, by depositing alternating layers of oppositely-charged hydrophobic polymers or polyelectrolytes on said nanofibrous or mesofibrous membrane, or by in situ polymerization of hydrophobic polymers on said nanofibrous or mesofibrous membrane.

88. The method of claim 69, wherein said solid support is of glass.

89. The method of claim 69, wherein said solid support is of a plastic that does not melt or soften when heated to 100 °C.

90. The method of claim 89, wherein said plastic that does not melt or soften when heated to 100 °C is a thermosetting polymer or a thermoplastic plastic.

91. The method of claim 90, wherein said thermosetting polymer or thermoplastic is a polyether ether ketone, polytetrafluoroethylene, polybenzimidazole, or polydicyclopentadiene.

92. The method of claim 69, wherein said solid support is of metal.

93. The method of claim 92, wherein said metal is of stainless steel.

94. A device for determining whether a sample contains an analyte of interest, said device comprising, in said order:(a) a solid support having an upper surface having an area,(b) a thermally stable film having (1) a lower surface adhered to said upper surface of said solid support, and (2) an upper surface, and(c) a wettable nanofibrous or mesofibrous membrane that is intermediate between being hydrophilic and being hydrophobic, which wettable nanofibrous or mesofibrous membrane is disposed on said upper surface of said thermally stable film.

95. The device of claim 94, wherein said thermally stable film is adhered to said upper surface of said solid support by an adhesive.

96. The device of claim 94, wherein said thermally stable film is adhered to said upper surface of said solid support by double-sided tape.

97. The device of claim 94, wherein said nanofibrous or mesofibrous membrane is adhered to said upper surface of said thermally stable film film by an adhesive.

98. The device of claim 94, wherein said nanofibrous or mesofibrous membrane is adhered to said upper surface of said thermally stable film by double-sided tape.

99. The device of claim 94, wherein said nanofibrous or mesofibrous membrane is of polytetrafluoroethylene, cellulose nitrite, or nylon.

100. The device of claim 94, wherein said nanofibrous or mesofibrous membrane is of poly tetrafluoroethyl ene .

101. The device of claim 94, wherein said nanofibrous or mesofibrous membrane is nanofibrous.

102. The device of claim 94, wherein said thermally stable film is of polyimide.

103. The device of claim 94, wherein said solid support is of glass.

104. The method of claim 94 wherein said solid support is of a plastic that does not melt or soften when heated to 100 °C.

105. The method of claim 104, wherein said plastic that does not melt or soften when heated to 100 °C is a thermosetting polymer or a thermoplastic plastic.

106. The method of claim 105, wherein said thermosetting polymer or thermoplastic is a polyether ether ketone, polytetrafluoroethylene, polybenzimidazole, or polydicyclopentadiene107. The method of claim 94, wherein said solid support is of metal.

108. The device of claim 94, wherein said nanofibrous or mesofibrous membrane is surrounded by a liquid-impermeable barrier.

109. The device of claim 94, further comprising at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample.

110. The device of claim 94, further comprising at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop.

111. The device of claim 94, further comprising a cover over said membrane.

110. The device of claim 94, wherein (1) said solid support is a thermosetting polymer or a thermoplastic, (2) said thermally stable film is polyimide and is adhered to said upper surface of said solid support by an adhesive, (3) said nanofibrous or mesofibrous membrane is nanofibrous, is made of polytetrafluoroethylene, and is surrounded by a liquid-impermeable barrier, and (4) comprises at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample.

111. The device of claim 110, further comprising (5) at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop.

112. The device of claim 110, further comprising a removable or slidable cover over said membrane.

113. The device of claim 94, wherein (1) said solid support is a thermosetting polymer or a thermoplastic, (2) said thermally stable film is polyimide and is adhered to said upper surface of said solid support by a double-sided tape, (3) said nanofibrous or mesofibrous membrane is nanofibrous, is made of polytetrafluoroethylene, and is surrounded by a liquid- impermeable barrier, and (4) comprises at least one mark, a physical guide, or both, showing where on said membrane to place a drop of a sample.

114. The device of claim 113, further comprising (5) at least one mark, a physical guide, or both, showing where on said membrane to place a plasmonic drop.

115. The device of claim 113, further comprising a removable or slidable cover over said membrane.

Images