I. Core Research Summary

• A systematic comparison of four modification strategies identifies gate modification as the optimal approach.

Four distinct systems were tested: an electrolyte‑free GOx system, a PEDOT‑channel‑immobilized GOx system, a sputtered Pt‑gate‑immobilized GOx system, and an aerosol‑printed nano‑Pt‑gate‑immobilized GOx system. Among these, the aerosol‑printed nano‑Pt‑gate configuration strikes an optimal balance between an ultra‑wide detection range and dual‑range high sensitivity, making it the only architecture well‑suited for sweat‑based sensing.

• Aerosol‑derived nano‑platinum with a highly rough surface significantly enhances enzyme immobilization capacity.

The aerosol‑printed Pt thin film exhibits a roughness of 13.8 nm, which is 16 times that of sputtered platinum (0.852 nm), enabling a larger specific surface area to accommodate more GOx and resulting in a markedly enhanced signal response to low‑concentration glucose; the detection range in PBS buffer spans 100 nM to 50 mM.

• Perfectly matches the human sweat concentration range.

ISO‑standard artificial sweat testing, with a linear detection range of 0.1–10 mM, fully covering the physiological glucose concentration range in both healthy individuals and patients with diabetes; sensitivity is 0.068 NR per tenfold change at low concentrations and 0.384 NR per tenfold change at high concentrations, with a linear fit R² = 1.

• All-aerosol multilayer monolithic fabrication,无需光刻 / 真空溅射

A single system switches between ultrasonic and pneumatic atomization modules, sequentially depositing silver traces, PEDOT channels, Pt sensing grids, and an insulating passivation layer, followed by in-situ laser sintering of nano‑platinum, thereby streamlining the fabrication process of microfluidic sensor chips and significantly reducing prototyping and iteration costs.

• The device exhibits excellent stability, with enzyme activity retained over the long term.

The physically adsorbed chitosan–GOx composite membrane exhibits excellent adhesion and maintains high signal reproducibility across multiple concentration-gradient tests; gate modification does not compromise the conductive PEDOT channel, and the transistor’s transconductance shows no significant degradation.

II. Core Working Principle and Complete Preparation Procedure of Aerosol Spraying

(1) Complete OECT Aerosol Printing Process Flow

• Kapton flexible PI substrate is pre-treated, silver nano‑ink is diluted by ultrasonic atomization, source–drain conductive traces are printed, and the structure is cured and formed in a 130°C oven.
• Ultrasonic atomization of PEDOT:PSS functional conductive ink for CNC-printed transistor channel layers;
• Pneumatic atomization UV‑PDMS insulating ink is used to print the passivation protective layer in a partitioned manner, leaving only the channel and gate windows exposed.
• Ultrasonic atomization was used to prepare Pt nanoparticle ink, which was then deposited via jet printing to form planar gate electrodes; subsequent rapid sintering with an 830 nm laser yielded a highly rough nanostructured platinum thin film.
• A chitosan–GOx mixed solution is drop‑coated onto the aerosol Pt‑grid surface and dried at low temperature, thereby achieving biofunctionalization and assembling an electrochemical detection platform.

(II) Five Key Advantages of Aerosol-Based Processes (Compared with Lithography, Magnetron Sputtering, and Inkjet Printing)

• Multi-material dual-atomization–compatible integrated molding

Ultrasonic processing is compatible with low-viscosity PEDOT, Pt, and silver inks, while pneumatic processing works well with high-viscosity PDMS insulating adhesives. A single piece of equipment can fabricate the complete structure—comprising the conductive layer, the insulating layer, and the sensing grid—eliminating the need for multiple vacuum‑processing systems.

• One-step direct writing of high-roughness nano-platinum gates

The sputtered platinum surface is exceptionally smooth and exhibits low enzyme loading; aerosol‑based layer‑by‑layer deposition of nanoparticles naturally forms a porous, rough film, enhancing biomolecule adsorption capacity without the need for additional etching.

• Digital, maskless, and freely adjustable gate dimensions.

CAD enables direct modification of gate area and array layout, allowing rapid iterative chip design for diverse samples such as sweat, blood, and cell culture media, thereby shortening the R&D cycle by 70%.

• Low-temperature laser sintering, with no deformation of the flexible substrate.

The platinum grid is subjected to pulsed laser–induced transient thermal treatment, eliminating the need for prolonged high‑temperature baking; this prevents warping and embrittlement of the thin PI flexible substrate, making it well suited for wearable flexible sensor patches.

• Consumables are deposited on demand, delivering a significant cost advantage in mass production.

The precious metal Pt is deposited only on the sensing area, avoiding wasteful full‑area sputtering; both small‑batch laboratory prototypes and large‑scale diagnostic chips can be fabricated at low cost.

III. Five Core Technological Innovations of This Study

Innovation 1: Aerosol‑derived nano‑platinum grids as the optimal bio‑modified substrate

Innovative 2 High-Roughness Aerosol-Deposited Platinum Thin-Film Sensing Interface

Innovation 3: A Complete Manufacturing Process for All-Printed Flexible OECTs

Innovation 4: A Bilinear Wide-Range Sweat Detection System

Innovative chitosan-assisted physical immobilization–based bio‑modification process

IV. Existing Process Challenges and Corresponding Solutions for Overcoming Them

Challenge 1: Conventional sputtered platinum surfaces are smooth, resulting in low enzyme loading and no significant signal for low-concentration glucose in sweat.

Challenge 2: Immobilizing enzymes in the PEDOT channel can degrade the conductive polymer, leading to a significant drop in transistor sensitivity.

Challenge 3: The combined lithography and sputtering process is highly complex, with a prototyping cycle lasting several days.

Challenge 4: Most printed electrodes can only detect high-concentration blood samples, leaving the sweat concentration range unaddressed.

Challenge 5: Enzymes tend to desorb in the electrolyte, resulting in poor sensor reproducibility.

V. Future Diversified Industrial Application Scenarios

1. Wearable non-invasive sweat glucose monitoring patch

2. Home-use rapid diabetes screening sensor chip

3. High-throughput platform for extracellular cellular glucose metabolism assays

4. Real-time monitoring device for human exercise metabolism

5. Point-of-care portable POCT biochemical testing devices

Figure 1. Glucose sensor based on printed OECTs developed in this study. (a) Micrograph of an unfunctionalized OECT printed on a Kapton substrate, showing the printed silver source and drain contacts, the PEDOT:PSS channel, and the as‑received in‑plane platinum gate; the metal interconnects were passivated with a PDMS mask to prevent short circuits during measurement. (b) Micrograph of a fully printed OECT on a Kapton substrate, with the channel functionalized by graphene oxide (GOx) for glucose sensing. (c) Micrograph of an OECT printed on a Kapton substrate, featuring a GOx‑functionalized printed platinum gate for glucose sensing. (d) Micrograph of an OECT printed on a SiO₂/Si substrate, with a sputtered platinum gate functionalized by GOx.

Figure 2. (a) shows the typical output characteristics of a printed OECT measured in PBS electrolyte at various gate voltages (0 to 1 V); (b) presents the transfer characteristics of an unfunctionalized printed OECT over a range of drain voltages (0.1 to 0.8 V); (c) displays the transconductance (gm) curve extracted for drain voltages between 0.1 and 0.4 V; (d) indicates that the drain current (ID) varies only slightly with increasing glucose concentration, suggesting that the OECT has not undergone any GOx functionalization.

Figure 3. ID responses of printed OECTs to the addition of glucose solutions under a constant bias (VG = 0.6 V, VD = 0.2 V). (a) ID response of an unfunctionalized OECT with a printed platinum gate and a suspended GOx‑modified channel (inset: ID responses at various glucose concentrations, from 100 nM to 100 mM); (b) ID response of an OECT with a GOx‑modified channel and a printed platinum gate (inset: ID responses at different glucose concentrations, from 1 mM to 500 mM); (c) ID response of an OECT with a GOx‑modified printed platinum gate (inset: ID responses at various glucose concentrations, from 100 nM to 10 mM); (d) ID response of a printed OECT with a functionalized sputtered platinum gate (inset: ID responses at different glucose concentrations, from 1 mM to 2 mM).

Figure 4. Normalized drain current (NR) responses of the OECT sensor to logarithmically spaced glucose concentrations. (a) For the OECT sensor employing a floating GOx enzyme and a printed platinum gate, the average NR values reveal two linear regions: a slope of 0.020 NR/decade in the range from 100 nM to 153 mM, and a slope of 0.283 NR/decade in the range from 153 mM to 50 mM. (b) For the OECT sensor with functionalized channels and a printed platinum gate, the average NR values show two linear regions: a slope of 0.019 NR/decade from 10 mM to 413 mM, and a slope of 0.254 NR/decade from 413 mM to 2 mM. (c) For the OECT sensor with a functionalized printed platinum gate, the average NR values indicate two linear regions: a slope of 0.022 NR/decade from 100 nM to 250 mM, and a slope of 0.255 NR/decade from 250 mM to 50 mM. (d) For the OECT sensor with a functionalized sputtered platinum gate, the average NR values display two linear regions: a slope of 0.015 NR/decade from 1 mM to 1.3 mM, and a slope of 0.314 NR/decade from 1.3 mM to 50 mM. (e) The average NR values obtained using all four functionalization methods are plotted on the same graph, with error bars representing the standard error of all data points.

Table 1. Comparison of Detection Ranges and Sensitivities for Different Types of OECT-Based Glucose Sensors

Figure 5. (a) shows the response curve of a printed OECT with a glucose oxidase (GOx)-functionalized printed platinum gate upon addition of a glucose solution dissolved in an artificial sweat buffer; (b) presents the average response values (NR) of printed OECTs with a functionalized printed platinum gate to logarithmically spaced concentrations of sweat‑borne glucose. Two linear regions were observed over the concentration ranges of 100–500 mM and 500 mM–10 mM, with slopes of 0.068 NR per decade and 0.384 NR per decade, respectively. Error bars indicate the standard error.

Table 2. Overview of Glucose Biosensors for Glucose Analysis in Artificial or Human Sweat