1. Introduction
Corrosion is an omnipresent problem affecting the integrity of metallic structures and is the most common cause of metal deterioration and failure. Effective prevention and maintenance are of paramount importance to maintaining structural health, safety, and durability [
1]. Corrosion in a structure can manifest itself in different forms: from an easily detectable uniform and exfoliating surface to a very hard-to-detect intragranular and pitting form. Furthermore, corrosion can be caused by a variety of physical and environmental factors. Physical factors include cracks and crevices, static and dynamic stress, and coating degradation, whereas environmental factors include exposure to high levels of humidity, temperatures, and salinity [
2].
Corrosion monitoring in aircraft structures is an important safety measure due to the nature of in-service conditions that provide auspicious environments for corrosion onset and progression. Currently, aircraft corrosion monitoring and management programs are based on “find and fix” protocols, where corrosion detection relies on visual inspection and mass loss sensors, such as corrosion coupons, electrical resistance, corrosion potential, ultrasonic and acoustic emission, etc. [
3]. While such methods provide valuable information about the presence and progression of structural damage, they fail to differentiate between corrosion and cracks, and do not specify the form nor the cause of the detected corrosion. They may detect corrosion damage, but they fail to identify early stages of corrosion and factors responsible for corrosion events, namely changes in pH, metal ions (Fe
2+, Co
2+, Al
3+, Cu
2+), changes in humidity, and chloride ions (Cl
−) [
1,
2]. Reliable structural health monitoring requires continuous and early on detection of environmental factors responsible for corrosion events using in situ, non-destructive sensing mechanisms able to detect minute changes in levels and concentrations of the aforementioned environmental conditions. Frequent exposure to chlorides and humidity provides the essential electrochemical conditions favorable to corrosion events [
4]. Chloride ion concentration thresholds for corrosion initiation has been identified to be within the range of 100 ppm and 3500 ppm depending on the type of metal alloy, temperature, and pH [
5,
6]. As such, there is an urgent need for Cl
− ion sensing in the aviation industry. Timely detection of chloride ions below corrosion initiation thresholds would allow for the implementation of preventive measures and avoid severe and costly damage. Most Cl
− ion detectors predominantly use silicon-based potentiometric sensors, such as ion selective electrodes (ISE) and electrochemical field effect transistors (ChemFET), which rely on ionophore membranes and electrolytic solutions [
7,
8]. Although such sensors have been able to reach detection limits as low as 3.5 ppb [
7], their implementation in environmental sensing, especially in aircraft corrosion monitoring, remains challenging. The main challenge to integrating potentiometric sensors in corrosion monitoring is to maintain the environmental stability of the aqueous/electrolytic solution or the membrane [
8,
9].
Organic thin film transistor (OTFT) sensors have the potential to fill in the gap in corrosion prevention strategy, improve corrosion management approaches, and safeguard the integrity of metallic structures while reducing repair and maintenance costs. OTFTs rely on carbon-based organic semiconductors and functional surface layers to perform as chemical sensors capable of targeting and detecting specific environmental conditions and chemical species responsible for and indicative of corrosion onset. Compared to other emerging sensing techniques, such as fiber optics and electrochemical sensors, OTFT chemical sensors are more advantageous in terms of robustness, small footprint, low-cost, and energy-efficient fabrication, which is adaptable for large scale production. Furthermore, they exhibit higher sensitivity, faster response time, and low power consumption [
10].
Organic semiconductors have been the focus of intensive interest in both the academic and industrial sectors, owing largely to their abundance, solution processability, large-area manufacturing capabilities, mechanical flexibility, and comparable electronic properties compared to their inorganic counterparts. While p-type organic semiconductors have had enormous success in various applications, n-type semiconductor progress has been substantially lagging owing to its environmental instability and low carrier mobility stemming from electron trapping. The performance of n-type OTFTs can be improved by both molecular design and device optimization in term of deposition techniques, passivation, and encapsulation [
11]. As a chemically sensitive functional layer, modified graphene has attracted significant attention in recent years owing to its high electrical and thermal conductivity, high carrier mobility, and mechanical flexibility [
12]. The solution processability and ease of chemical functionalization of graphene allows for low-cost, large-scale fabrication of thin films with high sensitivity to target analytes.
There have been little to no attempts thus far at designing chloride ion sensors using thin film sensors for structural health monitoring. Our reported work is motivated by this scarcity of reported chloride ion sensors designed specifically for monitoring aircraft corrosion, and utilizing advantages offered by graphene-derived materials in corrosion monitoring. In our work, OTFT humidity and chloride sensors are fabricated using reduced graphene oxide (rGO) and hydrogenated graphene coatings, respectively. rGO is the product of the reduction of graphene oxide, by which GO is partially reduced by the controlled removal of oxygen-containing groups from the surface of GO layers using simple treatments, namely chemical, thermal, and photochemical reduction [
13]. This process restores the inherent electrical conductivity of graphene while maintaining the solubility in water and in organic solvents and chemical reactivity characteristic to GO [
14]. The sensitivity of rGO to humidity stems from the presence of functional groups at the edges of the graphene layers that bind with water molecules and enhance the conductivity of rGO [
15]. Additionally, chemically functionalized hydrogenated graphene coating is used for Cl
− ions sensing. There are conceptually two main methods for the synthesis of hydrogenated graphene, namely gas phase approaches, which include plasma hydrogenation and thermal cracking, and wet chemical approaches, such as Birch and electrochemical reduction [
16,
17]. The hydrogen atoms disrupt the structure of the graphene matrix leading to the hybridization of sp
2 bonds into sp
3 and the formation of C-H bonds [
17]. Although the chemical affinity of hydrogenated graphene has not been considerably investigated, studies suggest that the C-H bonds in hydrogenated graphene may exhibit chemical reactivity and display selectivity toward halide ions, most specifically to Cl
− ions [
18,
19]. The rGO-coated OTFT humidity sensor is designed for qualitative measurements of surface moisture levels, and it is used as a peripheral sensor to ensure the hydrogenated graphene-coated OTFT Cl
− ion sensor operates in relatively dry conditions.
2. Materials and Methods
Fullerene C60 (sublimed, 99.9% purity), 1,2 dichlorobenzene, Poly(vinyl alcohol) (PVA), poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS), Sodium:Pottasium alloy (Na-K (1:3)), ethylenediamine, fluorographite, isopropanol, ethanol, and n-hexane were obtained from Sigma Aldrich. Graphene nanoplatelets (GnP) and rGO were acquired from Kennedy Labs and Poly(methyl methacrylate) (PMMA) from DuPont. De-ionized (DI) water was supplied by Carleton University microfabrication lab.
2.1. Preparation of OTFT Humidity and Chloride Ions Sensor
We fabricated a top gate, bottom contact OTFT configuration using fullerene (C
60) as the active material and functionalized graphene as the sensory layer, as illustrated in
Figure 1b,c.
The OTFT sensors were solution processed using layer-by-layer deposition performed under a nitrogen atmosphere in a spin-coating apparatus. The structure materials were spin coated on a passivated silicon substrate on which chromium source and drain electrodes were patterned via UV lithography and negative lift-off. A common core OTFT device structure was fabricated by depositing 0.9% wt. C
60 in 1,2-dichlorobenzene, followed by a dielectric stack made of PVA (125 nm) and PMMA (200 nm) and completed by a conductive multilayer composed of PEDOT:PSS (1 μm) and PEDOT:PSS doped with graphene (Gr:PEDOT:PSS) (500 nm). Sensitivity toward humidity and chloride ions was achieved by replacing Gr:PEDOT:PSS by 0.1% suspensions of rGO and hydrogenated graphene in PEDOT:PSS, respectively. Finally, chromium gate metal contacts were deposited through a shadow mask using thermal vapour deposition, as shown in
Figure 1a. The device patterning process yielded multiple devices with channel length and width ranging from 10 μm to 25 μm and 100 μm to 1000 μm, respectively, and were fabricated under the same conditions. This allowed for the investigation of device scalability and reproducibility.
2.2. Wet Chemical Synthesis of Hydrogenated Graphene
Hydrogenated graphene was chemically synthesized in-house using a modified Birch method [
20]. Briefly, 2.33 g of 1:3 Na-K alloy and 5 mL of ethylenediamine were added into a 50 mL three-necked round-bottomed flask under a nitrogen atmosphere and stirred until a blue mixture was obtained. Meanwhile, a solution of 43 mg of fluorographite in 5 mL of ethylenediamine was sonicated for 30 s. The fluorographite-ethylenediamine suspensions were then added to the flask through a gastight syringe and stirred under a nitrogen atmosphere at room temperature for 2 h. Isopropanol (10 mL) was added using a peristaltic pump at a rate of 5 mL/h until the blue colour disappeared. After the Na-K alloy was completely exhausted, the mixture was dispersed in 50 mL of de-ionized water and extracted with n-hexane, after which the mixture was filtered through a 0.22 μm nylon membrane. After washing the resulting brown cake product alternatively with n-hexane and alcohol, it was dried overnight under vacuum at 60 °C.
2.3. Chemical Characterization of Hydrogenated Graphene
The chemical characterization of hydrogenated graphene was performed using X-ray photoelectron spectroscopy (XPS) and Fourier Transform Infrared Spectroscopy (FTIR). X-ray photoelectron XPS data were collected using AlKα radiation at 1486.69 eV (150 W, 10 mA), a charge neutralizer, and a delay-line detector (DLD) consisting of three multi-channel plates. Survey spectra were recorded from −5 to 1200 eV at a pass energy of 160 eV (number of sweeps: 2) using an energy step size of 1eV and a dwell time of 100 ms to confirm the hydrogenation of the graphene. To further verify the hydrogenation, FTIR was performed on an ABB Bomem MB-Series FTIR spectrometer, with a resolution of 8 cm−1, using a KBr pellet method. Scanning was from 400 to 4000 cm−1.
2.4. Charge Carrier Mobility Characterization
The characterization of the charge carrier mobility of the semiconductor channel was performed using the Dark Injection Space Charge Limited Current (DI SCLC) method, shown in
Figure 2. In this configuration, the semiconductor is placed between a blocking electrode and an injecting electrode. A step function is then applied between the electrodes and the resulting transient current is measured. The carrier mobility is calculated using Equation (1), where t
tr is the free carrier transient time, V the applied voltage, and d the thickness of the semiconductor film. τ
DI corresponds to the time it takes for the current to peak and represents the time it takes for the first sheet of charge carriers to reach the counter electrode [
21].
The sample was prepared by depositing a thin film of C60 on a silicon substrate coated with nickel acting as the blocking electrode. A thin film of PEDOT:PSS (~5 nm) was deposited on the C60 film to enhance its conductivity, followed by a chromium layer acting as the charge injecting electrode.
2.5. OTFT Testing and Characterization
The OTFT individual layers thicknesses were measured using cross sectional Scanning Electron Microscopy (SEM) images of the deposited films. The morphology features and surface topology of the sensing films were investigated using Atomic Force Microscopy (AFM).
The OTFT electrical characterization was performed using a probe station and a semiconductor parameter analyzer. All measurements were obtained in a dark and ambient environment to minimize photodegradation. The output characteristic was extracted by sweeping the drain-source voltage VDS from −2 V to 30 V and measuring the drain-source current and the gate leakage current IGS at a constant gate bias of 80 mV, 500 mV and 1 V. The transfer characteristic was obtained by sweeping the gate-source voltage VGS from −1 V to 5 V and measuring IDS at constant drain-source voltages of 50 mV. Output and transfer sweeps were both performed at a step size of 50 mV.
The humidity sensor response was monitored by placing the sensor in a gas chamber equipped with nitrogen gas flow to remove ambient gases before and after exposure to moisture. The sensor was exposed to a high humidity above 80% RH to mimic condensation and surface wetness conditions, sufficient to test the peripheral sensor for wetness detection. The response of the chloride sensor was tested by depositing Cl− ion solutions with varying concentrations on the gate surface and monitoring the change in output current after the complete evaporation of water.