According to the World Health Organization, 15 million babies worldwide are born underweight by less than 2.5 kilograms every year [1]. Babies that are born underweight have a high mortality rate and are likely to develop diabetes and cardiovascular disease even after reaching adulthood. The main causes of these phenomena are delayed growth in the womb and premature birth. The premature birth rate, which means the proportion of all newborns born before 37 weeks, is also rising due to the increase in advanced maternal age and multifetal pregnancies due to in vitro fertilization [2]. In particular, premature birth within 32 weeks is known to severely affect the survival and health of the baby. Therefore, if there are signs of an impending premature birth, delaying delivery through treatment is important for the health of the fetus and mother. In previous studies, the amount of interleukin (IL) 13 in the amniotic fluid is known to have a significant correlation with premature birth [3–5]. The cutoff value of IL-13 to predict the risk of premature birth at less than 32 weeks is known to be approximately 50 pg mL⁻¹ [6]. For this reason, the fabrication of elaborate sensors for detecting IL-13 below this threshold value is a major factor for avoiding the risk of premature birth.

One ofthe most common detection systems for biomarker detection is enzyme-linked immunosorbent analysis (ELISA), which is well known in a wide range of medical applications in the field of diagnosis [7–10]. Although this technique can be used for IL-13 detection, ELISA has limitations, including high cost, long operation time, impossibility of performing real-time measurements, and the need for skilled technicians. To overcome these limitations, in this study, we used a graphene field-effect transistor (GFET) using the interaction between the antibody and the antigen. As a field-effect transistor (FET) channel, a two-dimensional conducting nanomaterial, graphene, was utilized for its excellent electrical properties and versatility through surface modification [11–13]. Chemical vapor deposition (CVD) was used to synthesize graphene, and a microelectromechanical system (MEMS) was used to fabricate the GFET. Transmission electron microscopy (TEM) and Raman spectroscopy were used to confirm the successful fabrication of mono-layered graphene [14].

By using the interfacing chemical PDI-diacid, an IL-13 antibody was used as a bioprobe to generate selective and sensitive responses to the target antigen. To test the performance of the antibody-conjugated graphene field-effect transistor (Immuno field-effect transistor, ImmunoFET), a source meter instrument was utilized to measure the transduced current from the chemical reaction. The ImmunoFET was exposed to various concentrations of IL-13 protein, and a response according to the concentration was shown. The limit of detection (LOD) of the sensor was 15 pg mL⁻¹, and the value was lower than the threshold value mentioned above. Additionally, a selectivity test was performed to indicate the accurate detection of IL-13 among other nontargets (IL-6 and IL-17) with the fabricated GFET platform. It is also known that high concentrations of IL-6 and IL-17 are correlated with premature birth [6]. Furthermore, compared to conventional analysis methods, our newly developed device showed advantages such as rapid detection, facile operation, and no need for skilled technicians. In addition, by varying the antibody as a bioprobe, the fabricated ImmunoFET can be adapted to many other fields, such as infectious disease diagnosis, water quality management, and industry.

2.1. Materials and chemicals

4-(4,6-Dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMTMM) and poly(methyl methacrylate) (PMMA) (950 PMMA A4 4 % in anisole) were purchased from MicroChem Co. IL-13 antibody and human IL-6, IL-13, and IL-17 protein were purchased from Sino Biological (Beijing, China). 10X Phosphatebuffered saline (PBS) was purchased from Welgene (Seoul, Korea). Ammonium persulfate (APS) 98 % was purchased from Samchum Chemicals (Seoul, Korea). Copper foil 0.025 mm thick was purchased from Alfa Aesar (Haverhill, USA).

2.2. Synthesis of PDI-diacid

When imidazole (1.6 g, 23.1 mM) became sticky while stirring at 95 °C, perylene-3,4,9,10-tetracarboxylic acid dianhydride (300 mg, 0.77 mM) and amino butyric acid (159 mg, 1.54 mM) were added to a 100 mL round-bottom flask. The reaction mixture obtained by stirring for 16 h under nitrogen was then cooled to room temperature. Thereafter, distilled water was added, and residual perylene-3,4,9,10- tetracarboxylic acid dianhydride was filtered through filter paper. The filtrate was slowly neutralized by dropping 1 M HNO₃ and washed with distilled water. Ultimately, water was evaporated in a vacuum to obtain a dark red solid (402 mg, 93 %) as the product.

2.3. Process of fabricating the ImmunoFET

Graphene was manufactured on Cu substrate by using CVD with process steps that included H₂, CH₄, and Ar exposure. After fabrication, PMMA A4 was spin-coated on graphene for further experiments. By using 0.2 M APS, the Cu substrate was etched, and graphene was transferred to a 4 inch wafer. The water was dried completely and placed in a vacuum oven overnight. The graphene-transferred wafer was washed with acetone, and the PMMA was removed. The GFET device was fabricated with a MEMS with process steps that included a spin-coater, aligner, etcher, e-beam evaporator, and dicing saw. The graphene channel was first treated with PDI-diacid and DMTMM every 2 h at room temperature and washed. Finally, the IL-13 antibody was bound to PDI-diacid on graphene, and the ImmonoFET was manufactured.

2.4. Preparing ImmunoFETs for detections

A small chamber for storing 50 µL of PBS buffer was prepared. Grease was applied to the bottom of the chamber, and the chamber was attached to the GFET electrode. The chamber was filled with PBS buffer, and the GFET electrode was prepared for measurement using a Keithley 2602A System Sourcemeter.

2.5. Instruments

Electrical measurements were performed by using KI2600S 4Channel FET Measurement Software (IVSolution). The ImmunoFET was connected to an external probe station connected to a Keithley 2602A System Sourcemeter. To verify the characteristics of graphene, high-resolution transmission electron microscopy (HR-TEM), Raman spectroscopy, and field-emission scanning electron microscopy were performed. The electrode was fabricated by using a dicing saw (DAD 522, Disco), e-gun evaporator (ZZZ550- 2/D, Maestech), ICP etcher (Hi-etch, BMR Technology Corp.), aligner (MA-6, Karl-suss), and spin-coater (TFT).

3.1. Real-time detection platform for IL-13

Figure 1 shows the overall schematic diagram of the developed equipment for detecting IL-13 quickly and easily. For premature birth diagnosis, amniotic fluid taken from the mother is injected into the sensor. Through antibody-antigen reactions, ImmunoFET detects the target IL-13 and transduces the signal into a current. Since the shifted current is target concentration-dependent, the amount of IL-13 contained in the sample can be specified. By comparing this with the threshold, a diagnosis of premature birth can be made.

Figure 1. Schematic diagram of IL-13 detection using the ImmunoFET. Maternal amniotic fluid is collected, injected into the developed ImmunoFET, and measured through signal changes.

3.2. Fabrication process of the ImmunoFET

Graphene is well known for its high electrical conductivity, heat transfer ability and strength. In particular, it has superior electrical properties, including a high carrier mobility and zero bandgap [15]. Consequently, mono-layered graphene has been adapted in various fields, such as in bio/chemosensors, solar cells, touchscreens and displays, and photodetectors. In this paper, mono-layered graphene was used as an FET channel and a substrate for attaching bioreceptors. Chemical vapor deposition was used as a method for synthesizing graphene on Cu foil [Fig. 2(a)]. To confirm the mono-layered nature of the fabricated graphene, an HR-TEM image was obtained [Fig. 2(b)]. The boundary of the mono-layer graphene was observed, presenting layer uniformity. Moreover, the clear hexagonal patterns of graphene showed a crystalline structure according to the selected-area electron diffraction (SAED) pattern [Fig. 2(b) inset]. In Fig. 2(c), Raman spectroscopy was used to verify the characteristics of the fabricated graphene based on the intensity of the 2D/G peak ratio. As shown in the Raman spectrum, the 2D peak and G peak were observed at 2,695 and 1,584 cm⁻¹, indicating the successful synthesis of graphene. The ratio of the 2D peak to the G peak (ca. 2.2) was also used to demonstrate the characteristics of the mono-layered graphene [16]. Figure 2(d) is a schematic diagram of the MEMS process. To fabricate micro-patterned graphene and a source-drain electrode, two consecutive photolithography processes with negative and positive photoresists (PRs) were conducted [17]. By differing the type of PR, developed areas were reversed. A developer was used to eliminate the unwanted areas, and the deposition process was maintained. For the electrode, bilayer metal (Cr/Au) was deposited for better adhesion. After metal deposition, the surface of the electrode, except for the source, drain, and graphene channel, was passivated with SiO₂ to decrease the noise current.

Figure 2. Fabrication of graphene and GFET. (a) Scheme of the CVD method for synthesizing graphene. (b) TEM image of monolayered graphene (inset: SAED pattern of graphene). (c) Raman spectrum of monolayered graphene. (d) Schematic diagram of MEMS technology used to fabricate GFETs.

3.3. Characteristics of the ImmunoFET

To detect the IL-13 protein, an experiment was conducted to attach a bioreceptor and confirm the attachment. A scanning electron microscopy (SEM) image of the graphene channel was obtained to demonstrate the size of the channel (width: 50 µm, height: 80 µm). Figure 3(a) shows the source and drain electrodes in addition to the graphene channel. To test the characteristics of the device, the output curves of the GFET were measured as the gate voltage (VG) differed over a range from 0 to -1.2 V in steps of -0.3 V [Fig. 3(b)] [18]. IDS increased negatively as VG increased negatively, which is a characteristic behavior of a p-type semiconductor. According to previous studies, for liquid-ion gated FETs, p-type semiconductors are suitable due to the electron transfer caused by oxygen in the liquid [19]. PDIdiacid was used as an interfacing compound to immobilize the antibody onto the graphene surface. The introduction of PDI-diacids and their conjugation with the antibody are of importance in our platform. Therefore, to verify the attachment steps of the bioreceptor, a current-voltage (I-V) curve in the range of -0.1 to +0.1 V was measured according to the treatment step. After PDI functionalization and Ab immobilization, the slope of the I-V curve was slightly decreased, maintaining linearity [Fig. 3(c)]. These changes showed clear ohmic behavior through the serial steps of antibody introduction and indicated that the antibody immobilization to the graphene channels occurred successfully [14]. Figure 3(d) shows the molecular structure of PDI-diacid and graphene. PDI-diacid has an aromatic ring backbone that interacts with graphene. Subsequently, by treating DMTMM with PDI-diacid, the carboxyl group of PDI-diacid was activated and facilitated amide bonding with the amine group of the antibody [Fig. 3(e)].

Figure 3. Characteristics of the fabricated GFET. (a) SEM image of GFET (S: source, D: drain); the red mark indicates the graphene channel (width: 50 µm, height: 80 µm). (b) Change in output curve of GFET by varying the gate voltage. (c) Changes in I_DS-V_DS after surface treatment. (d) 2D structural images of PDI-diacid and graphene. (e) Schematic diagram of amide bonding between PDI-diacid and antibodies.

3.4. Real-time detection of IL-13 protein

Liquid-ion gating is known as the most efficient methodology for examining the electrical characteristics of ImmunoFETs. Therefore, a chamber with PBS was installed on the ImmunoFET to build a liquidion gated FET system. A schematic illustration of the liquid-ion gated ImmunoFET consisting of a source, drain and gate electrodes is shown in Fig. 4(a). The gate was immersed in a chamber containing PBS buffer to apply voltages for the sensitive detection of IL-13. The gate voltage was applied to form an electrical double layer between the graphene channel and the liquid to promote the field effect. When the target was injected, the change in current between the source and drain was induced by the transistor, and the electrical signal was obtained. To evaluate the performance of the manufactured ImmunoFET, PBS buffer was first added to set the base current. After baseline correction for the calibration, the target protein diluted at various concentrations was injected from lower concentrations to higher concentrations. The response current was normalized using Eq. (1), where I₀ and I represent the initial current and response current, respectively.

Figure 4. Measurement data of IL-13 protein with the ImmunoFET. (a) Schematic image of the ImmunoFET. (b) Real-time response of the ImmunoFET after injecting various concentrations of IL-13. (c) Selectivity test of the fabricated ImmunoFET (IL-6, IL-17: 15 ng mL⁻¹; IL-13 150 pg mL⁻¹).

To verify the reaction between the ImmunoFET and the target, GFET with no antibody was tested as a control group. The devices without IL-13 antibody showed no significant current changes after the injection of various sample concentrations. However, with the ImmunoFET containing antibodies, a change in current occurred within 5 s after target injection, which is explained by electrostatic gating effects [20]. The detection range was from 15 pg to 1.5 µg mL⁻¹, and the response was concentration dependent [Fig. 4(b)]. The LOD of the device is lower than that of the conventional ELISA method (32 pg mL⁻¹). In addition, the reaction between the ligand and receptor occurred rapidly. There was no need for a complicated process, which made detection possible without an experienced operator. Moreover, the sensor was able to detect a concentration three times lower than 50 pg mL⁻¹, which is the standard value for the diagnosis of preterm birth. To determine the effects of interference from other similar cytokines, a selectivity test was conducted with IL-6 and IL-17. After confirming the base current with PBS, 1.5 ng mL⁻¹of IL-6 and IL- 17 were injected respectively, and no significant current change was observed. However, 150 pg mL⁻¹ of the target was injected, and an immediate current change was shown [Fig. 4(c)].

In this study, we fabricated a highly sensitive and selective sensor to detect IL-13 for the diagnosis of premature birth. By using graphene as a substrate, the sensor was able to detect minute current changes caused by antibody-antigen interactions. The CVD method and the MEMS process for platform fabrication were demonstrated, and the characteristics of the produced graphene and electrodes were analyzed. ImmunoFET production was completed by conjugating the antibody using PDI-diacid to the graphene channel. The sensor was able to detect the target from 15 pg to 1.5 ng mL⁻¹, and the LOD was lower than the threshold (50 pg mL⁻¹) for diagnosis of premature birth. Moreover, among other cytokines, IL-6 and IL-17, the ImmunoFET selectively detected IL-13. As a result, a real-time detection platform for selectively detecting IL-13 was manufactured. In addition, the platform introduced in this paper can be applied to other areas with the replacement of the bioprobe.

This work was supported by the Korea Instituteof Planning and Evaluation for Technology in Food, Agriculture and Forestry (IPET) and Korea Smart Farm R&D Foundation (KosFarm) through Smart Farm Innovation Technology Development Program, funded by Ministry of Agriculture, Food and Rural Affairs (MAFRA) and Ministry of Science and ICT (MSIT), Rural Development Administration (RDA) (421020-03).