Monitoring of Graphene Properties in the Process of Viral Biosensor Manufacturing

. The properties of graphene chips with low reproducibility (LR) after photolithography (PLG) and graphene functionalization have been studied. It is shown that the introduction of additional cleaning after PLG can significantly increase the reproducibility of the parameters of processed graphene in biosensors. The use of dilute PBS solutions for virus detection makes it possible to increase the relative concentration sensitivity of biosensors by several times.


Introduction
Graphene, a promising two-dimensional material, has been actively investigated for applications in viral biosensors, including for detection of influenza and Covid-19 viruses [1][2][3].Successful registration of viruses using graphene-based resistors or transistor has been demonstrated over the past ten years in many papers [1][2][3][4][5][6][7][8][9].However, in recent review papers [2] it is noted that the widespread use of viral biosensors based on graphene in clinical medicine has not yet been implemented.The difficulties that have arisen are largely associated with the insufficiently low reproducibility (LR) of graphene properties in biosensors [1,2].The LR of graphene properties may occur at any stage of complicated biosensor fabrication process [2,3].
A biosensor commonly has several key stages in a fabrication process: the formation of a graphene film, the formation of chips with contact pads (graphene resistors or field-effect transistor) by PLG, controlled treatment of the graphene surface in chips (functionalization), immobilization (attachment) of bioreceptors (antibodies) [1][2][3][4][5].The bioreceptor recognizes the target biomolecule (antigen) in an analyte solution via antibody-antigen immunoreaction and creates an electrical signal of the chip indicating the antigen detection.
Such a sophisticated biosensor fabrication technique is necessary to enhance the selectivity and sensitivity of biosensors.It should be noted that studies on the LR properties of graphene films after each of the stages are not numerous, and the LR parameters of graphene after one of the stages, namely after PLG, are most often considered [6][7][8][9][10].It is known that LR of biosensor parameters after PLG is due to the interaction of the photoresist with grapheme, creating local regions with resist residues (LRRs) that are not completely removed from the graphene surface.
A large number of works [11][12] are devoted to the study of the functionalization of graphene.The methods used are classified into two broad classes: covalent (CV) and non-covalent (NCV) functionalization.They differ in the mechanism of interaction between adsorbed molecules and graphene.The modification of graphene properties after functionalization is revealed by the change in the spectra of X-ray photoelectron spectroscopy (XPS), in Raman spectroscopy by the appearance of a defect line D, in atomic force microscopy by a change in the root mean square (RMS) roughness of the graphene surface, and also in a change in the form of frequency dependences power spectral density of low-frequency noise (LFN).
Among various approaches described in the available review papers to get a selective recognition reaction on the graphene surface with a target biomolecule [7,11,12], it is not completely clear which functionalization method is the most effective for detecting viruses, and data on the reproducibility of graphene properties in biosensor chips after functionalization are practically absent.In addition, in some works, the influence of deformations on the adsorption properties of graphene and the possibility of self-functionalization of graphene at the stage of its production were noted [12].The reproducibility of graphene properties after the subsequent stages of antibody attachment to the functionalized graphene surface, as well as antibody-antigen immunoreaction on the graphene surface for viruses' detection has not been practically studied, which is associated with small sizes of antibodies and difficulties in visualizing this process.As a result, they are guided by the results of the final stage of virus detection.The dependence of the electrical parameters of biosensors on the concentration of viruses in an analyte is weak and nonlinear, as shown in numerous publications [1,4.6].However, it is possible to fix the signal from the introduction of viruses in very low concentrations (less than 10 -13 g/ml).
The possible causes of nonlinear dependencies and weak changes in the electrical signals of the chips are diverse and are studied in this paper.In this work, an attempt was made to trace the change in the properties of graphene at all stages of fabricating biosensors for viral infections by means of Raman spectroscopy, atomic-force microscopy (AFM), XPS, and low-frequency noise (LFN).These methods provide information about the nature of the nanoorganization of the graphene nanomaterial, about the distribution of deformations, and about the properties of the defective system.The experiments were carried out on chips with the topography of processed graphene with root mean square (RMS) roughness measured by AFM practically coinciding with the RMS values on as-grown graphene/SiC films from which the chips were obtained.The reasons leading to the LR of properties graphene properties at the key stages of biosensors fabricating identified on the basis of the analysis of the obtained results are discussed.

Experimental
The graphene films were formed on semi-insulating 4H-SiC substrates by thermal decomposition of the (0001) Si surface [5].Substrates were purchased from TANKEBLUE Co., Ltd., (Beijing, China).The growth temperature was 1730 ± 20 °C, the argon pressure in the growth chamber was 750 ± 20 Torr.High-purity (99.9999%) argon was used in the growth process.
The ordinary photolithography process was applied to form a topological pattern of as-grown graphene/SiC films utilizing AZ1318 photoresist.The photoresist was removed in acetone.Аdditional cleaning based on a piranha solution H2O2 (30%) + H2SO4 (concentrated) 1:2.Details of the graphene films and the processing and mounting of chips on holders can be found elsewhere [***].Treated chips with two contact pads (graphene resistors) are used as biosensor.The size of the sensing area (active surface of graphene in the chip) was about 0.8 × 0.8 mm 2 .
The properties of graphene in chips at the biosensor fabrication was traced using the methods of AFM, Raman spectroscopy, LFN, and XPS.Specific and description of the equipment used in this work for first three methods can be found elsewhere [6,8].X-ray photoelectron spectroscopy (XPS) measurements were made using a SPECS GmbH Proven X-ARPES system equipped with ASTRAIOS 190 electron energy analyzer and a 2D-CMOS electron detector.

Results and Discussion
Low reproducibility (LR) of graphene parameters in chips after conventional PLG is observed due to the formation of LRRs.The presence of LRRs on the surface of processed graphene appeared as a spread in RMS values from 0.5 to 10 nm at AFM scan of 10 µm × 10 µm and also in the resistance from 1 to 10 kOhm of processed chips from the same as-grown graphene/SiC film.The use of a simpler method than in [9,10], namely additional washing of LRRs after their detection by AFM, made it possible to significantly increase the reproducibility of the properties of processed graphene in chips.As a result, RMS values turned to be closed to the values on as-grown graphene/SiC film for the vast majority of the chips (up to 80%) and the spread of chip resistance values decreased to 1-1.6 kOhm [8], Fig. 1 shows the AFM topology (scan 10 µm ×10 µm) of the as-grown graphene/SiC film, the processed chip after PLG, and the processed chip after PLG followed by additional cleaning of the graphene surface.At the subsequent stages, the experiments were carried out on chips with graphene topography corresponding to the one in Fig. 1c., having RMS values closed to the values on as-grown graphene/SiC film and resistance values to 1-1.6 kOhm.On such chips from the same film (EG423), experimental studies of graphene properties were carried out after the graphene surface was functionalized by an amino group (-NH2 group) using two functionalization methods: 1) covalent functionalization by two-stage cyclic voltammetry (CF) [7.8] and (2) non-covalent functionalization (NCF) based on sorption of pyrene derivatives from an aqueous solution of [12] (1% aqueous solution of 1-pyrenmethylamine hydrochloric acid for 20 min followed by washing with distilled water).
The properties of graphene in chips were controlled by Raman spectroscopy and AFM.Before and after functionalization, the Raman spectra in the region of 1300-2800 cm −1 were dominated by sharp G (~ 1600 cm -1 ) and 2D (~ 2700 cm -1 ) lines and wide asymmetric bands centered at approximately 1380 and 1550 cm −1 .The former are characteristic of graphene, and the symmetric shape of the 2D line allows us to attribute both spectra to monolayer graphene areas [13].The latter wide bands correspond to the buffer layer [14].Regardless of the functionalization method used, a new defect-related D line at ~1350 cm -1 appeared in the Raman spectra of graphene in chips processed from the same film (Fig. 2a).

Key Engineering Materials Vol. 984
It was also observed an increase in the RMS values by an average 1.5-2 times after functionalization.In Fig. 2b shows an AFM image of a functionalized graphene surface in an EG423-D2 chip with an RMS value of 0.81 nm, whereas before functionalization this value was 0.50 nm, typical for the surface of as -grown graphene in Fig. 1a.
As it was shown earlier [12], the appearance of a defective band in the Raman spectra and an increase in RMS value indicates the presence of functional groups over the entire graphene surface after functionalization.However, sometimes inhomogeneous graphene functionalization was observed for both covalent and non-covalent functionalization, as shown in Fig. 2c.In this case, the RMS values increase up to 5-8 times, and contrasting areas of non-functional graphene surface are observed on AFM images.
The AFM image in Fig. 3c is similar to the one given in [12] where it was attributed to the inhomogenous functionalization of the graphene surface.This type of functionalization can lead to low reproducibility of antibody-antigen immunoreaction and, as a consequence, to low reproducibility of antigen detection (i.e., target molecule detection) results.Control of graphene topography by the AFM method makes it possible to exclude the use of such chips at subsequent stages of biosensor manufacturing.The XPS data confirmed the appearance of a functionalized layer by reducing the intensity of the Si 2p line after functionalization as shown in Fig. 2d.At the same time, S3 component with binding energy of ~285.4 is less pronounced than in functionalized CVD graphene [11].This difference in the properties of graphene grown by the two aforementioned methods was noted elsewhere [12].
The study of the features of the defective graphene structure in the chip before and after functionalization was carried out using low-frequency noise (LFN) methods.Figure 3(a-c) shows the frequency dependence of the spectral density of voltage fluctuations (SU) in chips with homogeneous and inhomogeneous graphene functionalization.
It is known that the shape of noise spectra SU ~ 1/f γ (γ > 1) at frequency region f<50Gz reveals the information about inhomogeneous deformations in the material [15,16].The greater the value of γ , the higher the level of deformation.It was earlier observed that the dependence SU ~ 1/fγ (γ >1) in graphene films and the chips is consistent with the Raman spectra data that demonstrate an inhomogeneous distribution of compressive stresses in graphene [16].
A weaker dependence SU ~ 1/f γ (γ < 1) is observed for higher frequencies f > 50 Hz.We believe the SU frequency dependence at a higher frequency may be due to a superposition of 1/f noise and generation-recombination (GR) noise.These features mean that, in addition to the system of defects typical to low-dimensional materials, there are single Shockley-Reed-Hall defects.These features are observed in graphene in chips regardless of the functionalization method used as shown in Fig. 3a, b.Moreover, Fig. 3c shows that these features can be observed even in chips before graphene functionalization and can be due to the change of deformation in graphene.This observation may be related to the self-functionalization of graphene obtained by sublimation of SiC [12].Additional studies are needed to use this effect to improve the reproducibility of the properties of functionalized graphene in chips.The use of a diluted PBS solution for the detection of viruses can be promising to enhance the sensitivity of a graphene biosensor.Dilution lowers the ionic strength of the PBS solution, which may reduce the effect of Debye screening on biosensor sensitivity [17].In Fig. 3d the response of graphene sensor ((RO-R)/RO) on exposures to PBS solution with different concentration of influenza B virus is presented.In the sensor response, Ro in the chip resistance in pure PBS solution and R is the chip resistance in the PBS solution with viruses.The response of graphene sensor in diluted PBS solutions turned to be greater than that in undiluted PBS solutions in several times.The usage of the analyte (influenza virus B) in 1000 times diluted PBS solution allowed to improve the sensitivity of biosensors in the range of the analyte concentration 10 -14 -10 -9 g/ml.Visualization of influenza A and B viruses, as well as SARS-CoV-2 on the graphene surface of biosensors using AFM and scanning electron microscopy was presented elsewhere [8].

Summary
The results of monitoring the properties of graphene showed that the introduction of additional cleaning after PLG allows to significantly increasing the reproducibility of the parameters of processed graphene in biosensors.It is shown that the spectral density of voltage fluctuations (SU) and the features of the type of frequency dependence SU make it possible to control the degree of deformation and features of the defective system of processed graphene after each stage of biosensor manufacturing.It was found that the efficiency of graphene functionalization depends to a greater extent on the change in graphene deformation than on the chosen method of functionalization.Additional studies are needed to use this effect to improve the reproducibility of the properties of functionalized graphene in chips.
The usage of the analyte (influenza virus B) in 1000 times diluted PBS solution allows to improve the sensitivity of biosensors in the range of the analyte concentration 10 -14 -10 -9 g/ml.