Fully Printed One-Step Biosensing Device Using Graphene/AuNPs Composite

Driven by the growing need of simple, cost efficient and flexible sensing systems, we have designed here a fully printed Reduced Graphene Oxide (rGO) based impedimetric sensor for one step sensing of DNA. The DNA sensor was fabricated by stamping of layered rGO and rGO/gold nanoparticles/single stranded DNA (rGO/AuNPs/ssDNA) composites over PET substrates using wax-printing technique. rGO works as an excellent working electrode, while the AuNPs create a suitable environment for ssDNA immobilization. Counter and reference electrodes were previously screen-printed on the plastic substrate, making thus a compact and highly integrated sensing platform. The change in electron transfer resistance after hybridization with a target ssDNA specific of Coxsackie B3 virus was monitored using electrochemical impedance spectroscopy (EIS), finding a linear response in the range of concentrations 0.01-20 µM. The novel, simple and straightforward one-step printing process for fabrication of a biosensing device developed keeps in mind the growing need of large scale device manufacturing. The successful proof-of-concept for the detection of DNA hybridization can be extended to other affinity biosensors, taking advantage of the integration of the bioreceptor on the sensor surface. Such ready-to-use biosensor would lead to a one-step electrochemical detection.


INTRODUCTION
Graphene is a well-known one-atom thick two dimensional carbon layer that possess outstanding inherent properties like high mechanical and chemical stability, high thermal and outstanding electrical and thermal conductivity (Novoselov, K.S. et al, 2004;Geim, A. K. and Novoselov K. S.,2007) . It has a high theoretical surface area of about ~2600 m 2 g -1 that is higher than that of carbon or CNTs (10 or 1315 m 2 g -1 ). It is sp 2 hybridized and the out of plane π bonds are responsible for the high conductivity. Although these properties are interesting and highly useful for research purposes in different areas, they vary greatly with the quality of graphene Bollella, P et al., 2017) that is, number of layers, sheet size, degree of oxidation or defects, all of which depend in turn of the production procedures. Indeed, graphene materials with varied quality and price can be prepared in different ways considering the end application; broadly by mechanical or chemical exfoliation of graphite, vapor deposition or epitaxial growth etc. (Lee, H. C.2017). All this has made possible the use of graphene in broad range of applications (Randviir, E.P. et al, 2014) like solar cells (Roy-Mayhew, J. D. and I. A. Aksay, 2014) , energy storage (Dubal, D. P. et al, 2017;Li, X. and L. Zhi, 2018) , electronics (Lee, S.-M. et al, 2015) , (bio)sensing (Shao, Y. et al, 2010;Justino, C. I. L. et al, 2017) to name a few. Owing to such great properties of graphene, especially due to its high conductivity and high biomolecule loading as a result of high surface area, it is widely explored in the research of (electrochemical) biosensings (Pumera, M., 2011;Bo, X. et al, 2017 ), for instance, Tzu-YenHuang et al demonstrated the use of rGO composite with single-walled carbon nanotubes as an effective electrode material for electrochemical sening (Huang, T. U. et al, 2013). Apart from this, graphene can interact with the biomolecules using non covalent interactions like π-π or Hydrogen bonding making it more suitable for sensors application.

Field Cod
For Two formats are widely followed for DNA sensing taking advantage of rGO, either attaching (labelled or non-labelled) ssDNAs directly over Graphene sheet via π-π bonding or by modifying Graphene surface by using polymers, nanoparticles etc. that have available groups or sites for binding to the DNA. Most commonly used nanomaterial for this purpose are gold nanoparticles (AuNPs) which are typicaly attached to the modified ssDNA at one end via thiol-Au interactions (Lin, L. et al, 2011;Singh, A., 2013) .
Flexibility and portability are characteristics most expected for wearable electronic systems and is been used in fancy applications like implantable/wearable sensors (Ray, W. J. and Wang. J, 2013;Singh, E. et al, 2017) Joseph , OLEDs (Lee, J. Et al, 2016) , transistors  , energy devices (Nagar, B. et al, 2018) etc. Material of interest is usually printed over a flexible substrate (usually plastics) to make a cost efficient, compatible, flexible and easy to handle or use device. The ability to form nanoscale patterns and features over the flexible substrates offers us the advantage of more sensitivity and precision in the sensing devices along with low cost and easy handling. Several printing techniques like screen printing, inkjet printing, roll-to-roll, wax transfer/stamping technique etc. (Søndergaard, R. et al 2012;Li, J. Et al, 2015;Baptista-Pires, L. Et al, 2016;Mattana, G. and Briand D., 2016) are being used for the fabrication of electrodes. However, it is important to select a particular technique for a particular application. Particularly, our objective is the development of a Point-of-Care (POC) device, which requires a fabrication technique that doesn't need any post printing step for keeping the biomaterial unharmed, being wax stamping technique ideal in this case. POC have been widely used for decades now but is yet continuously refined according the user and manufacturer's needs. The important features to be fulfilled for POCs are (a) ease of fabrication, (b) cost effectiveness and (c) easy handling. The advancements in recent years have provided us with loads of new technological strategies to make improved devices for biorecognition, interactions and sensing (Quesada-Gonzalez, D., Merkoçi, A., 2018) .
In this work, we used chemically exfoliated graphite as it promotes cost effective, large scale production of Graphene Oxide (GO; graphene sheets containing Oxygenated groups such as epoxides, alcohols or carboxyl groups on the surface or the edges of the sheets). The presence of these groups makes GO water dispersible which further helps in making water based inks for printing purposes. Such material was later reduced to reduced Graphene Oxide (rGO) and mixed with already conjugated gold nanoparticle/single stranded DNA probe (AuNPs/ssDNA). Later, wax stamping technique was applied to create the working electrode pattern made of the rGO/AuNPs/ssDNA composite over already screen-printed counter (carbon) and reference (Ag/AgCl) electrodes, allowing us to have a ready-to-use one-step electrochemical biosensor This strategy for patterning rGOe composite doesn't require harsh or toxic solvents helping us to get rid of the post printing steps (annealing) that could affect the functioning of the biorecognition element (DNA in this case). At the end, we get a sensing platofrm with uniform patterns integrated with the biorector,which is ready to be tested without any further steps (Figure 1). Figure 1: DNA sensing principle. After stamping on the PET substrates, rGO/AuNPs /ssDNA is incubated with the target ssDNA. DNA duplex formation causes an increase in impedance that is related with the amount of analyte.

Oligonucleotides
Synthetic oligonucleotides were obtained for Sigma-Aldrich. The target sequence employed corresponds to a region characteristic of the ECHO virus (Coxsackie virus B3).

Chemicals and equipment
The GO (Graphenea Graphene Oxide, 4 mg/ml) was purchased from Graphenea Inc and Corel Draw software, Screen printing was performed using DEK 248 semi-automatic creenprinter (England) and the electrochemical studies were carried out using Autolab302 Potentiostat/galvanostat PGST30 with the software GPES for cyclic voltammetry and FRA for impedace measurements.

Synthesis of rGO
10 mg/mL Graphene oxide (GO) was provided by angstrom materials. It was reduced by exposing 100 mL of 1mg/mL GO (solution to 100 mg of ascorbic acid in an autoclave at 121 ͦ C for 45 min.
Briefly, 50 mL of 1% HAuCl4 solution was heated in an Erlynmeyer flask under vigorous stirring until boiling starts. Then, 1.25 mL of 1% sodium citrate was added while stirring continued. The reaction was kept under same conditions until 10 min and then the heating was stopped andthe reaction was let to cool. The color of the solution changes from deep blue to wine red, indicatingAuNPs formation. AuNPs suspension was stored at 4 ͦ C protected from light until further use.

Conjugation of AuNPs with probe DNA (ssDNA)
The particles were then conjugated with the thiol modified ssDNA probe sequence using the protocol described and pioneered by Mirkin et al (Mirkin, C. A. et al, 1996). For this, 190 µL of AuNPs were mixed with 10 µL of 1500 µg/mL thiolated sequence at 250 rpm for 20h at 25 ͦ C. Later 50 µL of 10 mM phosphate buffer (pH 7) / 0.1M NaCl was added to the above solution and let it stand for 44h. Finally a centrifugation step at 14000 rpm at 4 ͦ C for 20 min was carried out to extract the conjugated DNA which was reconstituted in 200 µL Mili-Q water for further use. Later, 0.1% BSA solution in milli-Q water was added to graphene and conjugated AuNPs in separate eppendorf tubes and mixed for 1h at 600rpm at room temperature and 4 ͦ C respectively.  Govindaraju, S. et al, 2017) . Figure 4 (d) shows the N 1s spectra of the mixture that corresponds to tertiary amines at 400.3 eV and protonated primary amines at 401.6eV (Zhang, F. and Srinivasan M. P., 2004;Yang, Y. et al, 2009;Stevens J. S. et al, 2013;Yu, B. et al, 2014) most likely coming from the BSA that is used to modify Graphene and AuNPs. Figure   4(f) shows the UV-vis spectra of GO, rGO and rGO/AuNPs with BSA. Reduction of GO to rGO is visible from the spectra where GO shows the typical humps at 230 and 300 nm that corresponds to π-π* transitions of C=C bonds and n-π* transitions of C=O bonds respectively and changes to only one peak at 273nm due to a red shift and the peak at 300nm is

Electrodes characterization using Cyclic voltammetry (CV)
Cyclic voltammetry (CV) was initially employed to test the performance of the electrodes using 5.0 mM [Fe(CN)6] 3-4-/0.1 M PBS as redox probe. At first, different concentrations of GO were tested in order to optimize the best thickness needed for conductivity and robustness. Figure 5 (a) shows the optical images (inset) and the CV for rGO filtered ranging from 0.1 to 1mg/mL. It was observed that at higher concentrations (>0.6mg/mL), the prints were non-uniform and uneven (inset) and the oxidation/reduction peaks shifted towards more positive/negative potential respectively without any considerable increase in the current

Field Cod
values. At low concentrations (<0.6mg/ml), the current values where extremely low making it inefficient for further measurements (CV in Figure 5a). Consequently. 0.6 mg/mL was chosen as optimum concentration for all the final measurements. Next, study on electrochemical behavior of the optimized electrode was continued using CV. Figure 5(b) shows that the peak currents (Ipa and Ipc) linearly increase with the increasing scan rate. This is indicative of the fact that the ongoing electrochemical process of Fe(II) to Fe(III) was diffusion controlled. Also, the shift in oxidation/reduction peak positions with increasing scan rates demonstrates that the process was quasi-reversible since in a reversible process, the peak position is independent of the scan rate. This linear dependency of peak potential was plotted against the square root of scan rate, given in Figure 5c. Figure 5(d) are the signals obtained from electrodes of different printed materials. The signal decreases after the presence of ssDNA due to electrostatic repulsion from negatively charged DNA strands. This was decreased further after modifying the material with BSA, evidencing successful blocking of free sites that could cause non-specific adsorption. Later, after addition of 1µM target ssDNA, the current was seen decreasing more due to enhanced repulsion between the negatively charged DNA and negative ions from the electrolyte that hinders the charge transfer at electrode/electrolyte interface.

DNA quantification using Electrochemical Impedance Spectroscopy (EIS)
After optimizing the graphene electrode and its quality testing, the electrodes were further used for performing impedance measurements. Electrochemical impedance spectroscopy (EIS) is very useful electrochemical technique that is generally performed to obtain more sensitive information about the changes and interactions going on at the interface. One of the way to interpret the data is through the Nyquist plot which typically consists of a semicircle at high frequency regions corresponding to blocking of the transferred charges at electrode/electrolyte interface. This is called the charge transfer resistance (RCT) which can be calculated by directly measuring the diameter of the semicircle. At high frequency regions it shows a line with a slope of around 45 ͦ which demonstrates a diffusion controlled process.
This typical behavior was exhibited by our graphene electrodes (as displayed in Figure 6a. Graphene electrodes showed a very low charge transfer resistance of 1.3kΩ which was seen increasing after modifying its surface by addition of conjugated AuNPs (rGO/AuNPs/ssDNA) to 2.7 kΩ . This correlates with the CV results and the explanation that the coated layer was blocking the electron transfer from [Fe(CN)6] 3-/4to the electrode surface as a result of electrostatic repulsion. It was observed that AuNPs treated with BSA exhibited larger Rct value than without the treated ones. This implied that the free sites on the synthesized AuNPs were successfully covered and the non-specific adsorption/attachment was minimized if not completely eliminated. to the blank signal + three times its standard deviation) of 2.5 nM with an average relative standard deviation (RSD) of 13% was calculated. A non-significant increase in resistance was observed with the addition of blank buffer and 0.1µM non complementary ssDNA, demonstrating the specificity ssDNA due to electrostatic repulsion from negatively charged DNA strands. This was decreased further after modifying the material with BSA, evidencing successful blocking of free sites that could cause non-specific adsorption. Later, after addition of 1µM target ssDNA (as an example), the current was seen decreasing more due to enhanced repulsion between the negatively charged DNA and negative ions from the electrolyte that hinders the charge transfer at electrode/electrolyte interface. Long-term stabilty study results suggest that the sensor performance remain unaffected during at least up to three weeks (see supporting information). Longer times were not evaluated in this preliminary work.
The proposed biosensor showed comparable or enhanced performance than recently reported DNA sensors on conventional screen printed electrodes, where detection limits at the nM levels are achieved i.e.4.7nM for ebola Virus (Ilkhani, H. and Farhad S., 2018) and 35 and 21nM for influenza genes (Subak, H. and Ozkan-Ariksoysal D., 2018). Although, modifying graphene surfaces prior to printing and later DNA detection has result in sensors with higher sensitivities, even at fM levels, (Chen, M. Et al, 2016;Chen, S. et al, 2016) ; , our approach exhibit clear advantages in terms of integration, simplicity and low time of analysis It must also be noted that in this work, no additional conducting layer has been utilised for transfer of charges:graphene is the sole carrier for conduction as well as for anchoring AuNPs and ssDNA. This also opens the possibility directly adsorbing probe ssDNA (labelled or un labelled) via π-π bonding onto graphene substrates and checking the elecrtrochemical or optical signals.

CONCLUSION
An innovative strategy of printed sensing platform that requires only one step modification of the working electrode has been proposed and demonstrated. A composite containing the electrode material (rGO) as well as AuNPs connected with the biorecognition element (ssDNA), was stamped onto a PET substrate.Such strategy eliminates the need of conventional modification steps of the electrode prior to testing. The electrode properties and materials were characterized and the printing procedure was carefully optimized. ssDNA characteristic of a virus was selected as model analyte, for the demonstration of the proof-ofconcept, based on changes in the impedimetric signal of the electrode, reaching a detection limit of 2.5 nM. The performance of the system is comparable with those previously reported using conventional screen-printed carbon electrodes, but with clear advantages in terms of simplicity, integration and time of analysis. This successful proof-of-concept can be extended to other affinity biosensors, taking advantage of the integration of the bioreceptor (antibody, enzyme, etc) on the sensor surface. Such ready-to-use biosensor would lead to a one-step electrochemical detection. .

ACKNOWLEDGMENTS
We