Facile fabrication of screen-printed MoS2 electrodes for electrochemical sensing of dopamine

Molybdenum disulfide (MoS2) screen-printed working electrodes were developed for dopamine (DA) electrochemical sensing. MoS2 working electrodes were prepared from high viscosity screen-printable inks containing various concentrations and sizes of MoS2 particles and ethylcellulose binder. Rheological properties of MoS2 inks and their suitability for screen-printing were analyzed by viscosity curve, screen-printing simulation and oscillatory modulus. MoS2 inks were screen-printed onto conductive FTO (Fluorine-doped Tin Oxide) substrates. Optical microscopy and scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM/EDX) analysis were used to characterize the homogeneity, topography and thickness of the screen-printed MoS2 electrodes. The electrochemical performance was assessed through differential pulse voltammetry. Results showed an extensive linear detection of dopamine from 1 µM to 300 µM (R2 = 0.996, sensitivity of 5.00 × 10–8 A μM−1), with the best limit of detection being 246 nM. This work demonstrated the possibility of simple, low-cost and rapid preparation of high viscosity MoS2 ink and their use to produce screen-printed FTO/MoS2 electrodes for dopamine detection.

After that, MoS 2 /EC inks were prepared by dispersing 25, 45 and 60 wt% of MoS 2 powder with particle size of ~ 6 µm (max. 40 µm) in a polymeric binder (samples Mo6-25; Mo6-45 and Mo6-60). In the case of MoS 2 nanopowder with particle size of 90 nm, only inks with 25 and 45 wt% of MoS 2 (samples Mo90-25 and Mo90- 45) were prepared. Inks with 90 nm particle content over 45 wt% were not miscible. All inks were homogenized in a "homemade" hand-held mixing unit 5 times for 30 s.
Screen-printing process. The MoS 2 inks were screen-printed onto glass substrates covered by a conductive FTO layer. Before printing, FTO substrates were gradually pretreated in four steps; cleaning in detergent solution (1); acetone (2); isopropyl alcohol (3) in an ultrasonic bath, followed by UV-C treatment (4) for 20 min. For the screen-printing of prepared MoS 2 inks, yellow high modulus polyester yarn mesh was used, with mesh count 71 cm −1 and 48 µm thread diameter (PME 71-48 Y, SEFAR) in combination with the stencil prepared by the photochemical way. Screen-printed MoS 2 layers were prepared using a manual, auxiliary guide arm (constant pressure) equipped screen-printing machine (Screen Printing Table P65-80 KN, Drucktech). The printed area was 6 × 6 mm 2 . After printing, MoS 2 layers were left for 10 min at laboratory temperature for levelling and then were dried in a laboratory oven at 120 °C for 30 min.
Characterization methods. The rheological properties of prepared MoS 2 inks were investigated using a rheometer (HAAKE, MARS iQ) equipped with parallel plate geometry with a diameter of 35 mm and a gap height of 0.4 mm. The temperature was set to 25 °C during all measurements. The steady-state rheological test was performed at shear rates [ γ(s -1 )] of 0.001-1000 s −1 to measure the dynamic viscosity [η (Pa.s)] and was obtained by controlling shear rates (CR). The time-dependent controlled-shear-rate tests were performed to simulate the screen-printing process with constant shear rates in three intervals: (1) 0.5 s −1 for 90 s; (2) 1,000 s −1 for 30 s and (3) 0.5 s −1 for 180 s. The oscillatory tests are helpful to understand the structural changes of the inks occurring in the screen-printing process. In the shear strain amplitude sweep, the applied shear strain ranged from 0.1 to 100% at a frequency of 1 Hz, which helped to characterize the viscoelastic behavior of the inks and determine the linear viscoelastic region (LVR). Elastic and storage moduli were measured as a function of shear strain/stress. The elastic or storage modulus ( G ′ ) is related to the ability of ink to store energy and represents the elastic portion of the viscoelastic behavior. The viscous or loss modulus ( G ′′ ) indicates the fluidity of ink 24  www.nature.com/scientificreports/ factor tanδ, which is G ′′ /G ′ gives an indication of material internal strength and helps to define viscoelasticity (Eq. (1)) 25 .
The thicknesses of screen-printed MoS 2 layers were analyzed by optical microscope (LEICA DM 2700 M) using the 3D image sequential recording method. The software (Leica Application Suite V4, LEICA) was used to evaluate the arithmetic average of thickness. The adhesion of printed MoS 2 layers to FTO substrates was tested by peel adhesion test using the adhesive tape (Scotch Crystal, 3M). Disturbance of the screen-printed layers after peeling off the adhesive tape was evaluated. The layers were backlit from below, and the images were made by optical microscopy. The topography and homogeneity of FTO/MoS 2 layers were characterized by scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM/EDX; JEOL JSM-IT500HR). The electron acceleration voltage was set to 20 kV.
All procedures for evaluation of the performance of FTO/MoS 2 electrodes for electrochemical sensing were run with a laboratory potentiostat/galvanostat Autolab PGSTAT 302 N with an impedimetric module (Ecochemie, Utrecht, Netherlands) in combination with NOVA 1.10 software. Differential pulse voltammetry (DPV) was used for the detection of dopamine. MoS 2 layers onto FTO substrates were used as working electrodes, the platinum wire was used as the counter-electrode and an argentochloride electrode (Ag/AgCl/3 M KCl) as the reference electrode. Five scans were run in a plain phosphate buffer pH 7.0 in a potential window of 0-1 V to stabilize FTO/MoS 2 electrodes. The parameters applied for the differential pulse voltammetry were as follows: 50 ms modulation time, 0.5 s interval time, 25 mV modulation amplitude, and 5 mV step. Measurements were run under Nova Software 1.10, and data acquired were evaluated using OriginPro 9.1. The limit of detection (LOD) was calculated as a signal-to-noise ratio (S/N) = 3.

Result and discussion
In order to investigate the influence of ink formulation on rheological properties, different screen-printable MoS 2 inks were prepared with various particle sizes and concentrations of MoS 2 . All prepared MoS 2 inks exhibited typical shear-thinning non-Newtonian rheological behavior (Fig. 1a). It is characterized by a decreasing dynamic viscosity with an increasing shear rate, and it is crucial for the screen-printing process 24 . The increasing shear rate adversely affects the cohesive internal forces between the particles and the binder and destroys the ink structure, which leads to an unsteady ink system 26 . When the shear rate increased from 0.001 to 1000 s −1 , the dynamic viscosity of all inks dropped gradually. It is obvious that dynamic viscosity increases with MoS 2 concentration. The viscosity of Mo6-60 sample with the highest MoS 2 content was slightly higher than what would be optimum for screen-printing. The maximum value was 105 Pa.s at 10 s −1 , and the minimum value was 0.02 Pa.s at 1000 s −1 . This rapid decrease indicates that a higher concentration of MoS 2 causes a faster breakdown of the structure. It is also clear that particle size does not significantly affect dynamic viscosity. Values of dynamic viscosity are identical (the maximum was ~ 9 Pa.s and the minimum was 3 Pa.s) for samples Mo6-25 (particle size ~ 6 µm) and Mo90-25 (particle size ~ 90 nm) containing 25 wt% of MoS 2 . The same values (the maximum was ~ 22 Pa.s and the minimum was ~ 2 Pa.s) were obtained also for samples Mo6-45 (particle size ~ 6 µm) and Mo90-45 (particle size ~ 90 nm) containing 45 wt% of MoS 2 .
The viscosity of screen-printable inks has to be low enough (under applied external force) to allow the squeegee to press the ink through the screen mesh, but also high enough (after releasing the force) to support ink to retain the geometry of printed patterns 24 . The rheological behavior of MoS 2 inks during the screen-printing process simulated by the time-dependent controlled-shear-rate tests is shown in Fig. 1b, and the characteristic parameters are listed in Supplementary Table S1. Screen-printing process can be divided into three steps. The first interval of the curve simulates the behavior of the ink at rest on the screen at a pre-set low shear rate (0.5 s −1 for 90 s). The second interval corresponds to the structural breakdown of the ink when it is pressing through the openings in the screen. The shear rate increases up to 1000 s −1 for 30 s. Finally, in the third interval, the shear rate decreases to 0.5 s −1 for 120 s to allow structural regeneration evaluation. During the shear rate increase from 0.5 to 1000 s −1 in the second interval, the viscosity value drops significantly ( Fig. 1b and Supplementary Tab. S1). The applied shear stress destroys the internal structure of the inks. Subsequently, the ink structure is rebuilt and restored when the shear rate returns to 0.5 s −1 in the third interval 26 . The viscosity of Mo6-25 and Mo90-25 was the lowest (8.8 and 8.9 Pa.s) at 0.5 s −1 , and these samples showed the highest recovery rate of 97% and 94% in 250 s, respectively. In addition, the rheological behavior during the process of the screen-printing simulation was the same for samples containing 25 wt% of MoS 2 (Mo6-25, Mo90-25) and also for the samples containing 45 wt% of MoS 2 (Mo6-45, Mo90-45). Thus, the particle size does not have a noticeable effect on the printing paste rheology within the printing process. The regeneration of these four samples was quick from the beginning, with the rise of the curve being slow afterwards. The recovery rate of sample Mo90-45 is slightly over 100% in 250 s (Fig. 1b). Thus, if the third interval was longer, viscosity would drop back to (or even under) the reference value 27 .
The result of the screen-printing simulation process for the sample Mo6-60 is different from the samples with lower MoS 2 concentrations. The initial viscosity is too high (170 Pa.s), and as the shear rate increased to 1000 s −1 , the viscosity decreased to 2.5 Pa.s. In the third interval, the regeneration is slower, indicating a longer time for levelling of the applied layers after printing, resulting in a smoother surface. Despite the recovery rate being only 60%, due to the slower increase in viscosity during the regeneration step, sample Mo6-60 can be considered the most suitable from a layer topography point of view.
To investigate the impact of MoS 2 concentration and particle size on the viscoelastic properties of the inks, oscillatory measurements were carried out for prepared inks, specifically shear strain amplitude sweep tests.  Table 1.
The LVR is defined as the region in which the ink can endure mechanical deformation without destroying the structure 28 . LVR values are determined from the G ′ curves. Beyond the LVR, the values of G ′ and G ′′ are continuously decreasing for all samples, indicating the gradual internal structure breakdown. The highest LVR had samples containing 25 wt% of MoS 2 (< 42%), followed by samples with 45 wt% of MoS 2 (< 11%), and the lowest  .59 and 28.76, respectively. Further investigation of the inks elastic behavior was carried out by analyzing the ratio of liquid-like to solid-like bahavior ( G ′′ /G ′ ) , the loss factor-tanδ within the LVR (Table 1). A tanδ < 1 demonstrates that the ink is elastic, cohesive or tacky 26 . All prepared MoS 2 inks have a loss factor tanδ > 1, which indicates viscous behavior 24 . Decreasing loss factor tanδ within the LVR indicates that the MoS 2 content can positively impact the inks' structural strength and elasticity 29 .
After rheology characterization, MoS 2 inks were screen-printed onto FTO substrates (Fig. 2a). The thickness of MoS 2 electrodes was evaluated for layers screen-printed onto Al 2 O 3 ceramic substrates. These substrates were chosen because of their smooth surface, opacity and, from our experience, are more suitable for thin layer thickness measurements. First, a needle scratch was made through the middle of the layers. After that, thicknesses were evaluated by optical microscope using the 3D image sequential recording method (Fig. 2b). As expected, the thickness increased with the concentration of MoS 2 in the inks, and the measured values ranged from 4 ± 1 to 18 ± 1 µm. Specifically, measured values were 4 ± 1, 11 ± 2 and 18 ± 1 µm for samples Mo6-25, Mo6-45, Mo6-60 and 4 ± 1, 10 ± 1 µm for Mo90-25 and Mo90-45, respectively. As it was seen, particle size did not influence the thickness, but layers based on smaller 90 nm MoS 2 particles had lower roughness.
Screen-printed MoS 2 layers have undergone an evaluation of adhesion to FTO substrate by a peeling test. Figure 2c shows optical microscopy images of samples after the peeling test. As it can be seen, layers containing 6 µm MoS 2 have lower adhesion than layers containing 90 nm MoS 2 . Also, a more significant part of the layer was pulled down with increasing MoS 2 concentration. Therefore, the highest adhesion to FTO substrate was evaluated for the sample based on 90 nm MoS 2 with the lowest particle content (Mo90-25). The adhesion to the FTO substrate and the cohesion of the layers are mainly affected by the size of the particle's phase interface, their surface energy, and the concentration of ethylcellulose. Therefore, the degree of adhesion could be increased by finding the ideal ratio between binder, ethylcellulose, and MoS 2 .
SEM measurements showed ( Supplementary Fig. S1) that the particle size in the layers prepared from MoS 2 particles, characterized by the manufacturer as particles with an average size of 90 nm, actually ranged up to www.nature.com/scientificreports/ micrometres. Therefore, we assume that the manufacturer's labelling refers more likely to the particles with the highest frequency of occurrence (90 nm) than their average size. Despite the discrepancy, it was evident from the SEM measurements that the average diameter of particles referred to as 90 nm is lower than the diameter of the second particle type used with a more correctly declared average size of 6 µm ( Supplementary Fig. S1). In addition, similar to the layer thickness measurements, it is evident that the layers prepared from the 90 nm MoS 2 inks have a smoother surface, with a higher frequency of submicron particles. Figure 3 shows SEM and EDX mapping results of selected samples Mo6-25, Mo6-45 and Mo90-25. The homogeneity of printed electrodes was insufficient in the case of the lowest MoS 2 concentrations, regardless of the particle size (Mo6-25 and Mo90-25). As a result, these electrodes did not cover FTO substrates homogenously, as shown in the SEM images or EDX element maps (measured Sn signal from the FTO). Insufficient homogeneity of these MoS 2 electrodes, related to the measured lowest thickness of 4 µm, can be problematic when applied to electrodes other than FTO. For example, when applied onto Ag electrodes in printed three-electrode systems, which are primarily intended to provide electron transport from the working electrode and can not interact with the analyte solution. Printed electrodes with 45 wt% and 60 wt% MoS 2 were homogenous and did not expose the surface of FTO substrates. Supplementary Tab. S2 shows the atomic representation of the detected elements, where the carbon intensity corresponds to the added binder ethylcellulose and varies depending on its concentration.  (Fig. 4). Dopamine molecule plays an important role in human metabolism, being a crucial neurotransmitter in the central nervous system maintaining neuro-physiological control of mental activities. The concentration of DA ranges between 1 and 2 mM in the intracellular fluids of the central nervous system. Any relevant deviation from an optimal concentration of dopamine in the body causes Parkinson's disease or schizophrenia. Since dopamine is electrochemically active, it may be detected by electrochemical oxidation. MoS 2 has been studied as a working electrode material for DA detection in many works by various authors. However, most of this work is devoted to sophisticated MoS 2 electrodes modified with other materials and prepared by different techniques, which are not as suitable for mass production of working electrodes as printing. Moreover, printing (especially screen-printing) brings several advantages, such as simple, cheap, fast and reproducible preparation of electrodes in large areas on various and flexible substrates. Below, for comparison, are briefly mentioned works of other authors dealing with the preparation of MoS 2 sensors for DA detection    43 . Zhang with co-workers found out that while MoS 2 nanosheets catalyzed the electro-oxidation process of Ru(bpy) 3 2+ to generate Ru(bpy) 3 3+ , DA inhibited the effect on electrogenerated chemiluminescence (ECL) intensity of Ru(bpy) 3 2+ -MoS 2 nanosheets through the energy transfer process. LOD of 8.5 × 10 −10 mol L −1 was obtained for DA 44 . All mentioned works are summarized in Table 2.

Conclusion
The screen-printed FTO/MoS 2 working electrodes for the electrochemical detection of dopamine were successfully prepared in this work for the first time. Screen-printable MoS 2 high viscosity inks contained MoS 2 with different particle sizes and concentrations, ethylcellulose and terpineol. The screen-printing simulation confirmed that MoS 2 inks are suitable for screen-printing process with corresponding flow behavior, and the viscosity recovery rate. Oscillatory modulus determined the linear viscoelastic region and helped to understand the solid-and liquid-like behavior of prepared MoS 2 inks during screen-printing. After rheological behavior analysis, MoS 2 inks were screen-printed onto FTO substrates. The differential pulse voltammetry measurements showed that the best limit of detection of 246 nM for DA (S/N = 3) was evaluated for the working electrode printed using the ink containing 45 wt% MoS 2 powder with an average particle size of 6 μm. Presented results showed that screenprintable MoS 2 inks could be prepared by a simple, time-efficient mixing process taking only a few minutes. Moreover, MoS 2 inks prepared in this way allow the fabrication of working electrodes for detection of dopamine or further target analytes by simple, low cost and for mass suitable screen-printing technique.